Osteoporosis is the most common bone disease and is characterized by low bone mass and weak bone structure, increasing the susceptibility to fractures. The risk of osteoporosis is much greater in women; of the estimated 10 million Americans with osteoporosis, 8 million are women (27). Therefore, finding preventive measures that are suitable to women is important to decrease the risk of this disease.
Mechanical loading is theorized to be the primary factor to which bone adapts its strength (8) and the loads produced by muscle forces from physical activity result in the largest strains on bone (15). Characteristics of strength training, including high strain rates, a variety of strain distributions, and a sufficient strain magnitude (60-80% of maximum force), have been found to be more osteogenic (14) than aerobic exercise, which has more regular strain distributions and loading patterns. Cross-sectional studies in females have shown higher bone mineral content (BMC) in competitive weight lifters compared with other types of competitive athletes (e.g., cross-county skiers) (9,10).
Studying the effect of physical activity on bone through mechanical loading may be more appropriate in premenopausal women than that in older women because peak bone mass occurs in approximately in the third or fourth decade of life (18). Restoration of bone loss, which accelerates with menopause and aging, is very difficult. Therefore, preventive approaches to maximize peak bone mass before menopause may be a more appropriate public health strategy.
Six randomized studies have previously assessed the effect of strength training on bone in premenopausal women. Results are inconsistent with three studies showing significant positive effects lumbar spine (7,16,23) and proximal femur (7,16) in areal bone mineral density (aBMD), whereas the others have shown no significant effect in proximal femur or spine aBMD (6,22,29). Only one of these studies reported measuring the effect of strength training on BMC, and the intervention included both strength training and impact exercises (29). Nonrandomized studies have similarly conflicting results with some studies showing positive effects of strength training on bone measures, and other studies result in no significant changes (30).
The lack of consensus in randomized studies among premenopausal women, as well as the lack of consistency between cross-sectional studies and randomized trials, point to the need for further studies to measure the effect of strength training on bone in this age group. Differences in duration, type, and intensity of the strength training make direct comparisons between previous randomized trials difficult. Additionally, attrition during the multiyear strength training interventions ranged from 25% to over 50%. Although guidelines suggest that strength training two times per week at moderate to high intensity will be useful for bone health in premenopausal women (13), to our knowledge no previous studies have used published guidelines to design the strength training program.
Therefore, the aims of these analyses are to determine the effect of twice-weekly strength training based on published guidelines by the US Department of Health and Human Services (28) in healthy, sedentary premenopausal women on proximal femur and lumbar spine BMC and aBMD compared with a standard care control group.
Study design and protocol.
The Strong, Healthy, Empowered (SHE) study was a 2-yr randomized controlled trial of healthy, sedentary, young (25-44 yr) women with a primary aim of assessing the effects of strength training on total and abdominal fat. The SHE cohort consisted of 164 women randomized to exercise (ST, n = 82) and control (CO, n = 82). The exercise group participated in a twice-weekly strength training program. The study protocol was approved by the University of Minnesota Institutional Review Board, and written informed consent was obtained from all participants. Informed consent asked that the women not be planning to move nor planning to become pregnant over the next 2 yr.
Participant recruitment, selection, and randomization.
The SHE study cohort and randomization procedures have been previously described (20). In brief, the cohort included sedentary, overweight to mildly obese, young (25-44 yr), premenopausal women. Phone screens were completed on 1721 women; 164 met the inclusion criteria and agreed to randomization and participation. Sixteen participants were excluded because they became pregnant (n = 7) or used corticosteroids during the study (n = 9), leaving 148 participants for these analyses.
Measurements were completed at baseline and at 1 and 2 yr after the start of the study using standard written protocols. The measurement staff was different from the intervention staff to allow blinding of the measurements.
