sports medicine Volume 39 Issue 6 June 2009

Sports Med 2009; 39 (6): 423-438 0112-1642/09/0006-0423/$49.95/0

LEADING ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Steps Per Day The Road to Senior Health? Yukitoshi Aoyagi1 and Roy J. Shephard2 1 Exercise Sciences Research Group, Tokyo Metropolitan Institute of Gerontology, Itabashi, Tokyo, Japan 2 Faculty of Physical Education and Health, University of Toronto, Toronto, Ontario, Canada

Abstract

In older adults, as in younger individuals, habitual moderate-intensity physical activity is associated with a reduced risk of various chronic health conditions, including certain types of cardiovascular and musculoskeletal disease and certain forms of cancer. However, the pattern of physical activity associated with such benefits remains unclear. One problem is that most investigators have examined patterns of physical activity using either subjective questionnaires or accelerometer or pedometer measurements limited to a single week, despite clear evidence of both the unreliability/invalidity of questionnaires and seasonal changes in activity patterns. Since 2000, we have thus conducted an interdisciplinary study examining the habitual physical activity and health of elderly people living in a mediumsized Japanese town (the Nakanojo Study). In about one-tenth of some 5000 available subjects aged ‡65 years, physical activity has already been assessed continuously for 24 h/day for >8 years using a specially adapted pedometer/ accelerometer. This device has a storage capacity of 36 days and can distinguish >10 intensities of physical activity (expressed in metabolic equivalents [METs]). Data have to date been summarized as daily step counts and daily durations of activity of <3 and >3 METs, averaged over a 1-year period. This article provides a detailed overview of both factors influencing habitual physical activity and relationships between such activity and health in an elderly population. To date, analyses have been cross-sectional in type. Substantial associations have been noted between the overall health of participants and both the daily duration of effort undertaken at an intensity of >3 METs and the daily step count. In men, the extent of health is associated more closely with the daily duration of activity of >3 METs than with the daily step count, whereas in women, the association is closer for the step count than for the duration of activity >3 METs. In both sexes, the threshold amount of physical activity associated with better health is greater for physical than for mental benefits: >8000 versus >4000 steps/day and/or >20 versus >5 min/day at an intensity >3 METs, respectively. In other words, better physical health is seen in those spending at least 20 min/day in moderate walking (at a pace of around 1.4 m/s [5 km/h]) and a further >60 min of light activity per day. In contrast, better mental health is associated with much smaller amounts of deliberate physical activity.

Aoyagi & Shephard

424

The daily step count and the daily durations of activity of <3 and >3 METs are all influenced by meteorological factors, particularly precipitation and mean ambient temperature. Activity decreases exponentially to about 4000 steps/day as precipitation increases. Excluding the influence of rainfall, the daily step count peaks at a mean outdoor temperature of around 17C; above and especially below such readings, physical activity decreases as a quadratic function of temperature. Seasonal changes in microclimate should thus be considered when designing interventions intended to increase the habitual physical activity of elderly people. The observed associations between physical activity and health outcomes point to a need for longitudinal analyses; these should examine potential causal interpretations of the current findings and elucidate possible additional mediating variables.

1. The Nakanojo Study in an International Perspective In older adults, as in younger individuals, habitual moderate-intensity physical activity is associated with a reduced risk of various chronic health conditions, including certain types of cardiovascular and musculoskeletal disease and certain forms of cancer.[1-5] However, the details of this association (intensity and total amount of activity) remain unclear.[6] One problem is that most of those investigating the associations have used questionnaires asking about the frequency and/or duration of a given type of physical activity during a typical recent week.[7-17] Such subjective responses necessarily provide only limited estimates of the volume and intensity of physical activity undertaken, particularly in older adults, many of whom have difficulties in recalling recent events and/or have some loss of cognitive function.[18] Pedometers and accelerometers have been employed more recently to assess physical activity patterns with greater accuracy.[19-37] Such measurements are objective, but unfortunately, data collection has usually covered no more than a single week, despite clear evidence of seasonal changes in activity, particularly among elderly people.[38-40] Recent observations[39] show that the numbers of days of recording needed to obtain reliable estimates of an individual’s habitual physical activity over an entire year are substantially fewer for random or seasonally distributed data than for consecutive sampling. However, ª 2009 Adis Data Information BV. All rights reserved.

even with random or seasonal data collection, the necessary observation period exceeds common practice. In an older adult with no formal employment, >10 random or seasonally selected days are needed to estimate a year’s activity with >90% reliability.[39] At least in short-term studies (£1 week), a further problem may arise from the individual’s reaction to wearing a recording device. Thus, Clemes et al.[22] noted that if people realized that they had been fitted with a pedometer, they walked some 13 000 additional steps in the first week of observation. Since 2000, we have been conducting an interdisciplinary study on the habitual physical activity and health of elderly people that addresses some of these problems (table I).[38-46] The test site is the medium-sized Japanese town of Nakanojo, located about 150 km northwest of Tokyo. National census data show a population of 17 491 (8501 men and 8990 women); 29.1% of these are aged ‡65 years (25.7% of men and 32.3% of women). Our subjects include all willing residents ‡65 years of age except those who are severely demented or bedridden (a total of almost 5000 participants). All have completed a simple conventional questionnaire[45] once a year, and in about one-tenth of the sample, physical activity has been assessed continuously 24 hours per day for >8 years using a specially adapted uniaxial pedometer/accelerometer (modified Kenz Lifecorder, Suzuken Co., Ltd, Nagoya, Aichi, Japan),[38-40] as recommended by Janz.[47] Technical details of this monitoring Sports Med 2009; 39 (6)

Steps Per Day for Senior Health?

425

Table I. Details of components of the Nakanojo Study cited in this report Research

Sample

Reference

total no. (male/female)

age range (mean – SD) [y]

Meteorology and habitual physical activity of the elderly (selected environmental factors [day length, mean ambient temperature, duration of bright sunshine, mean wind speed, mean relative humidity and precipitation] vs daily step count measured by pedometer/accelerometer over a 450-day period)

41 (20/21)

65–78 (71 – 4)

38

The minimum number of observation days needed to obtain reliable estimates of the annual habitual physical activity of an elderly individual (consecutive vs random or seasonal sampling of daily step count, based on observed intraindividual variations in pedometer/accelerometer data)

81 (37/44)

65–83 (71 – 4)

39

Sex, age, season and habitual physical activity of the elderly (comparisons of 1 y and/or 1 mo pedometer/accelerometer measurements [average daily step count and the daily durations of physical activity <3 and >3 METs] between men and women, between those aged <75 and ‡75 y, and over the course of a year)

95 (41/54)

65–83 (71 – 4)

40

A new field method of prescribing exercise for the elderly, based on walking speed measured over a 5 m distance (preferred and maximal walking speeds vs peak isometric . . knee extension strength and VO2max; determinations of %VO2max, %HRmax, . %VO2reserve, %HRreserve, and RPE when exercising at 30–70% of the maximal walking speed)

23 (10/13)

65–74 (68 – 3)

41

Habitual physical activity and bone health in the elderly (1 y pedometer/accelerometer measurements [average daily step count and the daily duration of physical activity >3 METs] vs calcaneal QUS parameters for bone density [SOS], structure [TI], and stiffness [OSI] and vs risks of osteoporosis and fracture)

172 (76/96)

65–83 (73 – 4)

42

Habitual physical activity and signs of metabolic syndrome in the elderly (1 y pedometer/ accelerometer measurements [average daily step count and the daily duration of physical activity >3 METs] vs the presence or absence of five diagnostic markers [BMI, TG, HDL-C, BP, and BG])

220 (91/129)

65–84 (72 – 4)

43

Habitual physical activity and HR-QOL in the elderly (1 y pedometer/accelerometer measurements [average daily step count and the daily duration of physical activity >3 METs] vs overall SF-36 score and its eight constituent dimensions [PF, RP, BP, GH, VT, SF, RE, and MH])

181 (73/108)

65–85 (73 – 6)

44

A new self-reported recall questionnaire for assessing four domains of habitual physical 3084 (1398/1686) activity (transportation, exercise/sports, housework, and labour) common among older Japanese (the first vs the second administrations of the PAQ-EJ; 1 mo pedometer/ accelerometer measurements [average daily step count and the daily durations of physical activity <3 and >3 METs] vs the corresponding scores for the PAQ-EJ; comparisons of PAQ-EJ scores between men and women, between independent and dependent individuals, and among those aged 65–74, 75–84 and 85–99 y)

65–99 (75 – 7)

45

Habitual physical activity and psychosocial status of the elderly (1 y pedometer/ accelerometer measurements [average daily step count and the daily duration of physical activity >3 METs] vs mood state [HADS], cognitive function [MMSE], and symptoms of anxiety, depression, and dementia)

65–85 (72 – 4)

46

184 (83/101)

BG = blood glucose; BMI = body mass index; BP = blood pressure[43] or bodily pain[44]; GH = general health; HADS = hospital anxiety and depression scale; HDL-C = high-density lipoprotein cholesterol; HR-QOL = health-related quality of life; METs = metabolic equivalents; MH = mental health; MMSE = mini-mental state examination; OSI = osteosonic index; PAQ-EJ = physical activity questionnaire for elderly . Japanese; %HRmax. = percentage of maximal heart rate; %HRreserve = percentage of heart rate reserve; %VO2max = percentage of maximal oxygen intake; %VO2reserve = percentage of oxygen intake reserve; PF = physical functioning; QUS = quantitative ultrasonic; RE = role limitations because of emotional problem; RP = role limitations because of physical health; RPE = rating of perceived exertion; SD = standard deviation; SF = social functioning; SF-36. = medical outcomes study 36-item Short-Form Health Survey; SOS = speed of sound; TG = triglycerides; TI = transmission index; VO2max = maximal oxygen intake; VT = vitality and energy.

device are widely available.[48-54] Motion sensors used by other investigators have differing sensitivity thresholds, filtering devices and/or ª 2009 Adis Data Information BV. All rights reserved.

attachments,[55-58] but of >10 monitors tested to date, our design of pedometer/accelerometer offers the most consistently accurate estimates of Sports Med 2009; 39 (6)

Aoyagi & Shephard

426

both step count (intramodel reliability 0.998; absolute accuracy –<3%) and the intensity of ambulatory activity under both controlled and freeliving conditions.[48,52,53] The device has a storage capacity of 36 days. Data have been downloaded and batteries replaced monthly throughout the >8 years of observation. To date, physical activity patterns have been summarized as the number of steps taken per day and the daily durations of activity in three intensity categories expressed in metabolic equivalents (<3, 3–6 and >6 metabolic equivalents [METs]). The primary aims of the Nakanojo Study so far have thus been to: (i) determine the overall patterns of daily physical activity that are most closely associated with good health and the absence of disease in the elderly; and (ii) identify personal, social and environmental factors that are important in allowing an amount of physical activity likely to maintain physical condition and delay the aging process. Secondary research objectives have included improving the effectiveness of interventions designed to promote habitual physical activity through the development of techniques for: (i) valid and reliable pedometer/accelerometer monitoring of such activity; and (ii) electronic evaluation and feedback of information about personal activity in the light of the various criteria noted above. Based on this extensive experience, the present article provides an overview of associations between patterns of habitual physical activity and the physical, psychosocial, mental and metabolic components of health in the elderly. It also explores possible interactions between step count and the amounts of light and moderately vigorous physical activity, and it examines psychological and meteorological factors that modulate the intensity and total volume of physical activity undertaken. Relevant evidence reviewed from other laboratories concerns mainly objective accelerometer and pedometer assessments of physical activity. 2. Characteristics of Habitual Physical Activity in the Elderly In characterizing habitual physical activity, it is important to recognize that the relative ª 2009 Adis Data Information BV. All rights reserved.

intensity of any absolute rate of working is age dependent.[59] Thus, the intensity of activities undertaken by an elderly person over a typical day (figure 1) can be divided into three categories: low (<3 METs), moderate (3–6 METs) and high (>6 METs).[1,2,4] Most national and international physical activity guidelines and position statements[1-5] recommend that the elderly participate in regular physical activity as a means to prevent disease, promote health and especially to delay functional loss. The moderate intensity commonly recommended for such individuals corresponds to 50% of the person’s oxygen intake reserve (50–60% of maximal oxygen intake . [VO2max]) or 50% of the heart rate reserve (60–70% of maximal heart rate).. [1,2,4] There are significant correlations between VO2max and both maximal (r > 0.80) and preferred gait speeds (r > 0.65) in the elderly, irrespective of sex.[41] Thus, 60% of the maximal walking velocity and/or 110–115% of the preferred walking velocity are appropriate exercise recommendations for an elderly person.[41] In our laboratory, we calculate these two velocities from the time that an individual takes to move as fast as possible and at a usual comfortable pace from the 3 m to the 8 m mark on a flat 11 m walkway. Pedometer cut points corresponding to the threshold of moderate-intensity walking are roughly 100 steps/ min in both men and women.[60] To date, we have examined associations between pedometer/accelerometer measurements of habitual physical activity and several aspects of physical and psychosocial health in older people (table I). In our first analysis, physical activity patterns recorded over an entire year were summarized as the average step count per day and the daily durations of low- (<3 METs) and moderate(>3 METs) intensity activity. The daily step count (x) and the daily duration of physical activity >3 METs (y) were significantly correlated with each other in both men and women (for pooled male and female data, y = [1.96 · 10-7]x2 + [1.16 · 10-3]x; r2 = 0.93; figure 2). The form of the relationship was such that almost no physical activity >3 METs was recorded in subjects who took <2000 steps/day. Above this threshold, each increment of 1000 steps/day up to 6000 steps/day Sports Med 2009; 39 (6)

Steps Per Day for Senior Health?

427

Activity intensity (METs)

Brisk walking

Housework

Errands

9 High 6 Moderate 3 Low 0 0:00

• Breakfast 2:00

4:00

6:00

8:00

• Lunch 10:00

12:00

• Dinner 14:00

16:00

18:00

20:00

22:00

24:00

Time (h) Getting up

Napping

Going to bed

Fig. 1. A typical example of daily physical activity in the elderly (as seen in 24-hour step count recordings from the Nakanojo Study).[38-40,42-46] Low-intensity effort (<3 METs [metabolic equivalents]) includes light sport and recreational activities such as leisurely walking for exercise or pleasure, a walk on specific errands such as going to/from a store or taking grandchildren to school, and light household chores such as sweeping or vacuuming, cooking and washing dishes. Moderate-intensity effort (3–6 METs) includes somewhat strenuous sport and recreational activities such as deliberate brisk walking for exercise, and somewhat heavy household chores such as outdoor gardening and yard care (including snow or leaf removal). High-intensity effort (>6 METs) would include strenuous sport and recreational activities such as jogging for exercise; however, study participants undertook such activities for only <1 min/day.

was associated with an additional 2.5 min/day of activity of >3 METs. From 6000 to 12 000 steps/ day, each 1000 steps/day increment added a further 5 min/day of activity of >3 METs; from 12 000 to 18 000 steps/day, each 1000 steps/day added another 7.5 min/day of >3 METs; and beyond 18 000 steps/day, each 1000 steps/day added another 10 min/day of >3 METs. In other words, those study participants who took <4000 steps/ day averaged <5 min/day at an intensity >3 METs, whereas those who were taking >10 000 steps/day averaged >30 min/day at >3 METs. Tudor-Locke and Bassett[61] proposed classifying pedometer determinations of physical activity in healthy younger adults according to the following schemes: (i) a ‘sedentary’ lifestyle (<5000 steps/day); (ii) ‘low active’, likely avoiding sports or deliberate exercise (5000–7499 steps/ day); (iii) ‘somewhat active’, likely including some volitional and/or significant occupational activity (7500–9999 steps/day); (iv) ‘active’ (10 000–12 499 steps/day); and (v) ‘highly active’ (‡12 500 steps/day). As a matter of convenience, they smoothed the five categories to uniform increments of 2500 steps/day. It remains unclear how far such a classification is appropriate when ª 2009 Adis Data Information BV. All rights reserved.

considering the health of elderly people. Other investigators have suggested that an increase of no more than 2000 steps/day over a sedentary baseline might yield important health outcomes such as a prevention of weight gain,[62] a reduction in waist circumference,[23] and an enhanced health-related quality of life (HR-QOL).[44] This view is supported in a systematic review by Bravata et al.,[63] who examined 26 studies (8 randomized controlled trials and 18 observational studies) involving a total of 2767 sedentary participants of 49 – 9 years. According to the findings of Bravata et al.[63] and data from the Nakanojo Study,[42-44] it seems logical to divide ostensibly healthy older people into quartiles of physical activity (Q1–Q4) on the bases of their year-averaged daily step counts and daily durations of activity of >3 METs (figure 2). The cut points are such that individuals falling in Q1 take 2000 to <5000 (mean 4000) steps/day and spend <7.5 (mean 5) min/day at an intensity >3 METs. The corresponding figures for the remaining quartiles are: Q2, 5000 to <7000 (mean 6000) steps/day and 7.5 to <15 (mean 10) min/day at >3 METs; Q3, 7000 to <9000 (mean 8000) steps/ day and 15 to <25 (mean 20) min/day at >3 METs; Sports Med 2009; 39 (6)

Aoyagi & Shephard

428

and Q4, ‡9000 (mean 10 000) steps/day and ‡25 (mean 30) min/day at >3 METs. Our data (figure 2) show that: individuals taking 0 step/day and spending 0 min/day at an intensity >3 METs are likely to be classed as ‘bedridden’ or a related condition; those taking <2000 steps/day and spending <2.5 min/day at >3 METs are typically ‘dependent’ or restricted in their physical activity; and at <4000 steps/day and <5 min/day of activity >3 METs, the usual classification is ‘housebound’ or almost housebound.[38]

Previous questionnaire and/or interview,[17,64-66] accelerometer[29] and pedometer studies[19] from both Europe and North America suggest that habitual physical activity is likely to be greater in men than in women. Our pedometer/accelerometer data[40] support this view; accurate objective yearlong estimates of physical activity show mean values 20–30% greater in men than in women of equivalent points in old age (7000–8000 vs 6000–7000 steps/day, 45–60 vs 30–45 min/day at an intensity <3 METs, and 15–20 vs 10–15 min/day at

35.0 Male-specific area including more exercise

Metabolic health (age <75 years)

Q4

30.0

25.0 Metabolic health (age ≥75 years)

Q3

20.0

Physical health

15.0

Q2

10.0

Psychosocial health

7.5

Mental health

Q1

Housebound

5.0 Female-specific area including more housework

Year-averaged duration of physical activity >3 METs (min/day)

40.0

2.5

Dependent Bedridden

0 0

1000

2000

3000

4000 5000 6000 7000 8000 Year-averaged step count (steps/day)

9000 10 000 11 000 12 000

Fig. 2. Schematic diagram showing categories of habitual physical activity in elderly Japanese and relationships between such activity and health (based on data from the Nakanojo Study).[38,42-46] Subjects are categorized by the shaded and crosshatched areas on the bases of step count and the duration of physical activity >3 METs (metabolic equivalents): bedridden, 0 step/day and 0 min/day; dependent, <2000 steps/day and <2.5 min/day; housebound, <4000 steps/day and <5 min/day; quartile (Q) 1, 2000 to <5000 (mean 4000) steps/day and <7.5 (mean 5) min/day; Q2, 5000 to <7000 (mean 6000) steps/day and 7.5 to <15 (mean 10) min/day; Q3, 7000 to <9000 (mean 8000) steps/day and 15 to <25 (mean 20) min/day; and Q4, ‡9000 (mean 10 000) steps/day and ‡25 (mean 30) min/day. The thresholds of step count and/or duration of physical activity >3 METs associated with avoidance of health problems are: >4000 steps/day and/or >5 min/day for a lower risk of mental health disorders including depression; >5000 steps/day and/or >7.5 min/day for lower risk of impaired psychosocial health including a poor health-related quality of life; at least 7000–8000 steps/day and/or at least 15–20 min/day for lower risks of aortic arteriosclerosis, osteoporosis and sarcopenia and a higher level of physical fitness (particularly lower-extremity strength and gait speed in ‡75-year-old adults); and >8000 and >10 000 steps/day and/or >20 and >30 min/day for better metabolic health including a lower risk of hypertension and hyperglycaemia in adults aged ‡75 and <75 years, respectively. An analogous matrix can be proposed for younger adults by increasing the MET values on the ordinates (e.g. >3, >4 and >5 METs for those aged ‡65, 45–64 and 25–44 years, respectively). Q1–Q4 = first through fourth quartiles of physical activity in study participants (n = about 50 for each quartile).

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (6)

Steps Per Day for Senior Health?

an intensity >3 METs, respectively). On the other hand, the distribution of time between efforts <3 and >3 METs showed a similar 3 : 1 ratio in men and women.[40] Habitual physical activity is known to decrease progressively with age.[17,19,20,27,29,36,37,64,66,67] Our objective data[40] confirm the continuation of such a trend into the retirement years, with an inverse association between age and the year-averaged daily step count. In men, this was attributable mainly to an age-related reduction in activity of >3 METs, but in women, the decrease was in activity of <3 METs.[40] The sex difference might be explained by at least two factors (figure 2):[45] (i) regardless of age, many traditional older Japanese women spend long periods performing lowintensity household tasks; and (ii) free-living older men progressively assume the burden of such tasks because of the infirmity or death of their spouses. More precise data on relationships between patterns of physical activity and the performance of household chores are needed, but our findings to date suggest that the extent to which physical activity patterns change with age depends substantially on the life circumstances of the individual. A power spectrum analysis of individual daily step count data for 81 elderly Japanese was obtained by fast Fourier transformation.[39] Individual habitual physical activity patterns showed peaks with periods of 2.3, 3.5 and 7.0 days and an aperiodic component, which increased in power at low frequencies and thus could not be dismissed as white noise.[39] However, group-averaged data did not show a corresponding pattern; probably, the peaks were masked by the averaging out of interindividual differences in the effects of day of the week and season of the year.[39] The 2.3- and/or 3.5-day cycles of activity may in part be harmonic components of weekly variations, but could also reflect the frequency of going outdoors, attending clubs or communal meals, and/or shopping. Future studies should examine these phenomena more closely. Based on the characteristics of the power spectrum, we have defined the number of days of pedometer/accelerometer monitoring needed to attain specified confidence levels when estimating ª 2009 Adis Data Information BV. All rights reserved.

429

the annual habitual physical activity of individual subjects.[39] The necessary observation period is surprisingly long, although obviously it can be shorter if the need is simply for averaged data on a large sample. In an elderly man with no permanent employment, 25 days of consecutive data collection are required to obtain estimates of yearly step count with 80% reliability.[39] Relative to younger individuals, the activity of seniors is likely more vulnerable to unfavourable exogenous factors such as an adverse climate.[38,40] Probably because women usually assume the main burden of low-intensity household tasks,[45] the activity pattern of an elderly woman is more regular, and the necessary period of data collection is shorter than for a man – just 8 days.[39] To reach 90% reliability, 105 and 37 days of consecutive observation are needed in a man and a woman, respectively.[39] By sampling on a random basis, the monitoring periods can be substantially reduced: 4 days in both sexes for 80% reliability and 11 and 9 days for 90% reliability in a man and a woman, respectively.[39] If sampling considers the effects of season and day of the week, the periods of monitoring for 80% reliability can be reduced to 8 and 4 days for a man and a woman, respectively (i.e. 2 and 1 consecutive days every 89 days); for 90% reliability, the corresponding periods are 16 and 12 days (i.e. 4 and 3 consecutive days every 89 days).[39] 3. Habitual Physical Activity and Health in the Elderly Cross-sectional data from our project[42-44,46] and other studies of older adults[24-26,28] indicate that many measures of physical and psychosocial health are related to both the intensity and the total volume of habitual physical activity undertaken (figure 2). In men, the degree of health is associated more closely with the daily duration of physical activity of >3 METs than with the daily step count, whereas in women, the closer association is with step count.[42,44] In the elderly, large fractions of the daily step count reflect minor movements (at an intensity <3 METs) rather than deliberate walking.[40] We have suggested that the typical older woman in our sample Sports Med 2009; 39 (6)

430

spends long periods performing low-intensity household chores.[45] Given that the volume of such low-intensity effort has a positive association with various health outcomes, it may thus be important to encourage older people to engage in regular physical activity, even if they only reach low intensities of effort for much of the time that they are active. Aspects of impaired mental and psychosocial health such as a depressive mood state[46] and a poor HR-QOL[44] are less likely to be seen in elderly individuals who meet even very modest minimum standards of habitual physical activity: at least 4000–5000 steps/day and/or at least 5–7.5 min/day at an intensity >3 METs (figure 2). However, there remains an urgent need for longitudinal research to determine whether activity influences mood or whether the converse is the case. Among those with a clinical diagnosis of depression (n = 8, corresponding to 4.3% of our sample),[46] all except one male were taking <4000 steps/day; the exception took 8057 steps/day but at a low intensity, spending only 6.6 min/day of physical activity of >3 METs. It seems likely that the depressed individuals rarely went outdoors. A 450-day analysis of 41 participants in the Nakanojo Study[38] demonstrated an exponential decrease in physical activity from 6600 to 4000 steps/day as precipitation increased, and from this we inferred that a count <4000 steps/day was presumptive evidence that a person was remaining indoors, with most of the recorded counts reflecting very low-intensity incidental movements. We certainly need more precise data on relationships between step count, exercise intensity and the frequency of going outdoors. Nevertheless, our findings suggest that at least in this population, outdoor activity is generally needed to accumulate substantial activity at an intensity >3 METs, and it is this type of activity that is associated with mental health. We also saw significant associations between daily step count, the daily duration of physical activity >3 METs and the overall HR-QOL as assessed by the medical outcomes study 36-Item Short-Form Health Survey (SF-36) scale (potential range of score 0–100)[68,69] [figure 2]. After co-varying for age, the overall HR-QOL in both men and ª 2009 Adis Data Information BV. All rights reserved.

Aoyagi & Shephard

women was substantially higher (>10 units) in the second through fourth quartiles (Q2–Q4) than in the first quartile (Q1) of physical activity, whether classified by step count or the duration of activity of >3 METs.[44] The threshold volume of habitual physical activity associated with many aspects of better physical health, ranging from higher levels of mucosal immune function such as salivary secretory IgA[26] to freedom from musculoskeletal diseases such as sarcopenia and osteoporosis,[42] is at least 7000–8000 steps/day and/or at least 15–20 min/day at an intensity >3 METs in both men and women (figure 2). This seems significantly greater than that associated with better psychological health. Among our subjects who met the minimum criteria for better physical health, all except a few females either showed no evidence of osteoporosis or exceeded the diagnostic T-score of -2.5 associated with an increased risk of fractures.[42] Part of any benefit to bone may come from ultraviolet exposure rather than from physical activity per se. The US National Institutes of Health[70] have recommended at least 15 min/day of exposure to sunlight in order to meet body vitamin D requirements and thus facilitate calcium absorption. Our observations[42] suggested associations between a larger daily volume of exercise at an intensity >3 METs (which we believe is indicative of outdoor activity in our sample) and better bone health (as indicated by higher calcaneal quantitative ultrasonic [QUS] scores). In contrast, all subjects who would be categorized as sedentary (those taking <4000 steps/day and spending <5 min/day at an intensity >3 METs) had relatively low QUS values;[42] the T-scores of their osteosonic index (OSI; a figure that provides information on bone stiffness) typically fell below the fracture threshold (a value of -1.5). When our physical activity data for the year were categorized into quartiles, a logistic regression calculation of the multifactoradjusted odds ratio showed that the estimated vulnerability to fracture was significantly related to the daily step count and/or the duration of physical activity >3 METs (figure 2). The estimated risk of fractures in the lowest two quartiles (Q1 and Q2) was several times higher than that in Sports Med 2009; 39 (6)

Steps Per Day for Senior Health?

the top quartile (Q4), although the risk for this last group did not differ significantly from that for the second highest quartile (Q3).[42] Further work is needed to explore how much of the lower predicted risk of fractures in more active individuals is attributable to sunlight exposure and how much is related to the mechanical effects of exercise. The evidence of higher immune function seen in individuals taking >7000 steps/ day was substantiated by Shimizu et al.,[26] who applied a similar methodology to 284 healthy Japanese adults of comparable age (71 – 5 years; table I) over 14 days. In terms of pedometer counts, the salivary concentration and secretion rate of IgA were significantly higher for Q3 than for Q1, although salivary flow rates showed no interquartile differences.[26] Our observations[43] suggest that the threshold volume of habitual physical activity associated with absence of the metabolic syndrome may be even greater (figure 2). The five diagnostic criteria[71,72] that we adopted were: (i) a body mass index ‡25 kg/m2; (ii) a fasting serum triglyceride ‡1.7 mmol/L (‡150 mg/dL); (iii) a fasting serum high-density lipoprotein cholesterol <1.0 mmol/L (<40 mg/dL) for men or <1.3 mmol/L (<50 mg/ dL) for women; (iv) a systolic blood pressure ‡130 mmHg and/or a diastolic blood pressure ‡85 mmHg; and (v) a fasting plasma glucose ‡6.1 mmol/L (‡110 mg/dL) and/or a haemoglobin A1c ‡5.5%. Among individuals aged 65–74 years, risks of hypertension and hyperglycaemia were lower only in those who took >10 000 steps/day and/or spent >30 min/day of physical activity at >3 METs, and in those aged 75–84 years, the corresponding threshold was >8000 steps/day and/or >20 min/day at an intensity >3 METs.[43] Few subjects who met such activity levels showed three or more of the selected metabolic risk factors while taking any prescribed medications.[43] These various observations seem in general keeping with current health guidelines for older adults as proposed by Health Canada and the US Centers for Disease Control and Prevention,[73] and the American College of Sports Medicine and the American Heart Association.[4] All of these groups recommend that in order to prevent disease and promote health, the elderly should ª 2009 Adis Data Information BV. All rights reserved.

431

undertake moderate-intensity aerobic activity totalling a minimum of 30 min on 5 days each week, with some recommendations offering the alternative of vigorous-intensity aerobic activity for a minimum of 20 min on at least 3 days per week. However, our observations also undeline recent WHO comments[74] that physical activity recommendations that have focused on other aspects of physical and psychosocial health may need to be increased substantially if the obesity epidemic is to be controlled and the spread of the metabolic syndrome countered. The greater levels of activity recommended by the WHO should also help to reduce the functional impairment and disability commonly found among older individuals with the metabolic syndrome.[75] Several small-scale interventional studies,[76-79] mostly on somewhat younger people, have demonstrated that at least in such individuals additional health benefits can accrue if habitual physical activity is increased further to 10 000–20 000 steps/day over a period of 6–24 weeks. In 30 hypertensive male industrial workers, specific benefits included a lowering of both systolic and diastolic blood pressure, an increased capacity for endurance exercise . (VO2max), and a reduced sympathetic nerve activity.[76] In 15 postmenopausal women with hypertension, both body mass and systolic blood pressure were reduced.[77] In nine middle-aged individuals with type 2 diabetes mellitus, waist girth and systolic blood pressure were decreased,[78] and in 14 hospitalized obese patients with type 2 diabetes, body mass was reduced and insulin sensitivity increased.[79] Although the immediate success of such interventions seems impressive, these findings are unlikely to be replicated in ‘real-world’ situations, where it is difficult to sustain high levels of physical activity. Given that many of the acute benefits are lost within a few weeks of ceasing exercise,[80] it is probably important that elderly people are encouraged to engage in modest levels of voluntary physical activity that they are likely to maintain, rather than focus on a short-term regimen of intensive and closely supervised training.[3,81,82] Applying exponential regression models to our year-long data,[42] the QUS parameter for Sports Med 2009; 39 (6)

432

the calcaneus appeared to maintain a positive association with physical activity up to a volume of 8000–10 000 steps/day (and/or 20–30 min/day at an intensity >3 METs). The OSI was only slightly greater in those who exceeded this step count;[42] we thus suspect that activity exceeding this limit is likely to add little to bone health in an elderly person. The possibility of a ceiling response is echoed in a study by Kitagawa et al.;[25] based on 7 days’ observation of 143 Japanese women aged 71 – 6 years, they reported a mean count of 8401 – 3404 steps/day, with a quadratic attenuation of increments in the calcaneal QUS parameter above 12 000 steps/day. Nevertheless, their estimates of both the mean level of activity and the turning point are at least 2000 steps/day greater than what we observed in Japanese women of similar age (table I; 6288 – 2556 and at most 10 000 steps/day, respectively). Possibly, the discrepancy reflects the brief unblinded observation period in the study of Kitagawa and associates[25] (see the 1-week study of Clemes et al.,[22] in which there was a reactive increase of pedometer counts [from 9541 – 3186 to 11 385 – 3763 steps/day] when subjects knew that their activity patterns were being recorded). Our current data[39] show that >37 consecutive days of observation are required to achieve >90% reliability when estimating the annual habitual physical activity of a female who no longer has formal employment. It is possible that Kitagawa et al.[25] would have reached similar conclusions to ours given a longer period of observation. Certainly, the existence and level of any such ceiling need further exploration. In summary, accurate year-long observations of habitual physical activity suggest that better overall health is seen in elderly individuals who take an average of >8000 steps/day and/or spend >20 min/day at an intensity >3 METs (figure 2). Moreover, the extent of physical and psychosocial health seems greater in older individuals who undertake a larger proportion of their daily activity at an intensity >3 METs (i.e. those who are on or above the dotted line in figure 2). Thus, it would be useful to test the causal nature of the inference that among those who currently take a baseline of <6000 or 6000–10 000 steps/day, an ª 2009 Adis Data Information BV. All rights reserved.

Aoyagi & Shephard

increase of 2000 steps/day and/or respective increments of >5 or >10 min/day of physical activity >3 METs would lead to practical and clinically significant health advantages (figure 2). 4. Factors Affecting Habitual Physical Activity in the Elderly Consistent participation in moderate levels of physical activity appears necessary to optimize health.[1-5] Various authors have identified the day of the week[31,39,83,84] and the month or season of the year[38-40,83-91] as periodic factors that commonly influence the physical activity of free-living humans. The periods of 2.3, 3.5 and 7.0 days observed in power spectrum analyses of our individual step count data appear unrelated to sex, age or exercise habit,[39] but they may be modulated substantially by symptoms of anxiety and depression. The data suggest that in those with symptoms of anxiety and/or depression, intraindividual variations within a 1-week period are less marked than in the remainder of our sample, whereas the variations over an approximately 3-month period are greater than in other subjects. The implication may be that a depressed person has less interest in and/or less concern for things, and in consequence his or her lifestyle becomes more monotonous. However, this tendency may vary with the season (particularly in those with seasonal affective disorder)[92] and other short-term fluctuations of mood state. Variability in habitual physical activity is linked not only to common endogenous factors such as mood state, but also to exogenous influences, particularly precipitation, day length, and the range of ambient temperatures.[38-40,83-91] Public health recommendations[86] have noted that a short day length and extremes of ambient temperature are potential barriers to healthenhancing outdoor physical activity. Such environmental factors contribute to commonly observed seasonal variations in habitual physical activity. Several cross-sectional[90,91] and longitudinal studies[83-89] of younger adults from both Europe and North America have reported that in most climates, a person’s physical activity peaks in summer and reaches its nadir in winter. Sports Med 2009; 39 (6)

Steps Per Day for Senior Health?

Looking at a large US population, the Centers for Disease Control and Prevention[86] reported that leisure-time physical activity was increased during the summer and reduced during the winter months; this trend seemed true for both sexes and most age and racial/ethnic groups. Similarly, Pivarnik et al.[89] demonstrated that the average weekly leisure-time energy expenditure of 2843 Michigan adults was 15–20% higher during the spring and summer months than at other seasons. Applying a 24-hour physical activity recall technique to 580 Massachusetts residents five times per year, Matthews et al.[88] demonstrated seasonal changes in total physical activity of 1.4 MET h/day (507 kJ/day [121 kcal/day]) in men and 1.0 MET h/day (293 kJ/day [70 kcal/day]) in women. Physical activity peaked in July, and variations seemed to be associated with environmental conditions, specifically precipitation (measured as the number of rainy days per month), mean ambient temperature, and number of hours of daylight before 8am and after 5pm.[88] However, detailed study of relationships between environmental factors and habitual physical activity has been hampered by limitations in the available data sets, particularly the frequency of collection and the reliability/validity of the data.[93] The full extent of differences in physical activity between the most and the least active months can be detected by accurate year-long pedometer/accelerometer measurements,[38,40] but may be missed by infrequent subjective questionnaire estimates of physical activity.[85,88-90] Thus, the magnitude of the seasonal differences (15–20%) reported by Pivarnik et al.[89] is only a half of that (30–40%) detected by our full-year observations.[40] We examined relationships between the daily physical activity of free-living elderly people and selected environmental factors (day length, mean ambient temperature, duration of bright sunshine, mean wind speed, mean relative humidity and precipitation).[38] Habitual physical activity decreased exponentially from 6600 to 4000 steps/ day with increasing precipitation.[38] We suggested that a count <4000 steps/day was a simple objective criterion that an elderly person had become housebound. Such restriction of activity may induce a vicious circle of depression and ª 2009 Adis Data Information BV. All rights reserved.

433

insomnia.[46,94-96] Excluding days when precipitation was ‡1 mm and thus the main adverse effects of rainfall, our data[38] suggested that physical activity was influenced more strongly by day length (time from sunrise to sunset) and particularly ambient temperature than by duration of bright sunshine, wind speed or relative humidity. The daily step count peaked at a mean outdoor temperature of around 17C.[38] Above and especially below this temperature, physical activity decreased in a quadratic fashion.[38] Seasonal variations in our study[38,40] amounted to some 1500 steps/day, but effects could well be larger in areas with more extreme climates.[83] For homeotherms, changes in ambient temperature threaten the ability of the organism to keep core body temperatures within an appropriate narrow range.[97,98] However, in most cultures, humans react to environmental disturbances not only by autonomic regulatory mechanisms but also by behavioural adaptations (changes in habitual physical activity and in the amount and type of clothing worn).[99-105] In this way, they minimize the adverse consequences of exposure to quite wide ranges of temperature.[97,98] In a hot environment, physical activity is commonly decreased in order to limit increases in core temperature.[38,40] Under cold conditions, physical activity may also be decreased for a variety of reasons:[38,40] the ground conditions may be dangerous, the cold may cause physical discomfort, and the benefits of increased internal heat production may be offset by changes in posture and thus an increased convective heat loss that worsen thermal equilibrium. Comparative biology[106] shows that if small rodents are allowed to move along a thermal gradient, they displace themselves in such a way as to gain their preferred ambient temperature. Likewise, freeliving older people may try to select an environment that facilitates the regulation of their core temperature and maximizes comfort and safety, thereby contributing to seasonal changes in physical activity. In many developed countries, typical outdoor temperatures fall outside the thermoneutral range during both summer and winter months, thus encouraging elderly people to remain indoors (where ambient temperatures Sports Med 2009; 39 (6)

Aoyagi & Shephard

434

can be controlled by heating and air-conditioning units). During the spring and autumn, in contrast, outdoor temperatures approximate the thermoneutral range, and outdoor physical activity is then at its greatest. The Nakanojo data[38] suggest that total daily physical activity peaks at a mean outdoor temperature of around 17C. In some climates, fear of falling on icy surfaces can be an important additional factor inhibiting the outdoor activity of older individuals during the winter months.[107] In the early morning, temperatures in many regions sometimes dip below 0C, causing slippery conditions. A preliminary 1-year pedometer self-monitoring study of 23 working adults[84] showed significant differences in daily step count with season (summer > winter), day of the week (weekday > weekend), type of day (workday vs non-workday), and participation in sport/exercise (day with > day without sport/exercise). Our continuous yearlong pedometer/accelerometer assessments of 95 seniors,[40] likewise, showed clear seasonal variations in the month-averaged daily step count and the daily durations of physical activity <3 and >3 METs, irrespective of sex or age. Our selected measurements of habitual physical activity peaked in spring and/or autumn and reached their nadir in the winter months.[40] Many of the acute health benefits of physical activity are lost within a few weeks of ceasing exercise.[80] It is therefore possible that environmental factors could cause seasonal variations in the immediate risk of various chronic health conditions. The incidence of and mortality from coronary heart disease do indeed exhibit winter peaks in countries both north and south of the equator,[108] although this is probably due more to the direct effects of cold on blood pressure and the effects of snow-shovelling than to short-term reductions in habitual physical activity. Likewise, unusually hot spells are associated with an increase in deaths,[109] in part due to the direct circulatory effects of heat strain, but possibly influenced also by a seasonal decrease in the volume of moderate physical activity. In summer, when our daily step count approximated the average for the year, the proportion of activity <3 METs was increased at the expense of that ª 2009 Adis Data Information BV. All rights reserved.

>3 METs,[40] probably a manifestation of behavioural thermoregulation as described above.[97,98] Behavioural thermoregulation might also explain our findings[40] that physical activity <3 METs peaked in May or June, when mean ambient temperature was on the increase, whereas that >3 METs peaked in November, when the temperature was decreasing. Such observations suggest that in order to optimize health-promotional initiatives and reduce both direct and indirect health risks, due account should be taken of seasonal changes in the microclimate of elderly people. In many parts of the world, there is a need to encourage indoor (air-conditioned or climate-controlled) physical activity, such as mall walking, during extremes of summer and winter weather in order to maintain participation in at least moderate levels of physical activity. In areas where indoor facilities are not readily available, the time of day when exercise is taken can be changed, or a less stressful seasonally appropriate climate-adjusted target chosen. For instance, in a region where mean ambient temperatures range from -5C to 30C, older adults might attain a yearly average of 8000 steps/day and 20 min/day of physical activity >3 METs (figure 2) by taking 9000 steps/day and spending 25 min/day at an intensity >3 METs in the spring, 8000 steps/day and 15 min/day at >3 METs in the summer, 9000 steps/day and 30 min/day at >3 METs in the autumn, and 6000 steps/day and 10 min/day at >3 METs in the winter. 5. Conclusions Current cross-sectional data indicate that the overall health of older people is associated with both the year-averaged daily duration of physical activity undertaken at an intensity >3 METs and the year-averaged daily step count. In men, the extent of health seems associated more closely with the daily duration of activity >3 METs than with the daily step count, whereas in women, the closer association is with step count. In both sexes, the threshold amount of physical activity associated with better physical health is >8000 steps/day and/or >20 min/day at an intensity Sports Med 2009; 39 (6)

Steps Per Day for Senior Health?

>3 METs. The corresponding figures for better mental health are >4000 steps/day and/or >5 min/day at >3 METs. The physical health data can be equated with a daily total of at least 20 min of moderate activity (at a pace of some 1.4 m/s [5 km/h]) and at least 60 min of light activity. In contrast, good mental health seems associated with quite limited amounts of deliberate physical activity (although it remains to be assessed how far poor mental health is responsible for a lack of physical activity, rather than the converse). Plainly, the observed associations between physical activity and health outcomes also need examining longitudinally in order to test causal inferences and elucidate mediating variables. Both the intensity and the total amount of physical activity are influenced by meteorological factors, particularly mean ambient temperature and the extent of precipitation. In our study, habitual physical activity decreased exponentially to approximately 4000 steps/day with significant precipitation. Excluding the influence of rainfall, the daily step count peaked at a mean outdoor temperature of around 17C; above and especially below this temperature, physical activity decreased in a quadratic fashion. Seasonal changes in microclimate need to be taken into account when designing interventions to increase the physical activity of elderly people throughout the entire year. Acknowledgements This article focuses particularly on data from an interdisciplinary study on the habitual physical activity and health of elderly people living in Nakanojo, Gunma, Japan (the Nakanojo Study). The Nakanojo Study was supported in part by grants (Grant-in-Aid for Encouragement of Young Scientists: 12770037 and Grant-in-Aid for Scientific Research [C]: 15500503, [C]: 17500493, and [B]: 19300235) from the Japan Society for the Promotion of Science. The authors gratefully acknowledge the expert technical assistance of the research and nursing staffs of the Tokyo Metropolitan Institute of Gerontology (particularly Mr Hyuntae Park, Mr Sungjin Park, Dr Fumiharu Togo, Dr Akitomo Yasunaga and Mr Eiji Watanabe), The University of Tokyo (especially Dr Kazuhiro Yoshiuchi), and the Nakanojo Public Health Center. We would also like to thank the subjects whose participation made the Nakanojo Study possible. No sources of funding were used to assist in the preparation of this manuscript. The authors have no conflicts of interest that are directly relevant to the content of this manuscript.

