Are you in Canada? Click here to proceed to the HK Canada website.

For all other locations, click here to continue to the HK US website.

Human Kinetics Logo

Purchase Courses or Access Digital Products

If you are looking to purchase online videos, online courses or to access previously purchased digital products please press continue.

Mare Nostrum Logo

Purchase Print Products or eBooks

Human Kinetics print books and eBooks are now distributed by Mare Nostrum, throughout the UK, Europe, Africa and Middle East, delivered to you from their warehouse. Please visit our new UK website to purchase Human Kinetics printed or eBooks.

Feedback Icon Feedback Get $15 Off

The effects of physical activity on osteoporosis

This is an excerpt from Physical Activity Epidemiology-3rd Edition by Rodney K. Dishman,Gregory W. Heath,Michael D. Schmidt & I-Min Lee.

Physical Activity and Osteoporosis: The Evidence

The scientific advisory committee of the 2008 Physical Activity Guidelines for Americans concluded that a year of exercise training in older adults can increase BMD or slow age-related declines in BMD by 1% to 2% in clinically important regions of the spine and hip. Based on rodent studies, it appears that even small improvements in BMD after mechanical loading (i.e., exercise) result in substantial resistance to fracture (Physical Activity Guidelines Advisory Committee 2008). More recently, the scientific advisory committee for the 2018 Physical Activity Guidelines for Americans, second edition (Physical Activity Guidelines Advisory Committee 2018), concluded that exercise programs started after hip fracture rehabilitation ends are effective for improving physical function in community-dwelling older adults. However, there was only limited evidence that exercise programs that combine strengthening and balance activities improve physical function in older people who are at risk of fractures because they have osteoporosis or osteopenia.

There is scientific consensus that physical inactivity is associated with decreased bone mass. Extreme losses of 2% to 10% have been found among young men after four months of bed rest (Buchner and Wagner 1992), but most of that loss is recovered soon after a return to upright posture and normal daily movement. The transience of this bone loss can be attributed to the reduced gravitational load, similar to the loss of bone mass that occurs in microgravity (e.g., during space travel). Nonetheless, there does seem to be a more insidious loss of bone mass among people who live sedentary lifestyles that compounds the natural history of bone loss with increasing age. An early review of the research literature concluded that physical activity involving high-intensity loading of bone promotes bone density and may help prevent osteoporosis—and that physical activity involving static muscle contractions or slow movements has no effect or smaller effects on bone mass than activities that involve rapidly applied forces (Vuori 2001).


A summary of 12 systematic reviews and meta-analyses concluded that peak bone mass in young girls can be improved with short bouts of high-impact exercise and that training interventions combining high-impact and resistance exercises are best for improving or preserving BMD in pre- and postmenopausal women (Xu et al. 2016).


Much of the evidence suggesting the benefits of exercise or sports on bone health initially came from cross-sectional studies that compared athletes or regular exercisers with sedentary people, or from poorly controlled clinical studies of exercise training and bone mass (Montoye 1984). A few cross-sectional studies are reviewed in the next section. Then we discuss population-based studies addressing whether physical activity boosts peak bone mass during adolescence and young adulthood. After that, we discuss clinical trials of the impact of endurance and resistance exercise training on BMD during childhood and early adolescence, young adulthood, middle age, and older age, including comparisons of pre- and postmenopausal women. Finally, we summarize the evidence about whether exercise training reduces the risk of osteoporotic fractures.

Exercise Recommendations for Bone Health in Adults

Mode: Intermittent weight-loading activities (tennis, stair climbing, jogging, jumping, walking) and resistance exercise (e.g., weightlifting)

Intensity: Bone-loading forces that have moderate-to-high magnitude

Frequency: Three to five days per week for endurance activities and two or three days per week for resistance activities

Duration: 30 to 60 min each session or day consisting of a combination of bone-loading activities that involve major muscle groups

Source: Kohrt et al., 2004.