Body composition, BMC (g), aBMD (g·cm−2), and bone area (cm2) were measured in the proximal femur and lumbar spine by dual x-ray absorptiometry (DXA) (Lunar Prodigy, Lunar Radiation Corp., Madison, WI). Anteroposterior spine was measured at the level of the first to fourth lumbar vertebrae (L1-L4), and dual proximal femur scans were analyzed for the femoral neck and trochanteric regions as well as the total femur. The DXA used in this study has a narrow angle fan beam, which eliminates the magnification of images with different distances from the scanner, and shown to have similar measurements of BMC and aBMD to those obtained with pencil beam DXA scanners (17). Body composition [total body lean and fat mass (g)] was measured with a DXA total body scan. Reproducibility with the instrument used is less than 1.0% for fat and lean mass and aBMD. For example, monthly measurement of aBMD with lumbar spine phantoms over a 12-month period showed a range of 1.246-1.266 g·cm−2 (coefficient of variation 0.8%).
Muscle strength was assessed with the maximum weight lifted once (1 repetition maximum) on both the bench press and the leg press using a previously described protocol (21). This testing was repeated at two separate appointments no more than 2 wk apart to ensure the best possible effort from each participant, and the higher of the two values was selected.
Prior year usual dietary intake was estimated using the National Institutes of Health Diet History Questionnaire (DHQ). The DHQ was shown to have correlations with three 24-h recalls of 0.48 for total energy intake and 0.66 for calcium intake, which was higher than other food-frequency questionnaires (25). Total kilocalorie, calcium, and vitamin D intake (from diet and supplements) were estimated as the average intake of the three annual questionnaires.
Physical activity was assessed at each measurement time point with an accelerometer (Actigraph LLC, model 7164, version 2.2, Fort Walton, FL), which was worn for at least 4 d (two weekdays and two weekends) during waking hours, except when was in contact with water (i.e., showering, swimming). The criterion of 60 min or more of consecutive zero accelerometer counts was used to identify period of nonwearing. Identified nonwearing minutes were excluded from analysis. A treadmill test was completed to determine an individually calibrated threshold above which participants could be considered nonsedentary. The participants walked on a treadmill for 7 min at 2 mph, and the mean accelerometer counts per minute for the 7 min were calculated to determine the threshold for nonsedentary activity. The physical activity level reported is the number of minutes spent above this individually calibrated threshold, with weekdays and weekend days weighted appropriately to reflect a 1-wk period. Body weight and height were measured using a digital stand-on scale and scale-mounted stadiometer (Scale-tronix 5005, Scale-tronix, White Plains, NY). Study staff administered questionnaires about menstrual history, medication use, and demographic information.
Adherence to strength training was assessed by dividing the number of strength training sessions attended by the total possible sessions for a given period. The total number of possible sessions per year was 100 (twice weekly for 52 wk, with 2 wk off per year).
The exercise program consisted of twice-weekly strength training of three sets of 8 to 10 repetitions using variable resistance machines and free weights that stressed all major muscle groups following published guidelines (28). The exercise sessions were supervised by certified fitness trainers (ratio of participants to trainers, ≤6:1) for the first 16 wk to instruct in technique and weight progression. After the first 16 wk, the participants exercised primarily unsupervised, but there was regular contact with the fitness trainers either in person or by telephone or email (20). The amount of weight lifted was minimal for the first 3 wk to allow participants to learn the exercise programs. After the initial 3 wk, the intensity of the strength training was set so that the participants were able to complete three sets of 10 repetitions. When each participant was able to complete four sessions (2 wk) of three sets of 10 repetitions, the weight was increased by the smallest increment possible. This progression continued throughout the first year. During the second year of the intervention, participants were allowed to maintain the highest weight lifted for each exercise, although some continued to increase. Other components of the intervention have been described elsewhere but included regular contact with the fitness trainers to promote adherence and attendance to the strength training (20). Additionally, the participants completed exercise logs, and these were checked weekly by the fitness trainers with follow-up phone calls when sessions were missed. The control group intervention consisted of mailed American Heart Association brochures that recommended 30 min of moderate intensity activity on most days of the week, consistent with published guidelines (28). This is standard care for a public health intervention.
Regardless of group assignment, participants were asked not to make any changes in their diet. Seasonal variations in their diet were expected and allowed. This message was communicated during recruitment, included in the informed consent document, and reiterated during the measurement visits and strength training sessions.