ª 2009 Adis Data Information BV. All rights reserved.

435

References 1. American College of Sports Medicine Position Stand. The recommended quantity and quality of exercise for developing and maintaining cardiorespiratory and muscular fitness, and flexibility in healthy adults. Med Sci Sports Exerc 1998 Jun; 30 (6): 975-91 2. American College of Sports Medicine Position Stand. Exercise and physical activity for older adults. Med Sci Sports Exerc 1998 Jun; 30 (6): 992-1008 3. Aoyagi Y, Shephard RJ. Aging and muscle function. Sports Med 1992 Dec; 14 (6): 376-96 4. Nelson ME, Rejeski WJ, Blair SN, et al. Physical activity and public health in older adults: recommendation from the American College of Sports Medicine and the American Heart Association. Med Sci Sports Exerc 2007 Aug; 39 (8): 1435-45 5. Shephard RJ. Aging, physical activity, and health. Champaign (IL): Human Kinetics, 1997 6. Blair SN, Haskell WL. Objectively measured physical activity and mortality in older adults. JAMA 2006 Jul 12; 296 (2): 216-8 7. Brahm H, Mallmin H, Michaelsson K, et al. Relationships between bone mass measurements and lifetime physical activity in a Swedish population. Calcif Tissue Int 1998 May; 62 (5): 400-12 8. Coupland CA, Grainge MJ, Cliffe SJ, et al. Occupational activity and bone mineral density in postmenopausal women in England. Osteoporos Int 2000 May; 11 (4): 310-5 9. Evans EM, Ross KM, Heinrichs KL, et al. Ultrasound of the calcaneus and bone mineral density differs in older black and white women but is not impacted by current physical activity. Osteoporos Int 2005 Dec; 16 (12): 1755-60 10. Greendale GA, Barrett-Connor E, Edelstein S, et al. Lifetime leisure exercise and osteoporosis: the Rancho Bernardo study. Am J Epidemiol 1995 May 15; 141 (10): 951-9 11. Kakiyama T, Matsuda M, Koseki S. Effect of physical activity on the distensibility of the aortic wall in healthy males. Angiology 1998 Sep; 49 (9): 749-57 12. Laaksonen DE, Lakka HM, Salonen JT, et al. Low levels of leisure-time physical activity and cardiorespiratory fitness predict development of the metabolic syndrome. Diabetes Care 2002 Sep; 25 (9): 1612-8 13. Lee C, Russell A. Effects of physical activity on emotional well-being among older Australian women: crosssectional and longitudinal analyses. J Psychosom Res 2003 Feb; 54 (2): 155-60 14. Medras M, Slowinska-Lisowska M, Jozkow P. Impact of recreational physical activity on bone mineral density in middle-aged men. Aging Male 2005 Sep-Dec; 8 (3-4): 162-5 15. Nguyen TV, Center JR, Eisman JA. Osteoporosis in elderly men and women: effects of dietary calcium, physical activity, and body mass index. J Bone Miner Res 2000 Feb; 15 (2): 322-31 16. Pescatello LS, Murphy DM, Anderson D, et al. Daily physical movement and bone mineral density among a mixed racial cohort of women. Med Sci Sports Exerc 2002 Dec; 34 (12): 1966-70 17. Washburn RA, McAuley E, Katula J, et al. The physical activity scale for the elderly (PASE): evidence for validity. J Clin Epidemiol 1999 Jul; 52 (7): 643-51

Sports Med 2009; 39 (6)

436

18. Stewart AL, Mills KM, King AC, et al. CHAMPS physical activity questionnaire for older adults: outcomes for interventions. Med Sci Sports Exerc 2001 Jul; 33 (7): 1126-41 19. Bennett GG, Wolin KY, Puleo E, et al. Pedometerdetermined physical activity among multiethnic lowincome housing residents. Med Sci Sports Exerc 2006 Apr; 38 (4): 768-73 20. Cavanaugh JT, Coleman KL, Gaines JM, et al. Using step activity monitoring to characterize ambulatory activity in community-dwelling older adults. J Am Geriatr Soc 2007 Jan; 55 (1): 120-4 21. Chan CB, Spangler E, Valcour J, et al. Cross-sectional relationship of pedometer-determined ambulatory activity to indicators of health. Obes Res 2003 Dec; 11 (12): 1563-70 22. Clemes SA, Matchett N, Wane SL. Reactivity: an issue for short-term pedometer studies? Br J Sports Med 2008 Jan; 42 (1): 68-70 23. Dwyer T, Hosmer D, Hosmer T, et al. The inverse relationship between number of steps per day and obesity in a population-based sample: the AusDiab study. Int J Obes 2007 May; 31 (5): 797-804 24. Gerdhem P, Dencker M, Ringsberg K, et al. Accelerometermeasured daily physical activity among octogenerians: results and associations to other indices of physical performance and bone density. Eur J Appl Physiol 2008 Jan; 102 (2): 173-80 25. Kitagawa J, Omasu F, Nakahara Y. Effect of daily walking steps on ultrasound parameters of the calcaneus in elderly Japanese women. Osteoporos Int 2003 May; 14 (3): 219-24 26. Shimizu K, Kimura F, Akimoto T, et al. Effect of freeliving daily physical activity on salivary secretory IgA in elderly. Med Sci Sports Exerc 2007 Apr; 39 (4): 593-8 27. Strycker LA, Duncan SC, Chaumeton NR, et al. Reliability of pedometer data in samples of youth and older women. Int J Behav Nutr Phys Act 2007 Feb 17; 4: 4 28. Sugawara J, Otsuki T, Tanabe T, et al. Physical activity duration, intensity, and arterial stiffening in postmenopausal women. Am J Hypertens 2006 Oct; 19 (10): 1032-6 29. Troiano RP, Berrigan D, Dodd KW, et al. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc 2008 Jan; 40 (1): 181-8 30. Trost SG, McIver KL, Pate RR. Conducting accelerometerbased activity assessments in field-based research. Med Sci Sports Exerc 2005 Nov; 37 (11 Suppl.): 531S-43S 31. Tudor-Locke C, Burkett L, Reis JP, et al. How many days of pedometer monitoring predict weekly physical activity in adults? Prev Med 2005 Mar; 40 (3): 293-8 32. Tudor-Locke C, Bassett Jr DR, Rutherford WJ, et al. BMIreferenced cut points for pedometer-determined steps per day in adults. J Phys Act Health 2008 Jan; 5 (1 Suppl.): 126S-39S 33. Walsh JT, Charlesworth A, Andrews R, et al. Relation of daily activity levels in patients with chronic heart failure to long-term prognosis. Am J Cardiol 1997 May 15; 79 (10): 1364-9 34. Ward DS, Evenson KR, Vaughn A, et al. Accelerometer use in physical activity: best practices and research recommendations. Med Sci Sports Exerc 2005 Nov; 37 (11 Suppl.): 582S-8S

ª 2009 Adis Data Information BV. All rights reserved.

Aoyagi & Shephard

35. Westerterp KR. Pattern and intensity of physical activity. Nature 2001 Mar 29; 410 (6828): 539 36. Wyatt HR, Peters JC, Reed GW, et al. A Colorado statewide survey of walking and its relation to excessive weight. Med Sci Sports Exerc 2005 May; 37 (5): 724-30 37. Yoshioka M, Ayabe M, Yahiro T, et al. Long-period accelerometer monitoring shows the role of physical activity in overweight and obesity. Int J Obes 2005 May; 29 (5): 502-8 38. Togo F, Watanabe E, Park H, et al. Meteorology and the physical activity of the elderly: the Nakanojo Study. Int J Biometeorol 2005 Nov; 50 (2): 83-9 39. Togo F, Watanabe E, Park H, et al. How many days of pedometer use predict the annual activity of the elderly reliably? Med Sci Sports Exerc 2008 Jun; 40 (6): 1058-64 40. Yasunaga A, Togo F, Watanabe E, et al. Sex, age, season, and habitual physical activity of older Japanese: the Nakanojo Study. J Aging Phys Act 2008 Jan; 16 (1): 3-13 41. Aoyagi Y, Togo F, Matsuki S, et al. Walking velocity measured over 5 m as a basis of exercise prescription for the elderly: preliminary data from the Nakanojo Study. Eur J Appl Physiol 2004 Oct; 93 (1-2): 217-23 42. Park H, Togo F, Watanabe E, et al. Relationship of bone health to yearlong physical activity in older Japanese adults: cross-sectional data from the Nakanojo Study. Osteoporos Int 2007 Mar; 18 (3): 285-93 43. Park S, Park H, Togo F, et al. Yearlong physical activity and metabolic syndrome in older Japanese adults: crosssectional data from the Nakanojo Study. J Gerontol A Biol Sci Med Sci 2008 Oct; 63 (10): 1119-23 44. Yasunaga A, Togo F, Watanabe E, et al. Yearlong physical activity and health-related quality of life in older Japanese adults: the Nakanojo Study. J Aging Phys Act 2006 Jul; 14 (3): 288-301 45. Yasunaga A, Park H, Watanabe E, et al. Development and evaluation of the physical activity questionnaire for elderly Japanese: the Nakanojo Study. J Aging Phys Act 2007 Oct; 15 (4): 398-411 46. Yoshiuchi K, Nakahara R, Kumano H, et al. Yearlong physical activity and depressive symptoms in older Japanese adults: cross-sectional data from the Nakanojo Study. Am J Geriatr Psychiatry 2006 Jul; 14 (7): 621-4 47. Janz KF. Physical activity in epidemiology: moving from questionnaire to objective measurement. Br J Sports Med 2006 Mar; 40 (3): 191-2 48. Crouter SE, Schneider PL, Karabulut M, et al. Validity of 10 electronic pedometers for measuring steps, distance, and energy cost. Med Sci Sports Exerc 2003 Aug; 35 (8): 1455-60 49. Kumahara H, Schutz Y, Ayabe M, et al. The use of uniaxial accelerometry for the assessment of physicalactivity-related energy expenditure: a validation study against whole-body indirect calorimetry. Br J Nutr 2004 Feb; 91 (2): 235-43 50. Kumahara H, Ishii K, Tanaka H. Physical activity monitoring for health management: practical techniques and methodological issues. Int J Sport Health Sci 2006; 4: 380-93 51. Rafamantanantsoa HH, Ebine N, Yoshioka M, et al. Validation of three alternative methods to measure total

Sports Med 2009; 39 (6)

Steps Per Day for Senior Health?

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

energy expenditure against the doubly labeled water method for older Japanese men. J Nutr Sci Vitaminol 2002 Dec; 48 (6): 517-23 Schneider PL, Crouter SE, Lukajic O, et al. Accuracy and reliability of 10 pedometers for measuring steps over a 400-m walk. Med Sci Sports Exerc 2003 Oct; 35 (10): 1779-84 Schneider PL, Crouter SE, Bassett DR. Pedometer measures of free-living physical activity: comparison of 13 models. Med Sci Sports Exerc 2004 Feb; 36 (2): 331-5 Warms C. Physical activity measurement in persons with chronic and disabling conditions: methods, strategies, and issues. Fam Community Health 2006 Jan-Mar; 29 (1 Suppl.): 78S-88S Le Masurier GC, Tudor-Locke C. Comparison of pedometer and accelerometer accuracy under controlled conditions. Med Sci Sports Exerc 2003 May; 35 (5): 867-71 Le Masurier GC, Lee SM, Tudor-Locke C. Motion sensor accuracy under controlled and free-living conditions. Med Sci Sports Exerc 2004 May; 36 (5): 905-10 McClain JJ, Sisson SB, Washington TL, et al. Comparison of Kenz Lifecorder EX and ActiGraph accelerometers in 10-year-old children. Med Sci Sports Exerc 2007 Apr; 39 (4): 630-8 Tudor-Locke C, Ainsworth BE, Thompson RW, et al. Comparison of pedometer and accelerometer measures of free-living physical activity. Med Sci Sports Exerc 2002 Dec; 34 (12): 2045-51 Bouchard C, Shephard RJ. The model and key concepts. In: Bouchard C, Shephard RJ, Stephens T, editors. Physical activity, fitness and health. Champaign (IL): Human Kinetics, 1994: 77-97 Tudor-Locke C, Sisson SB, Collova T, et al. Pedometerdetermined step count guidelines for classifying walking intensity in a young ostensibly healthy population. Can J Appl Physiol 2005 Dec; 30 (6): 666-76 Tudor-Locke C, Bassett Jr DR. How many steps/day are enough? Preliminary pedometer indices for public health. Sports Med 2004 Jan; 34 (1): 1-8 Hill JO, Wyatt HR, Reed GW, et al. Obesity and the environment: where do we go from here? Science 2003 Feb 7; 299 (5608): 853-5 Bravata DM, Smith-Spangler C, Sundaram V, et al. Using pedometers to increase physical activity and improve health: a systematic review. JAMA 2007 Nov 21; 298 (19): 2296-304 Caspersen CJ, Pereira MA, Curran KM. Changes in physical activity patterns in the United States, by sex and cross-sectional age. Med Sci Sports Exerc 2000 Sep; 32 (9): 1601-9 Martinez-Gonzalez MA, Varo JJ, Santos JL, et al. Prevalence of physical activity during leisure time in the European Union. Med Sci Sports Exerc 2001 Jul; 33 (7): 1142-6 Weiss DR, O’Loughlin JL, Platt RW, et al. Five-year predictors of physical activity decline among adults in lowincome communities: a prospective study. Int J Behav Nutr Phys Act 2007 Jan 18; 4: 2 Japan Ministry of Health, Labour and Welfare. The national nutrition survey in Japan 2001 [in Japanese]. Tokyo: Daiichishuppan, 2003

ª 2009 Adis Data Information BV. All rights reserved.

437

68. Fukuhara S, Suzukamo Y. Manual of SF-36v2 Japanese version [in Japanese]. Kyoto: Institute for Health Outcomes & Process Evaluation Research, 2004 69. Ware JE, Sherbourne CD. The MOS 36-item Short-Form Health Survey (SF-36): I, conceptual framework and item selection. Med Care 1992 Jun; 30 (6): 473-83 70. National Institutes of Health Osteoporosis and Related Bone Diseases, National Research Center. Bone health and osteoporosis: a guide for Asian women aged 50 and older. Bethesda (MD): National Institutes of Health, 2005 71. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001 May 16; 285 (19): 2486-97 72. World Health Organization Western Pacific Region, International Association for the Study of Obesity, International Obesity Task Force. The Asia-Pacific perspective: redefining obesity and its treatment. Sydney (NSW): Health Communications Australia Pty Limited, 2000 73. Shephard RJ. Whistler 2001: a Health Canada/CDC conference on ‘‘Communicating physical activity and health messages: science into practice’’. Am J Prev Med 2002 Oct; 23 (3): 221-5 74. Erlichman J, Kerbey AL, James WP. Physical activity and its impact on health outcomes: paper 2, prevention of unhealthy weight gain and obesity by physical activity – an analysis of the evidence. Obes Rev 2002 Nov; 3 (4): 273-87 75. Blaum CS, West NA, Haan MN. Is the metabolic syndrome, with or without diabetes, associated with progressive disability in older Mexican Americans? J Gerontol A Biol Sci Med Sci 2007 Jul; 62 (7): 766-73 76. Iwane M, Arita M, Tomimoto S, et al. Walking 10 000 steps/day or more reduces blood pressure and sympathetic nerve activity in mild essential hypertension. Hypertens Res 2000 Nov; 23 (6): 573-80 77. Moreau KL, Degarmo R, Langley J, et al. Increasing daily walking lowers blood pressure in postmenopausal women. Med Sci Sports Exerc 2001 Nov; 33 (11): 1825-31 78. Tudor-Locke CE, Myers AM, Bell RC, et al. Preliminary outcome evaluation of the First Step Program: a daily physical activity intervention for individuals with type 2 diabetes. Patient Educ Couns 2002 May; 47 (1): 23-8 79. Yamanouchi K, Shinozaki T, Chikada K, et al. Daily walking combined with diet therapy is a useful means for obese NIDDM patients not only to reduce body weight but also to improve insulin sensitivity. Diabetes Care 1995 Jun; 18 (6): 775-8 80. McArdle W, Katch F, Katch V. Exercise physiology: energy, nutrition, and human performance. 3rd rev. ed. Philadelphia (PA): Lea & Febiger, 1991 81. Aoyagi Y, Katsuta S. Relationship between the starting age of training and physical fitness in old age. Can J Sport Sci 1990 Mar; 15 (1): 65-71 82. Aoyagi Y, Katsuta S. The starting age of training and its effect on reduction in physical performance capability with aging. In: Kaneko M, editor. Fitness for the aged, disabled, and industrial worker. International series on

Sports Med 2009; 39 (6)

Aoyagi & Shephard

438

83.

84.

85.

86.

87.

88.

89.

90. 91.

92. 93.

94.

95.

96.

sports sciences, 20. Champaign (IL): Human Kinetics, 1990: 118-24 Chan CB, Ryan DA, Tudor-Locke C. Relationship between objective measures of physical activity and weather: a longitudinal study. Int J Behav Nutr Phys Act 2006 Aug 7; 3: 21 Tudor-Locke C, Bassett DR, Swartz AM, et al. A preliminary study of one year of pedometer self-monitoring. Ann Behav Med 2004 Dec; 28 (3): 158-62 Bergstralh EJ, Sinaki M, Offord KP, et al. Effect of season on physical activity score, back extensor muscle strength, and lumbar bone mineral density. J Bone Miner Res 1990 Apr; 5 (4): 371-7 Centers for Disease Control and Prevention. Monthly estimates of leisure-time physical inactivity – United States, 1994. MMWR Morb Mortal Wkly Rep 1997 May 9; 46 (18): 393-7 Haggarty P, McNeill G, Manneh MK, et al. The influence of exercise on the energy requirements of adult males in the UK. Br J Nutr 1994 Dec; 72 (6): 799-813 Matthews CE, Freedson PS, Hebert JR, et al. Seasonal variation in household, occupational, and leisure time physical activity: longitudinal analyses from the seasonal variation of blood cholesterol study. Am J Epidemiol 2001 Jan 15; 153 (2): 172-83 Pivarnik JM, Reeves MJ, Rafferty AP. Seasonal variation in adult leisure-time physical activity. Med Sci Sports Exerc 2003 Jun; 35 (6): 1004-8 Uitenbroek DG. Seasonal variation in leisure time physical activity. Med Sci Sports Exerc 1993 Jun; 25 (6): 755-60 US Department of Health and Human Services. Patterns and trends in physical activity. In: Physical activity and health: a report of the Surgeon General. Atlanta (GA): US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, 1996: 173-208 Lurie SJ, Gawinski B, Pierce D, et al. Seasonal affective disorder. Am Fam Physician 2006 Nov 1; 74 (9): 1521-4 Trost SG, Owen N, Bauman AE, et al. Correlates of adults’ participation in physical activity: review and update. Med Sci Sports Exerc 2002 Dec; 34 (12): 1996-2001 Czeisler CA, Dumont M, Duffy JF, et al. Association of sleep-wake habits in older people with changes in output of circadian pacemaker. Lancet 1992 Oct 17; 340 (8825): 933-6 Duffy JF, Dijk DJ, Klerman EB, et al. Later endogenous circadian temperature nadir relative to an earlier wake time in older people. Am J Physiol 1998 Nov; 275 (5 Pt 2): R1478-87 Togo F, Aizawa S, Arai J, et al. Influence on human sleep patterns of lowering and delaying the minimum core body temperature by slow changes in the thermal environment. Sleep 2007 Jun 1; 30 (6): 797-802

ª 2009 Adis Data Information BV. All rights reserved.

97. Aoyagi Y. Endurance training, heat acclimation, and protective clothing: the thermophysiology of exercising in a hot climate [dissertation]. Toronto (ON): University of Toronto, 1996 98. Aoyagi Y, McLellan TM, Shephard RJ. Interactions of physical training and heat acclimation: the thermophysiology of exercising in a hot climate. Sports Med 1997 Mar; 23 (3): 173-210 99. Aoyagi Y, McLellan TM, Shephard RJ. Effects of training and acclimation on heat tolerance in exercising men wearing protective clothing. Eur J Appl Physiol Occup Physiol 1994 Mar; 68 (3): 234-45 100. Aoyagi Y, McLellan TM, Shephard RJ. Effects of 6 versus 12 days of heat acclimation on heat tolerance in lightly exercising men wearing protective clothing. Eur J Appl Physiol Occup Physiol 1995 Mar; 71 (2-3): 187-96 101. Aoyagi Y, McLellan TM, Shephard RJ. Determination of body heat storage in clothing: calorimetry versus thermometry. Eur J Appl Physiol Occup Physiol 1995 Mar; 71 (2-3): 197-206 102. Aoyagi Y, McLellan TM, Shephard RJ. Determination of body heat storage: how to select the weighting of rectal and skin temperatures for clothed subjects. Int Arch Occup Environ Health 1996 Jun; 68 (5): 325-36 103. Aoyagi Y, McLellan TM, Shephard RJ. Residual analysis in the determination of factors affecting the estimates of body heat storage in clothed subjects. Eur J Appl Physiol Occup Physiol 1996 May; 73 (3-4): 287-98 104. Aoyagi Y, McLellan TM, Shephard RJ. Effects of endurance training and heat acclimation on psychological strain in exercising men wearing protective clothing. Ergonomics 1998 Mar; 41 (3): 328-57 105. McLellan TM, Aoyagi Y. Heat strain in protective clothing following hot-wet or hot-dry heat acclimation. Can J Appl Physiol 1996 Apr; 21 (2): 90-108 106. Gordon CJ. Relationship between preferred ambient temperature and autonomic thermoregulatory function in rat. Am J Physiol 1987 Jun; 252 (6 Pt 2): R1130-7 107. Bruce DG, Devine A, Prince RL. Recreational physical activity levels in healthy older women: the importance of fear of falling. J Am Geriatr Soc 2002 Jan; 50 (1): 84-9 108. Pell JP, Cobbe SM. Seasonal variations in coronary heart disease. QJM 1999 Dec; 92 (12): 689-96 109. Qiu D, Tanihata T, Aoyama H, et al. Relationship between a high mortality rate and extreme heat during the summer of 1999 in Hokkaido Prefecture, Japan. J Epidemiol 2002 May; 12 (3): 254-7

Correspondence: Dr Yukitoshi Aoyagi, Exercise Sciences Research Group, Tokyo Metropolitan Institute of Gerontology, 35-2 Sakaecho, Itabashi-ku, Tokyo 173-0015, Japan. E-mail: [email protected]

Sports Med 2009; 39 (6)

Sports Med 2009; 39 (6): 439-468 0112-1642/09/0006-0439/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Exercise and Bone Mass in Adults Amelia Guadalupe-Grau, Teresa Fuentes, Borja Guerra and Jose A.L. Calbet Department of Physical Education, University of Las Palmas de Gran Canaria, Las Palmas de Gran Canaria, Canary Islands, Spain

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Experiments with Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Studies with Humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Cross-Sectional Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Young Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 Young Women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Longitudinal Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Young Women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Young Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.3 Premenopausal Women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.4 Middle-Aged Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Postmenopausal Women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Older Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Practical Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

439 442 443 443 443 447 447 448 452 453 454 454 455 460 461 462

There is a substantial body of evidence indicating that exercise prior to the pubertal growth spurt stimulates bone growth and skeletal muscle hypertrophy to a greater degree than observed during growth in non-physically active children. Bone mass can be increased by some exercise programmes in adults and the elderly, and attenuate the losses in bone mass associated with aging. This review provides an overview of cross-sectional and longitudinal studies performed to date involving training and bone measurements. Crosssectional studies show in general that exercise modalities requiring high forces and/or generating high impacts have the greatest osteogenic potential. Several training methods have been used to improve bone mineral density (BMD) and content in prospective studies. Not all exercise modalities have shown positive effects on bone mass. For example, unloaded exercise such as swimming has no impact on bone mass, while walking or running has limited positive effects. It is not clear which training method is superior for bone stimulation in adults, although scientific evidence points to a combination of high-impact (i.e. jumping) and weight-lifting exercises. Exercise involving high impacts, even a relatively small amount, appears to be the most efficient for enhancing bone mass, except in postmenopausal women. Several types of resistance exercise have been tested also with positive results, especially when the

Guadalupe-Grau et al.

440

intensity of the exercise is high and the speed of movement elevated. A handful of other studies have reported little or no effect on bone density. However, these results may be partially attributable to the study design, intensity and duration of the exercise protocol, and the bone density measurement techniques used. Studies performed in older adults show only mild increases, maintenance or just attenuation of BMD losses in postmenopausal women, but net changes in BMD relative to control subjects who are losing bone mass are beneficial in decreasing fracture risk. Older men have been less studied than women, and although it seems that men may respond better than their female counterparts, the experimental evidence for a dimorphism based on sex in the osteogenic response to exercise in the elderly is weak. A randomized longitudinal study of the effects of exercise on bone mass in elderly men and women is still lacking. It remains to be determined if elderly females need a different exercise protocol compared with men of similar age. Impact and resistance exercise should be advocated for the prevention of osteoporosis. For those with osteoporosis, weight-bearing exercise in general, and resistance exercise in particular, as tolerated, along with exercise targeted to improve balance, mobility and posture, should be recommended to reduce the likelihood of falling and its associated morbidity and mortality. Additional randomized controlled trials are needed to determine the most efficient training loads depending on age, sex, current bone mass and training history for improvement of bone mass.

The most important function of bone tissue is to withstand and transmit forces without breaking. The strength of bone depends on the amount of tissue, its material composition and how bone material is organized microarchitecturally and geometrically (shape and size).[1,2] As summarized by Seeman and Delmas,[3] optimal bone tissue characteristics are defined by optimal levels of stiffness, flexibility and lightness. To efficiently withstand and transmit loads, bone must be stiff and able to resist deformation. However, it cannot be too stiff – i.e. unable to absorb some energy by shortening and widening when compressed, and by lengthening and narrowing when submitted to traction – otherwise the energy imposed during loading will be released by structural failure. Conversely, bone cannot be too flexible, because on loading it could easily deform beyond its peak strain, and fracture.[3] Bone must also have the ability to continually adapt to changes in physiological and mechanical environment. The mechanical properties of bone are determined by two major factors: the characteristics ª 2009 Adis Data Information BV. All rights reserved.

of the collagen matrix and the degree of mineralization, i.e. the amount of calcium hydroxyapatite crystals deposited on and between the collagen fibres. Bone strength is primarily determined by tissue mass and stiffness. While stiffness is mainly determined by the mineral phase,[4-7] the collagen matrix contributes primarily to bone toughness resilience (i.e. the ability to absorb energy without breaking).[8-10] Increasing bone mineral density (BMD) results in greater stiffness but lower flexibility.[11] Collagen, of which about 95% is type I collagen, comprises about 80% of the total protein in bone.[12] Collagen fibres are packed together by the formation of inter- and intramolecular crosslinks. Mature crosslinks such as pyridinoline (PYD) and deoxypyridinoline (DPD) reach a maximum concentration between 15 and 40 years of age, and their concentrations are lower in trabecular bone than in cortical bone.[13] If there are too many crosslinks, the ability to absorb energy diminishes, i.e. the bone becomes more brittle. Likewise, without the collagen matrix the bone becomes less elastic and more brittle.[14] In Sports Med 2009; 39 (6)

Exercise and Bone Mass in Adults

humans it has been shown that the compressive biomechanical ultimate strength of bone is correlated, independently of BMD, with the ratio PYD/DPD, but not with PYD, DPD or pyrrole separately.[15] Non-fibrillar organic matrix acts as the ‘glue’ that holds the mineralized fibrils together.[16] Bone strength also depends on the orientation of osteons (and thus collagen fibres) within the cortical bone.[17] Longitudinal fibres are found in regions supporting tensile loads, while transverse fibres predominate in regions under compressive loading.[2,18] Part of the bone plasticity in response to loading depends on its capacity to reorient its collagen fibres. For example, it has been reported in dogs that a 10% reduction in vertebral BMD elicited by a strenuous progressive running programme (up to 40 km/day) for 1 year did not change the bone mechanical properties.[19] These dogs, compared with their sedentary counterparts, showed reorganization of the collagen fibres in a more parallel manner without changes in the concentration of crosslinks, suggesting that collagen reorganization during exercise may contribute to the maintenance of bone strength despite decreased mineral density.[19] Bone mechanical properties are modified depending on loading, such that bone strength is enhanced or reduced in response to either increased or reduced mechanical loading.[3,20-22] The adaptive response is very complex and depends on the characteristics of loading history, but also on systemic and local factors, which include neuroendocrine, endocrine and paracrine changes in metabolites, cytokines, growth factors, hormones, vitamins and minerals.[23-27] Excellent reviews have been published recently on the molecular mechanisms that mediate adaptive responses of bone tissue to changes in loading, and the interested reader is referred to them.[24-26,28,29] The main signals for bone adaptation to mechanical loading are the rate and magnitude of strain, which should reach minimal levels or threshold to elicit structural modifications in bone.[30-33] To enhance bone mass or BMD in non-physically active humans, bone tissue must be submitted to mechanical strains above those experienced by daily living activities.[31,34] ª 2009 Adis Data Information BV. All rights reserved.

441

Although experimental evidence indicates that mechanical loads must be great to augment bone mass, to induce bone strains sufficient to cause microdamage and stimulate bone formation through the repair of damaged tissue,[35,36] the intensity of loading is not the only stimulus for bone accretion, as demonstrated by Rubin et al.[37,38] These authors demonstrated that high frequency vibration (20–50 Hz) of very low magnitude (<10 microstrain), continually present during even subtle activities such as standing, increases trabecular bone mass in weight-bearing regions of the skeleton in animals.[37-39] Mechanical loading triggers a cascade of cellular events that involve estrogen receptoralpha (ERa).[26] This may be the reason why the osteogenic effect of loading is greater when the estrogen receptor number is high, as during adolescence, and less when the estrogen receptor number is low, as occurs postmenopausally, during amenorrhoea, or after ovariectomy.[26] Signals from calcium channels, G-proteins, integrins and the cytoskeleton elicited by mechanical loading are conveyed in the activation of key intracellular enzymes leading to release of nitric oxide, prostaglandin E2, transforming growth factor b, insulin-like growth factor (IGF)-I or IGF-II, ultimately leading to bone formation.[28,29] Osteoporosis is a reduction in BMD 2.5 standard deviations below the mean for healthy young women at the age of attainment of peak bone mass (expressed as a T-score), in general using a reference population matched for age, sex and race (expressed as a Z-score).[40] Loss of BMD contributes to loss of mechanical strength and to bone fragility, and thus to predisposition for bone fractures, which may occur even under low loading conditions, as reviewed elsewhere.[40-43] This condition is a considerable worldwide concern and a cause of high healthcare costs.[44] Although some risk factors for osteoporosis, like genetics (sex, age, body size and ethnicity), cannot be modified, it is possible to change variables like lifestyle and physical activity to stimulate greater accumulation of peak bone mass.[45] Sports participation during childhood and adolescence,[46-50] especially before the pubertal growth spurt,[51-58] promotes bone mass accumulation, Sports Med 2009; 39 (6)

442

i.e. gain in total bone mineral content (BMC), and geometrical changes in bone size and shape leading to a higher bone mass and stronger bones in adult life.[40,56,59,60] In fact, epidemiological studies indicate that bone size is related to fracture risk when examined in relation to body size in children,[61] and participation in sports prior to puberty promotes bone hypertrophy, i.e. physically active pre-pubertal children appear to develop bone of greater size than their sedentary peers, although this effect is confined to the loaded regions.[59,62] Although well documented reviews have been published in postmenopausal women,[3,40,42,63] less is known about the effects that exercise programmes and sports participation may have on bone mass in young adult women at premenopausal age.[64] Thus, this article focuses on the influence of physical activity on BMC and BMD in premenopausal and postmenopausal women and in men. In addition, we review relevant studies in animals and humans, highlighting variables like mode of exercise, intensity, duration, endocrine and metabolic factors, and sex differences in the osteogenic response to training. A MEDLINE database search was conducted (from 1969 to June 2008, with a special focus on the latest publications), using the following systematic search terms: ‘bone’, ‘bone mineral density (BMD)’, ‘bone mineral content (BMC)’, ‘dual x-ray absorptiometry (DXA)’, ‘bone mass accrual’, combined with ‘exercise’, ‘sports participation’, ‘adults’. The abstract of studies resulting from this search were examined according to the following criteria: Inclusion criteria:  sports participation and bone mass measurements in adults;  effects of training on bone mass in adults;  effects of vibration on bone mass in adults;  sex differences in the osteogenic response to training; Exclusion criteria:  prepubertal subjects;  studies using single photon absorptiometry to measure bone mass gains/losses. The following information was extracted from each study: skeletal regions measured; experiª 2009 Adis Data Information BV. All rights reserved.

Guadalupe-Grau et al.

mental training protocol; training parameters including intensity, frequency and duration; and objective outcomes. This information was tabulated according to the experimental training protocol used and study design (cross-sectional or longitudinal studies). Studies that used more than one type of training protocol were included. The results of these studies were extracted and are summarized according to sex, study design and different life-stages: (i) young women and men; (ii) premenopausal women and middle-aged men; and (iii) older women and men. 1. Experiments with Animals Using animal in vivo models – in which mechanical loads have been specifically applied to the rat tibia,[65] rat tail,[66] rat ulna[67] and avian bone[37] – it has been shown that the effects of mechanical loading are dependent upon the magnitude, duration and frequency of the mechanical stimulus applied.[68] However, exercise not only consists of generation of mechanical loads, it also perturbs acid-base balance, stimulates sympathetic activity and influences production of several hormones, cytokines and adipokynes with known effects on bone metabolism.[69-73] Thus, information provided by in vivo mechanical models should be combined with the information gained with exercise models, bearing in mind that the response to exercise may differ between animal species and that it is affected by other factors such as age, sex, nutrition and genetics. The use of animal models to study bone adaptation to exercise is based on the fact that similar mechanisms control bone formation and resorption in animals and humans.[74] Experiments with rats have shown that running has osteogenic effects on loaded bones of young male[75] and female[76,77] rats, as well as in ovariectomized[78] and orchidectomized rats,[79] although the osteogenic effect of exercise appears to be less efficient in female ovariectomized rats.[78] The forces generated during running play a role in the osteogenic response, since the rats that run with a loaded backpack on top of their back show a greater gain in bone mass than the rats running without extra load.[80] In contrast, running Sports Med 2009; 39 (6)

Exercise and Bone Mass in Adults


at 80% of maximal oxygen load (VO2max) reduced longitudinal bone growth and induced bone loss, mainly due to decreased osteoblastic activity, in 5-week-old male rats[81] (table I). From these studies, it can be concluded that treadmill running may be useful to increase bone mass in young and adult rats of both sexes, especially in appendicular long bones at weightbearing sites; however, the increase in lumbar bone mass is absent or only detectable when longterm exercise is applied.[82] However, these studies should be interpreted cautiously, since rats do not stop growing throughout their lifespan. Also, in contrast to what is observed in humans, longitudinal bone growth in adult rats increases after ovariectomy, and estrogen replacement inhibits this growth.[83] In postpubertal female rats, bone is less responsive to loading than in ovariectomized rats or male rats of similar age.[84] Jarvinen et al.[84] have raised the question about the efficiency of bone loading during the estrogen-replete period in women, i.e. between puberty and menopause. Cross-sectional and observational longitudinal studies do indeed show a higher responsiveness to loading in human female bone when regular exercise starts before puberty than in adult life.[59,60,85] 2. Studies with Humans 2.1 Cross-Sectional Studies

In this section, recent cross-sectional studies measuring bone mass and/or BMD in recreational or professional athletes and sedentary adult women and men are reviewed, focusing on the osteogenic response to training as well as sex differences in the osteogenic response to training (table II). This issue has been previously reviewed by others,[22,94-97] and only some representative studies are commented on in this section. Inference from animal studies to humans[63] implies that strength training, instead of endurance training programmes such as running, should result in the greatest increase in skeletal density. Focusing our attention on the early stage of life, girls and adolescents who spend more time ª 2009 Adis Data Information BV. All rights reserved.

443

training than is proposed by physical activity curricula (<2 h/wk in most European countries) achieve better peak bone mass, especially if they start before the pubertal growth spurt.[57,58] Cross-sectional studies including sports participation and dual-energy x-ray absorptiometry (DXA) measurement at prepubertal ages show different results depending on the intensity and rate of the strains elicited. High strain-eliciting sports like gymnastics, which is thought to generate ground reaction forces close to ten times body mass in prepubertal children,[98] report 5.7% higher mean upper extremities BMD compared with sedentary subjects in 9-year-old girls.[49] Weight-bearing physical activities also improve bone mass in prepubertal subjects. In 9-year-old boys, soccer (football) practice (3 h/wk) has been associated with increased BMD compared with non-physically active boys at the lower limbs ( › 4%), lumbar spine ( › 2%), and femoral neck ( › 5%).[47,51,92] Similarly, premenarchial girls enrolled in handball practice showed enhanced BMC and BMD at the lumbar spine, pelvic region and lower extremities; they also showed greater BMC in the whole body and enhanced BMD in the right upper extremity and femoral neck than the control subjects.[48] In these sports, the potential osteogenic forces acting on the appendicular and axial bones are elicited by the high ground reaction forces evoked during jumping, kicking, sprinting, rapidly changing direction, starting, stopping, throwing, fall landings and blocks during defensive actions. Sports like swimming or rowing, without the action of gravitational forces, are beneficial for physical fitness but do not appear to have osteogenic benefits.[87,99] Swimmers, like astronauts,[100] operate in a low-gravity environment, with minimal impact on bone structures, and only relatively low tensions are transmitted by the muscular system to the bones during this type of exercise. This could explain why exercising in water early in life and regularly during the lifespan might abolish the osteogenic effect of sport. 2.1.1 Young Men

Studies performed in our laboratory have reported that adult (19–27 years old) amateur male Sports Med 2009; 39 (6)

Study

Subjects n

M/F

age (wk)

444

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Bone adaptation to treadmill training in rats Frequency

Training intensity

Protocol time

Other

Bone measurement site (s)

results

7% grade treadmill Ovariectomized rats

DXA femur, tibia, L4

Prevention of bone loss

OXE = 20 OXS = 20

F

36

4/wk 40 min/day

21 m/min

3 mo

Hagihara et al.[77] (2005)

ER = 5 ER2 = 5 ER3 = 5 C=5

F

8

4–7/wk 30 min/day

15 m/min

8 wk

QTC-trabecular L2, tibia, femur

› L2, tibia, femur ER, vs C

Bourrin et al.[75] (1995)

ER = 20

M

9

7/wk


60% VO2max

5 wk

HPC-tibia

› 27% volume › 8% trabeculae

Bourrin et al.[81] (1994)

ER = 20

M

5

7/wk


80% VO2max

11 wk

HPC-tibia, L2, T2

fl volume fl trabeculae

Horcajada et al.[79] (1997)

SHE = 20 CXE = 20 C = 20

M

6

7/wk 15–60 min/day

15–60 m/min
60% VO2max

3 mo

Orchidectomized rats

DXA-femur

› femur BMD in CXE, SHE vs C

Yeh et al.[76] (1993)

ER = 28 DR = 24 C = 30

F

6

5/wk

20 m/min

6 wk

Denervated rats

HPC-tibia

› cortical bone area in ER vs DR, C

Van der Wiel et al.[80] (1995)

ER1 = 10 ER2 = 10 ER3 = 10 C = 14

F

20

5/wk

20 m/min

17 wk

ER1: no extra load 30 min ER2: 5030 g load 30 min ER3: 50 g load 15 min

DXA-WB, LL

› 16% LL BMD ER1 vs C › 15% LL BMD ER2 vs C › 15% WB, 20% LL BMD ER3 vs C

BMD = bone mineral density; C = controls; CXE = castrated exercising rats; DR = sciatic denervated rats; DXA = dual x-ray absorptiometry; ER = exercising rats; HPC = histomorphometric analysis; L2 = second lumbar vertebra; L4 = fourth lumbar vertebra; LL = lower limbs; M/F = male/female; n = number of subjects; OXE = ovariectomized exercising rats; . OXS = ovariectomized sedentary rats; QTC = quantitative CT; SHE = sham-operated exercising rats; T2 = second thoracic vertebra; VO2max = maximal oxygen uptake; WB = whole body; › indicates significant increase p < 0.05; fl indicates significant decrease p < 0.05.

Guadalupe-Grau et al.

Sports Med 2009; 39 (6)

Barengolts et al.[78] (1993)

Study

Subjects n

Bone measurement site (s) DXA

results BMD

Soccer

WB, SP, FN, IT, GT, WT, WL

› › › › › ›

10% SP vs C 21% FN vs C 19% IT vs C 21% GT vs C 27% WT vs C 10% WL vs C

Running Rugby Fighting Body-building Swimming

WB, A, WL, SP, S

› › › fl › › › › fl

10% WB RGB vs OS 2% WB FS vs OS 2% WB BB vs OS 8% WB SW vs OS 5% A RGB vs OS 4% A FS vs OS 9% WL RGB vs OS 1% WL FS vs OS 5% WL BB vs OS

Sport M/F

age (y)

sports history; mean time (y)

training volume (mean time) 7 h/wk

BMC

Young men Calbet et al.[86] (2001)

EX = 33 C = 19

M

19–27

EX = 12

Morel et al.[87] (2001)

REX = 126 RGB = 110 FS = 44 BB = 28 SW = 14

M

25–40

REX = 22 RGB = 15 FS = 18 BB = 16 SW = 11

Calbet et al.[88] (1998)

EX = 9 C = 13

M

21–32

EX = 17

25 h/wk

Tennis

WB, A, SP, FN, WT, WL

› 15% SP EX vs C › 10–15% FN EX vs C

› 5% DAP vs NDAP

Wittich et al.[89] (1998)

EX = 24 C = 22

M

20–24

EX = 8

20 h/wk

Soccer

WB, WL, PR

› 11% WB vs C › 14% PR vs C › 14% WL vs C

› 15% WB vs C › 25% PR vs C › 20% WL vs C

Egan et al.[90] (2006)

REX = 11 RGB = 30 NB = 20 C = 25

F

19–23

REX = 9 RGB = 4 NB = 4

8.4 h/wk 4.1 h/wk 3.7 h/wk

Running Rugby Netball

WB, SP, LPF

› › › ›

Nichols et al.[91] (2007)

HOAEX = 21 F HEX = 72 ROAEX = 17 RNEX = 51

14–16

HOAEX; HEX = 6.5 ROAEX; RNEX = 6.1

8.6 h/wk 8.5 h/wk

WB, SP, WH, Soccer, FN, GT volleyball, softball, tennis, lacrosse, running, swimming

8.1 h/wk 8.7 h/wk 9.1 h/wk 8.1 h/wk 8.7 h/wk

› › › › › › ›

13% WB vs C 13% SP vs C 24% FN vs C 18% IT vs C 23% GT vs C 24% WT vs C 16% WL vs C

Exercise and Bone Mass in Adults

ª 2009 Adis Data Information BV. All rights reserved.

Table II. Effects of sports training on bone tissue adaptations in young adult men and women: cross-sectional studies

Young women WB all sports vs C 13.5% RGB WB vs C 16.5% SP RGB vs C 21.7% FN RGB vs C

Continued next page

445

Sports Med 2009; 39 (6)

› 4% HEX WH vs HOAEX › 7% HEX GT vs HOAEX › 10% HEX LS vs ROAEX › 5% HEX LS vs ROAEX

Guadalupe-Grau et al.