Cross-Sectional Studies

Cross-sectional comparisons generally show that, compared with nonathletes, athletes in high-load sports (e.g., weightlifting, gymnastics, basketball, volleyball) have 10% to 15% higher bone mass in the lumbar spine, femoral neck, pelvis, and arm but that bone mass of athletes who perform repetitive weight-bearing activities (e.g., distance running and Nordic skiing) that don’t involve a lot of peak loading is only about 3% to 8% higher (Montoye 1984; Taaffe et al. 1995, 1997). This ranking does not fully clarify whether high BMD resulted from the particular sport or whether athletes excelled in these sports because of their optimal levels of bone mass. However, a study of college-aged gymnasts found that bone density at relevant sites responded dramatically to high-impact loading, independent of reproductive hormone status and despite high initial BMD values (Taaffe et al. 1997). This provides evidence that mechanical loading, rather than selection bias, underlies the high BMD values characteristic of female gymnasts.


A meta-analysis of 14 studies of young adults (18-30 years) found that regular swimmers had BMD in the whole body, femoral neck, and lumbar spine that was statistically similar to nonathletes but about 1 to 1.5 standard deviations lower than other athletes engaged in high-impact sports (Gomez-Bruton et al. 2018).


Greater muscle strength is positively associated with BMD; increased BMD possibly results from stimulation of bone modeling by the transfer of force from muscle to bone. For example, bone mass was positively associated with grip strength and peak torques at the hip and knee joints independently of body weight (Bauer et al. 1993). In a recent cross-sectional sample of 147 college-aged adults, the association between customary moderate-to-vigorous physical activity measured by accelerometer and estimated bone strength of the tibia was mediated directly by thigh muscle mass and also indirectly by the influence of muscle mass on knee extension strength (Higgins et al. 2020). A study of elite junior tennis players estimated muscle size and bending, twisting, and compression strength in the racket and nonracket arms using peripheral quantitative computed tomography and concluded that muscle size and action on bone does influence bone strength. However, the investigators concluded that torsional loads (e.g., twisting) on the bone during a tennis stroke likely provided the largest osteogenic strain (Ireland et al. 2013). Other research on professional baseball players estimated that one-half of bone size and one-third of the bending strength of bone in the throwing arm developed during youth persists throughout adulthood (Warden et al. 2014).

NHANES III—United States

In a nationally representative sample of 4254 U.S. men aged 20 to 59 years from the third National Health and Nutrition Examination Survey (NHANES III), jogging was associated with higher BMD of the total femur (Mussolino, Looker, and Orwoll 2001). A total of 954 men (22.3%) reported jogging in the past month. Bone mineral density was 5% higher among joggers than among nonjoggers and was 8% higher than in 577 nonjoggers who also reported no other leisure activities. Those differences were 3% and 6.5% after adjustment for age at interview, race or ethnicity, BMI, dietary protein, calcium, total calorie intake, smoking, alcohol consumption, chronic health conditions, and weight change. Joggers were more likely to report also doing other weight-bearing activities, but their BMD was still higher than that of nonjoggers after further adjustment for total number of leisure physical activities.

The prevalence of reported osteoporosis in 8073 women 20 years of age or older who participated in NHANES during 1999 to 2004 was 8.9% (95% CI: 7.7-10.1) in sedentary women but 6.2% (95% CI: 4.4-8.5) in women who said they spent at least 30 MET-hours each week doing physical activity (Robitaille et al. 2008).

NHANES 2005 to 2006—United States

Moderate-to-vigorous physical activity and sedentary behavior were measured objectively by an accelerometer device in about 2000 U.S. adults aged 23+ years sampled in the 2005 to 2006 National Health and Nutrition Examination Survey (NHANES) (Chastin, Mandrichenko, Helbostadt, and Skelton 2014). Adjustment was made for any confounding effects of age, smoking, BMI, ethnicity, dietary calcium, alcohol consumption, vitamin D levels, use of a corticosteroid, family history of osteoporosis, and levels of parathyroid hormones. Among men, BMD at the hip (i.e., proximal femur) was 0.306 g/cm2 higher (95% CI: 0.02-0.59) for each 10-min increment in daily physical activity. There was not a significant association between time spent in sedentary behavior and BMD at the hip in men. In women, BMD at the hip was 0.159 g/cm2 lower (95% CI: −0.24 to 0.08) for each 10-min increment spent being sedentary each day. There was not a significant association of BMD with physical activity in women. No significant associations were found between physical activity or sedentary time with spinal BMD for either men or women. The results showed that repeated exposure to sitting during the day is a risk factor for lower BMD at the hip in women, independent of the time spent in moderate-to-vigorous physical activity. In contrast, sedentary time is not associated with BMD at either the hip or spine in men. Rather, moderate-to-vigorous physical activity conferred a protective effect on BMD at the hip in men.