With type 1 error for two-sided tests set at 0.05, an a priori power analysis showed that 70 participants per group would have 80% power to detect differences between groups of 0.65 g and 0.01 g·cm−2 in the femoral neck BMC and aBMD, respectively; 1.46 g and 0.009 g·cm−2 in the femoral trochanter BMC and aBMD, respectively; 2.41 g and 0.09 g·cm−2 in the total femur BMC and aBMD, respectively; and 5.8 g and 0.01 g·cm−2 in the lumbar spine BMC and aBMD, respectively. These are similar to the increases reported in previous trials (7,16,23).
Baseline characteristics across the two randomized groups were calculated as means for continuous variables and percentages for categorical variables. Differences between the strength training and control groups were assessed via Student t-test for continuous variables and χ2 tests for categorical variables. Changes in muscle strength, body composition, and physical activity were assessed between groups with t-tests at each time point.
The analysis was based on the principle of intention to treat. If a participant dropped out, the average experience of the control group was used as the outcome data. When a missing value occurred, the slope of each bone measure over time in the control group was used to calculate the change from the previous measurement point.
Repeated-measures ANCOVA was used to assess the effect of the strength training intervention on each of the bone measures. For these analyses, the dependent variables were the BMC and aBMD for the femoral neck and trochanter, total femur, and lumbar spine (L1-L4) for a total of eight outcome measures. The independent variable was the randomized condition (strength training or control group) as an interaction with time representing the three measurement time points (baseline, year 1, and year 2). Covariates considered were baseline age, height, weight, years of hormonal contraceptive use, change in physical activity and weight, and average kilocalorie, calcium, and vitamin D intake. All models were adjusted for baseline height and weight as continuous variables as the addition of these covariates improved the fit of the models. Average intake of calcium and vitamin D were assessed for possible effect modification, but results were similar regardless of level of calcium or vitamin D intake. SAS Version 8.2 was used for analysis (SAS Institute, Inc. Version 8.2 Cary, NC).
One hundred and forty-eight participants were included in this analysis (72 in the strength training group and 76 in the control group). The analyses were performed as intention to treat. Data for 19 participants in year 1 and 27 participants in year 2 were missing and substituted with data from the control experience. When the analyses were done per protocol, the number of study participants decreased to 121 (62 in the strength training and 59 in the control group). Despite the decrease in the number of participants, the results did not change substantially. The data from the intention-to-treat analyses are presented.
Baseline characteristics were similar between the intervention and the control groups (Table 1) with no statistically significant differences. All of the women (except nine) were considered overweight (BMI ≥ 25 kg·m−2), and 44.6% were classified as obese (BMI ≥ 30 kg·m−2). The distribution of women who were classified as overweight or obese did not differ by group (P = 0.30). The percentage of participants who reported dietary intake meeting the recommendations for calcium was 41% and 58% for vitamin D. Average adherence to the strength training intervention was 74% (range, 69-100%) for the first year, with highest adherence during the first 16 wk (supervised strength training). Average adherence decreased in the second year (average 60%; range, 30-100%) and was significantly different from the adherence in the first year (P < 0.0001). The analyses were repeated after excluding participants with less than 70% and 80% adherence, and similar results were found.
Table 2 shows muscle strength, body composition, and physical activity values, presented by treatment group for baseline, as well as changes from baseline to year 1 and baseline to year 2. As expected, the strength training group significantly increased bench (5%) and leg (10%) press muscle strength (within group P < 0.001 for both) with no significant change in muscle strength seen in the control group. There was a corresponding increase in total body lean mass in the exercise group that was statistically significantly higher than that in the control group. Physical activity changed minimally in both groups outside of the prescribed strength training sessions, with no significant between- or within-group differences.
Table 3 shows the mean BMC, aBMD, and area for each bone site by intervention group and measurement time point after adjustment for baseline height and weight. The P value corresponds to the difference between groups over the three measurement time points, explaining the effect of the strength training over time in the exercise group compared with the control group. In the femoral neck, the strength training group showed no change in the BMC over 2 yr whereas the control group showed a 1.5% decrease. The between-group effects were statistically different (P = 0.04). There were no other statistically significant between-group effects found in any of the other measurement sites for BMC or for aBMD and area (Table 3 and Fig. 1). The strength training intervention did not show an effect on BMC (P = 0.47) or aBMD (P = 0.24) when pooled across all sites (data not shown).