A = arms; BB = body-builders; BMC = bone mineral content; BMD = bone mineral density; C = control subjects; DAP = dominant arm tennis players; DXA = dual x-ray absorptiometry; EX = exercising subjects; FD = femur diaphysis; FN = femoral neck; FS = fighting sports subjects; GT = greater trochanter; H = humerus; HEX = high/odd impact eumenorroeic athletes (soccer, volleyball, softball, tennis and lacrosse); HOAEX = high/odd impact oligo/amenorroeic athletes (soccer, volleyball, softball, tennis and lacrosse); IT = intertrochanteric subregion; LPF = left proximal femur; MDR = mid-distal radius region; M/F = male/female; n = number of subjects; NB = netball players; NDAP = non-dominant arm tennis players; OS = other sports subjects; PR = pelvic region; PT = proximal tibia; R = radius; REX = running exercise subjects; RGB = rugby players; RNEX = repetitive/non-impact eumenorroeic athletes (runners, swimmers); ROAEX = repetitive/non-impact oligo/amenorroeic athletes (runners, swimmers); S = skull; SP = lumbar spine; SW = swimmers; T = tibia; TDR = thirddistal radius region; UDR = ultradistal radius region; WB = whole body; WH = whole hip; WL = whole leg; WT = Ward’s triangle; › indicates significant increase p < 0.05; fl indicates significant decrease p < 0.05.

› 4.8% TDR DAP vs NDAP R (UDR, MDR, TDR) Tennis EX = 47 C = 58 Ducher et al.[93] (2004)

M/F

20–25

EX = 14

3 h/wk

10.7% SP vs C 13.7% FN vs C 19.6% WT vs C 12.6% FD vs C 12% PT vs C › › › › › WB, SP, S, FN, WT, T, H, FD, PT 6 h/wk 18–27 EX = 16 C = 13 Alfredson et al.[92] (1996)

F

Subjects n Study

Table II. Contd

M/F

age (y)

sports history; mean time (y)

training volume (mean time)

Sport

Soccer

Bone measurement site (s) DXA

results BMD

BMC

› 15.6% MDR DAP vs NDAP › 13.3% TDR DAP vs NDAP

446

ª 2009 Adis Data Information BV. All rights reserved.

soccer players with a long training history (mean 12 years) have increased BMC and BMD at the lumbar spine (13% and 10% respectively), femoral neck (24% and 21%) and lower limbs (16% and 10%) compared with age-, height- and weight-matched sedentary controls of the same Caucasian population.[86] Similar results were reported by Wittich et al.[89] in 20- to 22-year-old soccer players. These adaptations are likely elicited by the ground reaction forces generated during jumping and sprinting with sudden changes in the direction of movement, combined with the high strains elicited during kicking.[101,102] With relatively low volumes of exercise (2–3 hours) per week it is possible to elicit increases in BMC and BMD in the loaded bones of prepubertal tennis players (Sanchis et al., unpublished observations). The magnitude of the local adaptation elicited by tennis participation is further enhanced for training volumes above 7 h/wk (Sanchis et al., unpublished observations). A direct comparison of prepubertal soccer players[51] or tennis players with professional adult players[86,88] indicates that part of the bone mass gained through sports participation is achieved before puberty. In contrast, compared with sedentary peers, a 20% lower lumbar spine BMC has been reported in 19- to 56-year-old longdistance runners performing a training volume close to 100 km/week.[103] Morel et al.[87] examined the influence of regular exercise (mean 9 h/wk) on BMD in 403 non-professional male subjects aged 30 years involved in different sports. This is an ideal age to compare BMD measurements because peak bone mass has been already reached and bone loss is still insignificant. Soccer, basketball, volleyball, gymnastics, weight-lifting and ice hockey were associated with a higher whole body BMD (WBBMD) whereas rowers and especially swimmers had a WBBMD similar to that of a sedentary group. Regional BMD comparison suggested that there may be site-specific responses due to the specific types of mechanical loading exerted through physical activity, because sportsmen involved in impact-loading sports (i.e. fighting sports) had a higher leg BMD than those in active loading sports (i.e. body-building) [see table II]. Sports Med 2009; 39 (6)

Exercise and Bone Mass in Adults

In the upper limbs the positive effect of mechanical loading via musculotendinous attachments is demonstrated in racquet sports such as tennis[88,104] and squash.[105] The asymmetrical nature of racket sports offers an interesting model to study the adaptability of both the skeletal and soft tissue of the upper limbs to physical stress, using the non-dominant arm as a control. Based on a side-to-side comparison, these studies enabled elimination of the confounding effects of genetic, hormonal and nutritional factors that are encountered in cross-sectional studies.[104] Elite young tennis players (mean age 26 years), with a high training load of 25 h/wk, show enhanced BMC (20%) and BMD (6%) in the dominant arm compared with the contralateral arm. These tennis players also had increased (10–15%) femoral neck and lumbar spine BMD compared with controls.[88] Similar effects have been described in the mid- and third-distal radius in young adult male and female recreational tennis players.[106] 2.1.2 Young Women

Second division female soccer players (aged 18–24 years) show bone adaptations similar to those reported in males,[92] with 11%, 15% and 20% higher BMD at lumbar spine, femoral neck and Ward’s triangle, and 8–13% higher BMD at the non-dominant femur and humerus, distal femur and proximal tibia compared with non-physically active women (table II). Egan et al.[90] compared BMD and body composition among young female athletes (rugby players, distance runners and netball players) with a mean age of 21 years versus sedentary control subjects. All sports groups had higher BMD values than the controls, but rugby players had the greatest BMD values at the lumbar spine ( › 16.5%), femoral neck ( › 21.7%) and hip ( › 13.5%). Moreover, significant correlations were observed between BMD and fat-free soft tissue mass, BMD and body mass, and BMD and training volume. Menstrual status has a major influence in the osteogenic response to exercise.[96] Cross-sectional data indicate that female oligo-/amenorrhoeic high school athletes (mean age 16 years) practicing ª 2009 Adis Data Information BV. All rights reserved.

447

impact sports may not be accruing the same bone mass as their eumenorrhoeic counterparts.[91] In summary, when normal menstrual status is reported in women, both sexes seem to benefit equally from sport participation (table II): the mean gains for total body BMD are about 10%, reaching 15–20% in the sport-specific loaded sites. According to the training load and intensity, athletes who train more (20 h/wk approximately) achieve greater bone mass gains. However, recreational athletes also benefit from a 5- to 6-h/wk training volume. If not accompanied by menstrual disturbances, high training volumes do not seem to negatively affect the osteogenic adaptive response to loading. Most studies report that the osteogenic response to exercise is specific to the loaded bones, and sports with higher impact and ground reaction forces elicit superior osteogenic responses. These effects are more marked in athletes who began their sport participation close to the pubertal growth spurt.[107-109] Cross-sectional investigations are suggestive of a relationship between training and bone metabolism; however, these studies compared independent samples and therefore were not able to establish a causal relationship between the variables of interest. In fact part of the effects described in the participants of different sports may have been caused by selection bias. Girls or boys with a larger musculoskeletal size and bone mass,[110] due to inherited characteristics, may choose to exercise because they may be more likely to be successful in competition and feel more rewarded by exercise practice. To rule out such a possibility, longitudinal studies and randomized clinical control trials are necessary. 2.2 Longitudinal Studies

Although evidence is accumulating suggesting that childhood is the best period of life to obtain osteogenic benefits from physical activity, modern children (especially girls) have become increasingly sedentary.[111,112] In addition, girls experience the pubertal growth spurt 1–2 years before boys do,[113] meaning that they should start regular activity even sooner than boys to Sports Med 2009; 39 (6)

Guadalupe-Grau et al.

448

achieve the full osteogenic benefit from physical activity. The strength of longitudinal studies is that randomized study designs reduce self-selection in group assignment, which is particularly important in exercise trials where individuals may be more or less predisposed to participate in physical activity. The majority of these studies have been performed in women. There are few longitudinal studies testing the response of men and women to the same training programme. A review of recent longitudinal studies performed with young adult men and women is depicted in table III. 2.2.1 Young Women

Peak bone mass is thought to be attained by the end of the third decade of life, therefore the early adult years may be the final opportunity for its augmentation.[127] As summarized previously in this review, athletes involved in high impact sports have higher bone mass than other athletes; consequently, longitudinal exercise programmes including exercises eliciting high impacts and forces on bones should evoke greater accumulation of bone mass and enhancement of BMD. However, the findings of studies performed in young adults have not consistently produced these findings. Standard resistance-training protocols enhance muscle mass,[128-131] and in several cases bone mass,[114,132] in young women. Friedlander et al.[114] reported significant increases after a 2-year programme of aerobics and weight-training in lumbar spine, intertrochanteric and calcaneal BMD. The addition of daily calcium supplementation did not add significant benefit to the intervention.[114] Also, Snow-Harter et al.[133] noted significant increases in lumbar BMD in young women completing either a progressive aerobic training programme (jogging) or a progressive resistance training programme, when compared with a control group. The resistance training group showed significant strength increases compared with the aerobictrained women; however, the increases in BMD were not significantly different between the two exercise groups. This result is consistent with the site-specific principles of mechanical loading, as both groups of women performed weight-bearing ª 2009 Adis Data Information BV. All rights reserved.

exercise, stressing the lower body and lumbar spine. Differences might not necessarily be expected between exercise groups in these studies due to the short duration of the protocol and the physiological limits of bone formation and remodelling. These conventional strength-training techniques involve both a concentric and an eccentric component. However, maximal skeletal muscle eccentric contractions develop greater tension than maximal concentric or isometric contractions,[134] and the magnitude of the load associated with maximal eccentric contraction is responsible for significant increases in bone mass in young women.[116] Schroeder et al.[115] further investigated this topic by testing the hypothesis that young women participating in high-intensity eccentric resistance training would have significantly greater increases in lean body mass and muscle strength and improved bone adaptations compared with low-intensity eccentric resistance training. These authors reported that low-intensity eccentric training (ET; 75% of concentric 1 repetition maximum [RM]) was as effective as high-intensity training (125% of a concentric 1 RM) and, surprisingly, there were no significant alterations in bone mass in the high-intensity group.[115] This finding is at odds with a previous study from this group reporting 3.9% greater BMD in the mid-femur after a similar ET protocol.[116] ET both at low and high intensities has been associated with elevated concentrations of osteocalcin in conjunction with decreased crosslinks, which suggests osteogenesis combined with reduced bone resorption.[115] Nickols-Richardson et al.[117] studied the effects of unilateral leg and arm high-intensity strength training lasting 5 months in young women. One group underwent ET and other group underwent concentric training (CT), with non-trained limbs serving as controls. The two kinds of exercise were similarly effective for improving muscular strength and bone mass, and density with gains in WBBMC (0.4% CT vs 0.6% ET), proximal femur BMD (0.5% CT vs 1.2% ET), total forearm BMD (0.6% CT vs 0.4% ET) and total forearm BMC (0.4% CT vs 0.5% ET). High impact training has also been demonstrated to yield positive results in young women.[118,119] Sports Med 2009; 39 (6)

Study

Subjects n

M/F

age (y)

Training

Frequency

Exercises

Protocol time (mo)

Friedlander et al.[114] (1995)

EX = 63

F

20–35

24

Schroeder et al.[115] (2004)

HRT = 14 LRT = 14C = 9

F

22–26 ERT

2/wk, 3 sets · 10 rep

Hawkins et al.[116] (1999)

EX = 8 C=8

F

19–23 CO-E RT

3/wk Isokinetic knee CO: 3 sets · 4 rep flexion and E: 3 sets · 3 rep extension

F

18–26 UET UCT

3/wk 1–3 sets · 6 rep 30 rep//limb/session

Training intensity

Other

Results Bone measurement site (s)

Young women

Nickols-Richardson et al.[117] EEX = 37 (2007) COEX = 33

Seated chest press, lat pulldown, biceps curl, triceps extension, single-leg extension, double leg curl

Isokinetic arm and leg resistance training

1500 mg Ca/S

HRT: 125% 1 RM LRT: 75% 1 RM

WB, SP, F

› 1.7% SPBMD in LRT

4.5

E: 1 RM CO: 1 RM

WB, SP, F

› 3.9% FBMD in ERT › FBMD ERT vs C

5

MVE

WB, TPF, DT, TF

› 1.2% WB, 5% TFBMC in EEX › 1.1% TPF, 0.5% TFBMD in EEX › 0.9% WB, 4% TFBMC in COEX › 0.6% TF, 0.5% TPFBMC in COEX

Continued next page

449

Sports Med 2009; 39 (6)

4

Exercise and Bone Mass in Adults

ª 2009 Adis Data Information BV. All rights reserved.

Table III. Effects of training protocols on bone tissue adaptations in young and middle-aged women: longitudinal studies

Study

450

ª 2009 Adis Data Information BV. All rights reserved.

Table III. Contd Subjects n

Training M/F

age (y)

Frequency

Exercises

Protocol time (mo)

Training intensity

Other

Bone Results measurement site (s)

6

MVE

300 mg Ca/S in EX, C

SP, FN, GT, WT

› 2.6% FN, 2.4% SPBMD in EX

R, SP, FN, GT, WT

› 4.1% GTBMD in EX

Kato et al.[118] (2006)

EX = 18 C = 18

F

20–22 HIT

3/wk 1 set · 10 rep

Countermovement jumps

Bassey et al.[119] (1994)

EX = 14 C = 13

F

27–34 HIT

3/wk

Jumping, skipping

12

MVE

Sinaki et al.[120] (1996)

EX = 60 C = 60

F

30–40 RT

3/wk 3 sets · 10 rep

Back extension and shoulder girdle weightlifting exercises

36

50–100% 10 RM

Chilibeck et al.[121] (1996)

EX = 20 C = 10

F

19–21 RT + ET

2/wk

Arm curl, bench press, leg press

4.5

WB, SP, L, A, NC FN, WT, GT, IT

Nindl et al.[122] (2000)

EX = 26 C=5

F

24–32 RT

5/wk 4–6 set · 10–12 rep

Squat, bench press, sit up, leg curl, row with elbows low + aerobics

6

WB, L, A, Tr

NC

EX = 22 C = 34

F

28–39 RT

3/wk 3 sets · 8–12 rep

Bicep curl, 18 bench press, supine flys, lat pulldown, leg curl, leg extension, leg press, military press, low rows, right wrist curl

WB, SP, FN, A, L

› SPBMD in EX › FNBMD in EX

Physically WB, SP, WH, R active women

NC

Premenopausal women

Lohman et al.[123] (1995)

75–80% 1 RM

500 mg Ca/S in RT, C

Guadalupe-Grau et al.

Sports Med 2009; 39 (6)

Continued next page

Study

Subjects n

Training

Frequency

Exercises

35–40 HIT

3/wk

Step patterns, 12 stamping, jumping, running, walking

MVE

SP, F, FN, IT, DF.

› 1.1 FNBMD EX vs C › 0.8 ITBMD EX vs C › 0.1 FBMD EX vs C › 2.2 L1BMD EX vs C

F

34–44 RT + HIT

3/wk 3–9 sets x 8–12 rep

12 LEX: jumps, squat, lunges, calf raises ULEX: LEX + upright row, one-arm row, lat dorsi pull down, chest press, chest fly, biceps curl, triceps extension

› 0–13% BW

WB, SP, FN, GT

› 2.6% GTBMD ULEX vs C › 2.2% GTBMD LEX vs C › 1.3% SPBMD ULEX vs C

F

35–45 HIT

3/wk

Jumping, callisthenics

2–6 times BW

SP, FN, GT, TDF, R

› 1.6% FNBMD in EX

M/F

age (y)

F

Winters-Stone and Snow[125] LEX = 19 (2006) ULEX = 16 C = 24

Heinonen et al.[126] (1996)

Vainionpa¨a¨ et al.[124] (2005)

EX = 39 C = 41

EX = 49 C = 49

Protocol time (mo)

18

Training intensity

Other

Bone Results measurement site (s)

451

Sports Med 2009; 39 (6)

1RM = 1 repetition maximum; A = arms; BMD = bone mineral density; C = control subjects; Ca/S = supplemental calcium; CO-E RT = concentric-eccentric resistance training; COEX = concentric exercising subjects; DF = distal forearm; DT = distal tibia; EEX = eccentric exercising subjects; ERT = eccentric resistance training; ET = endurance training; EX = exercising subjects; FD = femur diaphysis; FN = femoral neck; GT = greater trochanter; HIT = high-impact training; HRT = high-intensity eccentric resistance training subjects; IT = intertrochanteric subregion; L = whole leg; L1 = first lumbar vertebra; lat = latissimus; LEX = lower body exercising subjects; LRT = low-intensity eccentric resistance training; M//F = male/female; MVE = maximal voluntary effort; n = number of subjects; NC = no changes; R = radius; rep = repetition; Tr = trunk; RT = resistance training; SP = lumbar spine; TDF = total distal femur; TF = total forearm; TPF = total proximal femur; UCT = unilateral concentric training; UET = unilateral eccentric training; ULEX = upper + lower body exercising subjects; WB = whole body; WH = whole hip; WT = Ward’s triangle; %BW = percentage of bodyweight; › indicates significant increase p < 0.05; fl indicates significant decrease p < 0.05.

Exercise and Bone Mass in Adults

ª 2009 Adis Data Information BV. All rights reserved.

Table III. Contd

Guadalupe-Grau et al.

452

Recently, Kato and co-workers[118,119] tested the effect of a 6-month low repetition jump training programme (10 maximum vertical jumps/day, three times/week) on lumbar and hip BMD in young women (age 21 years). Despite the low number of jumps compared with other studies, based on jumping exercises in which subjects performed ten times more jumps per week,[119] BMD increased significantly in both regions, whereas no changes were observed in the control group. Although more studies are needed, these results suggest that the bones of young women respond to low repetition as well as high repetition jump training. On the other hand, resistance training has been reported to have either no effect or a negative impact on bone in a few studies in young women.[120-122] Sinaki et al.[120] found no significant effect of a 3-year non-strenuous, weightlifting exercise programme on BMD at lumbar spine, whole hip or mid-radius in active but not athletic premenopausal women 30–40 years of age. A particularity of this study is that only two different exercises were used: back extension and shoulder girdle weight-lifting exercises (as defined by the authors). The back extension exercise session consisted of three sets of 10 RM performed once a day for 3 days per week (one supervised session each week at the medical centre) with back extension. The shoulder exercise programme was adjusted every 3 months following the sequence 50% of 10 RM for the first month, 75% of 10 RM for the second month, and 100% of 10 RM for the third month, which was prior to the next visit and re-evaluation.[120] Chilibeck et al.[121] reported in 20-year-old women that after a 20-week weight-lifting training period, which increased muscle strength (23–73%) and lean tissue mass (3–10%) in the trunk and the extremities, BMC and BMD failed to increase in whole body, arm, leg, ribs, thoracic and lumbar spine, and pelvis segments. Similarly, hip BMC and BMD at femoral neck, trochanter, intertrochanter and Ward’s triangle sites and total hip did not increase with training.[121] In 28-year-old women, no changes in whole body or regional bone mass and density were observed following a 24-week training programme comª 2009 Adis Data Information BV. All rights reserved.

bining weight-lifting with endurance exercise (1.5 h/day for 5 days/wk), despite a 2.2% increase in whole body lean mass and a 5.5% increase in leg lean mass.[122] To this point, the studies reviewed indicate that strength training needs to involve highintensity exercises to enhance bone mass in young women. The osteogenic effect may be attenuated if endurance exercise is carried out in combination with strength training, and enhanced if the strength training programme is combined with high impact exercises, like jumping. Bone formation in weight-bearing regions of the skeleton can be stimulated by low-magnitude highfrequency strains, induced through vibration.[38,39] In young women aged 15–20 years with low BMD, low level whole body vibration (30 Hz; 0.3 g) applied daily (between 2 and 10 minutes) for 1 year increased cancellous bone in the lumbar vertebrae (2.1%) and cortical bone in the femoral midshaft (3.4%), respectively, measured by quantitative computer tomography (QCT), compared with controls. It is noteworthy that, in this study, these gains were not detected with DXA.[135] 2.2.2 Young Men

Randomized longitudinal studies on sedentary young men are scarce. A 4-month strength training programme (three times per week, at 60–80% of 1 RM) in Oriental men (aged 23–31 years) did not elicit changes in whole body or loaded bones BMC and BMD. However, serum osteocalcin and bone-specific alkaline phosphatase concentrations were increased 1 month after the start of the training, suggesting that bone markers are more efficient to detect changes in the bone remodelling than DXA.[136] Hartman et al.[137] did not find significant changes in WBBMC after 12 weeks of strength training (five times per week) in subjects who consumed fat-free milk, soy or carbohydrates after the training sessions despite substantial increases in lean body mass and strength. However, a regional analysis of BMD was not reported in that study. With 6 months of training (strength + aerobics) combined with either protein or carbohydrate supplementation, positive effects on BMC, BMD and geometrical variables of the Sports Med 2009; 39 (6)

Exercise and Bone Mass in Adults

tibia measured by peripheral QCT have been reported in both young men and young women.[138] In this study, all groups experienced significant increases in tibia cortical thickness and area, and decreased their endosteal circumference over the intervention period. Cortical thickness among females was greater if they were receiving supplemental protein, whereas among men the change was the greatest if they were receiving carbohydrate supplementation. No changes were observed on whole body measurements in any group. These results suggest that there may be sex-specific differences in the bone response to exercise when supplementing with protein. However, due to the absence of a non-exercising control group it is not clear if exercise was the responsible factor for the changes. This study contrasts with that of Ryan et al.,[139] who found no sex differences in the training response between men and women for any of the whole body, femoral neck and lumbar spine BMD measurements after a 6-month resistance training programme. The latter study is also affected by the lack of age- and sex-matched control groups. We have not found long-duration (more than 6 months) randomized longitudinal studies on the effects of strength training on bone mass in sedentary young men. 2.2.3 Premenopausal Women

A long-term (18-month) randomized, controlled, prospective study with high-intensity resistance training in adult women indicated that regional BMD at the femoral neck and trochanteric sites can be increased by resistance training exercise.[123] However, WBBMD did not change significantly over the 18 months of this trial, which could indicate that increases in strength and lean tissue may be greater than increases in BMD in premenopausal women and that in young women there may be a site-specific redistribution of BMD rather than a total body increase in BMC.[123] This hypothesis is supported in other studies, which evaluated regional body composition changes in women after periodized physical training.[122,129-131,140] Similarly, Vainionpa¨a¨ et al.[124] noted significant BMD inª 2009 Adis Data Information BV. All rights reserved.

453

creases in femoral and lumbar sites in premenopausal women after a progressive 12-month high impact training (jumping) programme; WB was not measured in this study but calcaneal broadband ultrasound attenuation also showed a significant increase in the exercise group compared with the control group,[124] suggesting enhanced bone quality. Winters-Stone and Snow[125] also support the site-specific response of lumbar spine and hip BMD to upper and lower body resistance exercise training, found in two groups of nonactive premenopausal women (aged 33–44 years). They performed a high impact training programme (lower limbs group) or high impact and resistance training in the upper limbs (upper + lower group) for 12 months, which resulted in an increased greater trochanter BMD in both groups and lumbar spine BMD only in the upper + lower group. An 18-month high impact training programme (jumps + calisthenics) without concurrent strength training has been reported to increase femoral neck BMD by 1.6% in women aged 35–45 years.[126] In combination, these results indicate that targeted training could be effective in women with low bone mass in an isolated bone site – a potentially inexpensive and safe way to prevent and/or treat osteoporosis later in life. In a mixed population (21 men and 35 women aged 19–38 years), whole-body vibration – administered during 8 months (4 min/day, three to five times per week) at 25–45 Hz, corresponding to estimated maximum vertical accelerations of 2–8 g – had no effect on mass, structure or estimated strength of bone at any skeletal site.[141] Serum markers of bone turnover did not change during the vibration intervention.[141] These findings contrast with those of Beck et al.,[142] who reported a 2% increase in BMD in the proximal femur of five premenopausal women (age 18–45 years) submitted to whole-body vibration (30 Hz; 2 · 10 min/day; 0.2 g stimulus) for 12 months. This study, however, must be interpreted cautiously because it lacked a control group and one of the women was on treatment with biphosphonates.[142] Thus, it remains to be determined whether whole-body vibration alone or in combination with strength training could be an efficient stimulus to enhance BMD in Sports Med 2009; 39 (6)

Guadalupe-Grau et al.

454

premenopausal women with either normal or reduced BMD at the start of the programme. 2.2.4 Middle-Aged Men

The effects of exercise on bone mass and structure in middle-aged men are controversial. Strength training (three times per week at 5–15 RM) in men aged 54–61 years for 4 months resulted in a 2.0% gain in lumbar spine and 3.8% gain in femoral neck BMD after this period.[143] On the other hand, regular aerobic (40–60%
VO2max) exercise in men (age 53–62 years) for a long time (48 months) does not appear to have beneficial effects on the age-related loss of femoral BMD, suggesting that starting regular aerobic exercise in middle age to prevent osteoporosis in men may not be efficacious.[144] However, when aerobic exercise was combined with strength training for 6 months, WBBMD was maintained.[145] A unique study that involved both sexes in the same exercise programme showed that 6 months of moderate strength training (50% 1 RM) and moderate aerobic training (60–90% maximal heart rate) maintained WBBMD in males but decreased WBBMD in females.[145] Both males and females experienced exercise-induced bodyweight and fat losses, suggesting that although lean mass was increased after the training period (3%), the negative effect of losing weight on BMD was stronger, although the age range of the participants was too great (55–75 years).[145] Nevertheless, a lack of randomized controlled trials testing any kind of training and bone measurements in middle-aged men precludes any definitive conclusion on the effects of exercise on bone mass in this segment of the population.[146] 2.3 Aging

Some authors support the hypothesis that the magnitude of the peak skeletal mass in the first three decades of life probably accounts for the variability in bone mass in elderly persons,[147] but it is currently impossible to prove or disprove this theory because there are no studies with randomized controlled trials from puberty to old age that have investigated bone fractures as an ª 2009 Adis Data Information BV. All rights reserved.

endpoint. These studies would require extremely large cohorts and a >50-year study duration before the research question could be answered. Furthermore, traits independent of BMD, such as skeletal architecture, bone size, balance, muscle strength and neuromuscular propioception, may also be affected by exercise, all of which could influence the fracture risk, and are not reviewed in this article.[148,149] The Leisure World Study reported that women with an activity level >1 hour a day had a reduced risk of hip fracture, but the beneficial effect was lost if the activity level was reduced.[150] In the Study of Osteoporotic Fracture, a longitudinal study including 9704 women >65 years of age and followed for about 8 years, women in the highest quintile of current activity level had a 42% lower hip fracture risk than the least active quintile of women, and self-reported walking time was associated with a 30% reduction in hip fracture risk during a 4.1-year follow-up.[43] Studies evaluating the question of whether bone mass is maintained after a reduction or cessation of exercise show contrasting results, irrespective of the level of BMD found in retired athletes. To evaluate the hypothesis that exercise during growth reduces the clinical problem of fragility fractures, it would be needed to demonstrate that retired athletes have fewer fractures than controls. Wyshak et al.[151] compared a large cohort (n = 10 796) of former female college athletes with sedentary controls aged 21–80 years. The number of former athletes with fractures after retirement was no different than among the controls. Among women aged ‡60 years, who were fracture-free up to the age of 40 years, the rate of any fracture at age ‡40 years was 29% for former athletes compared with 32% for non-athletes, a nonsignificant difference. Nordstro¨m et al.[41] measured BMD in two cohorts; the first comprised 65 young male ice hockey players, 73 young soccer players (two high-impact sports) and 61 age-matched controls. Measures were taken again after 5 years; at that time, 55 athletes had retired from their active sports career. The second cohort comprised 400 former soccer and ice hockey players and 800 age- and sex-matched controls. At baseline, Sports Med 2009; 39 (6)

Exercise and Bone Mass in Adults

all active groups had higher BMD values at whole body, femoral neck, lumbar spine and arms compared with controls; after 5 years the young retired athletes still had a 4–8% higher BMD than controls, whereas young athletes increased the difference in BMD compared with the controls at femoral neck and arms. These results suggest that higher BMD persists until several decades after retirement. Furthermore, retired athletes had fewer fractures than controls. Therefore, it seems that exercise during childhood and adolescence may be associated with lower risk of sustaining fragility fractures during old age in men,[94] but in women these beneficial results only persist if exercise practice is maintained. The effects of exercise protocols on bone density have also been reported in older populations; a review of recent longitudinal studies is provided in tables IV and V for women and men, respectively. 2.4 Postmenopausal Women

Nelson et al.[152] completed a 1-year randomized, controlled trial of high-intensity resistance training in postmenopausal women. The results of the study demonstrated that women in a 2 days/ week resistance training programme gained an average of 1% in BMD of the femoral neck and lumbar spine whereas the control group lost 2.5% and 1.8% at these sites, respectively. In addition, the resistance-trained women tended to maintain WBBMC of the skeleton whereas the women in the control group had a 1.2% decline in WBBMC. Also, the resistance-trained women had 35–76% increase in strength, 14% improvement in dynamic balance, and a 1.2 kg increase in total body lean mass and a 27% increase in physical activity unrelated to the intervention, whereas the control group showed declines in all of these parameters. In agreement with these results, Kerr et al.[153] reported that postmenopausal bone mass can be significantly increased by a strength regimen that uses high loads and a low number of repetitions (3 · 8 RM) but not by an endurance regimen that uses low loads and a high number of repetitions (3 · 20 RM). In 1–7 years, postmenopausal women, following ª 2009 Adis Data Information BV. All rights reserved.

455

9 months of strength training with intermediate loads (2 · 10–15 RM), lumbar spine BMD was enhanced by 1.6%. In this study, each subject performed one set of 10–12 RM (increasing progressively) for upper body training and one set of 10–15 RM for lower body training. In contrast, the women from the control group experienced a 3.6% decline in lumbar spine BMD.[159] Altogether, these studies show that the peak load is more important than the number of loading cycles in increasing bone mass in postmenopausal women. In late postmenopausal women (aged 60–72 years), 9 months of endurance training (mostly
running at 60–70% of V O2max, three to four times a week, for 35–50 min/session) either alone or in combination with hormone replacement therapy (HRT) resulted in significant increases in lumbar spine and femoral neck.[154] Exercise and HRT resulted in independent and additive effects on the BMD of the lumbar spine and Ward’s triangle, and a synergistic effect on whole body BMD. These effects were accompanied by a reduction in serum osteocalcin levels, indicating that increases in BMD in response to HRT and to exercise + HRT were due to decreased bone turnover.[154] The lack of change in serum osteocalcin and IGF-I in response to exercise alone suggests that the increases in BMD were due to decreased bone resorption and not to increased formation.[154] Other studies have reported just the maintenance of BMD in postmenopausal women with resistance training.[155,162] There is evidence that postmenopausal women respond differently to a resistance training programme than do premenopausal women.[163,164] Bassey et al.[163] studied the effects of a vertical jumping exercise regimen on BMD using randomized controlled trials in both pre- and postmenopausal women, the latter stratified for HRT. The exercise consisted of 50 vertical jumps on 6 days/week of mean height 8.5 cm, which produced mean ground reactions of 3.0 times bodyweight in the young women and 4.0 times in the older women. In the premenopausal women, the exercise resulted in a significant increase of 2.8% in femoral BMD after 5 months. In the postmenopausal women, there Sports Med 2009; 39 (6)

Study

Subjects n

M/F

age (y)

Training

Frequency

Exercises

Protocol time (mo)

456

ª 2009 Adis Data Information BV. All rights reserved.

Table IV. Effects of training protocols on bone tissue adaptations in young men, middle-aged men and age-specific sex comparisons: longitudinal studies Training intensity

Other

Bone measurement site (s)

Results

Young men EXM = 18 EXS = 19 C = 19

M

18–30

RT

5/wk 2–4 sets · 4–12 rep

Military press, bench press, seated chest fly, seated triceps extension, seated lateral pull down, seated wide grip row, seated reverse fly, seated biceps curl, abdominals, inclined leg press, 2-leg knee extension, 2-leg hamstring curl, seated calf raise

3

80% 1 RM

500 mL M/S in EXM; 500 mL S/S in EXS; 500 mL C/S in C

WB

NC

Ballard et al.[138] (2006)

YMEX = 13 YWEX = 12 CM = 12 CW = 11

M/F

20–22

RT + ET

5/wk 3 sets · 12failure rep

Bench press, inclined bench press, shoulder press, lat pulldown; cable rows, arm curl and extensions, hip sled, squats, calf raises + aerobics

6

70% 1 RM; 70%
VO2max

EX: 42 g P/S C: 70 g C/S

T, WB, A, L

› T vBMD in YMEX and YWEX › ABMC in YMEX and YWEX

Ryan et al.[139] (2004)

YWEX = 8 OWEX = 11 YMEX = 13 OMEX = 10

M/F

Y = 20–29 O = 65–74

RT

3/wk 2 sets, failure

Leg press, chest press, leg curl, lat pulldown; leg extension, military press, seated row, triceps pulldown, abdominal crunch, biceps curl, sit ups

6

12–15 RM

WB, SP, FN, WT, GT

› FNBMD in ESP

Fujimura et al.[136] (1997)

EX = 8 C=7

M

23–31

RT

3/wk 2–3 sets · 10 rep

Leg extension, leg curl, bench press, sit up, back extension, arm curl, wrist curl, half squat leg lunge, lateral pull down, back press

4

60–80% 1 RM

WB, FN, SP, R

NC

Continued next page

Guadalupe-Grau et al.

Sports Med 2009; 39 (6)

Hartman et al.[137] (2007)

ª 2009 Adis Data Information BV. All rights reserved.

457

1RM = 1 repetition maximum; A = arms; BMD = bone mineral density; C = control subjects; CM = male control subjects; C/S = supplemental carbohydrate; CW = female control subjects; ESP = entire study population; ET = endurance training; EX = exercising subjects; EXM = exercise + supplemental milk products; EXS = exercise + supplemental soy products; FN = femoral neck; GT = greater trochanter; L = whole leg; lat = latissimus; M/F = male/female; M//S = supplemental milk; N = number of subjects; NC = no changes; O = old; OMEX = older men exercising subjects; OWEX = older women exercising subjects; P/S = supplemental protein; R = radius; rep = repetitions; RT = resistance training; SP = lumbar
spine; S/S = supplemental soy; T = tibia; vBMD = volumetric BMD; V O2max = maximum oxygen consumption; WB = whole body; WT = Ward’s triangle; Y = young; YMEX = young men exercising subjects; YWEX = young women exercising subjects; › indicates significant increase p < 0.05; fl indicates significant decrease p < 0.05.

› 2.0% SP in EX › 3.8% FN in EX SP, FN 5–15 RM 4 Chest press, overhead press, lat pull, upper back row, leg press, leg extension 3/wk 1–2 sets · 15 rep RT 54–61 EX = 11 C=7 Menkes et al.[143] (1993)

M

EX = 70 C = 70 Huuskonen et al.[144] (2001)

Middle-aged men

Study

Table IV. Contd

Subjects n

M

M/F

50–60

age (y)

ET

Training

3–5/wk 30–60 min

Frequency

Exercises

Brisk walking

48

Protocol time (mo)

40–60%
VO2max

Training intensity

Other

SP, FN, GT, WT

Bone measurement site (s)

Results

fl SP, FN, GT, WT BMD in EX

Exercise and Bone Mass in Adults

was no significant difference between the exercise and control groups after 12 months (total n = 123) nor after 18 months (total n = 38). HRT status did not affect this outcome, at least up to 12 months. Sugiyama et al.[164] studied a group of Japanese female volunteers aged around 50 years divided into premenopausal women with a regular menstruation cycle and postmenopausal women within 5 years since menopause. About half of the subjects in each group chose to be non-exercisers. The remainder followed a 6-month training programme consisting of rope skipping (100 jumps/day, with an interval of 2–3 days). In total, they completed 10 days per month, 60 days during the study period. Among the premenopausal women, the BMD in the exercise group increased significantly compared with the controls for total hip (+1.6%) and femoral neck (2.4%), but changes at the whole body and lumbar spine levels were not significant. In contrast, there were no significant differences at any measurement sites among the postmenopausal women. Interestingly, in the premenopausal exercise group, the baseline value of urinary g-carboxyglutamate (Gla) residues (an indirect measure of osteocalcin carboxylation) was inversely correlated (r = -0.62) with the change in whole body BMD. The latter could indicate that bone gain induced by high impact exercise could become greater in proportion to the degree of deterioration in bone material properties.[164] Therefore, although optimum training strategies are still under discussion, it is generally acknowledged that the training should be population specific. Stengel et al.[155] tested the hypothesis that power training was more effective than conventional strength training for maintaining BMD at lumbar spine and hip. Forty-two postmenopausal women performed a 12-month training programme; the only difference between the two groups was the velocity at which movements were performed. The training protocol specified a 4-second concentric, 4-second eccentric sequence in the resistance training group, and a concentric fast/explosive 4-second eccentric sequence in the power training group. In addition, all women performed gymnastics and home training sessions. Women involved Sports Med 2009; 39 (6)

Study

Subjects n

Training

458

ª 2009 Adis Data Information BV. All rights reserved.

Table V. Effects of training protocols on bone tissue adaptations in older women, men and age-specific sex comparisons: longitudinal studies Frequency

Exercises

Protocol Training Other intensity time (mo)

Results Bone measurement site (s)

2/wk 3 set · 8 rep

5 weight-lifting exercises

12

WB, SP, FN

› › fl fl

M/F age (y)

Older women EX = 20 C = 19

Kerr et al.[153] (1996)

REX EEX C

Kohrt et al.[154] (1995)

REX C

F

RT

3/wk

Weight-bearing exercises

12

Stengel et al.[155] (2005)

PEX = 21 REX = 21

F

54–60 PT RT

4/wk

2 weight-lifting sessions 1 gymnastics session 1 home training session

12

70–90% 1 RM

1.500 mg Ca/S, 500 Vit-D/S

SP, WH, FN, T, IT

NC FNBMD in PEX NC SPBMD in PEX fl 0.9% SPBMD in REX fl 1.2% WHBMD in REX

Chien et al.[156] (2000)

EX = 22 C = 21

F

48–65 ET + HIT

3/wk 50 min

Treadmill walking + stepping exercise

6

70–85%
VO2max

Osteopenic subjects HRTh

WB, SP, FN

› 6.8% FNBMD in EX

Kohrt et al.[157] (1997)

GREX JREX C

F

60–74 ET RT + ET

3/wk GREX: 30–45 min JREX: 2–3 sets · 8–12 rep 15–20 min

GREX: walking, jogging, stair climbing JREX: overhead press, biceps curl; triceps extension, leg press, leg extension, leg flexion, bench press, squats

11

GREX: 60–85 MHR JREX: 8–12 RM, 60–85 MHR

WB, SP, FN, GT, W

› 2.0% WBBMD in GREX › 1.6% WBBMD in JREX › 1.8% SPBMD in GREX › 1.5% SPBMD in JREX › 6.1% GTBMD in GREX › 5.1% GTBMD in GREX

Verschueren et al.[158] (2004)

VEX: 25 EX: 22 C: 23

F

60–70 VEX: VEX, EX: VEX and EX: RT + WBV 3/wk leg extension, EX: RT 1–3 leg press sets · 10–15 rep

VEX, EX: 20–8 RM

WB, F, SP

› 0.9% FBMD in VEX NC in EX and C

F

50–70 RT

RT ET

80% 1 RM

1% SPBMD in EX 1% FNBMD in EX 1.8% SPBMD in C 2.5% FNBMD in C

12

6

HRTh

Continued next page

Guadalupe-Grau et al.

Sports Med 2009; 39 (6)

Nelson et al.[152] (1994)

Study

Pruitt et al.[159] (1992)

Subjects n EX: 17 C: 10

Training

Frequency

Exercises

M/F age (y) F

52–56 RT

3/wk Biceps curl, lat 1 set · 10–15 pulldown, bench press, rep wrist roller, leg press, leg ab/adduction, leg curl, leg extension, trunk extension, hip extension, lateral flexion

Protocol Training Other time intensity (mo)

Bone Results measurement site (s)

9

10–15 RM

SP, FN

› 1.6% SPBMD in EX fl 3.6% SPBMD in C

WH, FN, GT, SP

› 1.9% SPBMD HMEX › 1.3% GTBMD HMEX › 2.0% GTBMD HWEX

WB, SP, FN

› 2.8% FNBMD in EX

Hypertensive WB, SP, FN, subjects IT, WH

› 1.7% SPBMD in MRT

Older men Maddalozzo et al.[160] (2000)

MMEX = 12 M/F 50–60 MRT HRT HMEX = 12 MWEX = 9 HWEX = 9

MRT: 3/wk 3 sets · 10–13 rep HRT: 3/wk 3 sets · 2–10 rep

M: Leg press, leg extension hamstring curls, arm curl, triceps press, chest press, Pec deck, shoulder press, side lateral raise, lat pulldown, seated row, abdominal crunch, calf raise H: free weight back squat, deadlift, biceps curls, sit ups, triceps extension, chest press, incline shoulder press, high lat pull down, leg curl, gripper, calf raise.

6

MRT: 40–60% 1 RM HRT: 70–90% 1 RM

Ryan et al.[161] (1994)

EX = 21 C = 16

M

3/wk 2 sets · 15 rep

Leg press, chest press, leg curl, lat pull down, leg extension, military press, adductor, abductor, upper back, triceps, lower back, abdominals, biceps curl

4

5 RM

Stewart et al.[145] (2005)

MRT = 26 WRT = 31 C = 58

M/F 55–75 RT + ET

Bench press, shoulder press, 3/wk 2 sets · 10–15 seated mid-rowing, lat pulldown, leg extension, rep leg curl, leg press

6

51–71 RT

Dietary control

459

Sports Med 2009; 39 (6)

1RM = 1 repetition maximum; Ca/S = supplemental calcium; EEX = endurance training exercising subjects; ET = endurance training; EX = exercising subjects; F = femur; FN = femoral neck; GREX = ground reaction forces exercising subjects; GT = greater trochanter; H = high-intensity exercises; HIT = high-impact training; HMEX = men high-intensity exercising subjects; HRT = high-intensity resistance training; HRTh = hormone replacement therapy; HWEX = women high-intensity exercising subjects; IT = intertrochanteric subregion; JREX = joint reaction forces exercising subjects; lat = latissimus; M = moderate-intensity exercises; M/F = male/female; MHR = maximal heart rate; MMEX = men moderate-intensity exercising subjects; MRT = moderate resistance training; MWEX = women moderate-intensity exercising subjects; N = number of subjects; NC = no changes; pec = pectoralis; PEX = power exercising subjects; PT = power training; rep = repetition; REX = resistance training exercising subjects; RT = resistance training; SP = lumbar spine; T = tibia;
VEX = vibratory exercising subjects; Vit-D/S = supplemental vitamin D; V O2max = maximum oxygen consumption; W = wrist; WB = whole body; WBV = whole body vibration; WH = whole hip; › indicates significant increase p < 0.05; fl indicates significant decrease p < 0.05.

Exercise and Bone Mass in Adults

ª 2009 Adis Data Information BV. All rights reserved.

Table V. Contd

Guadalupe-Grau et al.