A separate analysis added self-reports of screen time (watching TV or using a computer) and frequency of vigorous playtime and strengthening activities among 671 males and 677 females aged from 8 to 22 years (Chastin, Mandrichenko, and Skelton 2014). Sedentary screen time was negatively associated with bone mineral content (BMC) at the hip. BMC at the hip was 0.77 g lower (95% CI: −1.31 to −0.22) in females and 0.45 g lower (95% CI: −0.83 to −0.06) in males, independently of device-measured moderate-to-vigorous physical activity. However, the risk of sedentary screen time on lower bone mineral was canceled by regular participation (five days per week) in vigorous exercise (or play) or strengthening exercise (males). In this analysis, the association of screen-based sedentary time and lower bone mineral content was independent of the amount of physical activity measured objectively by the accelerometer device but not independent of self-reported frequency of vigorous or strengthening activities.

Prospective Cohort Studies

There have not been many prospective cohort studies of the effect of physical activity or fitness on bone mass or risk of osteoporosis. Early studies included small samples of about 25 to 200 youths or young adults and lasted only 5 to 12 months. Despite the short time periods of the studies, most showed a 3% to 10% increase in BMD. Many of those studies did an incomplete job of controlling for diet and other potential confounders that influence bone mass. Nonetheless, the cumulative evidence is encouraging that vigorous physical activity, especially the types that involve high peak loads (e.g., resistance exercise or power sports), might promote higher peak bone mass (Modlesky and Lewis 2002).

A systematic review by the National Osteoporosis Foundation of research reported this millennium located 20 prospective observational cohort studies of physical activity during developmental years (Weaver et al. 2016). Eighteen of the 20 cohort studies reported statistically significant differences in bone mass or density between the most physically active and less active children and adolescents, particularly when children participated in organized sports. In addition, eight prospective observational cohort studies reported significant associations between physical activity and whole bone structure. For example, the University of Saskatchewan Pediatric Bone Mineral Accrual Study found 8% to 12% greater size of the hip (Jackowski et al. 2014) and 10% larger size and 13% greater bending strength of the tibia (Duckham et al. 2014) in young adults who were active as adolescents compared to those who were less active. Some other exemplary studies are discussed next.

Nord-Trøndelag Health Study

The Nord-Trøndelag Health Study (called HUNT) is an ongoing, population-based study in North Trøndelag, one of Norway’s 19 counties. More than 20,000 women completed an initial health survey during 1984 to 1986 (HUNT 1), and a follow-up survey was conducted during 1995 through 1997 (HUNT 2) (Augestad et al. 2004, 2006). Approximately 8000 women had a measure of BMD at the distal and ultradistal regions of the nondominant forearm at HUNT 2. There were 1396 premenopausal women 31 to 44 years of age after exclusion of those who reported diabetes, stroke, cancer, epilepsy, severe physical disability, rheumatoid arthritis, hyperthyroidism, perceived very poor health, severe disability, bilateral ovariectomy, or pregnancy. After further exclusion of women who also had osteoporosis or prior fractures, used calcium or vitamin D supplementation, or used asthma medication, there were 2924 postmenopausal women aged 55 to 98 years.

Premenopausal women who said they participated in high-intensity leisure-time physical activity and also did heavy physical occupational work in both 1984 and 1995 had about a 55% reduction in odds of low BMD (less than the lowest 20%) at the ultradistal radius at HUNT 2, with or without adjustments for age, smoking, amenorrhea, BMI, and daily milk consumption. Among postmenopausal women, odds (adjusted for age, BMI, age at menarche, years since menopause, smoking, milk consumption, and HRT) of having low BMD 10 years later at HUNT 2 at either forearm site were 30% lower in those who at HUNT 1 said they participated in high-intensity recreational physical activity.

The University of Saskatchewan Bone Mineral Study

The influence of physical activity on gains in bone mass during adolescence was observed for six years among 53 girls and 60 boys who ranged from 8 to 14 years old at the start of the study (Bailey et al. 1999). Measures of physical activity, diet, height, weight, and bone mineral content (BMC) by DXA were taken every six months. Peak rates of gain in BMC were computed for the total body, lumbar spine, and proximal femur. After adjustment for height and weight gains, the peak rate of gain in BMC in the femoral neck, lumbar spine, and total body was higher for both males and females who were more physically active than in the less active subjects. The BMC in the femoral neck measured one year after the peak rate of gain was 7% higher in boys and 9% in girls in the most active quartile compared with the least active quartile; total-body gains were 9% and 17% higher in the most active boys and girls, respectively.