Despite favorable changes in lean mass and muscle strength, a 2-yr strength training program based on published (US Department of Health and Human Services) guidelines (28) did not significantly impact aBMD in the proximal femur or the lumbar spine in premenopausal women.
Our finding of no differences in aBMD differ from previously published strength training trials in premenopausal women with a similar length intervention (7,16). In an 18-month study comparing thrice weekly strength training to a nonexercising control group in 56 premenopausal women, Lohman et al. (16) reported a significant aBMD increase of 2.3% in the lumbar spine and 1.8% in the femoral trochanter and a nonsignificant increase of 1.4% in the femoral neck in the strength training group relative to changes in the control group. In contrast to that study, we showed nonsignificant aBMD changes of 1.2% and 0.8% in the trochanter and lumbar spine, respectively, with no change in the femoral neck or total femur. The lower frequency of strength training in the current study compared with the previous trials may account for the differences seen. The frequency of strength training for this study was based on current guidelines (28).
Despite the lack of changes in aBMD in the intervention group, strength training did appear to prevent loss of bone mass at the femoral neck site. The control group in our study had a decrease in BMC of 1.5%, which is similar to that reported in other populations in this age group (2). In contrast, the strength training group had no change in femoral neck BMC, which suggests that the loading from muscle force may have preserved bone mass in the femoral neck. These findings are clinically important given the fact that hip fractures commonly occur at the femoral neck site, and the incidence, mortality, morbidity, and cost of repair is higher for hip fractures than for fractures at any other site (5).
Previous results from animal (12,19) and human (1) studies showed mechanical loading increased bone area or size (periosteal diameter) and bone strength with no change in bone volumetric density. These findings may translate to significant increases in bone strength that we were not able to measure with BMC or aBMD outcomes (12,26). This may be due, in part, to the choice to use DXA to measure bone. Bone measurements using DXA are limited in their ability to provide information about the bone geometry that may better describe the strength of the bone due to the planar nature of the measurement.
There are several reasons why the results of this study differed from other randomized trials conducted in premenopausal women. First, many of the previous randomized trials supplemented the premenopausal women in their studies with calcium (6,7,16,23), and calcium has been shown to interact with physical activity to optimize the bone response to loading (24). In the current study, just over 40% of the women met the recommended daily intake of calcium. Although it is possible that calcium supplementation would have resulted in a greater increase in the bone outcomes, our data show no evidence that calcium intake modified the effect in our sample.
A second difference between our study and previous studies is in the reported bone outcomes. None of the previous studies that tested the effect of an exercise intervention that included strength training only on aBMD in premenopausal women also reported area and BMC separately, so it is impossible to determine what component of aBMD changed (if any). Hui et al. (11) examined data from 466 women who had between 2 and 13 DXA measurements of femoral neck and lumbar spine bone to examine longitudinal changes in bone area, BMC, and aBMD. A small, but positive change in bone area persisted throughout adulthood, showing that the area of bone increases with age. The participants in the exercise group had significant increases in bone area at each of the measurement regions over the 2 yr of the study, except for a nonsignificant decrease the femoral neck in the control group.
Only one previous randomized study reported the effect of strength training on BMC in premenopausal women, but the intervention included both joint (strength training) and ground (jump roping) reaction force exercises. In a 2-yr trial (n = 22) comparing the effect of an intervention consisting of strength training and rope jumping with a control group, Weaver et al. (29) showed ∼2% increase in spine BMC in the intervention and control groups with no significant between-group differences. Therefore, the current study is the first randomized trial to report on the effect of strength training only on BMC in premenopausal women.