460

in the power training group maintained their BMD at lumbar spine and hip, whereas resistance training women experienced a significant decrease in lumbar spine and hip sites from baseline. These results indicate that to elicit a osteogenic response in older women, the strains and ground reaction forces required may be higher than those able to elicit a similar or even greater response in younger women. Palombaro[165] reviewed the effects of walkingonly programmes on BMD at various skeletal sites. This meta-analysis showed that walking interventions alone did not attenuate bone loss at the skeletal sites reported. Thus, other forms of exercise in addition to walking should be incorporated into training regimens for patients at risk for osteoporosis. Supporting this notion, Chien et al.[156] examined the efficacy of a 24-week aerobic plus high exercise programme for osteopenic postmenopausal women, and this appeared to be effective in offsetting the age-related decline of BMD, especially at the femoral neck, which showed a significant improvement of 6.8% in the exercise group. Kohrt et al.[157] applied two different training protocols to postmenopausal women at risk of osteoporosis. The first protocol consisted of exercises that stimulated the skeleton through ground-reaction forces (walking, jogging, stair climbing), while the second protocol included activities eliciting joint-reaction forces (weightlifting, rowing). The intensity was initially set at a low level (60% maximal heart rate, 12 RM) and progressed with training (to 85% maximal heart rate, 8 RM). After 11 months, BMD was increased at the whole body level and femoral neck in both groups, but the effects were greater in the ground reaction than the joint-reaction group.[157] These results could be explained by the combination of walking, jogging and stair climbing, which may generate ground reaction forces between 2.8–6 times bodyweight[166] in the ground reaction group, and the use of free weights in the resistance training combined with rowing in the joint reaction group. However, more randomized, controlled studies testing aerobic plus high-impact training in older adults are needed. ª 2009 Adis Data Information BV. All rights reserved.

Whole body vibration training in postmenopausal women has been shown to increase femoral neck BMD and balance more than walking.[167] Compared with resistance exercises progressing from low (20 RM) to high (8 RM) loading conditions, 6 months of static and dynamic knee extensor exercises on a vibration platform (35–40 Hz; 2.28–5.09 g) enhanced hip BMD by 0.9%.[158] In another study, whole body vibration inhibited bone loss in the spine and femur of postmenopausal women.[168] These authors performed a 1-year prospective, randomized, double-blind, placebo-controlled trial of 70 postmenopausal women who undertook brief periods (<20 minutes) of a low-level (30 Hz; 0.2 g) vibration applied during quiet standing. The efficacy of this intervention was enhanced in the women with significantly greater compliance, particularly in those subjects with lower body mass.[168] The studies in postmenopausal women indicate that BMD can be increased, or at least the decline in bone mass during the menopause attenuated, following weight training exercises. The osteogenic effects are site specific and can only be achieved with high loading intensities (>70% of 1 RM) with 3–4 sessions per week and 2–3 sets per session.[169] Although significant effects can be observed after 4–6 months in some locations, the efficacy of the training programme is greater when extended for ‡1 year. Combining strength training with aerobic exercise may also result in positive effects on BMD. Whole body vibration alone or in combination with exercise may help to increase or at least prevent BMD decline with aging in postmenopausal women. However, the gains in bone density and neuromuscular functions achieved by training are lost 5 years after cessation of training.[170] Continuous highintensity weight-loading physical activity is probably necessary to preserve bone density and neuromuscular function in older women. 2.5 Older Men

Older men have been much less studied than older women, possibly because of the lower osteoporotic fracture incidence in men.[171] One of Sports Med 2009; 39 (6)

Exercise and Bone Mass in Adults

these studies compared the effects of either a moderate (three sets of 10–13 repetitions at 40–60% of 1 RM) or high (three sets of 2–10 repetitions at 70–90% of 1 RM) intensity resistance training programme (with exercises involving all major muscle groups) in men and women aged 50–60 years.[160] Both older men and older women achieved significant increases in muscular strength and muscle mass regardless of intensity or training protocol.[160] The high-intensity training men experienced a significant increase in lumbar spine and greater trochanter BMD; however, women training with high-intensity increased greater trochanter BMD only slightly, maybe because these women were primarily early postmenopausal (within 36 months), a time during which there is accelerated bone loss of 2–6.5% per year. High-intensity free weight training was tolerated well by older adults and produced BMD changes in only 6 months. In older men, high-intensity training was more osteogenic at the lumbar spine than moderateintensity training. In agreement with Maddalozzo and Snow[160] Stewart et al.[145] reported no effect on BMD in men (55–75 years old) following 6 months of multistation machine at 50% of 1 RM followed by 45 minutes of aerobic training at 60–90% of their maximal heart rate. Nevertheless, this training programme resulted in other positive effects such as gains in lean mass, reduced fat mass and improved aerobic capacity.[160] Bone mass improvement has been observed in older men (mean age 61 years) after a relatively short training period.[161] In this study, femoral neck BMD was enhanced by 2.8% following 3 months of high-intensity training (5 RM; three times per week).[161] These results could be conflicting, because Frost[172] has argued that short-term increases in BMD measured by photon absorptiometry may reflect transient increases. In general, 1–3% BMD improvement in loaded bones can be achieved in old men with 6 months of strength training using heavy loads (above 70% of 1 RM, three times per week), while loads below 60% of 1 RM are unlikely to have a positive influence on bone mass.[173] ª 2009 Adis Data Information BV. All rights reserved.

461

3. Practical Recommendations Any prescription of exercise aiming to improve bone mass must take into consideration the following factors: (i) Age and sex of the subjects. At prepubertal and young adult ages, sex differences are not so important, but at middle and older age, evidence from the literature suggests that women have to train at higher intensities than men to improve their bone mass, always keeping a security range to avoid injuries. (ii) Choice and order of the exercises. Since bone adaptation is limited to loaded regions, exercise must be chosen to specifically act on the clinically relevant sites, i.e. lumbar and thoracic spine, whole hip, and especially greater trochanter, intertrochanteric and femoral neck regions. The easiest and safest way to load these regions is by using weight-lifting exercises like: leg press, leg extension, leg curl, squats, loaded back extensions, and some shoulder and arm exercises. If not contraindicated, the training programme should include impact exercises like jumping, jogging, stair climbing and sprinting. Impact exercises must be increased progressively to the maximal effort possible according to the subject’s specific capabilities. The kind of impact exercise included in the programme must be appropriate for the age of the participants, trying to keep the risk of fall as low as possible in the elderly. It must be taken into consideration that the osteogenic potential of jumping exercise is reduced in postmenopausal women, but postmenopausal women may respond well to strength training. (iii) Intensity. To enhance bone mass the threshold intensity must be reached. This level has not been unequivocally established and may vary from subject to subject, probably being lower for subjects with already reduced bone mass. Most strength training programmes showing positive effects on bone mass have used intensities of 70–90% 1 RM, always following an appropriate progression from lower to higher intensities. (iv) Frequency. Most studies with positive results have used 2–3 training days per week. However, good responses to jumping exercise sessions with Sports Med 2009; 39 (6)

Guadalupe-Grau et al.

462

frequencies up to 6 days per week have also been reported. Weight-bearing endurance exercise (30–60 minutes) can be carried out three to five times, or even on a daily basis, depending on training experience and tolerance. (v) Volume. In weight-lifting programmes, the major muscle groups of the upper and lower extremities must be trained in a balanced way, without creating imbalance between agonist and antagonist. The number of repetitions per exercise must be close to maximal that can be performed with a given load and 2–3 sets should be completed, with 1–3 minutes’ resting periods in between. With regard to high impact training, there is no consensus in the literature on how many jumps must be performed, but depending on the subject’s tolerance, 50–100 jumps should be carried out each training day. Strength training combined with high impact exercise could have additive effects in some subjects. (vi) Movement velocity. Although a progression from medium to high speed of movement is advocated at the start of the training programme, as soon as subjects are able to carry out the exercise safely, i.e. with proper biomechanical execution, movements must be performed focusing on achieving the maximal execution speed possible. Explosive muscle contractions are expected to elicit a greater osteogenic stimulus.[155] 4. Conclusions The research completed to date indicates that participation in high impact sports, especially prior to puberty, is important for maximizing bone mass accumulation and achieving a greater peak bone mass independent of sex. The effects of loading appear to be limited to the loaded bones. Starting the exercise before puberty has an additional benefit, since exercise elicits geometrical changes in bone, which in turn enhance mechanical competence. Continuing sport practice is associated with fewer bone fragility fractures in old age in both men and women. Several training methods have been used to improve BMD and content in prospective studies. Not all exercise modalities have positive effects on bone mass. For example, unloaded ª 2009 Adis Data Information BV. All rights reserved.

exercise, like swimming and cycling, has no impact on bone mass, while walking or running has limited positive effects. It is not clear which is the best training method for enhancing bone mass, although scientific evidence points to a combination of high impact exercises (i.e. jumping) with weight-lifting exercises. High impact exercise, even a limited amount, appears to be the most efficient to enhance bone mass except in postmenopausal women. Several types of resistance exercise have been tested with positive results when the intensity of the exercise was high and the speed of movement elevated. Resistance training is positively associated with high BMD in both young people and adults, and the effect of resistive exercise is relatively site specific to the working muscles and the bones to which they attach. However, more studies are needed to establish whether there are sex differences in the bone response to training. Although aerobic exercise and weight-bearing physical activity are important in maintaining overall health and healthy bones, resistance exercise has been shown to have a more potent effect on bone density. Studies performed in older adults show a sex discrepancy. Older men respond better to osteogenic training protocols than their female counterparts, although randomized longitudinal studies on the effects of exercise on bone mass in the elderly are still lacking. Old women show only mild increases or just a maintenance or attenuation of BMD losses. It remains to be determined if old women need a different exercise protocol to men of similar age. Impact and resistance exercise should be advocated for the prevention of osteoporosis. For those with osteoporosis, weight-bearing exercise in general, and resistance exercise in particular, as tolerated, along with exercise targeted to improve balance, mobility and posture, should be recommended to reduce the likelihood of falling and its associated morbidity and mortality. There is certainly a need for additional randomized, controlled trials in this research area, which will allow development of criteria for appropriate training loads according to age, sex, actual bone mass and past training history. Sports Med 2009; 39 (6)

Exercise and Bone Mass in Adults

463

Acknowledgements The study was supported financially by Ministerio de Educacio´n y Ciencia (DEP2006-56076-C06-04/ACTI). Consejerı´ a de Educacio´n, Cultura y Deportes del Gobierno de Canarias (2006/179 0001 and FEDER). Borja Guerra is a fellow of the ‘Recursos Humanos y Difusio´n de la Investigacio´n’ Program (ISCIII, MSC, Spain). The authors thank Jose´ Navarro de Tuero for his excellent technical assistance. The specialized advice from Tony Webster in editing the English version of the manuscript is also acknowledged. The authors have no conflicts of interest that are directly relevant to the content of this review.

References 1. Bouxsein ML, Karasik D. Bone geometry and skeletal fragility. Curr Osteoporos Rep 2006; 4 (2): 49-56 2. Viguet-Carrin S, Garnero P, Delmas PD. The role of collagen in bone strength. Osteoporos Int 2006; 17 (3): 319-36 3. Seeman E, Delmas PD. Bone quality: the material and structural basis of bone strength and fragility. N Engl J Med 2006; 354 (21): 2250-61 4. Ottani V, Raspanti M, Ruggeri A. Collagen structure and functional implications. Micron 2001; 32 (3): 251-60 5. Currey JD. Changes in the impact energy absorption of bone with age. J Biomech 1979; 12 (6): 459-69 6. Currey JD. The effect of porosity and mineral content on the Young’s modulus of elasticity of compact bone. J Biomech 1988; 21 (2): 131-9 7. Currey JD, Brear K, Zioupos P. The effects of ageing and changes in mineral content in degrading the toughness of human femora. J Biomech 1996; 29 (2): 257-60 8. Bailey AJ, Wotton SF, Sims TJ, et al. Post-translational modifications in the collagen of human osteoporotic femoral head. Biochem Biophys Res Commun 1992; 185 (3): 801-5 9. Boskey AL, Wright TM, Blank RD. Collagen and bone strength. J Bone Miner Res 1999; 14 (3): 330-5 10. Wang X, Bank RA, TeKoppele JM, et al. The role of collagen in determining bone mechanical properties. J Orthop Res 2001; 19 (6): 1021-6 11. Currey JD. The mechanical consequences of variation in the mineral content of bone. J Biomech 1969; 2 (1): 1-11 12. Niyibizi C, Eyre DR. Structural characteristics of crosslinking sites in type V collagen of bone: chain specificities and heterotypic links to type I collagen. Eur J Biochem 1994; 224 (3): 943-50 13. Eyre DR, Dickson IR, Van Ness K. Collagen cross-linking in human bone and articular cartilage: age-related changes in the content of mature hydroxypyridinium residues. Biochem J 1988; 252 (2): 495-500 14. Fantner GE, Birkedal H, Kindt JH, et al. Influence of the degradation of the organic matrix on the microscopic fracture behavior of trabecular bone. Bone 2004; 35 (5): 1013-22 15. Banse X, Sims TJ, Bailey AJ. Mechanical properties of adult vertebral cancellous bone: correlation with collagen

ª 2009 Adis Data Information BV. All rights reserved.

16.

17. 18.

19.

20. 21.

22. 23. 24.

25. 26.

27.

28.

29. 30.

31. 32.

33. 34.

intermolecular cross-links. J Bone Miner Res 2002; 17 (9): 1621-8 Fantner GE, Hassenkam T, Kindt JH, et al. Sacrificial bonds and hidden length dissipate energy as mineralized fibrils separate during bone fracture. Nat Mater 2005; 4 (8): 612-6 Hert J, Fiala P, Petrtyl M. Osteon orientation of the diaphysis of the long bones in man. Bone 1994; 15 (3): 269-77 Martin RB, Boardman DL. The effects of collagen fiber orientation, porosity, density, and mineralization on bovine cortical bone bending properties. J Biomech 1993; 26 (9): 1047-54 Puustjarvi K, Nieminen J, Rasanen T, et al. Do more highly organized collagen fibrils increase bone mechanical strength in loss of mineral density after one-year running training? J Bone Miner Res 1999; 14 (3): 321-9 Harada S, Rodan GA. Control of osteoblast function and regulation of bone mass. Nature 2003; 423 (6937): 349-55 Pavy-Le Traon A, Heer M, Narici MV, et al. From space to earth: advances in human physiology from 20 years of bed rest studies (1986-2006). Eur J Appl Physiol 2007; 101 (2): 143-94 Chilibeck PD, Sale DG, Webber CE. Exercise and bone mineral density. Sports Med 1995; 19 (2): 103-22 Frost HM. Bone’s mechanostat: a 2003 update. Anat Rec A Discov Mol Cell Evol Biol 2003; 275 (2): 1081-101 Rittweger J. What is new in neuro-musculoskeletal interactions: mechanotransduction, microdamage and repair? J Musculoskel Neuron Interact 2007; 7 (2): 191-3 Rosen CJ. Bone remodeling, energy metabolism, and the molecular clock. Cell Metab 2008; 7 (1): 7-10 Lee KC, Lanyon LE. Mechanical loading influences bone mass through estrogen receptor alpha. Exerc Sport Sci Rev 2004; 32 (2): 64-8 Skerry TM. One mechanostat or many? Modifications of the site-specific response of bone to mechanical loading by nature and nurture. J Musculoskel Neuron Interact 2006; 6 (2): 122-7 Scott A, Khan KM, Duronio V, et al. Mechanotransduction in human bone: in vitro cellular physiology that underpins bone changes with exercise. Sports Med 2008; 38 (2): 139-60 Rubin J, Rubin C, Jacobs CR. Molecular pathways mediating mechanical signaling in bone. Gene 2006; 367: 1-16 Saxon LK, Lanyon LE. Assessment of the in vivo adaptive response to mechanical loading. Method Molec Biol 2008; 455: 307-22 Frost HM. Bone ‘‘mass’’ and the ‘‘mechanostat’’: a proposal. Anat Rec 1987; 219 (1): 1-9 Frost HM. Wolff’s law: an ‘MGS’ derivation of gamma in the Three-Way Rule for mechanically controlled lamellar bone modeling drifts. Bone Miner 1993; 22 (2): 117-27 Bailey CA, Brooke-Wavell K. Exercise for optimising peak bone mass in women. Proc Nutr Soc 2008; 67 (1): 9-18 Shaw J. Exercise for skeletal health and osteoporosis prevention: ACSM’s resource manual for guidelines for exercise testing and prescription. Philadelphia (PA): Lippincott Williams & Wilkins, 1998: 288-93

Sports Med 2009; 39 (6)

464

35. Frost HM. Skeletal structural adaptations to mechanical usage (SATMU): 1, redefining Wolff’s law – the bone modeling problem. Anat Rec 1990; 226 (4): 403-13 36. Burr DB, Forwood MR, Fyhrie DP, et al. Bone microdamage and skeletal fragility in osteoporotic and stress fractures. J Bone Miner Res 1997; 12 (1): 6-15 37. Rubin CT, Lanyon LE. Regulation of bone formation by applied dynamic loads. J Bone Joint Surg Am 1984; 66 (3): 397-402 38. Rubin C, Turner AS, Bain S, et al. Anabolism: low mechanical signals strengthen long bones. Nature 2001; 412 (6847): 603-4 39. Rubin C, Turner AS, Muller R, et al. Quantity and quality of trabecular bone in the femur are enhanced by a strongly anabolic, noninvasive mechanical intervention. J Bone Miner Res 2002; 17 (2): 349-57 40. Borer KT. Physical activity in the prevention and amelioration of osteoporosis in women: interaction of mechanical, hormonal and dietary factors. Sports Med 2005; 35 (9): 779-830 41. Nordstro¨m A, Karlsson C, Nyquist F, et al. Bone loss and fracture risk after reduced physical activity. J Bone Miner Res 2005; 20 (2): 202-7 42. WHO Study Group. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. World Health Organ Tech Rep Ser 1994; 843: 1-129 43. Cummings SR, Nevitt MC, Browner WS, et al. Risk factors for hip fracture in white women: Study of Osteoporotic Fractures Research Group. N Engl J Med 1995; 332 (12): 767-73 44. Kanis JA, Melton 3rd LJ, Christiansen C, et al. The diagnosis of osteoporosis. J Bone Miner Res 1994; 9 (8): 1137-41 45. Mein AL, Briffa NK, Dhaliwal SS, et al. Lifestyle influences on 9-year changes in BMD in young women. J Bone Miner Res 2004; 19 (7): 1092-8 46. Vicente-Rodriguez G. How does exercise affect bone development during growth? Sports Med 2006; 36 (7): 561-9 47. Vicente-Rodriguez G, Jimenez-Ramirez J, Ara I, et al. Enhanced bone mass and physical fitness in prepubescent footballers. Bone 2003; 33 (5): 853-9 48. Vicente-Rodriguez G, Dorado C, Perez-Gomez J, et al. Enhanced bone mass and physical fitness in young female handball players. Bone 2004; 35 (5): 1208-15 49. Vicente-Rodriguez G, Dorado C, Ara I, et al. Artistic versus rhythmic gymnastics: effects on bone and muscle mass in young girls. Int J Sports Med 2007; 28 (5): 386-93 50. McVeigh JA, Norris SA, Pettifor JM. Bone mass accretion rates in pre- and early-pubertal South African black and white children in relation to habitual physical activity and dietary calcium intakes. Acta Paediatr 2007; 96 (6): 874-80 51. Vicente-Rodriguez G, Ara I, Perez-Gomez J, et al. High femoral bone mineral density accretion in prepubertal soccer players. Med Sci Sports Exerc 2004; 36 (10): 1789-95 52. Wang Q, Alen M, Nicholson P, et al. Weight-bearing, muscle loading and bone mineral accrual in pubertal girls: a 2-year longitudinal study. Bone 2007; 40 (5): 1196-202

ª 2009 Adis Data Information BV. All rights reserved.

Guadalupe-Grau et al.

53. Bradney M, Pearce G, Naughton G, et al. Moderate exercise during growth in prepubertal boys: changes in bone mass, size, volumetric density, and bone strength: a controlled prospective study. J Bone Miner Res 1998; 13 (12): 1814-21 54. Bass SL. The prepubertal years: a uniquely opportune stage of growth when the skeleton is most responsive to exercise? Sports Med 2000; 30 (2): 73-8 55. Bailey DA, McKay HA, Mirwald RL, et al. A six-year longitudinal study of the relationship of physical activity to bone mineral accrual in growing children: the university of Saskatchewan bone mineral accrual study. J Bone Miner Res 1999; 14 (10): 1672-9 56. Greene DA, Naughton GA. Adaptive skeletal responses to mechanical loading during adolescence. Sports Med 2006; 36 (9): 723-32 57. Bass SL, Saxon L, Daly RM, et al. The effect of mechanical loading on the size and shape of bone in pre-, peri-, and postpubertal girls: a study in tennis players. J Bone Miner Res 2002; 17 (12): 2274-80 58. Seeman E. An exercise in geometry. J Bone Miner Res 2002; 17 (3): 373-80 59. Kannus P, Haapasalo H, Sankelo M, et al. Effect of starting age of physical activity on bone mass in the dominant arm of tennis and squash players. Ann Intern Med 1995; 123 (1): 27-31 60. Heinonen A, Sievanen H, Kannus P, et al. High-impact exercise and bones of growing girls: a 9-month controlled trial. Osteoporos Int 2000; 11 (12): 1010-7 61. Manias K, McCabe D, Bishop N. Fractures and recurrent fractures in children; varying effects of environmental factors as well as bone size and mass. Bone 2006; 39 (3): 652-7 62. Laing EM, Wilson AR, Modlesky CM, et al. Initial years of recreational artistic gymnastics training improves lumbar spine bone mineral accrual in 4- to 8-year-old females. J Bone Miner Res 2005; 20 (3): 509-19 63. Lanyon LE. Using functional loading to influence bone mass and architecture: objectives, mechanisms, and relationship with estrogen of the mechanically adaptive process in bone. Bone 1996; 18 (1 Suppl.): 37S-43S 64. Martyn-St James M, Carroll S. Progressive high-intensity resistance training and bone mineral density changes among premenopausal women: evidence of discordant site-specific skeletal effects. Sports Med 2006; 36 (8): 683-704 65. Forwood MR, Turner CH. The response of rat tibiae to incremental bouts of mechanical loading: a quantum concept for bone formation. Bone 1994; 15 (6): 603-9 66. Pazzaglia UE, Andrini L, Di Nucci A. The effects of mechanical forces on bones and joints: experimental study on the rat tail. J Bone Joint Surg Br 1997; 79 (6): 1024-30 67. Robling AG, Duijvelaar KM, Geevers JV, et al. Modulation of appositional and longitudinal bone growth in the rat ulna by applied static and dynamic force. Bone 2001; 29 (2): 105-13 68. Ehrlich PJ, Lanyon LE. Mechanical strain and bone cell function: a review. Osteoporos Int 2002; 13 (9): 688-700

Sports Med 2009; 39 (6)

Exercise and Bone Mass in Adults

69. Pedersen BK, Hoffman-Goetz L. Exercise and the immune system: regulation, integration, and adaptation. Physiol Rev 2000; 80 (3): 1055-81 70. Van Hall G, Jensen-Urstad M, Rosdahl H, et al. Leg and arm lactate and substrate kinetics during exercise. Am J Physiol Endocrinol Metab 2003; 284 (1): E193-205 71. Mastorakos G, Pavlatou M. Exercise as a stress model and the interplay between the hypothalamus-pituitary-adrenal and the hypothalamus-pituitary-thyroid axes. Horm Metab Res 2005; 37 (9): 577-84 72. Calbet JA, Lundby C, Sander M, et al. Effects of ATPinduced leg vasodilation on VO2 peak and leg O2 extraction during maximal exercise in humans. Am J Physiol Regul Integr Comp Physiol 2006; 291 (2): R447-53 73. Ara I, Perez-Gomez J, Vicente-Rodriguez G, et al. Serum free testosterone, leptin and soluble leptin receptor changes in a 6-week strength-training programme. Br J Nutr 2006; 96 (6): 1053-9 74. Frost HM, Jee WS. On the rat model of human osteopenias and osteoporoses. Bone Miner 1992; 18 (3): 227-36 75. Bourrin S, Palle S, Pupier R, et al. Effect of physical training on bone adaptation in three zones of the rat tibia. J Bone Miner Res 1995; 10 (11): 1745-52 76. Yeh JK, Liu CC, Aloia JF. Effects of exercise and immobilization on bone formation and resorption in young rats. Am J Physiol 1993; 264 (2 Pt 1): E182-9 77. Hagihara Y, Fukuda S, Goto S, et al. How many days per week should rats undergo running exercise to increase BMD? J Bone Miner Metab 2005; 23 (4): 289-94 78. Barengolts EI, Curry DJ, Bapna MS, et al. Effects of endurance exercise on bone mass and mechanical properties in intact and ovariectomized rats. J Bone Miner Res 1993; 8 (8): 937-42 79. Horcajada M, Coxam V, Davicco M, et al. Influence of treadmill running on femoral bone in young orchidectomized rats. J Appl Physiol 1997; 83 (1): 129-33 80. van der Wiel HE, Lips P, Graafmans WC, et al. Additional weight-bearing during exercise is more important than duration of exercise for anabolic stimulus of bone: a study of running exercise in female rats. Bone 1995; 16 (1): 73-80 81. Bourrin S, Genty C, Palle S, et al. Adverse effects of strenuous exercise: a densitometric and histomorphometric study in the rat. J Appl Physiol 1994; 76 (5): 1999-2005 82. Iwamoto J, Takeda T, Sato Y. Effect of treadmill exercise on bone mass in female rats. Exp Anim 2005; 54 (1): 1-6 83. Wronski TJ, Yen CF. The ovariectomized rat as an animalmodel for postmenopausal bone loss. Cell Mater Suppl 1991; 1: 69-74 84. Jarvinen TL, Kannus P, Pajamaki I, et al. Estrogen deposits extra mineral into bones of female rats in puberty, but simultaneously seems to suppress the responsiveness of female skeleton to mechanical loading. Bone 2003; 32 (6): 642-51 85. Sanchis-Moysi J, Dorado C, Vicente-Rodriguez G, et al. Inter-arm asymmetry in bone mineral content and bone area in postmenopausal recreational tennis players. Maturitas 2004; 48 (3): 289-98

ª 2009 Adis Data Information BV. All rights reserved.

465

86. Calbet JA, Dorado C, Diaz-Herrera P, et al. High femoral bone mineral content and density in male football (soccer) players. Med Sci Sports Exerc 2001; 33 (10): 1682-7 87. Morel J, Combe B, Francisco J, et al. Bone mineral density of 704 amateur sportsmen involved in different physical activities. Osteoporos Int 2001; 12 (2): 152-7 88. Calbet JA, Moysi JS, Dorado C, et al. Bone mineral content and density in professional tennis players. Calcif Tissue Int 1998; 62 (6): 491-6 89. Wittich A, Mautalen CA, Oliveri MB, et al. Professional football (soccer) players have a markedly greater skeletal mineral content, density and size than age- and BMImatched controls. Calcif Tissue Int 1998; 63 (2): 112-7 90. Egan E, Reilly T, Giacomoni M, et al. Bone mineral density among female sports participants. Bone 2006; 38 (2): 227-33 91. Nichols JF, Rauh MJ, Barrack MT, et al. Bone mineral density in female high school athletes: interactions of menstrual function and type of mechanical loading. Bone 2007; 41 (3): 371-7 92. Alfredson H, Nordstro¨m P, Lorentzon R. Total and regional bone mass in female soccer players. Calcif Tissue Int 1996; 59 (6): 438-42 93. Ducher G, Prouteau S, Courteix D, et al. Cortical and trabecular bone at the forearm show different adaptation patterns in response to tennis playing. J Clin Densitom 2004; 7 (4): 399-405 94. Karlsson MK. Skeletal effects of exercise in men. Calcif Tissue Int 2001; 69 (4): 196-9 95. Layne JE, Nelson ME. The effects of progressive resistance training on bone density: a review. Med Sci Sports Exerc 1999; 31 (1): 25-30 96. Nichols DL, Sanborn CF, Essery EV. Bone density and young athletic women: an update. Sports Med 2007; 37 (11): 1001-14 97. Karlsson M. Has exercise an antifracture efficacy in women? Scand J Med Sci Sports 2004; 14 (1): 2-15 98. Daly RM, Rich PA, Klein R, et al. Effects of high-impact exercise on ultrasonic and biochemical indices of skeletal status: a prospective study in young male gymnasts. J Bone Miner Res 1999; 14 (7): 1222-30 99. Taaffe DR, Snow-Harter C, Connolly DA, et al. Differential effects of swimming versus weight-bearing activity on bone mineral status of eumenorrheic athletes. J Bone Miner Res 1995; 10 (4): 586-93 100. Hawkey A. The importance of exercising in space. Interdiscip Sci Rev 2003; 28 (2): 130-8 101. Perez-Gomez J, Olmedillas H, Delgado-Guerra S, et al. Effects of weight lifting training combined with plyometric exercises on physical fitness, body composition, and knee extension velocity during kicking in football. Appl Physiol Nutr Metab 2008; 33 (3): 501-10 102. Kellis E, Katis A, Vrabas IS. Effects of an intermittent exercise fatigue protocol on biomechanics of soccer kick performance. Scand J Med Sci Sports 2006; 16 (5): 334-44 103. Hetland ML, Haarbo J, Christiansen C. Low bone mass and high bone turnover in male long distance runners. J Clin Endocrinol Metab 1993; 77 (3): 770-5

Sports Med 2009; 39 (6)

466

104. Kannus P, Haapasalo H, Sievanen H, et al. The site-specific effects of long-term unilateral activity on bone mineral density and content. Bone 1994; 15 (3): 279-84 105. Haapasalo H, Kannus P, Sievanen H, et al. Long-term unilateral loading and bone mineral density and content in female squash players. Calcif Tissue Int 1994; 54 (4): 249-55 106. Ducher G, Tournaire N, Meddahi-Pelle A, et al. Shortterm and long-term site-specific effects of tennis playing on trabecular and cortical bone at the distal radius. J Bone Miner Metab 2006; 24 (6): 484-90 107. Vicente-Rodriguez G. How does exercise affect bone development during growth? Sports Med 2006; 36 (7): 561-9 108. Calbet JA, Dorado C, Diaz-Herrera P, et al. High femoral bone mineral content and density in male football (soccer) players. Med Sci Sports Exerc 2001; 33 (10): 1682-7 109. Egan E, Reilly T, Giacomoni M, et al. Bone mineral density among female sports participants. Bone 2006; 38 (2): 227-33 110. Ng MY, Sham PC, Paterson AD, et al. Effect of environmental factors and gender on the heritability of bone mineral density and bone size. Ann Hum Genet 2006; 70 (Pt 4): 428-38 111. Kemper HC, Post GB, Twisk JW, et al. Lifestyle and obesity in adolescence and young adulthood: results from the Amsterdam Growth And Health Longitudinal Study (AGAHLS). Int J Obes Relat Metab Disord 1999; 23 Suppl. 3: S34-40 112. Armstrong N, Welsman JR. The physical activity patterns of European youth with reference to methods of assessment. Sports Med 2006; 36 (12): 1067-86 113. Tanner JM. Principles of growth standards. Acta Paediatr Scand 1990; 79 (10): 963-7 114. Friedlander AL, Genant HK, Sadowsky S, et al. A twoyear program of aerobics and weight training enhances bone mineral density of young women. J Bone Miner Res 1995; 10 (4): 574-85 115. Schroeder ET, Hawkins SA, Jaque SV. Musculoskeletal adaptations to 16 weeks of eccentric progressive resistance training in young women. J Strength Cond Res 2004; 18 (2): 227-35 116. Hawkins SA, Schroeder ET, Wiswell RA, et al. Eccentric muscle action increases site-specific osteogenic response. Med Sci Sports Exerc 1999; 31 (9): 1287-92 117. Nickols-Richardson SM, Miller LE, Wootten DF, et al. Concentric and eccentric isokinetic resistance training similarly increases muscular strength, fat-free soft tissue mass, and specific bone mineral measurements in young women. Osteoporos Int 2007; 18 (6): 789-96 118. Kato T, Terashima T, Yamashita T, et al. Effect of lowrepetition jump training on bone mineral density in young women. J Appl Physiol 2006; 100 (3): 839-43 119. Bassey EJ, Ramsdale SJ. Increase in femoral bone density in young women following high-impact exercise. Osteoporos Int 1994; 4 (2): 72-5 120. Sinaki M, Wahner HW, Bergstralh EJ, et al. Three-year controlled, randomized trial of the effect of dose-specified loading and strengthening exercises on bone mineral density of spine and femur in nonathletic, physically active women. Bone 1996; 19 (3): 233-44

ª 2009 Adis Data Information BV. All rights reserved.

Guadalupe-Grau et al.

121. Chilibeck PD, Calder A, Sale DG, et al. Twenty weeks of weight training increases lean tissue mass but not bone mineral mass or density in healthy, active young women. Can J Physiol Pharmacol 1996; 74 (10): 1180-5 122. Nindl BC, Harman EA, Marx JO, et al. Regional body composition changes in women after 6 months of periodized physical training. J Appl Physiol 2000; 88 (6): 2251-9 123. Lohman T, Going S, Pamenter R, et al. Effects of resistance training on regional and total bone mineral density in premenopausal women: a randomized prospective study. J Bone Miner Res 1995; 10 (7): 1015-24 124. Vainionpa¨a¨ A, Korpelainen R, Leppaluoto J, et al. Effects of high-impact exercise on bone mineral density: a randomized controlled trial in premenopausal women. Osteoporos Int 2005; 16 (2): 191-7 125. Winters-Stone KM, Snow CM. Site-specific response of bone to exercise in premenopausal women. Bone 2006; 39 (6): 1203-9 126. Heinonen A, Kannus P, Sievanen H, et al. Randomised controlled trial of effect of high-impact exercise on selected risk factors for osteoporotic fractures. Lancet 1996; 348 (9038): 1343-7 127. Kohrt WM, Bloomfield SA, Little KD, et al. American College of Sports Medicine Position Stand: physical activity and bone health. Med Sci Sports Exerc 2004; 36 (11): 1985-96 128. Deschenes MR, Kraemer WJ. Performance and physiologic adaptations to resistance training. Am J Phys Med Rehabil 2002; 81 (11 Suppl.): S3-16 129. Kraemer WJ, Ratamess NA. Fundamentals of resistance training: progression and exercise prescription. Med Sci Sports Exerc 2004; 36 (4): 674-88 130. Kraemer WJ, Mazzetti SA, Nindl BC, et al. Effect of resistance training on women’s strength/power and occupational performances. Med Sci Sports Exerc 2001; 33 (6): 1011-25 131. Kraemer WJ, Nindl BC, Ratamess NA, et al. Changes in muscle hypertrophy in women with periodized resistance training. Med Sci Sports Exerc 2004; 36 (4): 697-708 132. Uusi-Rasi K, Sievanen H, Pasanen M, et al. Association of physical activity and calcium intake with the maintenance of bone mass in premenopausal women. Osteoporos Int 2002; 13 (3): 211-7 133. Snow-Harter C, Bouxsein ML, Lewis BT, et al. Effects of resistance and endurance exercise on bone mineral status of young women: a randomized exercise intervention trial. J Bone Miner Res 1992; 7 (7): 761-9 134. Mayhew TP, Rothstein JM, Finucane SD, et al. Muscular adaptation to concentric and eccentric exercise at equal power levels. Med Sci Sports Exerc 1995; 27 (6): 868-73 135. Gilsanz V, Wren TA, Sanchez M, et al. Low-level, highfrequency mechanical signals enhance musculoskeletal development of young women with low BMD. J Bone Miner Res 2006; 21 (9): 1464-74 136. Fujimura R, Ashizawa N, Watanabe M, et al. Effect of resistance exercise training on bone formation and resorption in young male subjects assessed by biomarkers of bone metabolism. J Bone Miner Res 1997; 12 (4): 656-62 137. Hartman JW, Tang JE, Wilkinson SB, et al. Consumption of fat-free fluid milk after resistance exercise promotes

Sports Med 2009; 39 (6)

Exercise and Bone Mass in Adults

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

149.

150.

151.

152.

greater lean mass accretion than does consumption of soy or carbohydrate in young, novice, male weightlifters. Am J Clin Nutr 2007; 86 (2): 373-81 Ballard TL, Specker BL, Binkley TL, et al. Effect of protein supplementation during a 6-month strength and conditioning program on areal and volumetric bone parameters. Bone 2006; 38 (6): 898-904 Ryan AS, Ivey FM, Hurlbut DE, et al. Regional bone mineral density after resistive training in young and older men and women. Scand J Med Sci Sports 2004; 14 (1): 16-23 Mullins NM, Sinning WE. Effects of resistance training and protein supplementation on bone turnover in young adult women. Nutr Metab (Lond) 2005; 2: 19 Torvinen S, Kannus P, Sievanen H, et al. Effect of 8-month vertical whole body vibration on bone, muscle performance, and body balance: a randomized controlled study. J Bone Miner Res 2003; 18 (5): 876-84 Beck BR, Kent K, Holloway L, et al. Novel, highfrequency, low-strain mechanical loading for premenopausal women with low bone mass: early findings. J Bone Mineral Metab 2006; 24 (6): 505-7 Menkes A, Mazel S, Redmond RA, et al. Strength training increases regional bone mineral density and bone remodeling in middle-aged and older men. J Appl Physiol 1993; 74 (5): 2478-84 Huuskonen J, Vaisanen SB, Kroger H, et al. Regular physical exercise and bone mineral density: a four-year controlled randomized trial in middle-aged men – the DNASCO study. Osteoporos Int 2001; 12 (5): 349-55 Stewart KJ, Bacher AC, Hees PS, et al. Exercise effects on bone mineral density relationships to changes in fitness and fatness. Am J Prev Med 2005; 28 (5): 453-60 Madeo B, Zirilli L, Caffagni G, et al. The osteoporotic male: overlooked and undermanaged? Clin Interv Aging 2007; 2 (3): 305-12 Hui SL, Slemenda CW, Johnston Jr CC, et al. The contribution of bone loss to postmenopausal osteoporosis. Osteoporos Int 1990; 1 (1): 30-4 Jones HH, Priest JD, Hayes WC, et al. Humeral hypertrophy in response to exercise. J Bone Joint Surg Am 1977; 59 (2): 204-8 Province MA, Hadley EC, Hornbrook MC, et al. The effects of exercise on falls in elderly patients: a preplanned meta-analysis of the FICSIT Trials. Frailty and injuries: cooperative studies of intervention techniques. JAMA 1995; 273 (17): 1341-7 Paganini-Hill A, Chao A, Ross RK, et al. Exercise and other factors in the prevention of hip fracture: the Leisure World study. Epidemiology 1991; 2 (1): 16-25 Wyshak G, Frisch RE, Albright TE, et al. Bone fractures among former college athletes compared with nonathletes in the menopausal and postmenopausal years. Obstet Gynecol 1987; 69 (1): 121-6 Nelson ME, Fiatarone MA, Morganti CM, et al. Effects of high-intensity strength training on multiple risk factors for osteoporotic fractures: a randomized controlled trial. JAMA 1994; 272 (24): 1909-14

ª 2009 Adis Data Information BV. All rights reserved.

467

153. Kerr D, Morton A, Dick I, et al. Exercise effects on bone mass in postmenopausal women are site-specific and loaddependent. J Bone Miner Res 1996; 11 (2): 218-25 154. Kohrt WM, Snead DB, Slatopolsky E, et al. Additive effects of weight-bearing exercise and estrogen on bone mineral density in older women. J Bone Miner Res 1995; 10 (9): 1303-11 155. Stengel SV, Kemmler W, Pintag R, et al. Power training is more effective than strength training for maintaining bone mineral density in postmenopausal women. J Appl Physiol 2005; 99 (1): 181-8 156. Chien MY, Wu YT, Hsu AT, et al. Efficacy of a 24-week aerobic exercise program for osteopenic postmenopausal women. Calcif Tissue Int 2000; 67 (6): 443-8 157. Kohrt WM, Ehsani AA, Birge Jr SJ, et al. Effects of exercise involving predominantly either joint-reaction or ground-reaction forces on bone mineral density in older women. J Bone Miner Res 1997; 12 (8): 1253-61 158. Verschueren SM, Roelants M, Delecluse C, et al. Effect of 6-month whole body vibration training on hip density, muscle strength, and postural control in postmenopausal women: a randomized controlled pilot study. J Bone Miner Res 2004; 19 (3): 352-9 159. Pruitt LA, Jackson RD, Bartels RL, et al. Weight-training effects on bone mineral density in early postmenopausal women. J Bone Miner Res 1992; 7 (2): 179-85 160. Maddalozzo GF, Snow CM. High intensity resistance training: effects on bone in older men and women. Calcif Tissue Int 2000; 66 (6): 399-404 161. Ryan AS, Treuth MS, Rubin MA, et al. Effects of strength training on bone mineral density: hormonal and bone turnover relationships. J Appl Physiol 1994; 77 (4): 1678-84 162. Ryan AS, Treuth MS, Hunter GR, et al. Resistive training maintains bone mineral density in postmenopausal women. Calcif Tissue Int 1998; 62 (4): 295-9 163. Bassey EJ, Rothwell MC, Littlewood JJ, et al. Pre- and postmenopausal women have different bone mineral density responses to the same high-impact exercise. J Bone Miner Res 1998; 13 (12): 1805-13 164. Sugiyama T, Yamaguchi A, Kawai S. Effects of skeletal loading on bone mass and compensation mechanism in bone: a new insight into the ‘‘mechanostat’’ theory. J Bone Miner Metab 2002; 20 (4): 196-200 165. Palombaro KM. Effects of walking-only interventions on bone mineral density at various skeletal sites: a metaanalysis. J Geriatr Phys Ther 2005; 28 (3): 102-7 166. Bergmann G, Graichen F, Rohlmann A. Hip joint loading during walking and running, measured in two patients. J Biomech 1993; 26 (8): 969-90 167. Gusi N, Raimundo A, Leal A. Low-frequency vibratory exercise reduces the risk of bone fracture more than walking: a randomized controlled trial. BMC Musculoskel Disord 2006; 7: 92 168. Rubin C, Recker R, Cullen D, et al. Prevention of postmenopausal bone loss by a low-magnitude, highfrequency mechanical stimuli: a clinical trial assessing compliance, efficacy, and safety. J Bone Miner Res 2004; 19 (3): 343-51

Sports Med 2009; 39 (6)

468

169. Zehnacker CH, Bemis-Dougherty A. Effect of weighted exercises on bone mineral density in post menopausal women: a systematic review. J Geriatr Phys Ther 2007; 30 (2): 79-88 170. Englund U, Littbrand H, Sondell A, et al. The beneficial effects of exercise on BMD are lost after cessation: a 5-year follow-up in older post-menopausal women. Scand J Med Sci Sports. Epub 2008 May 22 171. Mackey DC, Lui LY, Cawthon PM, et al. High-trauma fractures and low bone mineral density in older women and men. JAMA 2007; 298 (20): 2381-8

ª 2009 Adis Data Information BV. All rights reserved.

Guadalupe-Grau et al.

172. Frost HM. Some effects of basic multicellular unit-based remodelling on photon absorptiometry of trabecular bone. Bone Miner 1989; 7 (1): 47-65 173. Forwood MR, Burr DB. Physical activity and bone mass: exercises in futility? Bone Miner 1993; 21 (2): 89-112

Correspondence: Prof. Jose A.L. Calbet, Departamento de Educacio´n Fı´sica, Campus Universitario de Tafira, 35017 Las Palmas de Gran Canaria, Canary Islands, Spain. E-mail: [email protected]

Sports Med 2009; 39 (6)

Sports Med 2009; 39 (6): 469-490 0112-1642/09/0006-0469/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Lactate Threshold Concepts How Valid are They? Oliver Faude,1,2 Wilfried Kindermann2 and Tim Meyer1,2 1 Institute of Sports Medicine, University Paderborn, Paderborn, Germany 2 Institute of Sports and Preventive Medicine, University of Saarland, Saarbru¨cken, Germany

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Historical Remarks on Endurance Performance Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Incremental Exercise Testing and the Interpretation of Blood Lactate Curves . . . . . . . . . . . . . . . . . . . 2.1 The Entire Blood Lactate Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Test Design and Data Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 Methodology of Blood Lactate Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 A Framework for Endurance Diagnosis and Training Prescriptions. . . . . . . . . . . . . . . . . . . . . . . . . . 3. Validation of Lactate Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Competition Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 The Maximal Lactate Steady State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Lactate Threshold Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Located Lactate Threshold Concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.1 Aerobic Lactate Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.2 Anaerobic Lactate Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Lactate Thresholds and (Simulated) Competition Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Lactate Thresholds and Maximal Lactate Steady State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Conclusions and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

469 470 471 471 472 472 473 474 474 474 475 475 475 476 477 480 484

During the last nearly 50 years, the blood lactate curve and lactate thresholds (LTs) have become important in the diagnosis of endurance performance. An intense and ongoing debate emerged, which was mainly based on terminology and/or the physiological background of LT concepts. The present review aims at evaluating LTs with regard to their validity in assessing endurance capacity. Additionally, LT concepts shall be integrated within the ‘aerobic-anaerobic transition’ – a framework which has often been used for performance diagnosis and intensity prescriptions in endurance sports. Usually, graded incremental exercise tests, eliciting an exponential rise in blood lactate concentrations (bLa), are used to arrive at lactate curves. A shift of such lactate curves indicates changes in endurance capacity. This very global approach, however, is hindered by several factors that may influence overall lactate levels. In addition, the exclusive use of the entire curve leads to some uncertainty as to the magnitude of endurance gains, which cannot be precisely estimated. This deficiency might be eliminated by the use of LTs. The aerobic-anaerobic transition may serve as a basis for individually assessing endurance performance as well as for prescribing intensities in

Faude et al.