Amsterdam Growth and Health Longitudinal Study

Daily physical activity and fitness were monitored from age 13 to 29 years in a cohort of 182 males and females (Kemper et al. 2000). At a mean age of 28 years, BMD was measured by DXA at the lumbar spine, the femoral neck, and the distal radius. Physical activity during the previous three months was determined by interview when participants were between 13 and 16 years old and again between ages 21 and 27. Physical activity was expressed as energy expended (in MET-hours) per week or as peak intensity (in multiples of body mass), independently of the frequency and duration of activity. Physical fitness was measured with a neuromotor fitness test (a composite of six strength, flexibility, and speed tests) and as cardiopulmonary fitness (maximal oxygen uptake). After adjustment for sex, age, body composition, and dietary calcium, both measurements of physical activity and neuromotor fitness during adolescence and in young adulthood were positively related with the bone mass in the lumbar spine and femoral neck measured at a mean age of 28 years. Cardiorespiratory fitness was unrelated to bone mass.

Pelotas, Brazil Birth Cohort Study

The prospective association between physical activity and BMD measured at ages 18 and 30 years in the lumbar spine and femoral neck was examined in about 3500 young adults (Bielemann, Domingues, Horta, and Gigante 2014; Bielemann, Domingues, Horta, Menezes et al. 2014). Physical activity was measured by self-report at ages 11, 15, 18, and 23 years in males and ages 11, 15, and 23 in females.

At age 18, adjustments were made for skin color, family income at birth, and BMI. Boys who said they spent 5 h or more a week in physical activity (a little less than recommended) at age 11 or 15 had about 0.02 g/cm2 higher BMD at age 18 in the lumbar spine and 0.04 to 0.06 g/cm2 higher BMD at the femoral neck than inactive boys. After further adjustment for age at menarche, girls who said they spent 15 h or more a week in physical activity (about double the recommended amount) at age 11 or 15 had about 0.02 g/cm2 higher BMD at age 18 in the lumbar spine and 0.03 g/cm2 higher BMD at the femoral neck than inactive girls.

At age 30, adjustments were made for BMI, skin color, family income at birth, household wealth, birth weight, maternal smoking during pregnancy, breastfeeding, and smoking. Males who engaged in physical activity at least once per week at age 15 had 0.06 g/cm2 higher BMD at age 30 in the lumbar spine than those who were inactive. There were dose responses for males between weekly minutes of physical activity at age 18 and 23 and BMD at the lumbar spine and femoral neck. Those in the two highest quartiles of physical activity had about 0.04 to 0.06 g/cm2 higher BMD at age 30 in the lumbar spine and femoral neck than males in the lowest quartile. Females in the highest quartile of physical activity at age 23 had 0.020 g/cm2 higher BMD at age 30 in the femoral neck than the least active females.

Iowa Bone Development Study

Participants wore an accelerometer for three to five days to assess their customary participation in moderate-to-vigorous physical activity beginning at age 5 and then about every three years at ages 8, 11, 13, 15, and 17 years (Janz et al. 2014). Bone mineral content and bone size was assessed by DXA at age 17 in 364 participants. Geometric indexes of compressive strength (bone stress index) and bending strength (polar moment of inertia) of the tibia were assessed by peripheral computer quantitative tomography. The most active girls (average of 85 min/day at age 5 and 30 min/day at age 17) had 7% higher BMD at the hip (1.08 g/cm2 vs. ~1.01 g/cm2) than less active girls (less than 1 h/day at age 5 and 20-28 min/day at age 17). Similarly, BMD at the hip was 7% higher (1.20 g/cm2 vs. ~1.12 g/cm2) in the most active boys (average of 76 min/day at age 5 and 50 min/day at age 17) than in less active boys (about 45 min to 1 h/day at age 5 and 30-34 min/day at age 17). Also, measures of bone size at the femoral neck and bone strength of the tibia were about 5% to 13% greater in the most active girls and boys. However, only 6% of girls and 20% of boys maintained the higher levels of physical activity throughout childhood and adolescence.

More Excerpts From Physical Activity Epidemiology 3rd Edition