The results of this study showing only one significant effect of strength training on BMC are somewhat surprising, given the mechanism of bone formation. Mechanical loading applied to bone has a direct effect on bone remodeling (4), and physical activity, especially strength training, is a major factor by which mechanical loading occurs (14). Strength training causes muscle hypertrophy and increased contractile force from the tendon to the insertion on the bone, leading to mechanical loading on the insertion site. In the current study, lean mass increased more in the exercise group compared with the control group, showing that the intensity of the strength training was sufficient to produce muscle hypertrophy. One of the other characteristics for bone formation through mechanical loading is a progressive overload (13). In the current study, the participants progressively increased the weight lifted throughout the first year and then were able to maintain the weight lifted at the end of the first year for the second year. Although many participants indicated that they continued to increase the amount of weight lifted, not all did. This is evidenced by the fact that muscle strength and fat-free mass increased more from baseline to year 1 than from year 1 to year 2. This lack of progressive strength training throughout the full 2 yr of the intervention may have limited the ability to see an effect on BMC and aBMD.
The significant changes found in the femoral neck BMC may have been due to strength training maintaining bone during the initial period of bone loss. Hui et al. (11) found bone growth in the femoral neck ceased on average at age 24 (90% confidence interval [CI], 19-34) and bone loss in the femoral neck commenced on average at age 37 (90% CI, 20-42). The age range for the current study was 25-44 at baseline. Bone mass may have started decreasing as evidenced by decreasing BMC in the control group, which was not seen in the exercise group.
This study is the largest to date assessing the effect of strength training on bone measures in premenopausal women and has other design features worth highlighting. The frequency of strength training in the six previous randomized trials was thrice weekly. The current study included strength training 2 d·wk−1, which follows current published guidelines for strength training for overall health (28). The intensity of the strength training also followed published guidelines. The current study included one of the longest intervention periods to test the effect of strength training only on bone measures. Two previous studies in premenopausal women that included strength training of equal duration (2 yr) also included high impact exercise as part of the intervention (7,29). These combination intervention make it difficult to differentiate the effects of the strength training compared with impact exercises. Only one study has been completed with a longer intervention period of 3 yr, but the intervention included lumbar spine and shoulder exercise only at an intensity of 30% of 1RM, which is well below the recommended intensity (22). The current study included a total body strengthening program that stressed all major muscle groups, in line with current recommendations.
Notably, the women in this population had a BMI between 25 and 35 kg·m−2. Although body weight has been shown to be positively associated with changes in BMD (3), there are few studies exploring the effect of strength training on BMD in overweight women. It is possible that the higher weight in this population limited their bone response to loading given that BMD is already high in overweight women. In our population, baseline weight and weight gain were similar between groups, and adjusting for weight change did not influence the results.
In summary, strength training following published guidelines for 2 yr increased lean mass and muscle strength in premenopausal women. The strength training also resulted in statistically significant between-group changes in femoral neck BMC, a clinically relevant site for osteoporosis and osteoporotic fractures. No significant changes were seen in the proximal femur and the lumbar spine aBMD. Further trials are needed to understand the effects of strength training only or strength training combined with calcium on measures of bone strength. Further randomized controlled trials that assess change in bone dimensions and geometry, along with BMD, are needed to clarify the effects of strength training on bone strength and fracture risk in premenopausal women.