470

endurance training. Additionally, several LT approaches may be integrated in this framework. This model consists of two typical breakpoints that are passed during incremental exercise: the intensity at which bLa begin to rise above baseline levels and the highest intensity at which lactate production and elimination are in equilibrium (maximal lactate steady state [MLSS]). Within this review, LTs are considered valid performance indicators when there are strong linear correlations with (simulated) endurance performance. In addition, a close relationship between LT and MLSS indicates validity regarding the prescription of training intensities. A total of 25 different LT concepts were located. All concepts were divided into three categories. Several authors use fixed bLa during incremental exercise to assess endurance performance (category 1). Other LT concepts aim at detecting the first rise in bLa above baseline levels (category 2). The third category consists of threshold concepts that aim at detecting either the MLSS or a rapid/distinct change in the inclination of the blood lactate curve (category 3). Thirty-two studies evaluated the relationship of LTs with performance in (partly simulated) endurance events. The overwhelming majority of those studies reported strong linear correlations, particularly for running events, suggesting a high percentage of common variance between LT and endurance performance. In addition, there is evidence that some LTs can estimate the MLSS. However, from a practical and statistical point of view it would be of interest to know the variability of individual differences between the respective threshold and the MLSS, which is rarely reported. Although there has been frequent and controversial debate on the LT phenomenon during the last three decades, many scientific studies have dealt with LT concepts, their value in assessing endurance performance or in prescribing exercise intensities in endurance training. The presented framework may help to clarify some aspects of the controversy and may give a rationale for performance diagnosis and training prescription in future research as well as in sports practice.

1. Historical Remarks on Endurance Performance Diagnosis As early as 1808, Berzelius observed that lactic acid was produced in the muscles of hunted stags.[1] About a century later, several scientists studied the biochemistry of energy metabolism and muscle contraction in more detail. This led to a much deeper understanding of the formation of lactic acid (lactate and hydrogen ions) during intense exercise.[2-5] At that time, it was common belief that lactic acid is a waste product of glycolysis and will be formed when oxygen delivery to exercising muscles is not sufficient and muscle anaerobiosis occurs.[2,6,7] This view has been challenged considerably during the last two decª 2009 Adis Data Information BV. All rights reserved.

ades. Anaerobic glycolysis and, thus, lactate kinetics rather seem to be an ongoing process – even in the resting individual – which is highly related to the metabolic rate but not necessarily to oxygen availability (for detailed review see Gladden,[1,8] Brooks,[9] Robergs et al.[10]). In the first half of the 20th century the concept of maximum oxygen consumption as the first and probably most common means of evaluating aerobic endurance capacity was developed by the working group of Nobel Laureate AV Hill.[6]
maximal oxygen uptake (VO2max) has been established as a valuable tool to distinguish between fit and unfit subjects. However, several concerns were raised regarding the sensitivity of
VO2max. For instance, it is difficult to discriminate Sports Med 2009; 39 (6)

Validity of Lactate Thresholds

between subjects of homogenous performance
levels by means of VO2max.[11-18] In addition, sufficient effort during whole-body work and, therefore, adequate motivation of the investigated subject is necessary to appropriately de
termine VO2max. Particularly in clinical settings with diseased patients, whole-body exhaustion is difficult to attain or is even avoided because of the risk of adverse events.[19,20] Therefore, attempts have been made to establish sub-maximal parameters to assess cardiorespiratory fitness in patients and athletes. Early research by the working group of Hollmann established the so-called ‘point of optimum ventilatory efficiency’ corresponding to the first increase in the ventilatory equivalent of oxygen and of arterial lactate concentrations during incremental exercise.[19,21] A few years later, Wasserman and McIllroy[22] determined this intensity by plotting ventilation versus oxygen uptake in cardiac patients and named it the ‘anaerobic threshold’ (LTAn). At that time, routine determination of blood lactate concentrations (bLa) was associated with several difficulties and gas exchange measurements were more common – especially in clinical settings. Therefore, it became popular to detect the LTAn by means of gas exchange analysis. In the 1960s, the enzymatic method for measuring lactate concentrations from capillary blood samples was developed. This led to the increasing popularity of using bLa as a parameter to assess endurance capacity as well as for classifying work rate during exercise.[19,23,24] In the following years, numerous lactate threshold (LT) concepts were developed. The number of scientific studies on LTs has increased enormously up to now and the sub-maximal course of bLa during incremental exercise has probably become one of the most important means in the diagnosis of endurance performance in sports practice.[15,16,25,26] However, the variety of different threshold concepts has led to considerable confusion and misinterpretation. An intense and ongoing debate emerged, which was mainly based upon terminology and/or the physiological background of LT concepts.[27] Early assumptions on lactate producª 2009 Adis Data Information BV. All rights reserved.

471

tion and distribution in the organism have been challenged.[1,8-10,28] It has been argued that bLa increase continuously rather than show a clear threshold during incremental exercise. Furthermore, the contribution of aerobic and anaerobic pathways to energy production does not change suddenly but shows a continuous transition and, therefore, the term ‘threshold’ might be misleading.[29] Against this background and to unravel the confusion, it seems valuable to give a summary on published LT concepts. The present review is mainly aimed at evaluating the located LT concepts with regard to their validity in assessing aerobic endurance capacity and prescribing training intensity. A further aim was to try to integrate those concepts into a framework that was originally called the aerobic-anaerobic transition.[30-32] It has to be emphasized that this text focuses on LTs only. Although a close link between lactate and gas exchange markers has often been proposed,[21,31,33-36] there is still controversial debate with regard to the underlying physiological mechanisms.[37] A comprehensive review on gas exchange thresholds has recently been published.[31] Additionally, it is not within the scope of this article to exhaustively review the biochemistry of glycolysis and lactate metabolism. 2. Incremental Exercise Testing and the Interpretation of Blood Lactate Curves 2.1 The Entire Blood Lactate Curve

Usually, graded incremental exercise tests (GXTs) are used to evaluate aerobic endurance performance capacity. Typically, an exponential rise in bLa during incremental exercise testing can be observed (figure 1). The issue of interest is to interpret the resulting lactate curve with regard to endurance capacity. It is generally accepted that a rightward shift of the lactate curve (lower bLa at given workload) can be interpreted in terms of an improved endurance capacity[38-40] and, in contrast, a shift to the left (higher bLa at given workload) is usually considered to represent worsening endurance capacity.[41] Sports Med 2009; 39 (6)

Faude et al.

Blood lactate concentration (mmol/L)

472

10 8

Regenerative/ Moderate-/ Interval low-intensity high-intensity training endurance endurance sessions training training

6 Aerobicanaerobic transition

4 Aerobic threshold

MLSS = anaerobic threshold

2 0 Work intensity

Fig. 1. A typical lactate-workload plot including the aerobicanaerobic transition as a framework to derive endurance training intensities for different intensity zones. MLSS = maximal lactate steady state.

Overall lactate levels are known to be influenced by depleted glycogen stores (due to a low carbohydrate diet or preceding exhaustive exercise).[42-44] For instance, lower bLa at the same work rates have been reported in a glycogendepleted subject compared with a subject in normal condition. This may lead to a downward shift of the lactate curve and it is important that this is not falsely interpreted as an enhancement in endurance capacity.[45] Furthermore, several other factors like muscle fibre composition, glycolytic and lipolytic enzyme activity as well as capillary or mitochondrial density might influence blood lactate curves.[46] Additionally, the entire lactate curve is dependent on several other methodological issues, which should be taken into account when interpreting test results. 2.1.1 Test Design and Data Treatment

It is of note that the specific GXT protocol can vary considerably with regard to starting and subsequent work rates, work rate increments and stage duration. A recent review focused on the influence of varying test protocols on markers usually used in the diagnosis of endurance performance.[47] For instance, varying stage duration or work rate increments may lead to relevant differences in blood lactate curves and LTs.[48-50] A possible reason might be the time allowed for ª 2009 Adis Data Information BV. All rights reserved.

lactate diffusion in the blood until the next work rate increment.[47] In addition, there has been great debate on the best fitting procedure for the obtained bLa data set. For instance, a single-[51] or double-phase model[52] using two or three linear regression segments, a double-log model,[53] a third-order polymonial[54] or an exponential function[55] have been used in previous studies. Up to now, no generally accepted fitting procedure has been established.[47] Thus, it seems appropriate that test design as well as data fitting procedures should be chosen (and reported) as has been originally described for a certain LT. 2.1.2 Methodology of Blood Lactate Determination

From a methodological point of view, the site (earlobe, fingertip) as well as the method (venous, arterial, capillary) of blood sampling[56,57] and the laboratory methods (lactate analyser, analysed blood medium)[58-60] may also affect the test result. Samples taken from the earlobe have uniformly been shown to result in lower bLa than samples taken from the fingertip.[57,61,62] With regard to the analysed blood medium, plasma values were considerably higher than whole venous lactate concentrations, with capillary values lying in between.[48,56,63-65] In addition, several studies reported partly considerable differences between various lactate analysers (portable field vs laboratory analysers, amperometric vs photometric method) and under various climatic conditions.[58,66-69] The analysis of the whole blood lactate curve is a very global approach to evaluating endurance capacity. On the one hand, this approach is affected by the above-mentioned factors on overall lactate levels. On the other hand, the use of the entire curve leads to some uncertainty as to the magnitude of endurance gains that cannot be precisely estimated. However, the use of LTs enables a quantitative evaluation of changes in endurance performance. In addition, the ideal LT concept would not be affected by the abovementioned factors. There is evidence that approaches that analyse relative changes in bLa during GXTs may be favourable compared with the use of absolute lactate values in this regard.[56,67] Sports Med 2009; 39 (6)

Validity of Lactate Thresholds

2.2 A Framework for Endurance Diagnosis and Training Prescriptions

In 1979, Kindermann et al.[30] introduced the concept of the aerobic-anaerobic transition as a framework for performance diagnosis and training prescription in endurance sports (figure 1). Since then, this framework has been adopted, applied and refined by several scientists either using lactate or gas exchange markers.[16,26,31,33,34,46,70-75] This model consists of two typical breakpoints that are passed during incremental exercise. In the low intensity range, there is an intensity at which bLa begin to rise above baseline levels. This intensity was originally determined using gas exchange measurements,[21,22] and Wasserman called it the ‘anaerobic threshold’. This term has since been used for various LTs, particularly those with a different physiological background,[33,75] and, thus, has caused considerable confusion. Kindermann et al.[30] and Skinner and McLellan[34] suggested this intensity be called the ‘aerobic threshold’ (LTAer), because it marks the upper limit of a nearly exclusive aerobic metabolism and allows exercise lasting for hours. This intensity might be suitable for enhancing cardiorespiratory fitness in recreational sports, for cardiac rehabilitation in patients or for lowintensity and regenerative training sessions in high level endurance athletes.[16,25,26,32,70,76-81] Exercise intensities only slightly above the LTAer result in elevated but constant bLa during steady-state exercise and can be maintained for prolonged periods of time (~4 hours at intensities in the range of the first increase in bLa[82-84] and 45–60 minutes at an intensity corresponding to the maximal lactate steady state [MLSS][85,86]). Although anaerobic glycolysis is enhanced, it is speculated that such intensities may induce a considerable increase in the oxidative metabolism of muscle cells.[30,87] Theoretically, a high stimulation of oxidative metabolism for as long a period of time as is possible in this intensity range might be an appropriate load for endurance training. The highest constant workload that still leads to an equilibrium between lactate production and lactate elimination represents the MLSS. ª 2009 Adis Data Information BV. All rights reserved.

473

Some authors suggested that this intensity be called the ‘anaerobic threshold’.[27,30,49,88] It has been shown that the constant bLa at MLSS is not equal in all individuals and can vary considerably (values from 2 up to 10 mmol/L were reported in several studies).[50,72,86,89-93] Beneke and von Duvillard[94] as well as Beneke et al.[95] reported that bLa at MLSS is dependent on the motor pattern of exercise. Therefore, it was suggested that to determine the LTAn, individualized approaches rather than a fixed bLa should be used.[88,96,97] The MLSS represents the upper border of constant load endurance training.[30,49,71,95] Intensities above the MLSS have been used to guide interval training sessions in different endurance sports.[26,31,98-102] The intensity range between LTAer and LTAn is called the aerobic-anaerobic transition. The described thresholds (first increase in bLa and MLSS) have recently also been called ‘lactate threshold and lactate turnpoint’, ‘lactate threshold and anaerobic threshold’, or ‘anaerobic threshold 1 and 2’, respectively.[26,75,103,104] Within the present review, it was decided to stick to the originally introduced nomenclature.[30,31,34] There has been an exhaustive debate whether there exist clear breakpoints in the lactate/work rate relationship or whether lactate increase is rather a continuous function during incremental work.[47] Furthermore, the terms ‘aerobic’ and ‘anaerobic’ threshold may suggest clearly discernible physiological processes. However, these processes are rather of a transitional nature with aerobic and anaerobic energetic pathways always simultaneously contributing to energy production during both low- and high-intensity exercise. However, the proposed model seems appropriate both from a practical and from a didactical point of view. In addition, there is evidence that the described breakpoints may have some exercise physiological relevance. It has been shown that exercise above the MLSS is associated with an over-proportional excretion of stress hormones as well as of immunological markers during constant load exercise.[105,106] Furthermore, Lucia et al.[107] observed changes in electromyographical activity of the vastus lateralis and Sports Med 2009; 39 (6)

Faude et al.

rectus femoris that were coincidental with the aerobic-anaerobic transition in 28 elite male cyclists. The widespread use of this model as well as the absence of an accepted alternative was the rationale for using this framework in the present review to categorize published LT concepts. 3. Validation of Lactate Thresholds 3.1 Competition Performance

It is widely accepted that LTs (and the submaximal course of bLa during incremental exercise) are a criterion measure for aerobic endurance performance.[24,26,30,72,81,108] In particular, it has been shown that LTs are superior to maximal oxygen uptake when assessing endurance performance in homogenous groups of athletes.[11,12,109-111] The obvious gold standard to validate an LT concept is to compare it with the most recent competition performance in an endurance event (concurrent validity) or to assess its value in predicting endurance performance in future events (predictive validity). As an alternative to competition performance, the results of laboratory tests simulating an endurance event can be used. This might have the advantage of a higher standardization and, therefore, these test results may be more reliable. Correlations between the test value (LT) and the validity criterion (competition performance) can be dependent on several confounding factors such as, for example, the chosen competitive event (duration, laboratory or outdoor, athletic track or off-road), the sport that is evaluated as well as sex or age group and its heterogeneity in terms of endurance. 3.2 The Maximal Lactate Steady State

Endurance capacity can – from a metabolic point of view – be regarded as the highest steady state by energy supply from oxidative phosphorylation.[87] Therefore, another approach to assess aerobic endurance performance is the determination of the highest constant exercise intensity that can be maintained for a longer period of time ª 2009 Adis Data Information BV. All rights reserved.

Blood lactate concentration (mmol/L)

474

10 8 6 MLSS

4 2 0 Rest

10 20 Time (min)

30

40

Fig. 2. The blood lactate response to several constant workload exercises with different intensities. The highest workload during which blood lactate concentrations can be still accepted as being steady state is defined as the maximal lactate steady state (MLSS).

without a continuous rise in bLa. This intensity represents the MLSS, which has been shown to be highly related to competition performance in endurance events (r [correlation coefficient] = 0.92 with 8 km running, r = 0.87 with 5 km running and r = 0.84 with 40 km cycling time trial speed, respectively).[112-114] The MLSS has been defined by some authors as the ‘anaerobic threshold’ because it represents an exercise intensity that can be maintained without considerable contribution of anaerobic metabolism.[27,30,50,72,115] Each higher intensity results in a clearly identifiable increase in bLa with time during constant load work.[50,86,88] The gold standard for the determination of the MLSS is performing several constant load trials of at least 30 minutes’ duration on different days at various exercise intensities (in the range of
50–90% VO2max, figure 2).[49,50,86,116,117] An increase in bLa of not more than 1 mmol/L between 10 and 30 minutes during the constant load trials appears to be the most reasonable procedure for MLSS determination.[86,115] MLSS represents a steady state in several but not all physiological parameters. For instance, oxygen uptake, carbon dioxide output, respiratory exchange ratio and bicarbonate concentration were reported to remain nearly constant during constant load exercise at MLSS, but respiratory rate and heart rate significantly increased during this time.[85,118] Sports Med 2009; 39 (6)

Validity of Lactate Thresholds

In several endurance sports it is recommended to aim at a defined metabolic strain when a certain training stimulus is intended.[71,73,119,120] Therefore, it seems likely that training intensities for endurance training can be appropriately described when MLSS is known. For the purposes of this review based on the above-mentioned rationales, LTs are considered valid as performance indicators when there are high linear correlations with (simulated) endurance performance. In addition, a close relationship between LTs and MLSS suggests validity with regard to the prescription of training intensities. Therefore, it is desirable that LTs should fulfil both validity criteria. 4. Lactate Threshold Concepts For the purposes of the present paper, the MEDLINE database PubMed was searched for the search terms ‘lactate threshold’, ‘aerobic threshold’ and ‘anaerobic threshold’ combined with either ‘endurance performance’ or ‘maximal lactate steady state’. Additionally, the references of the selected articles were searched for further relevant papers. The located original publications were searched for papers describing different LT concepts (section 4.1), a correlation between LTs and (simulated) endurance performance (section 4.2) or the relationship between LTs and the MLSS (section 4.3). 4.1 Located Lactate Threshold Concepts

A total of 25 different LT concepts were located. Two studies were excluded from the present analysis because threshold determination was not solely based on bLa but also took gas exchange measurements into account.[121,122] All threshold concepts were divided into three categories. Several authors used so-called fixed blood lactate thresholds (LTfix) during incremental exercise to evaluate aerobic endurance performance. These fixed bLas were set at 2, 2.5, 3 or 4 mmol/L[24,108,123-125] with LT4 (4 mmol/L lactate threshold, originally described by Mader et al.[24] and by others later as the onset of blood ª 2009 Adis Data Information BV. All rights reserved.

475

lactate accumulation [OBLA][108]) being the most frequently used method. 4.1.1 Aerobic Lactate Thresholds

Table I shows an overview of LT concepts that could be categorized as the first rise in bLa above baseline levels (LTAer). Several researchers described the procedure to determine this threshold with terms like ‘‘the first significant/marked/ systematic/non-linear/sharp/abrupt sustained increase in bLa above baseline’’.[30,110,126-133,138] Although the visual determination of the first rise of bLa above baseline levels seems obvious and simple, in practice it is associated with considerable problems because of the only slight changes in bLa on the first steps during GXTs. Yeh et al.[142] demonstrated that the visual detection of the LTAer (in that study called ‘anaerobic threshold’) led to relevant differences between observers. Therefore, it does not seem appropriate to determine this threshold by simple visual inspection. Thus, other methods were developed to make threshold determination more objective. For instance, some authors took the typical error of their lactate analysers into account and

Table I. Lactate threshold concepts that were categorized in the aerobic threshold category. For further explanation see text Method and description Work intensity or oxygen uptake before/at which bLa begins to increase above baseline level[110,126] at which bLa exhibits a marked/systematic/significant/non-linear/ sharp/abrupt sustained increase above baseline value[30,110,127-133] first significant elevation of lactate level (approximately 2 mmol/L)[30,34] before an elevation in bLa above baseline (at least 0.2 mmol/L due to error of lactate analyser)[123,134] rise in delta lactate (onset of plasma lactate accumulation)[109] at minimum lactate equivalent (bLa divided by oxygen uptake or work intensity)[36,135-137] at which plasma lactate concentration begins to increase when log bLa is plotted against log (work intensity)[53] at which bLa increases 0.5 mmol/L above resting concentration[138] at which bLa increases 1 mmol/L above baseline (i.e. lactate at low
intensity corresponding to 40–60% VO2max)[111,139] preceding a bLa increase by 1 mmol/L or more[140,141]
bLa = blood lactate concentrations; VO2max = maximal oxygen uptake.

Sports Med 2009; 39 (6)

Faude et al.

476

Table II. Lactate threshold concepts that were categorized in the anaerobic threshold category. For further explanation see text Threshold concept

Method and description

IAT (Stegmann et al.)[88]

Tangent to bLa curve from recovery curve where bLa is equal to the value at end of GXT

IAT (Keul et al.)[96]

Tangent to bLa curve at 51

IAT (Simon et al.)[97]

Tangent to bLa curve at 45

IAT (Berg et al.)[137]

Intersection point between tangent for the minimum lactate equivalent and the linear function for the final 90 sec of GXT

IAT (Bunc et al.)[143]

Intersection between the exponential regression of the lactate curve and the bisector of the tangents of the upper and lower parts of the lactate curve

IAT (Dickhuth et al.)[36,136]

1.5 mmol/L above minimum lactate equivalent

IAT (Baldari and Guidetti)[144]

The second lactate increase of at least 0.5 mmol/L from the previous value

Dmax (Cheng et al.)[54]

Maximal distance from bLa curve to the line formed by its endpoints

Dmod (Bishop et al.)[140]

Maximal distance from bLa curve to the line formed by the point before the first rise in bLa and the value at cessation of exercise

Lactate turnpoint[103]

The final running velocity before the observation of a sudden and sustained increase in bLa
between LTAer and VO2max

Lactate minimum speed[145]

Minimum in bLa during GXT after high intensity exercise

bLa = blood lactate concentration; GXT = incremental exercise test; IAT = individual anaerobic threshold; LTAer = aerobic threshold;
VO2max = maximal oxygen uptake.

determined this LT as the workload 0.2 mmol/L above the lowest exercise lactate value.[123] Hughson and Green[138] arbitrarily chose an increase of 0.5 mmol/L above resting lactate concentrations. Another work group[111,139] chose a 1 mmol/L increment above lactate levels at low
intensity (~40% to 60% VO2max) because it could be determined objectively and in a standardized manner in all subjects. Furthermore, the lowest value when bLa is divided by work intensity or
VO2 has also been used as a marker for LTAer (minimum lactate equivalent).[36,135-137] Whereas Beaver and colleagues[53] used a log-log transformation to assess the first rise in bLa more objectively as the intersection of two linear regressions, Farrell et al.[109] plotted the difference in bLa between two consecutive stages against work intensity and determined the first rise of this relationship (onset of plasma lactate accumulation). 4.1.2 Anaerobic Lactate Thresholds

All threshold concepts that were assigned either to the MLSS or to a rapid/distinct change in the inclination of the blood lactate curve were categorized as LTAn (table II). Originally, the LT4 was established because it seemed to be the highest bLa that was sustainable for a longer duration and, therefore, was regarded ª 2009 Adis Data Information BV. All rights reserved.

as the upper border for constant load endurance training.[24] It was soon recognized that a fixed bLa does not take into account considerable interindividual differences and that LT4 may frequently underestimate (particularly in anaerobically trained subjects) or overestimate (in aerobically trained athletes) real endurance capacity.[88,96,97,146] Therefore, several so-called ‘individualized’ LT concepts were developed. For instance, Keul et al.[96] and Simon et al.[97] determined the individual anaerobic threshold (IAT) at a certain inclination of the lactate curve (tangent of 51 and 45, respectively). However, it seems questionable whether the use of a fixed inclination may reflect individual lactate kinetics better than a fixed bLa. Stegmann et al.[88] developed a more complicated model that is based on the exercise lactate curve as well as on the lactate course during the early recovery period. This model is based on several assumptions regarding lactate distribution in blood and muscle compartments, lactate diffusion through biological membranes and lactate elimination. However, some of these premises have been challenged.[8,147] Berg et al.[137] defined the LTAn as the intersection point between the tangent at the minimum lactate equivalent and the linear function Sports Med 2009; 39 (6)

Validity of Lactate Thresholds

for the final 90 seconds of GXT. Similarly, Bunc et al.[143] determined the LTAn as the intersection between the exponential regression of the lactate curve and the bisector of the tangents on the upper and lower parts of the regression. A comparable model was established by Cheng et al.[54] and called the Dmax method. Those authors determined the maximal perpendicular distance of the lactate curve from the line connecting the start with the endpoint of the lactate curve. It is obvious that these threshold models are dependent on the start intensity as well as the maximal effort spent by the subjects. To eliminate the influence of the start point of the GXT, Bishop et al.[140] connected the LTAer with the endpoint of the lactate curve and observed that this modified Dmax threshold (Dmod) was also highly correlated with performance during a 1-hour time trial in 24 female cyclists. Tegtbur et al.[145] developed the so-called lactate minimum test. This test starts with a short supramaximal exercise leading to high bLa. A short rest period (about 8 minutes)[145] should allow for an equilibrium between muscle and bLa. Immediately after this rest period, a standard incremental exercise test is conducted. After an initial fall of bLa at low exercise intensities, bLa begins to rise again. The lowest point of the lactate curve, the lactate minimum speed (LMS), is assumed to mark the LTAn. This procedure has recently been criticized because standardization is difficult.[112,148] For instance, the induced acidosis prior to the incremental test is unlikely to be uniform for different subjects. Additionally, initial intensity as well as stage increment and duration seem to affect LMS. Furthermore, supramaximal exercise might be contraindicated in some instances, for example in cardiac patients or in athletes at some time points during their training. Baldari and Guidetti[144] defined the IAT as the workload corresponding to the second lactate increase of at least 0.5 mmol/L with the second increase greater than or equal to the first one. A limitation to this approach is that only discrete stages according to the test protocol can be identified as threshold workload. Additionally, those authors determined the IAT by plotting each lactate value against the preceding workª 2009 Adis Data Information BV. All rights reserved.

477

load. This was claimed to be done because during 3-minute stages no steady-state lactate level could be reached[147] and, therefore, it was hypothesized that a lactate value at a given 3-minute stage would represent the realistic value of the previous stage. From empirical observations, the work group of Dickhuth et al.[36,135,136] estimated the IAT at a blood lactate concentration 1.5 mmol/L above the minimum lactate equivalent (i.e. above LTAer). Finally, the lactate turnpoint (LTP) has been defined as the final running velocity before the observation of a sudden and sustained increase in bLa between LTAer and
VO2max.[103] Reproducibility of the velocity or power output at LTs has been reported to be high (r > 0.9, independent of whether LTfix, LTAer or LTAn
were analysed).[52,149-152] For VO2 at LTs, reliability coefficients seem to be slightly lower (r = 0.8–0.9).[150,152,153] 4.2 Lactate Thresholds and (Simulated) Competition Results

Thirty-eight studies were located that compared LT values with performance in endurance events or simulated competitions. Six studies were excluded from the analysis. Three of these studies compared an LT obtained during cycling exercise with running performance,[110,154,155] two studies only reported LT as a fraction of
VO2max,[11,156] and one study reported correlations with time-to-exhaustion in an open-end interval programme.[157] A total of 32 studies were thus included in this analysis. Eighteen studies evaluated the correlation of
the work intensity (running velocity or VO2) at various LTs with performance in running competitions of different distances (800 m up to marathon; table III).[108,109,112,123,124,129-132,134,135,158-164] Competition distances from 0.8 to 3.2 km, from 5 km to 16.1 km and from 21.1 to 42.2 km were subsumed as correlates of short-, middle- and long-distance endurance events. The main result was that nearly all studies reported high correlation coefficients with (simulated) competition results. These results were confirmed by Weltman Sports Med 2009; 39 (6)

Faude et al.

478

Table III. Correlation coefficients between lactate thresholds and running performance over various distances Threshold concept

LTfix

0.8–3.2 km

5 km–16.1 km

19.3–42.2 km

v


VO2

v


VO2

v


VO2

0.82[135] 0.88[123] 0.86[123] 0.85[123] 0.87[134] 0.85[134] 0.84[134] 0.93[158] 0.78[132] 0.68[131] 0.85[131] 0.88[131]

0.79[123] 0.75[123] 0.75[123] 0.72[134] 0.74[134] 0.75[134] 0.73[158] 0.60[132] 0.51[131] 0.55[131] 0.69[131]

0.88[135] 0.91[135] 0.91[159] 0.93[159] 0.91[159] 0.84[159] 0.91[159] 0.94[159] 0.83[160] 0.81[112] 0.95[163] 0.94[163]

0.90[159] 0.92[159] 0.92[159] 0.83[159] 0.88[159] 0.93[159] 0.86[163] 0.74[163]

0.91[135] 0.81[135] 0.98[124] 0.98[124] 0.98[124] 0.68[129] 0.96[108] 0.91[163] 0.92[163]

0.76[161] 0.83[163] 0.73[163]

Median (min–max)

0.85 (0.68–0.93)

0.73 (0.51–0.79)

0.91 (0.81–0.95)

0.89 (0.74–0.93)

0.92 (0.68–0.98)

0.76 (0.73–0.83)

LTAer

0.74[135] 0.85[123] 0.70[134] 0.93[158] 0.77[132] 0.43[131] 0.65[131] 0.70[131] 0.91[109]

0.77[123] 0.61[134] 0.84[158] 0.69[132] 0.77[131] 0.66[131] 0.64[131] 0.85[109] 0.62[162] 0.66[162] 0.58[162]

0.73[135] 0.79[135] 0.78[160] 0.96[109] 0.97[109] 0.79[130] 0.83[130] 0.79[130] 0.84[130] 0.83[130] 0.81[130] 0.93[112] 0.94[163] 0.92[163] 0.92[163] 0.89[163] 0.87[163] 0.85[163]

0.89[109] 0.91[109] 0.84[162] 0.83[162] 0.79[162] 0.69[162] 0.92[162] 0.79[162] 0.76[130] 0.77[130] 0.84[130] 0.81[130] 0.82[130] 0.88[130] 0.72[163] 0.56[163] 0.66[163] 0.52[163] 0.81[163] 0.69[163]

0.76[135] 0.81[135] 0.78[129] 0.97[109] 0.98[109] 0.90[163] 0.91[163] 0.87[163] 0.86[163] 0.83[163] 0.77[163]

0.91[109] 0.89[109] 0.69[163] 0.52[163] 0.66[163] 0.42[163] 0.80[163] 0.65[163]

Median (min–max)

0.74 (0.43–0.93)

0.66 (0.58–0.85)

0.84 (0.73–0.97)

0.79 (0.45–0.92)

0.86 (0.76–0.98)

0.68 (0.42–0.91)

LTAn

0.88[135]

0.91[135] 0.92[135] 0.86[160] 0.83[112] 0.93[163] 0.91[163] 0.94[163] 0.90[163] 0.76[164] 0.73[164]

0.83[163] 0.70[163] 0.81[163] 0.66[163] 0.45[164] 0.45[164]

0.93[135] 0.93[135] 0.90[163] 0.91[163] 0.90[163] 0.89[163]

0.68[161] 0.83[163] 0.71[163] 0.81[163] 0.67[163]

Median (min–max) 0.88 0.91 (0.83–0.94) 0.76 (0.66–0.83) 0.91 (0.89–0.93)
LTfix = fixed lactate threshold; LTAer = aerobic threshold; LTAn = anaerobic threshold; v = velocity; VO2 = oxygen uptake.

et al.,[123,134] who cross-validated the obtained regression equations and found high correlation coefficients between actual and predicted scores. There is a tendency for higher correlations with longer endurance events (0.43–0.93 in short-term ª 2009 Adis Data Information BV. All rights reserved.

0.71 (0.67–0.83)

events vs 0.68–0.98 over the long-distance competitions). Additionally, correlations tended to be higher for LTfix and LTAn compared with LTAer. This might be due to the average intensity in running events being higher than the intensity Sports Med 2009; 39 (6)

Validity of Lactate Thresholds

479

corresponding to the first increase in bLa. In total, the results of the analysed studies point to a common variance of LTs and competition results in running events between 55% and 85%. In cycling, a total of eight studies evaluated the relationship between LTs and (simulated) cycling time trial performance (table IV).[12,89,140,141,165-168] Only one study analysed the correlation with short-duration time trial performance (4000 m individual pursuit) and found a high correlation coefficient of r = 0.86 in 18 male high-performance track cyclists.[167] Four studies evaluated distances between 13.5 and 20 km or time trial durations between 20 and 30 minutes.[89,165,166,168] The correlation coefficients in these studies were in most cases higher (between 0.8 and 0.9) than for the longer time trials (40 km or 60–90 minutes, r ~ 0.7).[140,141,165] Overall, the results of these studies were more heterogeneous. Correlation coefficients between LTs and (simulated) competition performance varied between r = 0.23[165] and r = 0.93.[89] In total, the results of the analysed studies point to a common variance of LTs and competition results between 35% and 85% in cycling events. However, the low number of studies and the heterogeneous results point to the need for further carefully designed studies to

arrive at more comprehensive conclusions with regard to the relationship of LTs and time trial performance in cycling. Two studies were found that analysed the relationship of LT markers with mountain bike cross-country race performance.[169,170] Such races are usually conducted on a graded terrain with considerable time spent ascending and descending. Impellizzeri et al.[170] observed high correlations between LTAer as well as OBLA and race time during a 31 km mountain bike race. Whereas correlations were about 0.7 when LT was expressed in absolute terms, correlations became considerably higher (~0.9) when power output at LT was expressed relative to body mass. Similarly, Gregory et al.[169] reported higher correlations between LTAer and a crosscountry time trial in 11 male mountain bikers when LTAer was expressed as related to body mass (r ~ 0.5 in absolute terms vs r ~ 0.8 relative to body mass). This finding can be explained with the considerable influence of bodyweight and body composition on performance capacity in cyclists during ascents.[171-173] In addition to the studies in running and cycling, another four studies were detected that evaluated LTs and (simulated) competition

Table IV. Correlation coefficients between lactate thresholds and cycling time trial events over various distances and times Threshold concept

4 km PO

13.5–20 km; 20–30 min
VO2

[165]

LTfix

0.23 0.82[166] 0.90[166]

Median (min–max) LTAer

0.86[167]

Median (min–max)

0.86

LTAn

40 km; 60–90 min
VO2

PO


V O2

PO [165]

0.54 0.60[141] 0.81[140]

0.82 (0.23–0.90)

0.60 (0.54–0.81)

0.67[165] 0.88[166] 0.86[166] 0.91[168] 0.88[168]

0.91[165] 0.59[141] 0.61[140] 0.69[140] 0.65[140]

0.93[12]

0.88 (0.67–0.91)

0.65 (0.59–0.91)

0.93

0.45[165] 0.89[166] 0.91[166] 0.93[89]

0.77[165] 0.58[141] 0.52[141] 0.72[141] 0.84[140] 0.83[140]

Median (min–max) 0.90 (0.45–0.93) 0.75 (0.52–0.84)
LTAer = aerobic threshold; LTAn = anaerobic threshold; LTfix = fixed lactate threshold; PO = power output; VO2 = oxygen uptake.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (6)

Faude et al.

480

performance. Two of these studies analysed competitive race walkers. Yoshida et al.[174] found correlation coefficients for OBLA as well as for LTAer of 0.94 and 0.85, respectively, with walking pace during a 5 km road race in eight female race walkers. Similar results were observed by Hagberg and Coyle[111] in a heterogeneous group of race walkers with correlation coefficients of 0.94 and 0.82 for velocity and oxygen uptake at LTAer in a 20 km race walking performance. Two studies dealt with rowing performance and LTs. Whereas Ingham et al.[175] observed high correlations (r = 0.86–0.92) between work rate at fixed and aerobic LTs and 2000 m ergometer performance in 41 rowers of different categories, Cosgrove et al.[176] found considerably lower correlations (r = 0.39–0.73) in 13 male rowers. To summarize, the overwhelming majority of published studies on the relationship between LTs and endurance performance showed strong correlations, particularly for running events. This supports findings of earlier training studies that found training-induced improvements in competitive performance significantly correlated with improvements in LTs.[130,162] Although it seems likely that other influences such as central nervous system processes may have regulatory and decisive characteristics in endurance events as it was recently claimed,[177] peripheral metabolic adaptations highly related to the LT[46] seem to be a necessary and important prerequisite for aerobic endurance performance.

4.3 Lactate Thresholds and Maximal Lactate Steady State

MLSS determination has become very popular in performance diagnosis in several endurance sports. Thus, numerous studies have dealt with the problem of an adequate estimation of MLSS during one single laboratory visit. For instance, some authors tried to estimate MLSS from performance during all-out time trials (5 km or 40 km)[114,178] from physiological strain (bLa, heart rate, ratings of perceived exertion) during standardized sub-maximal constant-load exercise[179-182] or from gas exchange measurements.[183-189] ª 2009 Adis Data Information BV. All rights reserved.

However, an overview of those studies is beyond the scope of the present review. There are several studies that examined the metabolic responses during steady-state exercise intensities related to LTs but did not analyse exercise intensities slightly above or below. Schnabel et al.[190] observed average steady-state lactate concentrations (~4.5 mmol/L) during 50-minute runs at the IAT according to Stegmann et al.[88] However, no other intensity was analysed in this investigation. Stegmann and Kindermann[146] compared 50-minute cycling exercise in 19 subjects at the IAT as well as at LT4 and found steady-state lactate levels (~4 mmol/L) during IAT trials, whereas exercise at LT4 resulted in continuously rising bLa (up to 9.6 mmol/L) and a premature cessation. This is in line with findings of OyonoEnguelle et al.,[191] who similarly reported no lactate steady state in three out of five subjects during exercise at LT4. In contrast, Loat and Rhodes[189] found continuously increasing bLa (on average from 3.4 mmol/L after 15 minutes to 4.6 mmol/L after 45 minutes) and premature fatigue during 60-minute constant load trials at the IAT. However, those authors did not use the originally described test protocol and Heck[50] has shown that IAT determination is dependent on the protocol used. Baldari and Guidetti[144] compared steadystate running at their IAT determined when lactate values were plotted against the corresponding exercise intensity (IATm) and against the preceding intensity (IATa) and found steady-state lactate levels for IATa (~4 mmol/L-1) but not for IATm. However, due to the determination procedure, the difference between both thresholds was exactly one stage increment and no other intensities in between were evaluated. Ribeiro et al.[192] assessed a 40-minute steady-state cycling exercise at LTAer, between LTAer and LTAn (LTP), at LTAn as well as between LTAn and maximum. Those authors found on average steadystate lactate levels up to LTAn (~5 mmol/L-1), whereas at the highest intensity, bLa increased continuously and exercise had to be terminated prematurely. Bacon and Kern[193] and Tegtbur et al.[145] compared constant load trials at LMS and 5% or 0.2 m/s, respectively, above the LMS. Those Sports Med 2009; 39 (6)

Validity of Lactate Thresholds

authors found that LMS intensity but not the higher intensity on average resulted in a lactate steady state. However, in the study of Bacon and Kern,[193] the average blood lactate increase between minutes 12 and 28 during the constant load trial at the LMS +5% intensity was 1.2 mmol/L, and in four out of ten subjects a lactate steady state according to the recommended criterion[72,115] was present. A total of 11 studies evaluated the relationship between one or more LT concepts and MLSS using the recommended procedure, including several constant load trials of at least 30 minutes’ duration to determine the MLSS (table V). One study determined MLSS with 20-minute constant load trials.[113] Most researchers analysed the relationship of LT4 with MLSS.[49,72,90,92,112,117] For instance, Heck and colleagues[49,50,72] found strong correlations between LT4 and MLSS during running as well as during cycling exercise. However, the fitness level of their subjects was quite heterogeneous and, therefore, the high correlations to some extent might be spurious. Additionally, they observed that the velocity at LT4 was higher than MLSS velocity when stage duration during the GXT was 3 minutes, whereas this was not the case with 5-minute stages. Therefore, these authors concluded that LT4 gives a valuable estimate of the MLSS when stage duration is at least 5 minutes. Also, Jones and Doust[112] found a high correlation between LT4 and the MLSS in a homogenous group of trained runners with LT4 being higher than MLSS (3-minute stages). Lower correlations were found by van Schuylenbergh et al.[92] in elite cyclists as well as by Beneke[117] in a homogenous group of rowers. Also, LT4 and MLSS did not differ significantly with 6-minute stages,[92] whereas LT4 was considerably higher than MLSS with 3-minute stages.[117] Lajoie et al.[90] evaluated whether the intensity corresponding to 4 mmol/L lactate during a GXT with 8-minute stages and 30 W increments is appropriate to estimate the MLSS in nine cyclists. Average power output at MLSS and LT4 was not significantly different. However, because bLa at MLSS differed considerably between subjects, the authors concluded ª 2009 Adis Data Information BV. All rights reserved.

481

that it is unrealistic to rely on a blood lactate value of 4 mmol/L as a universal criterion for MLSS. Unfortunately, a more detailed analysis regarding the correlation or individual differences between LT4 and MLSS was not reported. Heck et al.[49,50] observed high correlations between MLSS and the IAT according to Stegmann et al.[88] In addition, running velocity was not significantly different between IAT and MLSS independent of stage duration (3 or 5 minutes), whereas in cycling IAT was about 8% higher than MLSS. Urhausen et al.[86] found in runners as well as in cyclists that constant load trials at IAT resulted on average in a lactate steady state, whereas a 5% higher intensity led to a continuous rise in bLa. Similarly, McLellan and Jacobs[91] arrived at the conclusion that the IAT is a valid estimate for the MLSS in most subjects, although there exists a considerable difference in a few cases. Unfortunately, these studies reported no measure of correlation between IAT and MLSS or no quantitative data on individual differences between IAT and MLSS. In contrast to the previously mentioned studies, Beneke[117] found the IAT to be considerably higher than MLSS in nine rowers. Additionally, the correlation in this study was lower than was observed by Heck et al.[49] This finding might be due to the more homogenous performance level of the rowers as well as to the slow increment in the chosen test protocol.[50] Heck et al.[49] and Heck[50] found high correlations between the IAT according to Keul et al.[96] and Bunc et al.[143] and the MLSS in running and cycling. However, the high correlations might be partly accounted for by the heterogenous endurance level of the subjects. Furthermore, both thresholds were dependent on the test protocol during the running tests (3-minute vs 5-minute stages). The LMS was evaluated in two studies.[89,112] The results of these studies were contradictory. Jones and Doust[112] found only a low correlation between LMS and MLSS. Additionally, LMS was considerably lower than MLSS. In contrast, LMS was not significantly different from MLSS in the study of MacIntosh et al.[89] These contrasting observations might have been due to Sports Med 2009; 39 (6)

Faude et al.

482

Table V. Comparison of lactate threshold concepts with MLSS determined by several constant load trials of different intensity Threshold concept

Subjects

Main outcome

Reference

LT4, OBLA

16 healthy males (running)

High correlation between LT4 and MLSS (r = 0.98) LT4 on average 0.12 m/s higher than MLSS with 3 min stages but not with 5 min stages during GXT Heterogenous endurance level

Heck et al.[49,72]

22 healthy subjects (cycling)

Significant correlation between LT4 and MLSS (r = 0.92) LT4 on average 19.9 W higher than MLSS Heterogenous endurance level, slow increase in power output (+6 W/min)

Heck[50]

8 trained male runners

High correlation (r = 0.93) between OBLA and MLSS OBLA on average 0.4 km/h higher than MLSS

Jones and Doust[112]

21 elite cyclists

Low correlation (r = 0.71) between LT4 and MLSS No significant difference between LT4 and MLSS (MLSS 15 W higher) Homogenous endurance level

Van Schuylenbergh et al.[92]

9 male rowers

Significant correlation (r = 0.82) between LT4 and MLSS LT4 significantly higher (32 W) than MLSS Homogenous endurance level

Beneke[117]

10 well trained cyclists

Average power output at LT4 and MLSS was not significantly different (282 W vs 277 W) Strong MLSS criterion (<0.75 mmol/L from 10–60 min) No further data on correlations or intraindividual differences between LT4 and MLSS High correlation between IAT and MLSS (r = 0.96–0.98) IAT velocity on average similar to MLSS for 3 min as well as 5 min stages during GXT Heterogenous endurance level of subjects

Lajoie et al.[90]

22 healthy subjects (cycling)

Significant correlation between IAT and MLSS (r = 0.87) IAT on average 15.1 W higher than MLSS Heterogenous endurance level, slow increase in power output (+6 W/min) not corresponding to the originally described test protocol

Heck[50]

16 trained cyclists 14 trained runners

CLT at and below IAT resulted on average in LSS but not CLT at 105% IAT 100% IAT does not in all individuals exactly represent MLSS LSS was found in 6 (of 14 runners) and 9 (of 16 cyclists) at 105% IAT No further data on correlations or intraindividual differences between IAT and MLSS CLT at LT4 (cycling at 104% IAT) resulted on average not in a LSS
No LSS during CLT at IAT +5% VO2max; only 1 LSS during CLT at IAT
+2.5% VO2max
Two subjects showed no LSS during CLT at IAT -7.5% VO2max, all
other subjects showed LSS during CLT at IAT -2.5% VO2max No further data on correlations or intraindividual differences between IAT and MLSS

Urhausen et al.[86]

IAT (Stegmann et al.[88])

16 healthy males (running)

11 males (cycling)

9 male rowers IAT (Keul et al.[96])

16 healthy males (running) 22 healthy subjects (cycling)

Heck et al.[49]

McLellan and Jacobs[91]

Significant correlation (r = 0.81) between IAT and MLSS IAT significantly higher (32 W) than MLSS High correlation between IAT and MLSS (r = 0.98) IAT velocity on average 0.2 m/s higher than MLSS with 3 min stages and slightly lower with 5 min stages during GXT Heterogenous endurance level of subjects

Beneke[117]

Significant correlation between IAT and MLSS (r = 0.94) IAT on average 21.0 W higher than MLSS Heterogenous endurance level, slow increase in power output (+6 W/min-1)

Heck[50]

Heck et al.[49]

Continued next page

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (6)

Validity of Lactate Thresholds

483

Table V. Contd Threshold concept

Subjects

Main outcome

Reference

IAT (Bunc et al.[143])

16 healthy males (running)

High correlation between IAT and MLSS (r = 0.98–0.99) IAT velocity on average considerably higher than MLSS for 3-min (+0.31 m/s) as well as 5 min stages (+0.14 m/s) during GXT Heterogenous endurance level of subjects

Heck et al.[49]

22 healthy subjects (cycling)

Significant correlation between IAT and MLSS (r = 0.89) IAT on average 71.5 W higher than MLSS Heterogenous endurance level, slow increase in power output (+6 W/min)

Heck[50]

10 trained male runners

Low correlation (r = 0.61) between LMS and MLSS LMS on average 0.8 km/h lower than MLSS

Jones and Doust[112]

14 cyclists or triathletes

LMS on average not different from MLSS No good estimate of MLSS by LMS in three subjects MLSS criterion: <0.7 mmol/L during last 20 min No further data on correlations or intraindividual differences between LMS and MLSS

MacIntosh et al.[89]

Dmod

21 elite cyclists

Significant correlation (r = 0.85) between Dmod threshold and MLSS Dmod threshold significantly lower (-23 W) than MLSS

Van Schuylenbergh et al.[92]

LTP

8 males (running)

No correlation between LTP and MLSS (r = 0.18) On average no difference between LTP and MLSS (13.7 vs 13.8 km/h) 95% LoA[194] = –1.8 km/h

Smith and Jones[103]

LTAer

10 trained male runners

High correlation (r = 0.94) between LTAer and MLSS LTAer on average 0.6 km/h lower than MLSS

Jones and Doust[112]

11 male recreational runners


No correlation of LTAer with MLSS (speed: r = -0.01; VO2: r = -0.47) LTAer on average 1.1 km/h lower than MLSS 20 min CLT, but strong MLSS criterion (<0.2 mmol/L)

Haverty et al.[113]

LMS

CLT = constant load trial; Dmod = maximal distance from blood lactate concentration (bLa) curve to the line formed by the point before the first rise in bLa and the value at cessation of exercise; GXT = incremental exercise test; IAT = individual anaerobic threshold; LMS = lactate minimum speed; LoA = limits of agreement; LSS = lactate steady state; LT4 = 4 mmol/L lactate threshold; LTAer = aerobic threshold; LTP = lactate turnpoint; MLSS = maximal lactate steady state; OBLA = onset of blood lactate accumulation; r = correlation coefficient;
VO2max = maximal oxygen uptake.

the considerably different test protocols used in both studies. This is in line with the findings of Carter et al.[148] showing that LMS is highly dependent on the test protocol. For other threshold concepts, scientific data regarding the relationship of the threshold and MLSS are scarce. Van Schuylenbergh et al.[92] found a significant correlation between the Dmodthreshold and MLSS, although Dmod was significantly lower than MLSS. In contrast, LTP was found to be not different from MLSS on average, but it was not correlated to MLSS and the 95% limits of agreement (LoA)[194] of the difference between LTP and MLSS were wide.[103] There were also two studies that analysed the relationship between MLSS and LTAer.[112,113] As could be expected, LTAer was situated considerably below the MLSS in both studies. Whereas Jones and Doust[112] reported a high ª 2009 Adis Data Information BV. All rights reserved.

correlation between LTAer and MLSS, Haverty et al.[113] did not. This might be due to short constant load trials (20 minutes) and the strict MLSS criterion (<0.2 mmol/L increase during the last 10 minutes) in the latter study, which does not sufficiently consider the time course of bLa changes and may have led to an underestimation of the real MLSS.[115] To summarize, there is evidence that some LT concepts might be able to estimate the MLSS. In particular, the IAT according to Stegmann et al.,[88] and LT4 were repeatedly examined. Mostly linear regressions or average lactate courses were reported. Correlations and regressions determine relative reliability of two methods but do not assess systematic bias or absolute agreement. Furthermore, they depend greatly on the range of values in the analysed sample.[195] Thus, from a practical and statistical point of Sports Med 2009; 39 (6)

Faude et al.

484

Table VI. Mean bias (difference maximal lactate steady state [MLSS]-LT) and 95% limits of agreement (LoA) for four different lactate threshold concepts during treadmill (n = 16) and cycle ergometry (n = 22). Results calculated from raw data reported by Heck et al.[49,50,72] (with permission) Lactate threshold concept

Treadmill ergometry 3 min stages, +0.4 m/s mean bias LoA (m/s) (m/s)

LoA (%)

Treadmill ergometry 5 min stages, +0.4 m/s mean bias LoA (m/s) (m/s)

LoA (%)

Cycle ergometry 2 min stages, +25 W mean bias LoA (W) (W)

LoA (%)

LT4

-0.13

–0.35

–8

0.02

–0.39

–9

-19.8

–28.4

–14

IAT (Keul et al.[96])

-0.20

–0.39

–9

0.06

–0.35

–8

-21.0

–22.4

–11

IAT (Stegmann et al.[88])

-0.03

–0.51

–12

-0.03

–0.37

–9

-15.0

–35.0

–18

IAT (Bunc et al.[143])

-0.33

–0.33

–8

-0.14

–0.37

–9

-71.4

–52.8

–27

IAT = individual anaerobic threshold; LT4 = 4 mmol/L threshold.

view it would be of interest to know the absolute variability of individual differences between the LT and MLSS. An appropriate means to report this variability may be the mean bias and the 95% LoA as it was described by Bland and Altman.[194] There is only one study available that applied this procedure.[103] Such a procedure would also allow for assessing heteroscedasticity (i.e. whether the differences depend on the magnitude of the mean or – in this case – endurance capacity).[195] Table VI shows an example calculation of the mean bias and the 95% LoA for four different LT concepts from raw data reported by Heck et al.[49,50,72] These data show a mean bias between 0.5% and 8%, with LoA of about 10% in a running exercise. This means that for each new subject within the study population it could be expected (with a 95% probability) that the difference between MLSS and the respective LT is within these LoA.[195] For the cycling exercise the results are more heterogenous with greater mean bias and LoA. However, due to the limited data points these observations are preliminary and should be confirmed by further research. 5. Conclusions and Perspectives In conclusion, it can be stated that a huge amount of evidence exists that LT concepts are of considerable importance for the diagnosis as well as the prediction of aerobic endurance performance. The concept of the aerobic-anaerobic transition may serve as a reasonable means for ª 2009 Adis Data Information BV. All rights reserved.

performance diagnosis and intensity prescription in endurance sports. However, there are several open questions that should be appropriately addressed by future research. These are:  Whereas the relationship of LTs with competition performance is well established in running events and less strongly in cycling, there is lack of evidence for most other endurance sports.  Scientific studies comparing LTs with MLSS are rare and the results are partially conflicting. This might be due to different methodological approaches. It is suggested that the MLSS be assessed by the established procedure using several constant load trials with different intensities[72,115] and that the MLSS be compared with a chosen LT. To do so, measures of absolute agreement between LTs and MLSS should be reported according to the method introduced by Bland and Altman.[194]  In this context, it is important to know the basic variability and reproducibility of the MLSS. Up to now, no scientific data addressing this question exist. Therefore, it is recommended to evaluate the variability of MLSS in future research. Of note, this may enable an evaluation of the differences between LT and MLSS compared with the basic variability of the MLSS and, thus, give more detailed information on the quality of the MLSS estimate. Although there has been much and controversial debate on the LT phenomenon during the last three decades, many scientific studies have dealt with LT concepts, their value in asSports Med 2009; 39 (6)

Validity of Lactate Thresholds

sessing endurance performance or in prescribing exercise intensities in endurance training. It might be speculated that a considerable part of the debate has to be attributed to the misinterpretation of the physiological basis of the phenomenon. The presented framework may help to clarify the controversy and may give a rational basis for performance diagnosis and training prescriptions in future research as well as in sports practice. Acknowledgements No sources of funding were used to assist in the preparation of this review. The authors have no conflicts of interest that are relevant to the content of this manuscript.

References 1. Gladden LB. Lactate metabolism: a new paradigm for the third millennium. J Physiol 2004 Jul 1; 558 (Pt 1): 5-30 2. Fletcher WM, Hopkins FG. Lactic acid in amphibian muscle. J Physiol (London) 1907; 35: 247-309 3. Meyerhof O. Untersuchung u¨ber die Wa¨rmestro¨mung der vitalen Oxydationsvorga¨nge. Biochem Z 1911; 5: 246-328 4. Douglas CG, Haldane JS. The regulation of normal breathing. J Physiol 1909; 38: 420-40 5. Ryffel GH. Lactic acid metabolism: a critical review. Quart J Med 1910; 3: 221-3 6. Hill AV, Lupton H. Muscular exercise, lactic acid and the supply and utilization of oxygen. Quart J Med 1923; 16: 135-71 7. Margaria R, Edwards HT, Dill DB. The possible mechanism of contracting and paying the oxygen debt and the role of lactic acid in muscular contraction. Am J Physiol 1933; 106: 689-714 8. Gladden LB. Muscle as a consumer of lactate. Med Sci Sports Exerc 2000 Apr; 32 (4): 764-71 9. Brooks GA. The lactate shuttle during exercise and recovery. Med Sci Sports Exerc 1986 Jun; 18 (3): 360-8 10. Robergs RA, Ghiasvand F, Parker D. Biochemistry of exercise-induced metabolic acidosis. Am J Physiol Regul Integr Comp Physiol 2004 Sep; 287 (3): R502-16 11. Coyle EF, Coggan AR, Hopper MK, et al. Determinants of endurance in well-trained cyclists. J Appl Physiol 1988; 64 (6): 2622-30 12. Coyle EF, Feltner ME, Kautz SA, et al. Physiological and biomechanical factors associated with elite endurance cycling performance. Med Sci Sports Exerc 1991 Jan; 23 (1): 93-107 13. Lucı´ a A, Pardo J, Dura´ntez A, et al. Physiological differences between professional and elite road cyclists. Int J Sports Med 1998; 19: 342-8 14. Impellizzeri FM, Marcora SM, Rampinini E, et al. Correlations between physiological variables and performance in high level cross-country off-road cyclists. Br J Sports Med 2005 Oct; 39 (10): 747-51

ª 2009 Adis Data Information BV. All rights reserved.

485

15. Sjodin B, Svedenhag J. Applied physiology of marathon running. Sports Med 1985 Mar-Apr; 2 (2): 83-99 16. Faria EW, Parker DL, Faria IE. The science of cycling: physiology and training, part 1. Sports Med 2005; 35 (4): 285-312 17. Jacobs I. Blood lactate: implications for training and sports performance. Sports Med 1986; 3 (1): 10-25 18. Conley DL, Krahenbuhl GS. Running economy and distance running performance of highly trained athletes. Med Sci Sports Exerc 1980; 12 (5): 357-60 19. Hollmann W. 42 years ago: development of the concepts of ventilatory and lactate threshold. Sports Med 2001; 31 (5): 315-20 20. Meyer T, Scharhag J, Kindermann W. Peak oxygen uptake: myth and truth about an internationally accepted reference value. Z Kardiol 2005 Apr; 94 (4): 255-64 21. Hollmann W. Ho¨chst- und dauerleistungsfa¨higkeit des sportlers. Mu¨nchen: Barth, 1963 22. Wasserman K, McIlroy MB. Detecting the threshold of anaerobic metabolism in cardiac patients. Am J Cardiol 1964; 14: 844-52 23. Wells JG, Balke B, Van Fossan DD. Lactic acid accumulation during work: a suggested standardization of work classification. J Appl Physiol 1957 Jan; 10 (1): 51-55 24. Mader A, Liesen H, Heck H, et al. Zur Beurteilung der sportartspezifischen ausdauerleistungsfa¨higkeit im labor. Sportarzt Sportmed 1976; 27: 80-8, 109-12 25. Atkinson G, Davison R, Jeukendrup A, et al. Science and cycling: current knowledge and future directions for research. J Sports Sci 2003 Sep; 21 (9): 767-87 26. Jones AM. The physiology of the world record holder for the women´s marathon. Int J Sports Sci Coaching 2006; 1 (2): 101-16 27. Svedahl K, MacIntosh BR. Anaerobic threshold: the concept and methods of measurement. Can J Appl Physiol 2003 Apr; 28 (2): 299-323 28. Brooks GA. Anaerobic threshold: review of the concept and directions for future research. Med Sci Sports Exerc 1985; 17 (1): 22-34 29. Myers J, Ashley E. Dangerous curves: a perspective on exercise, lactate, and the anaerobic threshold. Chest 1997 Mar; 111 (3): 787-95 30. Kindermann W, Simon G, Keul J. The significance of the aerobic-anaerobic transition for the determination of work load intensities during endurance training. Eur J Appl Physiol 1979; 42: 25-34 31. Meyer T, Lucia A, Earnest CP, et al. A conceptual framework for performance diagnosis and training prescription from submaximal gas exchange parameters: theory and application. Int J Sports Med 2005 Feb; 26 Suppl. 1: S38-48 32. McLellan TM, Skinner JS. The use of the aerobic threshold as a basis for training. Can J Appl Sport Sci 1981; 6 (4): 197-201 33. McLellan TM. The anaerobic threshold: concept and controversy. Austr J Sci Med Sport 1987; 19 (2): 3-8 34. Skinner JS, McLellan TH. The transition from aerobic to anaerobic metabolism. Res Q Exerc Sport 1980; 51 (1): 234-48

Sports Med 2009; 39 (6)

486

35. Wasserman K, Whipp BJ, Koyl SN, et al. Anaerobic threshold and respiratory gas exchange during exercise. J Appl Physiol 1973; 35 (2): 236-43 36. Dickhuth HH, Yin L, Niess A, et al. Ventilatory, lactatederived and catecholamine thresholds during incremental treadmill running: relationship and reproducibility. Int J Sports Med 1999 Feb; 20 (2): 122-7 37. Peronnet F, Aguilaniu B. Lactic acid buffering, nonmetabolic CO2 and exercise hyperventilation: a critical reappraisal. Respir Physiol Neurobiol 2006 Jan 25; 150 (1): 4-18 38. Yoshida T, Udo M, Chida M, et al. Specificity of physiological adaptation to endurance training in distance runners and competitive walkers. Eur J Appl Physiol Occup Physiol 1990; 61 (3-4): 197-201 39. Acevedo EO, Goldfarb AH. Increased training intensity effects on plasma lactate, ventilatory threshold, and endurance. Med Sci Sports Exerc 1989; 21 (5): 563-8 40. Bosquet L, Leger L, Legros P. Methods to determine aerobic endurance. Sports Med 2002; 32 (11): 675-700 41. Mujika I, Padilla S. Cardiorespiratory and metabolic characteristics of detraining in humans. Med Sci Sports Exerc 2001 Mar; 33 (3): 413-21 42. McLellan TM, Gass GC. The relationship between the ventilation and lactate thresholds following normal, low and high carbohydrate diets. Eur J Appl Physiol Occup Physiol 1989; 58 (6): 568-76 43. Reilly T, Woodbridge V. Effects of moderate dietary manipulations on swim performance and on blood lactateswimming velocity curves. Int J Sports Med 1999 Feb; 20 (2): 93-7 44. Yoshida T. Effect of dietary modifications on lactate threshold and onset of blood lactate accumulation during incremental exercise. Eur J Appl Physiol 1984; 53 (3): 200-5 45. Maassen N, Busse MW. The relationship between lactic acid and work load: a measure for endurance capacity or an indicator of carbohydrate deficiency? Eur J Appl Physiol Occup Physiol 1989; 58 (7): 728-37 46. Midgley AW, McNaughton LR, Jones AM. Training to enhance the physiological determinants of long-distance running performance: can valid recommendations be given to runners and coaches based on current scientific knowledge? Sports Med 2007; 37 (10): 857-80 47. Bentley DJ, Newell J, Bishop D. Incremental exercise test design and analysis: implications for performance diagnostics in endurance athletes. Sports Med 2007; 37 (7): 575-86 48. Foxdal P, Sjodin B, Sjodin A, et al. The validity and accuracy of blood lactate measurements for prediction of maximal endurance running capacity: dependency of analyzed blood media in combination with different designs of the exercise test. Int J Sports Med 1994; 15 (2): 89-95 49. Heck H, Hess G, Mader A. Comparative study of different lactate threshold concepts [Vergleichende Untersuchung zu verschiedenen Laktat-Schwellenkonzepten]. Dtsch Z Sportmed 1985; 36 (1+2): 19-25, 40-52 50. Heck H. Laktat in der Leistungsdiagnostik. Schorndorf: Hofmann, 1991

ª 2009 Adis Data Information BV. All rights reserved.

Faude et al.

51. Lundberg MA, Hughson RL, Weisiger KH, et al. Computerized estimation of lactate threshold. Comput Biomed Res 1986 Oct; 19 (5): 481-6 52. Grant S, McMillan K, Newell J, et al. Reproducibility of the blood lactate threshold, 4 mmol.l(-1) marker, heart rate and ratings of perceived exertion during incremental treadmill exercise in humans. Eur J Appl Physiol 2002 Jun; 87 (2): 159-66 53. Beaver WL, Wasserman K, Whipp BJ. Improved detection of lactate threshold during exercise using a log-log transformation. J Appl Physiol 1985; 59 (6): 1936-40 54. Cheng B, Kuipers H, Snyder AC, et al. A new approach for the determination of ventilatory and lactate thresholds. Int J Sports Med 1992; 13 (7): 518-22 55. Hughson RL, Weisiger KH, Swanson GD. Blood lactate concentration increases as a continuous function in progressive exercise. J Appl Physiol 1987; 62 (5): 1975-81 56. Robergs RA, Chwalbinska-Moneta J, Mitchell JB, et al. Blood lactate threshold differences between arterialized and venous blood. Int J Sports Med 1990; 11 (6): 446-51 57. Feliu J, Ventura JL, Segura R, et al. Differences between lactate concentration of samples from ear lobe and the finger tip. J Physiol Biochem 1999 Dec; 55 (4): 333-9 58. McNaughton LR, Thompson D, Philips G, et al. A comparison of the lactate Pro, Accusport, Analox GM7 and Kodak Ektachem lactate analysers in normal, hot and humid conditions. Int J Sports Med 2002 Feb; 23 (2): 130-5 59. Thin AG, Hamzah Z, FitzGerald MX, et al. Lactate determination in exercise testing using an electrochemical analyser: with or without blood lysis? Eur J Appl Physiol Occup Physiol 1999 Jan; 79 (2): 155-9 60. Buono MJ, Yeager JE. Intraerythrocyte and plasma lactate concentrations during exercise in humans. Eur J Appl Physiol Occup Physiol 1986; 55 (3): 326-9 61. Forsyth JJ, Farrally MR. A comparison of lactate concentration in plasma collected from the toe, ear, and fingertip after a simulated rowing exercise. Br J Sports Med 2000 Feb; 34 (1): 35-8 62. Draper N, Brent S, Hale B, et al. The influence of sampling site and assay method on lactate concentration in response to rock climbing. Eur J Appl Physiol 2006 Nov; 98 (4): 363-72 63. Hildebrand A, Lormes W, Emmert J, et al. Lactate concentration in plasma and red blood cells during incremental exercise. Int J Sports Med 2000 Oct; 21 (7): 463-8 64. Foxdal P, Sjodin B, Rudstam H, et al. Lactate concentration differences in plasma, whole blood, capillary finger blood and erythrocytes during submaximal graded exercise in humans. Eur J Appl Physiol Occup Physiol 1990; 61 (3-4): 218-22 65. Foxdal P, Sjodin A, Ostman B, et al. The effect of different blood sampling sites and analyses on the relationship between exercise intensity and 4.0 mmol.l-1 blood lactate concentration. Eur J Appl Physiol Occup Physiol 1991; 63 (1): 52-4 66. Medbø JI, Mamen A, Holt Olsen O, et al. Examination of four different instruments for measuring blood lactate concentration. Scand J Clin Lab Invest 2000 Aug; 60 (5): 367-80 67. Buckley JD, Bourdon PC, Woolford SM. Effect of measuring blood lactate concentrations using different auto-

Sports Med 2009; 39 (6)

Validity of Lactate Thresholds

68.

69. 70. 71. 72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

mated lactate analysers on blood lactate transition thresholds. J Sci Med Sport 2003 Dec; 6 (4): 408-21 van Someren KA, Howatson G, Nunan D, et al. Comparison of the Lactate Pro and Analox GM7 blood lactate analysers. Int J Sports Med 2005 Oct; 26 (8): 657-61 Bishop D. Evaluation of the Accusport lactate analyser. Int J Sports Med 2001 Oct; 22 (7): 525-30 Lucı´ a A, Hoyos J, Chicarro JL. Physiology of professional road cycling. Sports Med 2001; 31 (5): 325-37 Weltman A. The blood lactate response to exercise. Champaign IL: Human Kinetics, 1995 Heck H, Mader A, Hess G, et al. Justification of the 4-mmol/l lactate threshold. Int J Sports Med 1985; 6 (3): 117-30 Seiler KS, Kjerland GO. Quantifying training intensity distribution in elite endurance athletes: is there evidence for an ‘‘optimal’’ distribution? Scand J Med Sci Sports 2006 Feb; 16 (1): 49-56 Esteve-Lanao J, San Juan AF, Earnest CP, et al. How do endurance runners actually train? Relationship with competition performance. Med Sci Sports Exerc 2005 Mar; 37 (3): 496-504 Bourdon P. Blood lactate transition thresholds: concepts and controversies. In: Gore J, editor. Physiological tests for elite athletes/Australian Sports Commission. Champaign (IL): Human Kinetics, 2000: 50-65 Meyer T, Faude O, Urhausen A, et al. Different effects of two regeneration regimens on immunological parameters in cyclists. Med Sci Sports Exerc 2004 Oct; 36 (10): 1743-9 Meyer T, Gorge G, Schwaab B, et al. An alternative approach for exercise prescription and efficacy testing in patients with chronic heart failure: a randomized controlled training study. Am Heart J 2005 May; 149 (5): e1-7 McConnell TR, Clark BA, Conlin NC, et al. Gas exchange anaerobic threshold: implications for prescribing exercise in cardiac rehabilitation. J Cardiopulmonary Rehabil 1993; 13: 31-6 Faude O, Meyer T, Urhausen A, et al. Recovery training in cyclists: ergometric, hormonal and psychometric findings. Scand J Med Sci Sports. Epub 2008 Apr 23 Weltman A, Seip RL, Snead D, et al. Exercise training at and above the lactate threshold in previously untrained women. Int J Sports Med 1992; 13 (3): 257-63 Londeree BR. Effect of training on lactate/ventilatory thresholds: a meta-analysis. Med Sci Sports Exerc 1997 Jun; 29 (6): 837-43 Scharhag J, Meyer T, Gabriel HH, et al. Does prolonged cycling of moderate intensity affect immune cell function? Br J Sports Med 2005 Mar; 39 (3): 171-7 Meyer T, Gabriel HH, Auracher M, et al. Metabolic profile of 4 h cycling in the field with varying amounts of carbohydrate supply. Eur J Appl Physiol 2003 Jan; 88 (4-5): 431-7 Meyer T, Faude O, Gabriel H, et al. Ventilatory threshold and individual anaerobic threshold are reliable prescriptors for intensity of cycling training [abstract]. Med Sci Sports Exerc 2000; 32 Suppl. 5: S171 Baron B, Noakes TD, Dekerle J, et al. Why does exercise terminate at the maximal lactate steady state intensity? Br J Sports Med 2008 Oct; 42 (10): 528-33

ª 2009 Adis Data Information BV. All rights reserved.

487

86. Urhausen A, Coen B, Weiler B, et al. Individual anaerobic threshold and maximum lactate steady state. Int J Sports Med 1993; 14 (3): 134-9 87. Mader A, Heck H. A theory of the metabolic origin of ‘‘anaerobic threshold’’. Int J Sports Med 1986 Jun; 7 Suppl. 1: 45-65 88. Stegmann H, Kindermann W, Schnabel A. Lactate kinetics and individual anaerobic threshold. Int J Sports Med 1981; 2: 160-5 89. MacIntosh BR, Esau S, Svedahl K. The lactate minimum test for cycling: estimation of the maximal lactate steady state. Can J Appl Physiol 2002 Jun; 27 (3): 232-49 90. Lajoie C, Laurencelle L, Trudeau F. Physiological responses to cycling for 60 minutes at maximal lactate steady state. Can J Appl Physiol 2000 Aug; 25 (4): 250-61 91. McLellan TM, Jacobs I. Reliability reproducibility and validity of the individual anaerobic threshold. Eur J Appl Physiol 1993; 67 (2): 125-31 92. Van Schuylenbergh R, Vanden Eynde B, Hespel P. Correlations between lactate and ventilatory thresholds and the maximal lactate steady state in elite cyclists. Int J Sports Med 2004 Aug; 25 (6): 403-8 93. Beneke R, Hutler M, Leithauser RM. Maximal lactatesteady-state independent of performance. Med Sci Sports Exerc 2000 Jun; 32 (6): 1135-9 94. Beneke R, von Duvillard SP. Determination of maximal lactate steady state response in selected sports events. Med Sci Sports Exerc 1996 Feb; 28 (2): 241-6 95. Beneke R, Leithauser RM, Hutler M. Dependence of the maximal lactate steady state on the motor pattern of exercise. Br J Sports Med 2001 Jun; 35 (3): 192-6 96. Keul J, Simon G, Berg A, et al. Bestimmung der individuellen anaeroben Schwelle zur Leistungsbewertung und Trainingsgestaltung. Dtsch Z Sportmed 1979; 30: 212-8 97. Simon G, Berg A, Dickhuth H-H, et al. Bestimmung der anaeroben Schwelle in Abha¨ngigkeit von Alter und von der Leistungsfa¨higkeit. Dtsch Z Sportmed 1981; 32: 7-14 98. Coen B. Individuelle anaerobe Schwelle-Methodik und Anwendung in der sportmedizinischen Leistungsdiagnostik und Trainingssteuerung leichtathletischer Laufdisziplinen. Ko¨ln: Sport und Buch Strauss, 1997 99. Coen B, Schwarz L, Urhausen A, et al. Control of training in middle- and long-distance running by means of the individual anaerobic threshold. Int J Sports Med 1991; 12 (6): 519-24 100. Niess AM, Fehrenbach E, Strobel G, et al. Evaluation of stress responses to interval training at low and moderate altitudes. Med Sci Sports Exerc 2003 Feb; 35 (2): 263-9 101. Niess AM, Ro¨cker K, Baumann I, et al. Laktatverhalten bei extensiven Tempolaufbelastungen unter Flachlandund moderaten Ho¨henbedingungen. Leistungssport 1999; (3): 49-53 102. Faude O, Meyer T, Scharhag J, et al. Volume vs intensity in the training of competitive swimmers. Int J Sports Med 2008 Nov; 29 (11): 906-12 103. Smith CG, Jones AM. The relationship between critical velocity, maximal lactate steady-state velocity and lactate turnpoint velocity in runners. Eur J Appl Physiol 2001 Jul; 85 (1-2): 19-26

Sports Med 2009; 39 (6)

488

104. Davis HA, Bassett J, Hughes P, et al. Anaerobic threshold and lactate turnpoint. Eur J Appl Physiol 1983; 50 (3): 383-92 105. Urhausen A, Weiler B, Coen B, et al. Plasma catecholamines during endurance exercise of different intensities as related to the individual anaerobic threshold. Eur J Appl Physiol 1994; 69: 16-20 106. Gabriel H, Kindermann W. The acute immune response to exercise: what does it mean? Int J Sports Med 1997; 18 Suppl. 1: S28-45 107. Lucia A, Sanchez O, Carvajal A, et al. Analysis of the aerobic-anaerobic transition in elite cyclists during incremental exercise with the use of electromyography. Br J Sports Med 1999 Jun; 33 (3): 178-85 108. Sjodin B, Jacobs I. Onset of blood lactate accumulation and marathon running performance. Int J Sports Med 1981; 2 (1): 23-6 109. Farrell PA, Wilmore JH, Coyle EF, et al. Plasma lactate accumulation and distance running performance. Med Sci Sports 1979; 11 (4): 338-44 110. Yoshida T, Chida M, Ichioka M, et al. Blood lactate parameters related to aerobic capacity and endurance performance. Eur J Appl Physiol 1987; 56 (1): 7-11 111. Hagberg JM, Coyle EF. Physiological determinants of endurance performance as studied in competitive racewalkers. Med Sci Sports Exerc 1983; 15: 287-9 112. Jones AM, Doust JH. The validity of the lactate minimum test for determination of the maximal lactate steady state. Med Sci Sports Exerc 1998 Aug; 30 (8): 1304-13 113. Haverty M, Kenney WL, Hodgson JL. Lactate and gas exchange responses to incremental and steady state running. Br J Sports Med 1988; 22 (2): 51-4 114. Harnish CR, Swensen TC, Pate RR. Methods for estimating the maximal lactate steady state in trained cyclists. Med Sci Sports Exerc 2001 Jun; 33 (6): 1052-5 115. Beneke R. Methodological aspects of maximal lactate steady state: implications for performance testing. Eur J Appl Physiol 2003 Mar; 89 (1): 95-9 116. Billat VL, Sirvent P, Py G, et al. The concept of maximal lactate steady state: a bridge between biochemistry, physiology and sport science. Sports Med 2003; 33 (6): 407-26 117. Beneke R. Anaerobic threshold, individual anaerobic threshold, and maximal lactate steady state in rowing. Med Sci Sports Exerc 1995; 27 (6): 863-7 118. Baron B, Dekerle J, Robin S, et al. Maximal lactate steady state does not correspond to a complete physiological steady state. Int J Sports Med 2003 Nov; 24 (8): 582-7 119. Urhausen A, Coen B, Kindermann W. Individual assessment of the aerobic-anaerobic transition by measurements of blood lactate. In: Garrett Jr WE, Kirkendall DT, editors. Exercise and sport science. Philadelphia (PA): Lippincott Williams & Wilkins, 2000: 267-75 120. Mujika I, Busso T, Geyssant A, et al. Modeling the effects of training in competitive swimming. In: Troup JP, Hollander AP, Strasse D, editors. Biomechanics and medicine in swimming. London: E & FN Spon, 1996: 221-8 121. Kumagai S, Tanaka K, Matsuura Y, et al. Relationships of the anaerobic threshold with the 5 km, 10 km, and 10 mile races. Eur J Appl Physiol 1982; 49 (1): 13-23

ª 2009 Adis Data Information BV. All rights reserved.

Faude et al.

122. Aunola S, Rusko H. Does anaerobic threshold correlate with maximal lactate steady-state? J Sports Sci 1992; 10 (4): 309-23 123. Weltman A, Snead D, Seip R, et al. Prediction of lactate threshold and fixed blood lactate concentrations from 3200-m running performance in male runners. Int J Sports Med 1987 Dec; 8 (6): 401-6 124. Fo¨hrenbach R, Mader A, Hollmann W. Determination of endurance capacity and prediction of exercise intensities for training and competition in marathon runners. Int J Sports Med 1987; 8: 11-8 125. Hurley BF, Hagberg JM, Allen WK, et al. Effect of training on blood lactate levels during submaximal exercise. J Appl Physiol 1984; 56 (5): 1260-4 126. Ivy JL, Withers RT, Van Handel PJ, et al. Muscle respiratory capacity and fiber type as determinants of the lactate threshold. J Appl Physiol 1980 Mar; 48 (3): 523-7 127. Caiozzo VJ, Davis JA, Ellis JF, et al. A comparison of gas exchange indices used to detect the anaerobic threshold. J Appl Physiol 1982; 53 (5): 1184-9 128. Tanaka H. Predicting running velocity at blood lactate threshold from running performance tests in adolescent boys. Eur J Appl Physiol 1986; 55 (4): 344-8 129. Tanaka K, Matsuura Y. Marathon performance, anaerobic threshold, and onset of blood lactate accumulation. J Appl Physiol 1984; 57 (3): 640-3 130. Tanaka K, Matsuura Y, Matsuzaka A, et al. A longitudinal assessment of anaerobic threshold and distance-running performance. Med Sci Sports Exerc 1984 Jun; 16 (3): 278-82 131. Yoshida T, Udo M, Iwai K, et al. Significance of the contribution of aerobic and anaerobic components to several distance running performances in female athletes. Eur J Appl Physiol Occup Physiol 1990; 60 (4): 249-53 132. Yoshida T, Udo M, Iwai K, et al. Physiological characteristics related to endurance running performance in female distance runners. J Sports Sci 1993 Feb; 11 (1): 57-62 133. Davis JA, Vodak P, Wilmore JH, et al. Anaerobic threshold and maximal aerobic power for three modes of exercise. J Appl Physiol 1976 Oct; 41 (4): 544-50 134. Weltman J, Seip R, Levine S, et al. Prediction of lactate threshold and fixed blood lactate concentrations from 3200-m time trial running performance in untrained females. Int J Sports Med 1989 Jun; 10 (3): 207-11 135. Roecker K, Schotte O, Niess AM, et al. Predicting competition performance in long-distance running by means of a treadmill test. Med Sci Sports Exerc 1998 Oct; 30 (10): 1552-7 136. Dickhuth HH, Huonker M, Mu¨nzel T, et al. Individual anaerobic threshold for evaluation of competitive athletes and patients with left ventricular dysfunctions. In: Bachl N, Graham TE, Lo¨llgen H, editors. Advances in ergometry. Berlin: Springer, 1991: 173-9 137. Berg A, Stippig J, Keul J, et al. Zur Beurteilung der Leistungsfa¨higkeit und Belastbarkeit von Patienten mit coronarer Herzkrankheit. Dtsch Z Sportmed 1980; 31: 199-205 138. Hughson RL, Green HJ. Blood acid-base and lactate relationships studied by ramp work tests. Med Sci Sports Exerc 1982; 14 (4): 297-302

Sports Med 2009; 39 (6)

Validity of Lactate Thresholds

139. Coyle EF, Martin WH, Ehsani AA, et al. Blood lactate threshold in some well-trained ischemic heart disease patients. J Appl Physiol 1983 Jan; 54 (1): 18-23 140. Bishop D, Jenkins DG, Mackinnon LT. The relationship between plasma lactate parameters, Wpeak and 1-h cycling performance in women. Med Sci Sports Exerc 1998 Aug; 30 (8): 1270-5 141. Amann M, Subudhi AW, Foster C. Predictive validity of ventilatory and lactate thresholds for cycling time trial performance. Scand J Med Sci Sports 2006 Feb; 16 (1): 27-34 142. Yeh MP, Gardner RM, Adams TD, et al. ‘‘Anaerobic threshold’’: problems of determination and validation. J Appl Physiol 1983 Oct; 55 (4): 1178-86 143. Bunc V, Heller J, Novack J, et al. Determination of the individual anaerobic threshold. Acta Univ Carol, Gymnica 1985; 27: 73-81 144. Baldari C, Guidetti L. A simple method for individual anaerobic threshold as predictor of max lactate steady state. Med Sci Sports Exerc 2000 Oct; 32 (10): 1798-802 145. Tegtbur U, Busse MW, Braumann KM. Estimation of an individual equilibrium between lactate production and catabolism during exercise. Med Sci Sports Exerc 1993 May; 25 (5): 620-7 146. Stegmann H, Kindermann W. Comparison of prolonged exercise tests at the individual anaerobic threshold and the fixed anaerobic threshold of 4 mmol/l lactate. Int J Sports Med 1982; 3 (2): 105-10 147. Orok CJ, Hughson RL, Green HJ, et al. Blood lactate responses in incremental exercise as predictors of constant load performance. Eur J Appl Physiol 1989; 59 (4): 262-7 148. Carter H, Jones AM, Doust JH. Effect of incremental test protocol on the lactate minimum speed. Med Sci Sports Exerc 1999 Jun; 31 (6): 837-45 149. Coen B, Urhausen A, Kindermann W. Individual anaerobic threshold: methodological aspects of its assessment in running. Int J Sports Med 2001 Jan; 22 (1): 8-16 150. Weltman A, Snead D, Stein P, et al. Reliability and validity of a continuous incremental treadmill protocol for the determination of lactate threshold, fixed blood lac
tate concentrations, and VO2max. Int J Sports Med 1990; 11 (1): 26-32 151. Aunola S, Rusko H. Reproducibility of aerobic and anaerobic thresholds in 20-50 year old men. Eur J Appl Physiol 1984; 53 (3): 260-6 152. Pfitzinger P, Freedson PS. The reliability of lactate measurements during exercise. Int J Sports Med 1998 Jul; 19 (5): 349-57 153. Zhou S, Weston SB. Reliability of using the D-max method to define physiological responses to incremental exercise testing. Physiol Meas 1997 May; 18 (2): 145-54 154. Takeshima N, Tanaka K. Prediction of endurance running performance for middle-aged and older runners. Br J Sports Med 1995 Mar; 29 (1): 20-3 155. Tanaka K, Matsuura Y, Kumagai S, et al. Relationships of anaerobic threshold and onset of blood lactate accumulation with endurance performance. Eur J Appl Physiol Occup Physiol 1983; 52 (1): 51-6

ª 2009 Adis Data Information BV. All rights reserved.

489

156. Bourdin M, Messonnier L, Hager JP, et al. Peak power output predicts rowing ergometer performance in elite male rowers. Int J Sports Med 2004 Jul; 25 (5): 368-73 157. Bjorklund G, Pettersson S, Schagatay E. Performance predicting factors in prolonged exhausting exercise of varying intensity. Eur J Appl Physiol 2007 Mar; 99 (4): 423-9 158. Grant S, Craig I, Wilson J, et al. The relationship between 3 km running performance and selected physiological variables. J Sports Sci 1997 Aug; 15 (4): 403-10 159. Fay L, Londeree BR, LaFontaine TP, et al. Physiological parameters related to distance running performance in female athletes. Med Sci Sports Exerc 1989 Jun; 21 (3): 319-24 160. Nicholson RM, Sleivert GG. Indices of lactate threshold and their relationship with 10-km running velocity. Med Sci Sports Exerc 2001 Feb; 33 (2): 339-42 161. Lehmann M, Berg A, Kapp R, et al. Correlations between laboratory testing and distance running performance in marathoners of similar performance ability. Int J Sports Med 1983 Nov; 4 (4): 226-30 162. Tanaka K, Watanabe H, Konishi Y, et al. Longitudinal associations between anaerobic threshold and distance running performance. Eur J Appl Physiol Occup Physiol 1986; 55 (3): 248-52 163. Tokmakidis SP, Leger LA, Pilianidis TC. Failure to obtain a unique threshold on the blood lactate concentration curve during exercise. Eur J Appl Physiol Occup Physiol 1998 Mar; 77 (4): 333-42 164. Stratton E, O’Brien BJ, Harvey J, et al. Treadmill velocity best predicts 5000-m run performance. Int J Sports Med 2009; 30 (1): 40-5 165. Bentley DJ, McNaughton LR, Thompson D, et al. Peak power output, the lactate threshold, and time trial performance in cyclists. Med Sci Sports Exerc 2001 Dec; 33 (12): 2077-81 166. McNaughton LR, Roberts S, Bentley DJ. The relationship among peak power output, lactate threshold, and shortdistance cycling performance: effects of incremental exercise test design. J Strength Cond Res 2006 Feb; 20 (1): 157-61 167. Craig NP, Norton KI, Bourdon PC, et al. Aerobic and anaerobic indices contributing to track endurance cycling performance. Eur J Appl Physiol Occup Physiol 1993; 67 (2): 150-8 168. Nichols JF, Phares SL, Buono MJ. Relationship between blood lactate response to exercise and endurance performance in competitive female master cyclists. Int J Sports Med 1997 Aug; 18 (6): 458-63 169. Gregory J, Johns DP, Walls JT. Relative versus absolute physiological measures as predictors of mountain bike cross-country race performance. J Strength Cond Res 2007 Feb; 21 (1): 17-22 170. Impellizzeri FM, Rampinini E, Sassi A, et al. Physiological correlates to off-road cycling performance. J Sports Sci 2005 Jan; 23 (1): 41-47 171. Coyle EF. Improved muscular efficiency displayed as Tour de France champion matures. J Appl Physiol 2005 Jun; 98 (6): 2191-6

Sports Med 2009; 39 (6)

490

172. Lucia A, Earnest C, Arribas C. The Tour de France: a physiological review. Scand J Med Sci Sports 2003 Oct; 13 (5): 275-83 173. Impellizzeri FM, Marcora SM. The physiology of mountain biking. Sports Med 2007; 37 (1): 59-71 174. Yoshida T, Udo M, Iwai K, et al. Physiological determinants of race walking performance in female race walkers. Br J Sports Med 1989 Dec; 23 (4): 250-4 175. Ingham SA, Whyte GP, Jones K, et al. Determinants of 2000 m rowing ergometer performance in elite rowers. Eur J Appl Physiol 2002 Dec; 88 (3): 243-6 176. Cosgrove MJ, Wilson J, Watt D, et al. The relationship between selected physiological variables of rowers and rowing performance as determined by a 2000 m ergometer test. J Sports Sci 1999 Nov; 17 (11): 845-52 177. Noakes TD. The central governor model of exercise regulation applied to the marathon. Sports Med 2007; 37 (4-5): 374-7 178. Swensen TC, Harnish CR, Beitman L, et al. Noninvasive estimation of the maximal lactate steady state in trained cyclists. Med Sci Sports Exerc 1999 May; 31 (5): 742-6 179. Snyder AC, Woulfe T, Welsh R, et al. A simplified approach to estimating the maximal lactate steady state. Int J Sports Med 1994; 15 (1): 27-31 180. Kilding AE, Jones AM. Validity of a single-visit protocol to estimate the maximum lactate steady state. Med Sci Sports Exerc 2005 Oct; 37 (10): 1734-40 181. Billat V, Dalmay F, Antonini MT, et al. A method for determining the maximal steady state of blood lactate concentration from two levels of submaximal exercise. Eur J Appl Physiol Occup Physiol 1994; 69 (3): 196-202 182. Palmer AS, Potteiger JA, Nau KL, et al. A 1-day maximal lactate steady-state assessment protocol for trained runners. Med Sci Sports Exerc 1999 Sep; 31 (9): 1336-41 183. Loat CE, Rhodes EC. Relationship between the lactate and ventilatory thresholds during prolonged exercise. Sports Med 1993; 15 (2): 104-15 184. Simon J, Young JL, Gutin B, et al. Lactate accumulation relative to the anaerobic and respiratory compensation thresholds. J Appl Physiol 1983; 54 (1): 13-7 185. Yamamoto Y, Miyashita M, Hughson RL, et al. The ventilatory threshold gives maximal lactate steady state. Eur J Appl Physiol 1991; 63 (1): 55-9

ª 2009 Adis Data Information BV. All rights reserved.