The SHE study was funded by a grant from NIDDK to Dr. Schmitz (NIH R01-DK060743) as well as the University of Minnesota's General Clinical Research Center grant (NIH M01-RR00400). Fitball USA is gratefully acknowledged for providing balls for participant incentives. The study sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript. The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Adami S, Gatti D, Braga V, Bianchini D, Rossini M. Site-specific effects of strength training on bone structure and geometry of ultradistal radius in postmenopausal women. J Bone Miner Res
2. Bainbridge KE, Sowers MF, Crutchfield M, Lin X, Jannausch M, Harlow SD. Natural history of bone loss over 6 years among premenopausal and early postmenopausal women. Am J Epidemiol
3. Blum M, Harris SS, Must A, et al. Leptin, body composition and bone mineral density in premenopausal women. Calcif Tissue Int
4. Chamay A, Tschantz P. Mechanical influences in bone remodeling: experimental research on Wolff's Law. J Biomech
5. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet
6. Dornemann TM, McMurray RG, Renner JB, Anderson JJ. Effects of high-intensity resistance exercise on bone mineral density and muscle strength of 40-50 year-old women. J Sports Med Phys Fitness
7. Friedlander AL, Genant HK, Sadowsky S, Byl NN, Gluer CC. A two-year program of aerobics and weight training enhances bone mineral density of young women. J Bone Miner Res
8. Frost HM. Bone "mass" and the "mechanostat": a proposal. Anat Rec
9. Heinonen A, Oja P, Kannus P, Sievanen H, Manttari A, Vuori I. Bone mineral density of female athletes in different sports. Bone Miner
10. Heinrich CH, Going SB, Pamenter RW, Perry CD, Boyden TW, Lohman TG. Bone mineral content of cyclically menstruating female resistance and endurance trained athletes. Med Sci Sports Exerc
11. Hui SL, Zhou L, Evans R, et al. Rates of growth and loss of bone mineral in the spine and femoral neck in white females. Osteoporosis Int
12. Jarvinen TL, Kannus PK, Sievanen H. Have the DXA-based exercise studies seriously underestimated the effects of mechanical loading on bone? J Bone Miner Res
13. Kohrt WM, Bloomfield SA, Little KD, Nelson ME, Yingling VR. American College of Sports Medicine Position Stand: physical activity and bone health. Med Sci Sports Exerc
14. Lanyon LE. Functional strain in bone tissue as an objective, and controlling stimulus for adaptive bone remodeling. J Biomech
15. 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
16. Lohman TG, Going SB, Pamenter RW, et al. Effects of resistance training on regional and total bone mineral density in premenopausal women: a randomized prospective study. J Bone Miner Res
17. Nord RH, Homuth JR, Hanson JA, Mazess RB. Evaluation of a new DXA fan-beam instrument for measuring body composition. Ann N Y Acad Sci
18. Recker RR, Davies KM, Hinders SM, Heaney RP, Stegman MR, Kimmel DB. Bone gain in young adult women. JAMA
19. Robling AG, Hinant FM, Burr DB, Turner CH. Improved bone structure and strength after long-term mechanical loading is greatest if loading is separated into short bouts. J Bone Miner Res
20. Schmitz KH, Hannan PJ, Stovitz S, Bryan CJ, Warren M, Jensen MD. Effect of strength training on total percent, and intra-abdominal adiposity: the Strong, Healthy, and Empowered Study. Am J Clin Nutr
21. Schmitz KH, Jensen MD, Kugler KC, Jeffery RW, Leon AS. Strength training for obesity prevention in midlife women. Int J Obes Relat Metab Disord
22. 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
23. Snow-Harter C, Bouxsein ML, Lewis BT, Carter DR, Marcus R. Effects of resistance and endurance exercise on bone mineral status of young women: a randomized exercise intervention trial. J Bone Miner Res
24. Specker BL. Evidence for an interaction between calcium intake and physical activity on changes in bone mineral density. J Bone Miner Res
25. Subar AF, Thompson FE, Kipnis V, et al. Comparative validation of the Block, Willett, and National Cancer Institute food frequency questionnaires: the Eating at America's Table Study. Am J Epidemiol
26. Turner CH, Robling AG. Mechanisms by which exercise improves bone strength. J Bone Miner Metab
27. US Department of Health and Human Services and Office of the Surgeon General. Bone Health and Osteoporosis. A Report of the Surgeon General
. Rockville, MD: US Department of Health and Human Services, Office of the Surgeon General;2004. pp. 68-87.
28. US Department of Health and Human Services. Physical Activity and Health. A Report of the Surgeon General
. Washington DC: National Center for Chronic Disease Prevention and Health Promotion;1996. pp. 11-56.
29. Weaver CM, Teegarden D, Lyle RM, et al. Impact of exercise on bone health and contraindication of oral contraceptive use in young women. Med Sci Sports Exerc
30. Wolff I, van Croonenborg JJ, Kemper HC, Kostense PJ, Twisk JW. The effect of exercise training programs on bone mass: a meta-analysis of published controlled trials in pre- and postmenopausal women. Osteoporosis Int
Keywords:©2008The American College of Sports Medicine
DXA; EXERCISE; OSTEOPENIA; MECHANICAL LOADING