Faude et al.

186. Scheen A, Juchmes J, Cession-Fossion A. Critical analysis of the ‘‘anaerobic threshold’’ during exercise at constant workloads. Eur J Appl Physiol Occup Physiol 1981; 46 (4): 367-77 187. Laplaud D, Guinot M, Favre-Juvin A, et al. Maximal lactate steady state determination with a single incremental test exercise. Eur J Appl Physiol 2006 Mar; 96 (4): 446-52 188. Dekerle J, Baron B, Dupont L, et al. Maximal lactate steady state, respiratory compensation threshold and critical power. Eur J Appl Physiol 2003 May; 89 (3-4): 281-8 189. Loat CER, Rhodes EC. Comparison of the lactate and ventilatory thresholds during prolonged work. Biol Sport 1996; 13 (1): 3-12 190. Schnabel A, Kindermann W, Schmitt WM, et al. Hormonal and metabolic consequences of prolonged running at the individual anaerobic threshold. Int J Sports Med 1982; 3 (3): 163-8 191. Oyono-Enguelle S, Heitz A, Marbach J, et al. Blood lactate during constant-load exercise at aerobic and anaerobic thresholds. Eur J Appl Physiol 1990; 60 (5): 321-30 192. Ribeiro JP, Hughes V, Fielding RA, et al. Metabolic and ventilatory responses to steady state exercise relative to lactate thresholds. Eur J Appl Physiol Occup Physiol 1986; 55 (2): 215-21 193. Bacon L, Kern M. Evaluating a test protocol for predicting maximum lactate steady state. J Sports Med Phys Fitness 1999 Dec; 39 (4): 300-8 194. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986 Feb 8; 1 (8476): 307-10 195. Atkinson G, Nevill AM. Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med 1998 Oct; 26 (4): 217-38

Correspondence: Dr Oliver Faude, Institute of Sports and Preventive Medicine, University of Saarland, Campus Bldg. B 8.2, 66123 Saarbru¨cken, Germany. E-mail: [email protected]

Sports Med 2009; 39 (6)

Sports Med 2009; 39 (6): 491-511 0112-1642/09/0006-0491/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

The Antidepressive Effects of Exercise A Meta-Analysis of Randomized Trials Chad D. Rethorst, Bradley M. Wipfli and Daniel M. Landers Arizona State University, Department of Kinesiology, Tempe, Arizona, USA

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Literature Search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Inclusion Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Data Extraction and Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Statistical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Overall Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Moderating Variables: Population Characteristics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Clinical Depression versus Non-Clinical Depression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Moderating Variables: Exercise Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Intervention Duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2 Exercise Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.3 Exercise Bout Duration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Moderating Variables: Methodological Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1 Concealment of Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.2 Intent to Treat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.3 Clinical Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Treatment Adherence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Exercise Compared with Other Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.1 Exercise versus Psychotherapy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.2 Exercise versus Antidepressant Medication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7 Dose Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Mechanisms of Antidepressive Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Using Exercise as a Treatment for Depression in Clinically Depressed Populations . . . . . . . . . . . . 4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

491 494 494 494 494 495 495 495 496 496 496 496 496 496 499 499 499 499 499 499 499 499 499 500 502 504 506

Several meta-analyses examining the effects of exercise on depression have been criticized for including studies of poor methodological integrity. More recent meta-analyses addressed the most common criticism by including only randomized control trials; however, these analyses suffer from incomplete literature searches and lack of moderating variable analyses. Using a more extensive search procedure, the current meta-analysis examines the effects of exercise on depressive symptoms in 58 randomized trials (n = 2982). An overall effect size of -0.80 indicates participants in the exercise treatment had significantly lower depression scores than those receiving the control treatment.

Rethorst et al.

492

This 3/4 SD advantage represents level 1, Grade A evidence for the effects of exercise upon depression. Analysis of moderating variables examined the influence of population characteristics, exercise characteristics and methodological characteristics. Examination of clinical significance in 16 trials with clinically depressed patients found 9 of 16 exercise treatment groups were classified as ‘recovered’ at post-treatment, with another three groups classified as ‘improved’. Analysis showed dropout rates for the exercise treatment were similar to those found in psychotherapeutic and drug interventions.

Depression disorders have become a widespread health concern throughout the world. In 1990, The Global Burden of Disease project ranked depressive disorders fourth in terms of global burden.[1] The worldwide prevalence of depression has been estimated at 10.4%.[2] Along with the prevalence of depressive disorders, the cost to treat these disorders has grown. In 2000, the National Institute of Mental Health (NIMH) estimated the cost to treat depressive disorders in the US at $US26 billion annually. In addition to the escalating costs associated with treatment, the accessibility and effectiveness of these treatments limit their impact. Only 55% of people afflicted with a depressive disorder are receiving treatment,[3] while alleviation of depressive symptoms was seen in only 32% of those receiving treatment. More recent research evidence has found that only 27.5% of depressed participants go into remission with initial medication treatment.[4] Of those who did not respond to initial treatment, between 17.6% and 24.8% responded to a switch in medication[5] and 30% responded to augmented medication.[6] These statistics indicate the need for more cost-effective, accessible and alternative treatments for depressive disorders. One such adjunct or alternative treatment that has been proposed is exercise. In Australia, the government has recently included the services of an exercise physiologist under the nation’s Medicare programme, allowing general practitioners to refer patients for a number of medical conditions including depression.[7] A similar movement has begun in the UK where in 2005 the Mental Health Foundation released a report encouraging general practitioners to use exercise as a frontline treatment for mild to moderate depression.[8] ª 2009 Adis Data Information BV. All rights reserved.

The effects of exercise on depression have been examined in hundreds of studies since the early 1900s. In an effort to reach a consensus finding for these studies, several meta-analyses have been conducted in the area.[9-11] The effect sizes of these meta-analyses ranged from 0.53 to 0.88, indicating a moderate to large effect. These results were consistent across all ages, sexes and various modes of exercise. Examination by North et al.[11] of moderating variables indicated larger effects associated with longer intervention durations and more exercise sessions,[11] while Craft and Landers[9] found similar results associated with longer intervention durations, and also found larger effects for groups with higher initial depression scores.[9] While these results provide evidence demonstrating the alleviation of depressive symptoms through exercise, these studies can be criticized for including studies of poor methodological integrity, such as quasi-experimental trials and cross-sectional studies. Guyatt et al.[12] developed guidelines for assessing the quality of research evidence based on the strength and consistency of results, methodology, sample size and cost/ benefit ratio.[12] According to these guidelines, Level 1, Grade A evidence, the highest level of recommendation, is the result of strong, clear-cut results from randomized controlled trials with very large sample sizes. These studies greatly reduced the likelihood of type I and type II errors through rigorous methodological procedures and a large sample size, respectively. Level 1, Grade A evidence can be provided either through one large randomized controlled trial or through a metaanalysis of smaller (Level II, Grade B) randomized controlled trials. Level II, Grade B studies Sports Med 2009; 39 (6)

The Antidepressive Effects of Exercise

are similar to Level 1, Grade A studies in that they are randomized controlled trials. However, these studies have smaller sample sizes and thus are susceptible to type II errors. A recent meta-analysis addressed this criticism by including only randomized controlled trials. Examining ten randomized trials (n = 479), Lawlor and Hopker[13] found an effect size of -1.1 (95% CI -0.6, -1.5), indicating that participants in exercise experimental groups had postintervention depression scores 1.1 standard deviation units lower than those receiving no treatment or a wait-list control treatment. However, Lawlor and Hopker argue that this evidence does not support the use of exercise in the treatment of depression, due to methodological weaknesses, including the lack of treatment concealment, the lack of intent to treat, and the lack of a clinical interview to confirm the diagnosis of the included studies. Lawlor and Hopker[13] characterized a study as using intent-to-treat analysis ‘‘if all the patients were analysed in the groups to which they were randomly allocated. If only those who started treatment or only those who completed treatment were included in the analysis we defined the study as not using [intent-to-treat] analysis’’ (p. 2). Likewise, a study was identified as using adequate concealment of allocation if treatment group assignment included ‘‘central randomisation at a site remote from the study, computerised allocation in which records are in a locked, unreadable file that can be accessed only after entering patient details, or the drawing of sealed and opaque sequentially numbered envelopes, and inadequately concealed if assignment procedures included open list or tables of random numbers, open computer systems or drawing of non-opaque envelopes, and unclear if no information in report, and the authors either did not respond to requests for information or were unable to provide information’’ (p. 2). Finally, Lawlor and Hopker believe that these results cannot be deemed clinically significant without a change in depression diagnosis as measured by a clinical interview by a psychologist. Lawlor and Hopker do not empirically derive these potential methodological problems through ª 2009 Adis Data Information BV. All rights reserved.

493

examination of moderating variables within their own analysis. Instead, they cite a study of pregnancy and childbirth data[14] that identifies potential methodological weaknesses to argue that a lack of treatment concealment and intent to treat might exaggerate the effectiveness of interventions by 20–40%. In other words, Lawlor and Hopker’s conclusion is speculative and the generalization of the findings may not be appropriate to a study of depression. Instead, these methodological weaknesses should be examined as moderating variables within a meta-analysis on exercise and depression. A narrative review by Brosse et al.[15] concurs with the conclusions drawn by Lawlor and Hopker. Brosse et al. conclude that while research ‘suggests’ that exercise is an effect treatment for depression, ‘‘the majority of studies suffer from significant methodological shortcomings’’. However, not all scholars are as quick to dismiss the meta-analytical results. Biddle[16] states that ‘‘even though the research designs included in the analysis were weak, the effect size was large’’ and in the same direction as found in narrative reviews[17] of experimental and epidemiological studies on this topic. Mutrie[18] argues that Lawlor and Hopker’s large effect size provides evidence of the effectiveness of exercise in decreasing depressive symptoms. Even if the methodological weaknesses are present in many of the studies, Callaghan[19] maintains that the Lawlor and Hopker results would not disappear, but only be weakened. According to Callaghan, a 40% reduction in effect size of 1.10 would still result in an overall effect size of 0.66 or a decrease of 7.3 on the Beck Depression Inventory (BDI). Stathopoulou et al.[20] conducted a similar meta-analysis, which included 11 randomized, controlled trials examining the effects of exercise on participants with affective disorders, and found an overall effect size of 1.39 (95% CI 0.89, 1.88). The larger effect size of this analysis compared with that of Lawlor and Hopker is most likely due to the exclusion of studies not published in peer-reviewed journals. Also, Stathopoulou et al. failed to address the methodological criticisms presented by Lawlor and Hopker, analyse moderating variables or conduct an Sports Med 2009; 39 (6)

Rethorst et al.

494

analysis of participant dropout rates. A final short-coming of this analysis is that Stathopoulou et al. base their recommendations for exercise dose on the few studies that directly compare varying doses of exercise rather than an examination of exercise dose as a moderating variable. The purpose of the current meta-analysis was to provide Level 1, Grade A evidence for the relationship between exercise and depression, marking the first time Level 1, Grade A evidence has been produced relative to the exercisedepression relationship. Analysis of moderating variables includes examination of exercise dose (e.g. intensity, duration and weeks of training) and is examined in the overall study sample and separately within those studies consisting of clinically depressed patients. In addition, this analysis examines the methodological characteristics of the included studies (e.g. lacking clinical interviews, treatment concealment and intent to treat) to determine the validity of the criticisms raised in the Lawlor and Hopker meta-analysis. Finally, the viability of the use of exercise in the treatment of clinically depressed populations and the potential mechanisms responsible for exercise reducing depression are discussed. Based on the results of previous metaanalyses, it is hypothesized that aerobic and resistance exercise programmes significantly alleviate depressive symptoms. In addition, participants with clinical levels of depression show improvements greater than those of the general population. Finally, intervention duration is hypothesized to be positively correlated with improvements in depression. 1. Methods 1.1 Literature Search

An electronic literature search was conducted for all articles published in 2005 or prior using PubMed, PsycINFO, SportDiscus and Dissertations Abstracts International using the terms ‘exercise’, ‘physical activity’, ‘running’, ‘jogging’, ‘walking’, ‘weight lifting’, ‘weight training’, ‘depression’ and ‘mental health’ to identify potential studies for inclusion. In addition, bibliographies ª 2009 Adis Data Information BV. All rights reserved.

of all articles found through electronic search on the topic were also examined to identify additional studies. 1.2 Inclusion Criteria

To be included in the analysis, a trial must have used a form of moderate to vigorous exercise (aerobic or resistance) as a treatment condition and must have measured depression as a dependent variable. Only randomized controlled trials were included in the analysis. Studies were included if the control conditions were a notreatment or wait-list control. Studies that used control groups who participated in light exercise, such as stretching or walking, were excluded from the analysis. 1.3 Data Extraction and Collection

Data were extracted from the included studies entered into an electronic database by the two lead authors. In cases where all relevant data were not included in the published manuscript, the authors were contacted via postal mail. If no correspondence was received from the authors, contact was attempted through phone or electronic mail. At this time, authors were also asked to report if they were aware of any other studies that examined exercise and depression. Studies were first coded based on the population of the study (general population vs clinically depressed population). Studies were coded as using a clinically depressed population if they were identified by the original studies as being clinically depressed, either through a clinical interview or a screening and the depression was not associated with other physical or psychological ailments such as fibromyalgia, cardiac infarction or chemical dependence. Studies were then coded for moderating variables including sex, dependent variable measurement, fitness improvement, intervention duration, exercise bout intensity, frequency and duration, and were examined in the overall population and within the clinically depressed population. The first author coded the studies and the second author coded a random sample of ten studies to examine coding reliability. Sports Med 2009; 39 (6)

The Antidepressive Effects of Exercise

1.4 Statistical Analysis

The effect sizes (ESs) for each study were calculated using Hedges’ g, g = (MT – MC)/SDpooled, where MT is the post-treatment mean of the treatment group and MC is the post-treatment mean of the control group.[21] In the 31 studies that contained multiple treatment groups or used multiple measures of depression, a single average effect was calculated. ESs were corrected for sample size according to procedures outlined by Hedges and Olkin.[21] Finally, based on the recommendations of Hedges and Olkin and Hunter and Schmidt,[22] a random effects model was used, and ESs were weighted by the inverse of the variance to calculate the overall ES.[23] Additionally, gains ESs were calculated for exercise and control groups, in which case the calculation for ES becomes MPost – MPre/SDPooled, where MPost is the post-treatment mean of the intervention group and MPre is the pre-treatment mean of the intervention group. A one-sample t-test was used to compare the overall effect size with zero. An overall Q value and I2 value were calculated to test for homogeneity of variance among the ESs. This Q value represents the total amount of variance among the set of ESs and was tested against a w2 distribution, in which df = k - 1, where k = number of ESs. I2 is calculated using the formula I2 = 100% · (Q – df)/Q. According to Higgins et al.,[24] this provides a precise, easily interpreted measure of heterogeneity. I2 values of 25%, 50% and 75% represent low, medium and high heterogeneity, respectively. A significant Q value indicates that the data are heterogeneous, and would warrant the examination of moderator variables by dividing the variance into Qw and Qb. Qb values are also tested against a w2 distribution, in which df = number of categories of the moderator variable – 1. A significant Qb value indicates that the moderator variable contributes to the variance among ESs. Weighted ESs and standard errors were then calculated for each category within the moderator variables, and 95% confidence intervals were calculated to determine whether or not each ES was significantly different from zero.[25] To determine ª 2009 Adis Data Information BV. All rights reserved.

495

significant differences between levels of each moderating variable, planned comparisons were calculated for all moderating variables that yielded a significant Qb.[21] The significance test for each planned comparison compared w2 to a Bonferroni-adjusted critical value (0.05/number of comparisons within moderating variable). Moderating variable analysis was conducted for the entire sample of ESs and also across the clinical population. The volume of exercise used in each study was estimated in units of energy (kcal/kg) expended during exercise, assuming 5 kcal/mL O2 consumed. Total O2 consumed/kg/wk was calculated as: . Total O2 consumed ¼ ðVO2 max Þ  ðintensityÞ ðbout durationÞ  ðfrequencyÞ . where total O2 units are mL/kg/wk, VO2max (maximum oxygen uptake) units are mL/kg/min, . average exercise bout intensity is in %VO2max, average bout duration is in minutes, and exercise frequency is number of exercise bouts per week. For determining volume of exercise, only those . were used, and pre- and studies reporting VO . 2max post-intervention VO2max measurements were averaged in the calculation.[26] After calculating exercise dose, the data were examined for outliers. Any exercise dose >3 standard deviations away from the mean were considered outliers, and were removed from the analysis.[27] Pearson’s product-moment correlations were then calculated between exercise dose and the corresponding ESs for the exercise groups. Additionally, a multiple regression analysis was used to examine a potential nonlinear relationship between exercise dose and effect size. 2. Results 2.1 Overall Results

The literature search resulted in 149 articles that were examined for inclusion criteria, with 75 of the studies meeting all inclusion criteria. Reasons for exclusion included the lack of a notreatment control group, experimental groups receiving exercise in combination with another treatment, or the lack of true randomization. Sports Med 2009; 39 (6)

Rethorst et al.

496

Of these studies, sufficient information to calculate effect sizes was obtained from 58 of the studies,[28-91] with a total population of 2982. Participants in the exercise treatment had significantly lower depression scores than those receiving the control treatment (ES = -0.80, 95% CI -0.92, 0.67). Fifty-six studies contained sufficient information to calculate gains effect sizes. Analysis of gains effect sizes revealed the average effect size for exercise treatment groups was 1.07, which is significantly different from zero (t(55) = -5.505; p < 0.05; 95% CI -1.458, -0.680). The average effect size for the control groups was -0.20, which was found to be not statistically different from zero (t(55) = -1.413; p > 0.05; 95% CI -0.481, 0.083). The average difference in BDI between treatment and control groups in the overall samples was 3.83 and the average difference on the Hamilton Rating Scale for Depression (HRSD) was 3.43. The test for homogeneity of variance was found to be significant (Q = 364.25; p < 0.001), warranting the examination of moderating variables. I2 was calculated to be 84.35%. Inter-rater reliability for the coding of moderating variables was found to be 0.93. Results from analysis of moderating variables can be found in table I and table II. Significant Qb values were found for intervention duration, exercise type, bout duration, exercise frequency, clinical diagnosis and exercise intensity (tables III–VI). 2.2 Moderating Variables: Population Characteristics 2.2.1 Clinical Depression versus Non-Clinical Depression

Seventeen studies examined the effect of exercise on clinically depressed participants (n = 574). Clinically depressed participants in the exercise treatment had significantly lower depression scores than those receiving the control treatment (ES = -1.03). Within the clinically depressed population, the average change in BDI was 10.60, while the average change in HRSD was 8.11. In the 40 studies that examined non-clinical samples (n = 2408), participants had significantly lower depression scores than those receiving the control ª 2009 Adis Data Information BV. All rights reserved.

treatment (ES = -0.59). Within the non-clinical population, the average BDI change was 2.64 and no studies used the HRSD. A planned comparison revealed the effect size of the clinical population was significantly larger than that of the general population (w2[1, n = 58] = 19.20; p < 0.05; 95% CI 0.246, 0.645). 2.3 Moderating Variables: Exercise Characteristics 2.3.1 Intervention Duration

Within the overall population, planned comparisons revealed that the effects of interventions lasting 4–9 weeks were significantly greater than the effects of interventions lasting 17–26 weeks (w2[1, n = 24] = 7.72; p = 0.0055; 95% CI -0.677, -0.117). Interventions lasting 10–16 weeks resulted in significantly larger effects compared with interventions lasting 17–26 weeks (w2[1, n = 64] = 13.22; p = 0.0003; 95% CI -0.765, -0.229) and >26 weeks (w2[1, n = 31] = 7.53; p = 0.0061; 95% CI -0.543, -0.090). Within the clinical population, a planned comparison revealed that interventions of 10–16 weeks in duration resulted in significantly larger effects compared with interventions of 4–9 weeks in length (w2[1, n = 15] = 4.87; p = 0.0273; 95% CI 0.102, -0.870). 2.3.2 Exercise Type

Within the overall population, planned comparisons revealed that a regimen of combined aerobic and resistance exercise resulted in significantly larger effects than aerobic exercise (w2[1, n = 52] = 22.93; p < 0.0001; 95% CI 0.643, 1.533) and resistance exercise (w2[1, n = 54] = 20.71; p < 0.0001; 95% CI 0.674, 1.694). However, within the clinically depressed population, no significant differences were found between exercise types. 2.3.3 Exercise Bout Duration

Within the overall population, planned comparisons revealed that bout durations of 20–29 minutes resulted in significantly larger effects than exercise bouts of 45–59 minutes (w2[1, n = 20] = 6.70; p = 0.0096; 95% CI -0.601, -0.083) and ‡60 minutes (w2[1, n = 24] = 7.64; p = 0.0057; 95% CI -0.558, -0.095). Within the clinical Sports Med 2009; 39 (6)

Study

Depression ES measure

N

Population

Sex

Age (y) [mean Exercise or range] type

Intervention length (wk)

Exercise frequency

Bout duration (min)

Intensity

GDS

-1.48

46

NC

M

66.97

Aerobic

24

3/wk

20–60

Bartholomew et al. (2005) Berger et al.[30] (1988)

POMS-D POMS-D

-0.26 -0.12

40 305

C NC

X X

38.10 20.00

Aerobic Aerobic

Acute 12

NA 3/wk

30 20+

. 50–60% VO2max 60–70% HRmax 65–80% HRmax

Blumenthal et al.[31] (1991)

BDI

-0.18

101

NC

X

67.00

Aerobic

16

3/wk

30

70% HRmax

Blumenthal et al.[32] (2005)

BDI

-3.14

134

NC

X

63.00

Aerobic

16

3/wk

35

70–85% HRmax

Broocks et al.[33] (1998)

Multiplea

-1.24

37

C

X

46.00

Aerobic

10

3–4/wk

NA

NA

Brown et al.[34] (2001)

Multiplea

-0.63

104

NC

F

42.00

Aerobic

8

5/wk

20

60% HRmax

Burrus[35] (1984)

DACL

-0.32

45

C

X

16.20

Aerobic

9

4/wk

35

NA

Castro et al.[36] (2002)

BDI

-0.30

85

NC

F

62.00

Aerobic

52

3–4/wk

30–40

60–75% HRmax NA

Antunes et al.[28] (2005) [29]

Chin A Paw et al.[37] (2004)

GDS

0.18

173

NC

X

64–94

Resistance

2/wk

45–60

Cramer et al.[38] (1991)

POMS-D

-0.44

35

NC

X

34.00

Aerobic

15

5/wk

45

60% HRmax

Crews et al.[39] (2004)

BDI

-0.89

66

NC

X

9.00

Aerobic

6

3/wk

20

134 bpm

Deivert[40] (1990)

Multiplea

-1.90

40

NC

X

31.00

Aerobic

8

3/wk

20

NA

BDI

-0.55

77

NC

X

20.27

Aerobic

12

3/wk

20–40

NA

DiLorenzo et al.[42] (1998)

Multiplea

-0.70

111

NC

X

32.00

Aerobic

12

4/wk

24 or 48

Dugmore et al.[43] (1999)

TAS-D

-4.07

124

NC

X

55.00

Combined

52

3/wk

NA

70–85% HRmax . 65–80% VO2max

DePalma

[41]

(1989)

6 mo

Dunn et al.[44] (2005)

HRSD

-0.64

80

C

X

35.00

Aerobic

12

3/wk

NA

NA

Eby[45] (1984)

Zung

-0.07

39

NC

X

19–31

Combined

NA

3/wk

60–90

NA

Emery et al.[46] (1998)

Multiplea

-0.42

74

NC

X

66.60

Aerobic

10

NA

NA

NA

Epstein[47] (1986)

Multiplea

-0.84

26

C

X

39.42

Aerobic

8

3/wk

30

NA

GDS

-0.57

64

NC

X

73.00

Resistance

6

3/wk

NA

80% 1 RM

Gowans et al.[49] (2001)

Multiplea

-0.83

31

NC

X

47.00

Aerobic

23

3/wk

30

60–75% HRmax

Gowans et al.[50] (2002)

Multiplea

-0.75

31

NC

X

47.00

Aerobic

23

3/wk

30

60–75% HRmax

Hembree et al.[51] (2000)

BDI

-0.69

53

NC

F

79.83

Combined

4

5/wk

NA

NA

Hilyer et al.[52] (1982)

Multiplea

-1.14

43

NC

M

15–18

Combined

20

3/wk

90

NA

Jorgensen[53] (1986)

SCL-90-D

0.00

11

C

X

36.40

Aerobic

6

3/wk

Kanner[54] (1990)

Multiplea

-0.73

68

C

X

13.32

Aerobic

8

3/wk

60

70–85% HRmax

King et al.[55] (1989)

Own

-0.16

113

NC

X

48.00

Aerobic

24

5/wk

50

65–77% HRmax

King et al.[56] (1993)

BDI

-0.12

300

NC

X

50–65

Aerobic

52

3/wk

40

Koukouvou et al.[58] (2004)

Multiplea

-1.66

26

NC

M

52.50

Aerobic

24

3–4/wk

60

73–88% HRmax . 50–75% VO2max

1h

TZHR

Lennox et al.[59] (1990)

DACL

-0.29

47

NC

X

45.00

Combined

13

3/wk

50–60

NA

Levin[60] (1983)

Multiplea

-0.67

38

C

X

26–35

Aerobic

10

3/wk

60

NA

Martinsen et al.[61] (1985)

BDI

-1.14

43

C

X

40.00

Aerobic

9

3/wk

60

50–70% HRmax Continued next page

497

Sports Med 2009; 39 (6)

Favilla[48] (1992)

The Antidepressive Effects of Exercise

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Methodological details of studies comparing exercise vs a no-treatment control group

498

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Contd Study

Depression ES measure

N

Population

Sex

Age (y) [mean Exercise or range] type

Intervention length (wk)

Exercise frequency

Bout duration (min)

Intensity

McCann and Holmes[62] (1984)

BDI

-1.01

43

C

F

NA

Aerobic

10

2/wk

60

NA

McNeil et al.[63] (1991)

BDI

-1.02

30

C

X

72.50

Aerobic

6

3/wk

40

NA

Mutrie[64] (1998)

Multiplea

-1.51

24

C

X

42.10

Combined

4

3/wk

20

NA

Neidig et al.[65] (2003)

Multiplea

-0.53

48

NC

X

36.00

Aerobic

12

3/wk

60

. 60–80% VO2max

a

-0.48

22

NC

X

<70

Aerobic

10

3/wk

60

60–80% HRmax

0.10

80

NC

X

16.50

Aerobic

10

2/wk

25–30

70–75% HRmax

Newton et al.[66] (1991)

Multiple

Norris et al.[67] (1992)

MAACL

Norvell and Belles[68] (1993)

SCL-90-D

-1.17

29

NC

M

32.84

Resistance

16

3/wk

20

NA

Petajan et al.[69] (1996)

POMS-D

-1.41

54

NC

X

40.00

Aerobic

15

3/wk

50

NA

Pierce et al.[70] (1993)

STAI-D

-0.10

99

NC

X

45.00

Aerobic

16

3/wk

50

. 70% VO2max

[71]

HRSD

-3.52

18

C

X

35.20

Aerobic

1

7/wk

60

75% HRmax

Roth et al.[72] (1987)

BDI

-0.20

36

NC

X

18.90

Aerobic

11

3/wk

30

75% HRmax

Roth and Homes[73] (1989)

POMS-D

-2.66

80

NC

X

20.80

Aerobic

NA-Acute

NA-acute

20

115–160 bpm

Setaro[74] (1985)

MMPI-D

-1.08

75

C

X

18–35

Aerobic

10

2/wk

NA

NA NA

Pinchasov et al.

(2000)

Simons and Birkimer[75] (1988)

POMS-D

-0.10

128

NC

X

43.20

Aerobic

8

2/wk

90

Singh et al.[76] (1997)

Multiplea

-2.20

32

C

X

71.00

Resistance

10

3/wk

45

80% 1 RM

Singh et al.[77] (2001)

BDI

-0.67

32

C

X

71.30

Resistance

20

3/wk

45

80% 1 RM

Singh et al.[78] (2005)

Multiplea

-0.58

60

C

X

60+

Resistance

8

3/wk

60

80% 1 RM

GHQ

-0.28

219

NC

X

44.90

Aerobic

52

3/wk

60

60–80% HRmax

Taylor[80] (1991)

BDI

-0.49

102

NC

F

39.10

Aerobic

6

3/wk

20

NA

van den Berg et al.[81] (2004)

HADS-D

-0.38

34

NC

X

58.60

Aerobic

3 mo

2/wk

Veale et al.[82] (1992)

Multiplea

-3.03

83

C

X

35.50

Aerobic

Wigers et al.[83] (1996)

VAS-D

-0.31

48

NC

X

44.00

Aerobic

14

3/wk

45

60–70% HRmax

Williams and Lord[84] (1997)

DASS

-0.22

187

NC

F

71.70

Aerobic

42

2/wk

60

NA

[79]

Sorensen et al.

(1999)

12

3/wk

1h NA

60% HR reserve NA

Wilson[85] (1985)

Zung

-1.30

34

NC

F

31.00

Aerobic

16

3/wk

40

70–85% HRmax

Zentner[86] (1981)

POMS-D

-0.75

80

NC

X

41.00

Aerobic

10

3/wk

60

NA

a

Broocks – MADRS, BDI; Brown – CES-D, GWB-D, POMS-D; Deivert – POMS-D, BDI; DiLorenzo – BDI, POMS-D; Emery – CES-D, SCL-90-D; Epstein – BDI, Zung; Gowans – BDI, MHI-D; Gowans – BDI, CES-D, FIQ-D, MHI-D; Hilyer – BDI, POMS-D; Kanner – CDI, ISC; Koukouvou – BDI, HADS-D; Levin – POMS-D, SCL-90-D; Mutrie – BDI, POMS-D; Neidig – CES-D, BDI, POMS-D; Newton – POMS-D, BDI; Singh – BDI, HRSD, GDS, DSM-IV; Singh – GDS, HRSD; Veale – CIS, BDI.

Rethorst et al.

Sports Med 2009; 39 (6)

BDI = Beck Depression Inventory; bpm = beats/min; C = clinical; CDI = Children’s Depression Inventory; CES-D = Center for Epidemiologic Studies Depression Scale; CIS = Clinical Interview Schedule; DACL = Depression Adjective Checklist; DASS = Depression Anxiety Stress Scales, Depression subscale; DSM-IV = Diagnostic and Statistical Manual, Depression symptoms; ES = effect size; F = female; FIQ-D = Fibromyalgia Impact Questionnaire, Depression subscale; GDS = Geriatric Depression Scale; GHQ = General Health Questionnaire, Depression subscale; GWB-D = General Well-Being Schedule, Depression subscale; HADS-D = Hospital Anxiety and Depression Scale, Depression subscale; HRmax = maximum heart rate; HRSD = Hamilton Rating Scale for Depression; ISC = Interview Schedule for Children; M = male; MAACL = Multiple Affect Adjective Checklist, Depression subscale; MADRS = Montgomery-Asberg Depression Rating Scale; MHI-D = Mental Health Inventory, Depression subscale; MMPI-D = Minnesota Multiphasic Personality Inventory, Depression subscale; NA = not available; NC = non-clinical; POMS-D = Profile of Mood States, Depression subscale; RM = repetition maximum; SCL90-D = Symptom Checklist-90, Depression subscale; STAI-D = State Trait Anxiety. Inventory, Depression subscale; TAS-D = Toronto Attitude Scale, Depression subscale; TZHR = theoretical zone heart rate; VAS-D = Visual Analogue Scales of Depression; VO2max = maximum oxygen uptake; X = mixed sex; Zung = Zung Self-Rating Depression Scale.

The Antidepressive Effects of Exercise

499

Table II. Methodological characteristics of clinically depressed population studies Study

Concealment

Intent to treat

Clinical interview

Bartholomew et al.[29] (2005)

No

No

Yes

Broocks et al.[33] (1998)

Yes

Yes

Yes No

Burrus[35] (1984)

No

No

Dunn et al.[44] (2005)

Yes

Yes

Yes

Epstein[47] (1986)

No

No

Yes

Kanner[54] (1990)

No

No

Yes

Levin[60] (1983)

No

No

No

Martinsen et al.[61] (1985)

Yes

No

No

McCann and Holmes[62] (1984) No

No

Yes

McNeil et al.[63] (1991)

No

No

Yes

Mutrie[64] (1998)

No

No

Yes

Pinchasov[71] et al. (2000)

No

No

Yes

Setaro[74] (1985)

No

No

No

Singh et al.[76] (1997)

No

No

No

Singh et al.[77] (2001)

Yes

No

Yes

Singh et al.[78] (2005)

No

No

No

Veale et al.[82] (1992)

Yes

Yes

Yes

population, planned comparisons revealed a significantly larger effect for exercise bouts of 45–59 minutes compared with exercise bouts of 30–44 minutes (w2[1, n = 7] = 10.75; p = 0.0010; 95% CI 0.443, 1.762) and ‡60 minutes (w2[1, n = 9] = 6.29; p = 0.0122; 95% CI -1.42, -0.219). 2.4 Moderating Variables: Methodological Characteristics

The methodological characteristics of the studies with a clinically depressed population are described in table II. 2.4.1 Concealment of Treatment

In the clinically depressed population, the effect sizes of studies that used adequate concealment were significantly larger than the effect sizes of studies that did not use adequate concealment (w2[1, n = 17] = 5.45; p = 0.0196; 95% CI 0.106, 0.893). 2.4.2 Intent to Treat

In the clinically depressed population, a planned comparison revealed that studies that used adequate intent to treat resulted in significantly ª 2009 Adis Data Information BV. All rights reserved.

higher effect sizes than those that did not (w2[1, n = 17] = 10.76; p = 0.0010; 95% CI 0.357, 1.340). 2.4.3 Clinical Interview

A planned comparison indicated the effect sizes of studies that used a clinical interview to confirm a depression diagnosis did not significantly differ from those studies that did not conduct a clinical interview (w2[1, n = 17] = 0.43; p = 0.3818; 95% CI -0.484, 0.242). 2.5 Treatment Adherence

Of the 18 studies that included clinically depressed patients, 16 studies reported participant dropout rates. In those studies, the average percentage of dropout in the exercise group was 14.6%, while the percentage of dropout in control groups was 11.4%. The dropout rates of each group were not significantly different from each other (F[1, 13] = 0.272; p > 0.05). 2.6 Exercise Compared with Other Treatments 2.6.1 Exercise versus Psychotherapy

Four studies[47,57,74,87] compared exercise with psychotherapy, resulting in an overall effect size of -0.26, indicating that exercise resulted in larger antidepressive effects than psychotherapy. However, this difference was not significant (t = 1.686; p > 0.05; 95% CI -0.628, 0.116). 2.6.2 Exercise versus Antidepressant Medication

Three studies[32,86,87] compared exercise versus antidepressants and found an overall effect size of 0.02. This difference was not significant (t = 0.223; p > 0.05; 95% CI -0.152, 0.184). 2.7 Dose Response

Twelve studies provided adequate data to calculate exercise volume. One study yielded an exercise dose that was >3 standard deviations above the mean of all exercise doses and was removed from the analysis. Some of the remaining 11 studies included multiple exercise groups, resulting in 25 exercise groups. The Pearson productmoment correlation for exercise dose and effect sizes was found to be nonsignificant (r = -0.040; p > 0.05). Multiple regression analysis also Sports Med 2009; 39 (6)

Rethorst et al.

500

Table III. Moderating variables for the overall population Category

df

Qb

Level

k

ES

SE

Upper CI

Lower CI

Population

1

19.19883*

Non-clinical

41

-0.58685

0.043453

-0.50049

-0.67082

Clinical Age (y)

3

5.108655

Sex

Intervention duration (wk)

2

4

4.339101

45.9213*

17

-1.03137

0.091975

-0.85110

-1.21164

<21

9

-0.67545

0.095481

-0.48831

-0.86259

21–34

8

-0.88846

0.116993

-0.65915

-1.11776

35–55

22

-0.58707

0.064817

-0.46003

-0.71411

>55

17

-0.65607

0.069295

-0.52026

-0.79189

Male

7

-0.72193

0.128623

-0.46982

-0.97403

Female

7

-0.49163

0.093083

-0.30918

-0.67407

Mixed

44

-0.7031

0.046652

-0.61166

-0.79453

Acute

2

-1.50348

0.220743

-1.07082

-1.93613 -0.79534

4–9

16

-0.64662

0.075876

-0.49791

10–16

26

-0.74666

0.063409

-0.62238

-0.87094

17–26

8

-0.24955

0.121108

-0.01218

-0.48692 -0.61906

>26 Exercise type

2

23.85307*

5

-0.43008

0.096417

-0.24111

48

-0.64056

0.042285

-0.55768

-0.72344

6

-0.54462

0.133631

-0.2827

-0.80653

Combined

4

-1.72853

0.22322

-1.29102

-2.16604

20–29

9

-0.79508

0.092961

-0.61287

-0.97728

30–44

14

-0.65653

0.083749

-0.49238

-0.82068

Aerobic Resistance

Bout duration (min)

Exercise frequency (/wk)

2

2

10.3643*

28.75021*

45–59

11

-0.45277

0.093997

-0.26854

-0.637

‡60

15

-0.46854

0.072793

-0.32586

-0.61121

7

-0.24831

0.089114

-0.07365

-0.42298

44

-0.78172

0.04868

-0.6863

-0.87713

5

4

-0.52009

0.12201

-0.28095

-0.75923

50–60

3

-0.76479

0.154123

-0.16667

-1.06687

61–74

11

-0.33339

0.08506

-0.16667

-0.50011

‡75

10

-0.8478

0.09814

-0.65544

-1.04015

2 3–4

Exercise intensity (%)

2

17.28126*

df = degrees of freedom; ES = effect size; k = number of effect sizes; Qb = measure of homogeneity, see text; SE = standard error; * p < 0.05.

revealed a nonsignificant relationship (F[1,23] = 0.037; p > 0.05). 3. Discussion Aerobic and resistance exercise programmes were hypothesized to significantly alleviate depressive symptoms. The overall effect size indicates an improvement in depression scores of 0.80 standard deviation units following an exercise programme. Analysis of moderating variables indicates aerobic and resistance exercises are equally effective. This finding supports the initial hypothesis and is consistent with the findings of Craft and Landers.[9] In the overall samª 2009 Adis Data Information BV. All rights reserved.

ple, analysis of moderating variables indicated that exercise interventions that combined aerobic and resistance exercise resulted in larger effects than aerobic or resistance exercise alone. It should be noted, however, that this effect size is based on only four trials, and further research must be conducted to confirm this finding. Additionally, within the clinically depressed population, aerobic and resistance exercises were found to be equally effective in alleviating depressive symptoms. The second hypothesis of this study stated that participants with clinical levels of depression would show improvements greater than those of the general population. Improvements within Sports Med 2009; 39 (6)

The Antidepressive Effects of Exercise

501

the clinically depressed population were 1.01 standard deviation units compared with 0.59 in the general population. This difference was found to be statistically significant and in line with the initial hypothesis and the previous findings of Craft and Landers.[9] The final hypothesis of this study was that longer intervention durations would result in greater improvements in depressive symptoms. This hypothesis was based on the findings of North et al.[11] and Craft and Landers[9] in which longer intervention durations resulted in larger decreases in depression scores. Analysis of moderating variables supports this hypothesis within the clinically depressed population, where interventions of 10–16 weeks resulted in larger effects than interventions lasting 4–9 weeks. However, within the overall population, interventions of 4–9 weeks resulted in significantly larger effects than interventions of 17–26 weeks, while interventions of 10–16 weeks resulted in significantly

larger effects than interventions of 16–26 weeks and >26 weeks. One possible explanation for these differences is a floor effect in the general population, where maximum improvements were achieved in the first 16 weeks of exercise training. Analysis also indicated significant differences in moderating variable categories that were not hypothesized. Within the overall population, exercise bouts of 20–29 minutes resulted in larger effects than bouts of ‡45 minutes, while within the clinically depressed population, exercise bouts of 45–49 minutes resulted in larger effects than bouts of 30–44 minutes and of ‡60 minutes. Once again, however, the analysis of moderating variables within the clinically depressed population includes a small number of trials, and more research must be done before conclusions can be drawn on the optimal exercise bout duration. Significant differences were also present across categories of exercise intensity across the overall population, with exercise of 61–74% maximum

Table IV. Planned comparisons for moderating variables (overall population) Category

Comparison

w2

p-Value

Population

Clinical vs non-clinical

19.199**

<0.0001**

4–9 vs 10–16

1.0234

0.3118

4–9 vs 17–26

7.7196

0.0055**

Intervention duration (wk)

4–9 vs >26 10–16 vs 17–26

Exercise type

Bout duration (min)

Exercise frequency (/wk)

Exercise intensity (%)

0.7339 13.223

0.3916 0.0003**

10–16 vs >26

7.5258

0.0061**

17–16 vs >26

1.3601

0.2435

Aerobic vs resistance

0.4686

0.4934

Aerobic vs combined

22.933**

<0.0001**

Resistance vs combined

20.709**

<0.0001**

20–29 vs 30–44

1.2261

0.2682

20–29 vs 45–59

6.7042**

0.0096*

20–29 vs ‡ 60

7.6487**

0.0057**

30–44 vs 45–59

2.6195

0.1055

30–44 vs ‡60

2.8703

0.0902

45–49 vs ‡60

0.0176

0.8945

27.593**

<0.0001**

2 vs 5

3.2356

0.0721

3–4 vs 5

3.9667

0.0464*

50–60 vs 61–74

6.0056*

0.0143**

2 vs 3–4

50–60 vs ‡75

0.2064

0.6496

61–74 vs ‡75

15.689**

<0.0001**

*p < 0.05, ** p < Bonferroni-adjusted p-value.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (6)

Rethorst et al.

502

Table V. Moderating variables for clinical population Category

df

Age (y)

3

Qb 5.251954

Sex

1

0.000269

Intervention duration (wk)

2

13.29871**

Exercise type

1

0.001189

Bout duration (min)

3

46.11954**

Exercise frequency (/wk)

2

16.68409**

Level <21

1

51.69517**

Clinical interview

1

0.428708

Intent to treat

1

11.45184**

Concealment

1

6.180855*

2

ES

SE

Upper CI

Lower CI

-0.56676

0.236152

-0.1039

-1.02962

21–34

2

-0.89392

0.228353

-0.44635

-1.34150

35–55

8

-1.27989

0.146603

-0.99254

-1.56722

>55

4

-1.0051

0.199561

-0.61396

-1.39624

Female

1

-1.01316

0.325082

-0.3760

-1.65032

Mixed

16

-1.03296

0.095894

-0.8450

-1.22091

Acute

1

-0.26461

0.317609

4–9

8

-0.90384

0.142851

-0.62385

10–16

7

-1.34226

0.138093

-1.0716

-1.61292

14

-1.03779

0.101263

-0.83931

-1.23626

Resistance

3

-1.00114

0.219851

-0.57023

-1.43205

20–29

1

-1.50875

0.570179

-0.3912

-2.6263

30–44

4

-0.49985

0.198906

-0.11

-0.88971

45–59

3

-1.60254

0.271284

-1.07082

-2.13426

‡60

5

-0.82846

0.216866

-0.40340

-1.25352

2

2

-1.04986

0.221579

-0.61557

-1.48416

13

-1.06424

0.107732

-0.85308

-1.27539

5

1

-3.51607

0.752085

-2.04198

-4.99015

61–74

1

-0.26461

0.317609

‡75

4

-0.78516

0.189078

-0.41456

-1.15575

Yes

11

-1.09942

0.120473

-0.86330

-1.33555

No

6

-0.93628

0.142408

-0.65716

-1.21540

No

14

-0.90001

0.100313

-0.70340

-1.09662

Yes

3

-1.72456

0.230435

-1.27291

-2.17622

No

12

-0.89263

0.109516

-0.67798

-1.10728

Yes

5

-1.36348

0.169438

-1.03139

-1.69558

Aerobic

3–4 Exercise intensity (%)

k

0.357902

0.357902

-0.88712 -1.18383

-0.88712

df = degrees of freedom; ES = effect size; k = number of effect sizes; Qb = measure of homogeneity, see text; SE = standard error; * p < 0.05, ** p < 0.01.

heart rate resulting in lower effects than exercise of lower intensity (50–60%) and higher intensity (‡75%). Within the clinically depressed population, no significant differences were found. This finding was also supported by the separate doseresponse analysis that indicated there was not a significant relationship between energy expenditure and changes in depression score. This is another area that will require further research because only 24 studies reported adequate information pertaining to exercise intensity, while only five studies within the clinically depressed population reported sufficient exercise intensity. In addition, all studies that reported exercise intensity used intensities ranging from 50% to 85%. Future research must also examine exercise intensities outside of this range. ª 2009 Adis Data Information BV. All rights reserved.

Finally, significant differences in effects were found for exercise frequency within the clinically depressed population, as exercising five times per week resulted in a significantly larger effect than exercise two to four times per week. However, only one study utilized a protocol of five exercise sessions per week. 3.1 Mechanisms of Antidepressive Effects

It has been hypothesized that decreased rates of adult neurogenesis are, at least in part, responsible for depressed mood.[92] Recent research has found that current antidepressant medications result in hippocampal neurogenesis in laboratory animals.[93] From these results, Ernst et al.[94] hypothesize that antidepressive effects of Sports Med 2009; 39 (6)

The Antidepressive Effects of Exercise

exercise are due to physiological changes that result in hippocampal neurogenesis. Ernst et al. identify four mechanisms by which exercise could potentially facilitate this neurogenesis. Firstly, an increase in b-endorphins, which have been linked to neurogenesis,[95] have also been found to be increased following exercise.[96] Similarly, vascular endothelial growth factor (VEGF) increases during exercise[97] and has been linked to hippocampal neurogenesis in adult rats,[98,99] while blockage of VEGF eliminated exercise-induced neurogenesis.[100] A third potential mechanism identified by Ernst et al. is brain-derived neurotrophic factor (BDNF), which has been shown to be increased by exercise in a number of studies,[101] while Wozniak[102] identified BDNF as having a crucial role in neuronal development and survival. Furthermore, when exercise is combined with antidepressants, BDNF levels were found to increase in as little as 2 days, compared with 2 weeks with antidepressants alone.[103] Finally, Ernst et al. examined the role of serotonin in the antidepressive effects of exercise. Jacobs[104] reported depressed patients typically have lower levels of serotonin. In fact, most current antidepressant medications target the release and reuptake of serotonin. Exercise increases tryptophan hydroxylase,[105] which is necessary for serotonin synthesis, while Brezun and Daszuta[106] linked increases in serotonin to neurogenesis, and decreases in serotonin with decreased neurogenesis in adult rats. Results of animal studies indicate that physical activity can increase the neural discharge of serotonin.[104,105] Serotonin levels may also be influenced by exercise-induced changes in sleep. Meta-analytical reviews have shown that exercise results in increased total sleep, increased slow-wave sleep and decreased REM sleep.[107,108] Serotonin discharge decreases significantly during REM sleep,[109] meaning that decreasing REM sleep will limit the times in which serotonin discharge is at its lowest. In addition to the hippocampal neurogenesis hypothesis put forth by Ernst et al., other physiological and psychological changes may be responsible for the antidepressive effects of exercise. ª 2009 Adis Data Information BV. All rights reserved.

503

Dietrich and McDaniel[110] explored the role of endocannabinoids in exercise-induced analgesia, sedation and anxiolysis. The endocannabinoid system consists of two receptors, CB1 and CB2, which are activated by two naturally occurring ligands, anandamide and 2-arcachidonylglycerol. Activation of the endocannabinoid system has been found to have an analgesic effect.[111] Sparling et al.[112] discovered increased anandamide levels following a bout of exercise, suggesting that exercise activates the endocannabinoid system. Other potential physiological mechanisms include altered hypothalamic-pituitary-adrenal (HPA) axis functioning and increased levels of noradrenaline (norepinephrine). Altered regulation of the HPA axis, resulting in increased secretion of corticotrophin-releasing hormone, has been associated with depression,[113] while exercise has been found to produce changes in basal HPA function.[114] A review of animal studies[115] concluded that exercise training results in a delayed HPA axis response to stress. In humans, it has been found that exercise-trained individuals exhibit an attenuated HPA axis response to physical and mental stress.[116-118] Depression has also been linked with lowered levels of noradrenaline. Current antidepressant medications have been shown to increase noradrenaline levels,[119] while Dishman[120] reports lower levels of noradrenaline metabolites in the urine of depressed patients and increased noradrenaline levels following exercise.

Table VI. Planned comparisons for moderating variables (clinical population) Category

Comparison

Intervention duration (wk)

4–9 vs 10–16

Bout duration (min)

30–44 vs 45–59

w2 4.8690 10.745

p-Value 0.0273** 0.0010**

30–44 vs ‡60

1.7788

0.1823

45–59 vs ‡60

6.2864

0.0122**

Exercise frequency (/wk)

2 vs 3–4

0.0034

0.9535

Clinical interview

Yes vs no

0.7649

Intent to treat

Yes vs no

10.7641**

0.0010**

Concealment

Yes vs no

5.4470*

0.0196**

0.3818

* p < 0.05, ** p < Bonferroni-adjusted p-value.

Sports Med 2009; 39 (6)

Rethorst et al.

504

In addition to these physiological mechanisms, psychological factors have the potential to influence the relationship between physical activity and depression. In a study of clinically depressed women, Ossip-Klein et al.[121] found that both aerobic and resistance training resulted in enhanced self-esteem, which was attributed to improved body image and an increased sense of mastery. This enhanced self-esteem was accompanied by a decrease in depressive symptoms, suggesting that an enhancement of selfesteem may be responsible for the alleviation of depressive symptoms. 3.2 Using Exercise as a Treatment for Depression in Clinically Depressed Populations

The overall findings of the current metaanalysis are similar to those of past meta-analyses that used primarily cross-sectional or correlational data and those meta-analyses consisting of only randomized trials, including Lawlor and Hopker.[13] However, Lawlor and Hopker dismissed their findings based on perceived methodological weaknesses including concealment of treatment, lack of a clinical interview, and intent to treat. The judgment of methodological weakness relied on a study of pregnancy and childbirth data,[14] which may not be applicable to the study of depression. Our analysis of moderating variables included examination of these methodological characteristics and found no significant difference in effect size due to the presence or absence of a clinical interview. The significant effects of concealment of treatment and intentto-treat analysis are actually the opposite of that predicted by Lawlor and Hopker, with those studies using concealment of treatment and intent-to-treat analysis achieving larger effects than those that did not. These results suggest that the findings of Schulz et al.[14] cannot be generalized to research studying treatments for major depression. Lawlor and Hopker’s second criticism of their results was based on the use of self-report measures of depression, stating that such instruments are ‘‘difficult to interpret clinically’’. They believe ª 2009 Adis Data Information BV. All rights reserved.

that a dichotomous result (depressed vs not depressed) would be more easily understood and more meaningful than reporting changes in depression scale scores. However, a growing base of literature has been devoted to the development of procedures that can assess the clinical significance of changes in self-report measures. Jacobsen et al.[122] proposed a two-step method for analysing clinical significance by examining a change in score from pre-test to post-test. The first step is to establish a cutoff point that must be crossed in moving from the depressed population to the general population. The second criterion used to establish clinical significance is a change in score greater than a pre-established reliable change index (RCI). According to Jacobsen et al., individuals can be classified as ‘recovered’ if they pass both criteria, ‘improved’ if they meet one of the criteria, ‘unchanged’ if they do not improve on either, or ‘deteriorated’ if their scores are higher post-treatment. Seggar et al.[123] examined existing literature to establish a cutoff point (14.29) and RCI (8.46) for the BDI. Examination of the mean pre- and postBDI scores of studies included in this metaanalysis that used clinically depressed samples (table VII) showed that six of the nine treatment groups crossed over the cutoff and exceeded the required RCI, indicating that the average score of the sample could be classified as ‘recovered’ following treatment. Of the three treatment groups that did not reach ‘recovered’ criteria, two met the criteria for ‘improvement’ by crossing the cutoff point, but not exceeding the required RCI. This could be attributed to lower pretreatment scores in these two samples.[33,63] Only one treatment group[64] could be classified as ‘unchanged’ at post-treatment. The average score for this study fell just short of meeting both criteria, with a post-treatment BDI score of 14.63 and an RCI of 7.23. Similarly, Grundy et al.[124] established a cutoff point (11.78) and RCI (7.74) for the HRSD. In examining the mean pre- and post-HRSD scores of studies that used clinical samples in this meta-analysis (table VIII), three of the seven treatment groups crossed over the cutoff and exceeded the required RCI, indicating that the Sports Med 2009; 39 (6)

The Antidepressive Effects of Exercise

505

Table VII. Assessing clinical significance in studies using the Beck Depression Inventory Study

Treatment group

Martinsen et al.[61] (1985)a

Exercise Control Exercise

21

Singh et al.[77] (2001)a Singh et al.[76] (1997)a Veale et al.[82] (1992)a McNeil et al.[63] (1991) Epstein[47] (1986)a Broocks et al.[33] (1998) Mutrie[64] (1998)a

a

Pre

Post

Change

25.2

12.1b

13.1c

31.5

22.8 9.2b

8.7 11.8c

Control

18.4

Exercise

21.3

11

Control

18.4

13.8

4.6

Exercise

22.91

13.94b

8.97c

Control

26.66

17.79

8.87

Exercise

16.6

11.1b

5.5

Control

15.2

14.7

0.5

Exercise

25.29

9b

16.29c

Control

22

Exercise

15.2

Control

18.3

Exercise

22.44

9.8b

7.4 11.5c

16.3 7.2b 15.8

5.7 8 2.5

9.46b 12.98c

Exercise

21.86

14.63

7.23

Control

23

21.42

1.58

Clinically significant change in depression scores.

b

Post-treatment scored crossing the cutoff point of 14.29.

c

Pretreatment to post-treatment change >8.46.

average score of the sample could be classified as ‘recovered’ following treatment. Of the four treatment groups that did not reach ‘recovered’ criteria, one met the criteria for ‘improvement’ by crossing the cutoff point, but not exceeding the required RCI.[76] This could be attributed to lower pretreatment scores in this sample. The three treatment groups that could be classified as ‘unchanged’ at post-treatment were all comparison groups that used lower intensity exercise than treatment groups that showed ‘recovery’. The two treatment groups that were ‘unchanged’ in the Dunn et al.[44] study exercised at an intensity of 7 kcal/wk compared with 17.5 kcal/wk for those groups that were ‘recovered’ at post-treatment measurement. In the study by Singh et al.,[78] the ‘unchanged’ group participated in resistance exercise using 20% of one-repetition maximum compared with the ‘recovered’ group exercising using 80% of one-repetition maximum. Ideally, the clinical significance protocol would be applied to individuals rather than group means. Therefore, some caution must be taken in interª 2009 Adis Data Information BV. All rights reserved.

preting the clinical significance results. A group treatment score reaching the ‘recovered’ criteria does not indicate that all participants in that group had reached the ‘recovered’ criteria. However, assuming a normal distribution of depression scores, the majority of participants in these groups would have reached ‘recovered’ criteria. A second cause for caution is the terms used in labelling the observed changes. Even though a change in depression score is labelled as ‘recovered,’ this does not necessarily mean that these individuals are free of depressive symptoms. Despite this shortcoming, previous research has found significant differences in those individuals reaching ‘recovered’ status. Using the same statistical method and a similar BDI cutoff point of 14.47, McGlinchey et al.[125] found that those patients classified as ‘recovered’ following treatment were less likely to relapse than those who did not reach ‘recovered’ status. In the analysis of clinical significance, three of the four ‘unchanged’ treatment groups exercised at a lower intensity, suggesting a potential doseresponse relationship. In addition, within the clinically depressed population there was a trend towards greater effect sizes with longer durations. However, our analysis of the dose-response relationship resulted in a nonsignificant finding, and examination of moderating variables revealed a nonsignificant effect for exercise intensity. One

Table VIII. Assessing clinical significance in treatment groups using the Hamilton Rating Scale for Depression Study

Treatment group

Pre

Singh et al.[78] (2005)

Exercise

18

Singh et al.[76] (1997) Dunn[44] (2005)

Post 8.5a

Change 9.5b

Exercise

19.5

12.4

7.1

Control

18.7

14.4

4.3

Exercise

12.3

5.3a

7

Control

11.4

8.5

2.9

Exercise

19.3

11.7

7.6

Exercise

19.2

12.8

6.4

Exercise

19.1

9a

10.1b

Exercise

19.1

10

a

9.1b

Control

20.5

14

6.5

a

Post-treatment scored crossing the cutoff point of 11.28.

b

Pretreatment to post-treatment change >7.74.

Sports Med 2009; 39 (6)

Rethorst et al.

506

potential explanation for the nonsignificant finding is the small sample size. Of the 58 studies included in the meta-analysis, only 12 studies provided adequate information to calculate an exercise dose. In addition, only three studies in the meta-analysis compared different intensities of exercise.[44,55,78] Future research should focus on examining the effects of various exercise intensities, durations and frequencies. Lawlor and Hopker assumed that the dropout rate for exercise would be ‘similar or worse’ than that of medication or psychotherapy. This is a long-standing assumption based on Zipf’s ‘principle of least effort’,[126] that postulates that people’s choices are made with an emphasis on minimizing effort over time. Following this logic, people would likely choose medication or psychotherapy as a treatment because they require less physical effort than exercise. The current analysis found that the dropout rates of exercise treatment groups were no different than the dropout rates for control groups. The exercise dropout rate of 14.6% is at least equivalent to, if not lower than, the dropout rates that have been calculated for antidepressant treatment, which has been found to be 21–33%,[127-130] and psychotherapy, which has been reported to range from 17.2% to 26%.[131] Finally, to completely assess the quality of a potential treatment, one must examine the costbenefit ratio of the intervention. The costs and health risks associated with exercise include time, cost of gym memberships, muscle soreness, perspiration, fatigue and effort expenditure. Conversely, the costs and health risks of the most recommended treatments for depression – psychotherapy and drug treatments – include high monetary costs, and in the case of drug treatments include mania, lethargy, sleep disturbances, weight loss, bleeding, sexual dysfunction, seizures and suicidal ideations,[132] while the benefits of these treatments typically do not expand beyond alleviation of depression. Additional benefits of exercise include: (i) reduced risk of cardiovascular disease, high blood pressure, colon, breast and prostate cancers, and noninsulin-dependent diabetes; (ii) improved mortality rates and cognitive functioning; and ª 2009 Adis Data Information BV. All rights reserved.

(iii) maintenance of normal strength, joint structure and function, and peak bone mass.[133-136] In Australia, referral to an exercise physiologist is now reimbursable under the nation’s Medicare system.[137] The Mental Health Foundation of the UK recommends that general practitioners should be offering an exercise programme to all patients with depression. Despite this recommendation, only 5% of general practitioners use exercise referral as one of the three most recommended treatment options, while 92% use antidepressants as one of the three most recommended treatments. This disparity is explained by the fact that only 41% of general practitioners believe exercise to be ‘very effective’ or ‘quite effective’ in treating depression.[8] In the US, neither the American Psychological Association nor the NIMH recognize exercise as a treatment for depression. 4. Conclusion In order for exercise to become a recommended form of treatment for depressive disorders in the US, researchers must provide conclusive results that demonstrate the effectiveness of exercise in treating depression. The current meta-analysis is the first step in providing those conclusive results. By including only randomized trials with a cumulative sample of nearly 3000 participants, the results of this meta-analysis provide Level 1, Grade A evidence supporting the use of exercise for the alleviation of depressive symptoms. Furthermore, a very large effect size was found within the clinically depressed population of over 500 participants, with the majority of those studies showing improvement that was clinically significant. These findings support the use of exercise in the treatment of major depression. However, further research is needed in a number of areas. First, while our analysis suggests the possibility of a dose-response relationship in terms of intervention duration, exercise intensity, bout duration and exercise frequency, further research is needed to confirm these findings. Other individual studies[44,78] have found significant differences between exercises of different intensities, yet these findings are not Sports Med 2009; 39 (6)

The Antidepressive Effects of Exercise

enough to draw concrete conclusions regarding a dose-response relationship. Secondly, the current analysis found no significant differences between exercise and psychotherapy or antidepressant medications. However, this is based on only a few studies that have directly compared the effects of exercise to these recognized treatments for depression. In addition to comparing exercise with antidepressant medications and psychotherapy, future research should also examine the use of exercise as an adjunct to these recognized treatments. Finally, the studies included in this analysis focus on the immediate effects of exercise on depression. Follow-up research is needed that examines the sustainability of these effects after exercise has ceased. Acknowledgements No sources of funding were used to assist in the preparation of this review. The authors have no conflicts of interest that are directly relevant to the content of this review.

507

9.

10.

11. 12.

13.

14.

15.

16.

References 1. Murray CJL, Lopez AD, editors. The global burden of disease: a comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020. Global Burden of Disease and Injury Series, Vol. 1. Cambridge (MA): Harvard University Press, 1996 2. Ustun T, Satorisus N. Mental illness in general health care: an international study. New York: John Wiley, 1995: 323-34 3. Andrews G, Sanderson K, Corry J, et al. Using epidemiological data to model efficiency in reducing the burden of depression. J Mental Health Pol Econ 2000; 3 (4): 175-86 4. Trivedi M, Rush A, Wisniewski S, et al. Evaluation of outcomes with citalopram for depression using measurementbased care in STARD: implications for clinical practice. Am J Psychiatr 2000; 163: 28-40 5. Rush A, Trivedia M, Wisniewski S, et al. Bupropion-SR, sertraline, or venlafaxine-XR after failure of SSRIs for depression. New Engl J Med 2006; 354: 1231-42 6. Trivedi M, Fava M, Wisniewski M, et al. Medication augmentation after the failure of SSRIs for depression. New Engl J Med 2006; 354: 1243-52 7. Minister for Health and Ageing. Exercise physiologists eligible to provide services under Medicare (ABB106/05) 2005, September 6 [online]. Available from URL: http:// www.sport.gov.au/internet/ministers/publishing.nsf/con tent/ [Accessed 2009 Apr 1] 8. Mental Health Foundation. Up and running? Exercise therapy and the treatment of mild or moderate depression in primary care [online]. Available from URL: http:// www.mentalhealth.org.uk/EasysiteWeb/getresource.axd?

ª 2009 Adis Data Information BV. All rights reserved.

17.

18. 19.

20.

21. 22.

23.

24. 25.

AssetID=38660&type=Full&servicetype=Attachment [Accessed 2006 Nov 1] Craft L, Landers D. The effect of exercise on clinical depression and depression resulting from mental illness: a meta-analysis. J Sport Exerc Psychol 1998; 20 (4): 339-57 Stich F. A meta-analysis of physical exercise as a treatment for symptoms of anxiety and depression [dissertation]. Madison (WI): University of Wisconsin-Madison, 1999 North T, McCullagh P, Tran Z. The effect of exercise on depression. Exerc Sport Sci Rev 1990; 18: 379-415 Guyatt G, Gutterman D, Baumann M, et al. Grading strength of recommendations and quality of evidence in clinical guidelines: report from an American College of Chest Physicians Task Force. Chest 2006; 129 (1): 174-81 Lawlor DA, Hopker SW. The effectiveness of exercise as an intervention in the management of depression: a systematic review and meta-regression analysis of randomized controlled trials [unabridged electronic version]. BMJ 2001; 322: 1-8 [online]. Available from URL: http://www. bmj.com/cgi/content/full/322/7289/ Schulz KF, Chalmers I, Hayes RJ, et al. Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA 1995; 273 (5): 408-12 Brosse A, Sheets E, Lett H, et al. Exercise and the treatment of clinical depression in adults: recent findings and future directions. Sport Med 2002; 32 (12): 741-60 Biddle S. The effectiveness of exercise as an intervention in the management of depression: systematic review and meta-regression analysis of randomised controlled trials [online]. Available from URL: http://bmj.bmjjournals. com/cgi/eletters/322/7289/763 [Accessed 2006 Jul 13] Mutrie N. The relationship between physical activity and clinically defined depression. In: Biddle S, Fox K, Boutcher S, editors. Physical activity and psychological well-being. London: Routledge, 2001: 46-63 Mutrie N. Healthy body, healthy mind? Psychologist 2002; 15 (8): 412-3 Callaghan P. Exercise: a neglected intervention in mental health care? J Psychiatr Mental Health Nurs 2004; 11 (4): 476-783 Stathopoulou G, Power M, Berry A, et al. Exercise interventions for mental health: a quantitative and qualitative review. Clin Psychol: Sci Pract 2006; 13 (2): 179-93 Hedges LV, Olkin I. Statistical methods for meta-analysis. Orlando (FL): Academic Press, 1985 Hunter JE, Schmidt FE. Methods of meta-analysis: correcting error and bias in research findings. Thousand Oaks (CA): Sage, 2004 Shadish WR, Haddock CK. Combining estimates of effect size. In: Cooper H, Hedges LV, editors. The handbook of research synthesis. New York: Russell Sage Foundation, 2004: 231-44 Higgins JPT, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analysis. BMJ 2003; 327: 557-60 Hedges LV. Fixed effects models. In: Cooper H, Hedges LV, editors. The handbook of research synthesis. New York: Russell Sage Foundation, 1994: 285-99

Sports Med 2009; 39 (6)

508

26. Powers SK, Howley ET. Exercise physiology: theory and applications to fitness and performance. New York: McGraw-Hill, 2003 27. Cohen J, Cohen P, West S, et al. Applied multiple regression/ correlation analysis for the behavioral sciences. Mahwah (NJ): Lawrence Erlbaum Associates, 2003 28. Antunes HKM, Stella SG, Santos RF, et al. Depression, anxiety, and quality of life scores in seniors after an endurance exercise program. Rev Brasil Psiquiatr 2005; 27 (4): 266-71 29. Bartholomew J, Morrison D, Ciccolo J. Effects of acute exercise on mood and well-being in patients with major depressive disorder. Med Sci Sport Exerc 2005; 37 (12): 2032-7 30. Berger B, Friedmann E, Eaton M. Comparison of jogging, the relaxation response, and group interaction for stress reduction. J Sport Exerc Psychol 1988; 10 (4): 431-47 31. Blumenthal J, Emery C, Madden D, et al. Long-term effects of exercise on psychological functioning in older men and women. J Gerontol 1991; 46 (6): 352-61 32. Blumenthal J, Sherwood A, Babyak M, et al. Effects of exercise and stress management training on markers of cardiovascular risk in patients with ischemic heart disease: a randomized controlled trial. JAMA 2005; 293 (13): 1626-34 33. Broocks A, Bandelow B, Pekrun G, et al. Comparison of aerobic exercise, clomipramine, and placebo in the treatment of panic disorder. Am J Psychiatr 1998; 155 (5): 603-9 34. Brown MA, Goldstein-Shirley J, Robinson J, et al. The effects of a multi-modal intervention trial of light exercise and vitamins on women’s mood. Women Health 2001; 34 (3): 93-112 35. Burrus M. The effects of a running treatment program on depressed adolescents [dissertation]. Miami (FL): University of Miami, 1984 36. Castro CM, Wilcox S, O’Sullivan P, et al. An exercise program for women who are caring for relatives with dementia. Psychosom Med 2002; 64 (13): 458-68 37. Chin A Paw M, van Poppel NMN, et al. Effects of anaerobic exercise and all-around, functional training on quality of life, vitality, and depression of older adults living in long-term care facilities: a randomized controlled trial [abstract]. Biomed Central Geriatr 2004; 4: 5 38. Cramer SR, Nieman DC, Lee JW. The effects of moderate exercise training on psychological well-being and mood state in women. J Psychosom Res 1991; 35 (4-5): 437-49 39. Crews DJ, Lochbaum MR, Landers DM. Aerobic physical activity effects on psychological well-being in low-income Hispanic children. Percept Motor Skills 2004; 98 (1): 319-24 40. Deivert RG. Efficacy of an aerobic exercise program as treatment for depression and anxiety in alcohol and chemically dependent adults [dissertation]. State College (PA): Pennsylvania State University, 1990 41. DePalma MT. The effects of exercise on anxiety, depression, and type A behavior [dissertation]. Ithaca (NY): Cornell University, 1989 42. DiLorenzo TM, Bargman EP, Stucky-Ropp R, et al. Longterm effects of aerobic exercise on psychological outcomes. Prevent Med 1999; 28 (1): 75-85

ª 2009 Adis Data Information BV. All rights reserved.

Rethorst et al.

43. Dugmore LD, Tipson RJ, Phillips MH, et al. Changes in cardiorespiratory fitness, psychological wellbeing, quality of life, and vocational status following a 12 month cardiac exercise rehabilitation programme. Heart 1999; 81 (4): 359-66 44. Dunn AL, Trivedi MH, Kampert JB, et al. Exercise treatment for depression: efficacy and dose response. Am J Prevent Med 2005; 28 (1): 1-8 45. Eby JM. An investigation into the effects of aerobic exercise on anxiety and depression [dissertation]. Toronto (ON): University of Toronto, 1984 46. Emery CF, Schein RL, Hauck ER, et al. Psychological and cognitive outcomes of a randomized trial of exercise among patients with chronic obstructive pulmonary Disease. Health Psychol 1998; 17 (3): 232-40 47. Epstein D. Aerobic activity versus group cognitive therapy: an evaluative study of contrasting interventions for the alleviation of clinical depression [dissertation]. Reno (NV): University of Nevada, 1986 48. Favilla G. The role of self-efficacy as a mediator in the relationship between muscle strength training and mood in the elderly [dissertation]. San Diego (CA): California School of Professional Psychology-San Diego, 1992 49. Gowans S, DeHueck A, Abbey S. Measuring exerciseinduced mood changes in fibromyalgia: a comparison of several measures. Arthritis Rheum 2002; 47 (6): 603-9 50. Gowans SE, DeHueck A, Voss S, et al. Effects of a randomized, controlled trial of exercise on mood and physical function in individuals with fibromyalgia. Arthritis Care Res 2001; 45 (6): 519-29 51. Hembree L. Exercise and its effect on hopelessness and depression in an aging female population in Eastern Oklahoma [dissertation]. Fayetteville (AR): University of Arkansas, 2000 52. Hilyer JC, Wilson DG, Cillon C, et al. Physical fitness training and counseling as treatment for youthful offenders. J Counsel Psychol 1982; 29 (3): 292-303 53. Jorgensen C. Aerobic conditioning in the therapeutic treatment of chronic schizophrenia [dissertation]. Flagstaff (AZ): Northern Arizona University, 1986 54. Kanner KD. High versus low intensity exercise as part of an inpatient treatment program for childhood and adolescent depression [dissertation]. San Diego (CA): California School of Professional Psychology, 1990 55. King A, Taylor C, Haskell W. Effects of differing intensities and formats of 12 months of exercise training on psychological outcomes in older adults. Health Psychol 1993; 12 (4): 292-300 56. King AC, Taylor C, Haskell WL, et al. Influence of regular aerobic exercise on psychological health: a randomized, controlled trial of healthy middle-aged adults. Health Psychol 1989; 8 (3): 305-24 57. Klein MH, Greist JH, Gurman AS, et al. A comparative outcome study of group psychotherapy vs exercise treatments for depression. Int J Mental Health 1985; 13 (3-4): 148-76 58. Koukouvou G, Kouidi E, Iacovides A, et al. Quality of life, psychological, and physiological changes following exercise training in patients with chronic heart failure. J Rehabili Med 2004; 36 (1): 36-41

Sports Med 2009; 39 (6)

The Antidepressive Effects of Exercise

59. Lennox S, Bedell J, Stone A. The effect of exercise on normal mood. J Psychosom Res 1990; 34 (6): 629-36 60. Levin S. The effects of a ten-week jogging program as an adjunctive treatment for patients in a social rehabilitation clinic [dissertation]. Garden City (NY): Adelphi University, 1983 61. Martinsen E, Medhus A, Danvik L. Effects of aerobic exercise on depression: a controlled study [abstract]. BMJ 1985; 291 (6488): 109 62. McCann IL, Holmes DS. Influence of aerobic exercise on depression. J Personality Soc Psychol 1984; 46 (5): 1142-7 63. McNeil JK, LeBlanc EM, Joyner M. The effect of exercise on depressive symptoms in the moderately depressed elderly. Psychol Aging 1991; 6 (3): 487-8 64. Mutrie N. Exercise as a treatment for depression within the UK Health Service. Paper presented at Proceedings of the Sport, Health, Psychology and Exercise Symposium. London: The Sports Council, 1988 65. Neidig JL, Smith BA, Brashers DE. Aerobic exercise training for depressive symptom management in adults living with HIV infection. J Assoc Nurse AIDS Care 2003; 14 (2): 30-40 66. Newton M, Mutrie N, McArthur JD. The effects of exercise in a coronary rehabilitation programme. Scot Med J 1991; 38 (2): 38-41 67. Norris R, Carrol D, Cochrane R. The effects of physical activity and exercise training on psychological stress and well-being in an adolescent population. J Psychosom Res 1992; 36 (1): 55-65 68. Norvell N, Belles D. Psychological and physical benefits of circuit weight training in law enforcement personnel. J Consult Clin Psychol 1993; 61 (3): 520-7 69. Petajan JH, Gappmaier E, White AT, et al. Impact of aerobic training on fitness and quality of life in multiple sclerosis. Ann Neurol 1996; 39 (4): 432-41 70. Pierce TW, Madden DJ, Siegel WC, et al. Effects of aerobic exercise on cognitive and psychosocial functioning in patients with mild hypertension. Health Psychol 1993; 12 (4): 286-91 71. Pinchasov BB, Shurgaja AM, Grischin OV, et al. Mood and energy regulation in seasonal and non-seasonal depression before and after midday treatment with physical exercise or bright light. Psychiatr Res 2000; 94 (1): 29-42 72. Roth D, Homes D. Influence of aerobic exercise training and relaxation training on physical and psychologic health following stressful life events. Psychosom Med 1987; 49 (4): 355-65 73. Roth DL. Acute emotional and psychophysiological effects of aerobic exercise. Psychophysiology 1989; 25 (5): 593-602 74. Setaro JL. Aerobic exercise and group counseling in the treatment of anxiety and depression [dissertation]. College Park (MD): University of Maryland, 1985 75. Simons CW, Birkimer JC. An exploration of factors predicting the effects of aerobic conditioning on mood state. J Psychosom Res 1988; 32 (1): 63-75 76. Singh NA, Clements KM, Fiatarone MA. A randomized controlled trial of progressive resistance training in depressed elders. J Gerontol 1997; 52A (1): M27-35 77. Singh N, Clements KM, Fiatrone-Singh MA. The efficacy of exercise as a long-term antidepressant in elderly sub-

ª 2009 Adis Data Information BV. All rights reserved.

509

78.

79.

80.

81.

82.

83.

84.

85.

86. 87.

88.

89.

90.

91.

92. 93.

94.

95.

jects: a randomized controlled trial. J Gerontol 2001; 56A (8): M497-504 Singh NA, Stavrinos TM, Scarbek Y, et al. A randomized controlled trial of high versus low intensity weight training versus general practitioner care for clinical depression in older adults. J Gerontol 2005; 60A (6): 768-76 Sorenson M, Anderssen S, Hjerman I, et al. The effect of exercise and diet on mental health and quality of life in middle-aged individuals with elevated risk factors for cardiovascular disease. J Sports Sci 1999; 17 (5): 369-77 Taylor MW. Effects of initial stress level, social support, and participation in an exercise or music condition on the post-treatment stress, depression, and anxiety of nurses [dissertation]. New York: St. John’s University, 1991 van den Berg-Emons R, Balk A, Bussmann H, et al. Does aerobic training lead to a more active lifestyle and improved quality of life in patients with chronic heart failure? Eur J Heart Fail 2004; 6: 95-100 Veale D, LeFevre K, Pantelis C, et al. Aerobic exercise in the adjunctive treatment of depression: a randomized controlled trial. J Royal Soc Med 1992; 85 (9): 541-4 Wigers S, Stiles T, Vogel P. Effects of aerobic exercise versus stress management treatment in fibromyalgia. Scand J Rheum 1996; 25 (2): 77-86 Williams P, Lord S. Effects of group exercise on cognitive functioning and mood in older women. Aust N Z J Pub Health 1997; 21 (1): 45-52 Wilson LM. The effects of an exercise conditioning program on reducing the stress response in nurses [dissertation]. Detroit (MI): Wayne State University, 1985 Zentner R. Psychological effects of a running program [dissertation]. Eugene (OR): University of Oregon, 1981 VandenHoek D. Long slow distance running as a treatment for moderate depression of outpatients [master’s thesis]. Kalamazoo (MI): Western Michigan University, 1983 Blumenthal J, Babyak M, Moore K. Effects of exercise on training on older patients with major depression. Arch Intern Med 1999; 159 (19): 2349-56 Herman S, Blumenthal JA, Babyak M, et al. Exercise therapy for depression in middle-aged and older adults: predictors of early dropout and treatment failure. Health Psychol 2002; 21 (6): 553-63 Fremont J, Craighead LW. Aerobic exercise and cognitive therapy in the treatment of dysphoric moods. Cognit Ther Res 1987; 11 (2): 241-51 Partonen T, Leppamaki S, Hurme J, et al. Randomized trial of physical exercise alone or combined with bright light on mood and health-related quality of life. Psychol Med 1998; 28 (6): 1359-64 Duman R, Heninger G, Nestler E. A molecular and cellular theory of depression. Arch Gen Psychiatr 1997; 54: 597-606 Malberg J, Eish A, Nestler E, et al. Chronic antidepressant treatment increases neurogenesis in adult rat hippocampus. J Neurosci 2000; 20 (24): 9104-10 Ernst C, Olson A, Pinel J, et al. Antidepressant effects of exercise: evidence for an adult-neurogenesis hypothesis? J Psychiatr Neurosci 2006; 31 (2): 84-92 Persson A, Thorlin T, Bull C, et al. Mu- and delta-opioid receptor antagonists decrease proliferation and increase

Sports Med 2009; 39 (6)

Rethorst et al.

510

96. 97.

98.

99.

100.

101.

102.

103.

104. 105.

106.

107.

108.

109. 110. 111. 112.

113. 114.

neurogenesis in cultures of rat adult hippocampal progenitors. Eur J Neurosci 2003; 17: 1159-72 Colt E, Wardlaw S, Frantz A. The effect of running on plasma delta-endorphin. Life Sci 1981; 28 (14): 1637-40 Schobersberger W, Hobisch-Hagen P, Fries D, et al. Increase in immune activation, vascular endothelial growth factor and erythropoietin after an ultramarathon run at moderate altitude. Immunobiology 2000; 201 (5): 611-20 Pereira A, Huddleston D, Brickman A, et al. An in vivo correlate of exercise-induced neurogenesis in the adult dentate gyrus. Proc Natl Acad Sci U S A 2007; 104 (13): 5638-43 Jin K, Zhu Y, Sun Y, et al. Vascular endothelial growth factor (VEGF) stimulates neurogenesis in vitro and in vivo. Proc Natl Acad Sci 2003; 99 (18): 11946-50 Fabel K, Fabel K, Tam B, et al. VEGF is necessary for exercise-induced adult hippocampal neurogenesis. Eur J Neurosci 2003; 18 (10): 2803-12 Cotman C, Berchtold N. Exercise: a behavioral intervention to enhance brain health and plasticity. Trends Neurosci 2002; 25 (6): 295-301 Wozniak W. Brain-derived neurotrophic factor (BDNF): role in neuronal development and survival. Folia Morphol 1993; 52 (4): 173-81 Russo-Neustadt A, Ha R, Ramirez R. Physical activityantidepressant treatment combination: impact on brainderived neurotrophic factor and behavior in an animal model. Behav Brain Res 2001; 120: 87-95 Jacobs B. Serotonin motor activity, and depression-related disorders. Am Sci 1994; 82: 456-63 Chaouloff F, Laude D, Elghozi J. Physical exercise: evidence for differential consequences of tryptophan on 5-HT synthesis and metabolism in central serotonergic cell bodies and terminals. J Neural Transmission 1989; 78 (2): 1435-63 Brezun J, Daszuta A. Serotonin may stimulate granule cell proliferation in adult hippocampus, as observed in rats grafted with foetal raphe neurons. Eur J Neurosci 2000; 12: 391-6 Kubitz K, Landers D, Petruzzello S, et al. The effects of acute and chronic exercise on sleep. Sports Med 1996; 21 (4): 277-91 Youngstedt S, O’Connor P, Dishman R. The effects of acute exercise on sleep: a quantitative synthesis. Sleep 1997; 20 (3): 203-14 McGinty D, Harper R. Dorsal raphe neurons: depression of firing during sleep in cats. Brain Res 1976; 101 (3): 569-75 Dietrich A, McDaniel W. Endocannabinoids and exercise. Br J Sport Med 2004; 3 (8): 536-41 Meng I, Manning B, Martin W, et al. An analgesia circuit activated by cannabinoids. Nature 1998; 395: 381-3 Sparling P, Giuffrida A, Piomelli D. Exercise activates the endocannabinoid system. Cognit Neurosci Neuropsychol 2003; 14 (17): 2209-11 Holsboer F. The corticosteroid receptor hypothesis of depression. Neuropsychopharmacology 2000; 23 (5): 477-501 Wittert G, Livesey J, Espiner E, et al. Adaptation of the hypothalamopituitary adrenal axis to chronic exercise stress in humans. Med Sci Sport Exerc 1996; 28 (8): 1015-9

ª 2009 Adis Data Information BV. All rights reserved.

115. Dienstbier RA. Behavioral correlates of sympathoadrenal reactivity: the toughness model. Med Sci Sport Exerc 1991; 23 (7): 846-52 116. Luger A, Deuster PA, Kyle SB, et al. Acute hypothalamicpituitary-adrenal responses to the stress of treadmill exercise: physiologic adaptations to physical training. New Engl J Med 1987; 316 (21): 1309-15 117. Wittert GA, Livesey JH, Espiner EA, et al. Adaptation of the hypothalamopituitary adrenal axis to chronic exercise stress in humans. Med Sci Sport Exerc 1996; 28 (8): 1015-9 118. Blumenthal JA, Fredrikson M, Matthews KA, et al. Stress reactivity and exercise training in premenopausal and postmenopausal women. Health Psychol 1991; 10 (6): 384-91 119. Beyer C, Boikess S, Luo B, et al. Comparison of the effects of antidepressants on norepinephrine and serotonin concentrations in the rat frontal cortex: an in-vivo microdialysis study. J Psychopharmacol 2002; 16 (4): 297-304 120. Dishman R. Brain monoamines, exercise, and behavioral stress: animal models. Med Sci Sport Exerc 1997; 29 (1): 63-74 121. Ossip-Klein DJ, Doyne EJ, Bowman ED, et al. Effects of running or weight lifting on self-concept in clinically depressed women. J Consult Clin Psychol 1989; 57 (1): 158-61 122. Jacobsen N, Follette W, Revenstorf D. Psychotherapy outcome research: methods for reporting variability and evaluating clinical significance. Behav Ther 1984; 15 (4): 336-52 123. Seggar L, Lambert M, Hansen N. Assessing clinical significance: application to the Beck Depression Inventory. Behav Ther 2002; 33 (2): 253-69 124. Grundy C, Lambert M, Grundy E. Assessing clinical significance: application to the Hamilton Rating Scale for Depression. J Mental Health 1996; 5 (1): 25-33 125. McGlinchey JB, Atkins DC, Jacobsen NS. Clinical significance methods: which one to use and how useful are they. Behav Ther 2002; 33 (4): 529-50 126. Zipf G. Human behavior and the principle of least effort: an introduction to human ecology. Cambridge (MA): Addison-Wesley, 1949 127. Anderson I, Tomenson B. Treatment discontinuation with selective serotonin reuptake inhibitors compared with tricyclic antidepressants: a meta-analysis. BMJ 1995; 310: 1433-8 128. Montgomery S, Kasper S. Comparison of compliance between serotonin reuptake inhibitors and tricyclic antidepressants: a meta-analysis. Int Clin Psychopharmacol 1995; 9 (4): 33-40 129. Steffens D, Krishnan K, Helms M. Are SSRIs better than TCAs? Comparison of SSRIs and TCAs: a meta-analysis. Depress Anxiety 1997; 6: 10-8 130. Anderson I. SSRIs versus tricyclic antidepressants in depressed inpatients: a meta-analysis of efficacy and tolerability. Depress Anxiety 1998; 7 Suppl. 1: 11-7 131. Gloaguen V, Cottraux J, Cucherat M, et al. A metaanalysis of the effects of cognitive therapy in depressed patients. J Affect Disord 1998; 49 (1): 59-72 132. Food and Drug Administration. Sertaline hydrochloride patient information sheet 2006 [online]. Available from

Sports Med 2009; 39 (6)

The Antidepressive Effects of Exercise

URL: http://www.fda.gov/cder/drug/InfoSheets/patient/ sertralinePT.pdf [Accessed 2007 Jan 25] 133. US Department of Health and Human Services. Physical activity and health: a report of the Surgeon General. Atlanta (GA): US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, 1996 134. Warburton D, Nicol C, Bredin S. Health benefits of physical activity: the evidence. CMAJ 2006; 174 (6): 801-9 135. Taylor A, Cable N, Faulkner G, et al. Physical activity and older adults: a review of health benefits and the effectiveness of interventions. J Sport Sci 2004; 22 (8): 703-25

ª 2009 Adis Data Information BV. All rights reserved.

511

136. Kelly G, Kelly K, Tran Z. Exercise and bone mineral density in men: a meta-analysis. J Appl Physiol 2000; 88 (5): 1730-60 137. Department of Health and Aging. Exercise physiologists eligible to provide services under Medicare, 2005 [online]. Available from URL: http://www.health.gov.au/internet/ ministers/publishing.nsf/Content/healthmediarel-yr2005-ta abb106.htm [Accessed 2007 Jan 27]

Correspondence: Dr Chad D. Rethorst, Department of Psychiatry, University of Rochester Medical Center, Box PSYCH, Rochester, NY 14642, USA. E-mail: [email protected]

Sports Med 2009; 39 (6)