The benefits of impact exercise on bone health have been clearly established (1,5,15), whereas nonweight-bearing sports such as swimming and cycling confer little to no benefit to bone (21,22). However, most data that have examined these effects have come from cross-sectional studies in adolescent and young adult athletes. Limited evidence exists regarding older adult athletes, particularly those who engage in a single sport throughout their adult years.
Several studies of cyclists have reported lower regional bone mineral density (BMD) values compared to runners, nonathletes, or reference values matched for ethnicity and age (2,16-18,21). In a study of male master cyclists, we reported that the cyclists had 10% lower lumbar spine and 11% lower total hip BMD compared to a group of nonathletes matched for age and body weight (16). Using the World Health Organization (WHO) criteria of T-score ≤−2.5 (24), 15% of the master cyclists had osteoporosis and an additional 52% had low bone mass at the hip and/or spine (i.e., T-score −1 to −2.5) (16). Because the study was cross-sectional in design, we were unable to determine any causal mechanisms for the high prevalence of low BMD. However, some of our findings suggested possible explanations. First, although the cyclists had been training and racing for an average of 20 years, most reported they had not participated in weight-bearing activities considered beneficial for bone health. Anecdotally, most cyclists reported avoiding weight-bearing activity as a means to aid in recovery from heavy cycle training. Thus, relative to the control subjects in that study (16) and to the dual-energy X-ray absorptiometry (DXA) reference population, it is likely that the master cyclists had chronically (∼20 y) undergone less-than-normal weight-bearing activity. Second, the low BMD in the cyclists may have been a result of a calcium imbalance. Although the group mean calcium intake of the master athletes met the dietary recommended intake (DRI) for men (16), it is possible that calcium requirements of cyclists may be greater than in nonathletes or nonendurance athletes. Recent reports have shown that dermal calcium loss from sweating can be high during sports participation (2,13). Thus, it is possible that the long-term effects of a lack of impact activity coupled with possible chronic calcium loss through sweat contributed to the poor bone health we observed in male master cyclists.
Currently, little is known regarding bone health of cyclists who train exclusively on their bikes and thus experience little, if any, beneficial mechanical loading effects on their bones. Only 1 prospective study has reported on bone health in competitive cyclists (2). Barry and Kohrt (2) reported a decrease in regional BMD during a competitive road cycling season in adult male cyclists. The BMD loss was only partially restored during the postseason. To our knowledge, no study has reported longitudinal changes in BMD in master cyclists. This information is important because male master cyclists are at a stage in life when they have achieved and are maintaining peak bone mass or are just beginning to lose bone mass; thus, they may benefit from information that has tracked BMD changes in this population.
According to 2007 data provided by the USA Cycling Association, which is the sanctioning organization for competitive cycling in the United States, 66.6% (22,541) of its male membership are ≥35 years and 31.4% are ≥45 years of age and presently competing in masters' cycling events (23) (note that in cycling, masters' competition begins at age 35). The popularity of masters' competition, along with its nonimpact nature and high risk of falls and fractures associated with crashes common in competitive cycling, warrants study of bone health in this population. Our initial findings of a high percentage of cyclists with low BMD prompted us to conduct a follow-up study with this population. Therefore, the purpose of the present study was to examine changes in total and regional BMD over a 7-year period in a group of highly trained male cyclists and a control group of age and weight-matched nonathletes initially studied in 2001. A secondary aim was to explore possible changes in behaviors related to bone health subsequent to each participant receiving results and interpretation of his BMD scan at the conclusion of the 2001 study.
Experimental Approach to the Problem
This study was designed to determine differential changes in BMD in competitive male master cyclists and nonathletes matched at baseline by age and body weight. Participants were assessed 7 years following the initial assessment in 2001. We hypothesized that the lower BMD observed among cyclists compared to nonathletes at baseline would persist at follow-up, with greater decline in cyclists who continued to train exclusively in the sport of cycling and engaged in minimal weight-bearing exercise.
A convenience sample of 27 competitive cyclists recruited from racing clubs in southern California and 24 nonathletic but active male controls participated in the baseline study in 2001. At baseline, inclusion criteria for the master cyclists included age 40-60 years; year-round cycling training consistently for a minimum of 10 hours per week, at least 150 miles per week, for at least 10 years; and competing in United States Cycling Federation (USCF) races for a minimum of 10 years. The study was approved by the University's Institutional Review Board, and all subjects gave written informed consent to participate.
The physical characteristics of the subjects at baseline and follow-up are shown in Table 1. By design, there were no baseline differences in age or body weight between the master cyclists and nonathletes. However, the cyclists had significantly more lean tissue mass and lower percent body fat than the nonathletes at both baseline and follow-up.
At baseline (2001), all subjects completed a health history questionnaire and were screened for conditions and medications known to affect bone health. Individuals were excluded from the study if they were taking or had ever taken medications or had any condition known to affect bone mass and/or bone metabolism, including inhaled steroids, anticonvulsants, calcitonin, alendronate, thyroid hormone, corticosteroids, cyclosporine, and anabolic steroids; thyroid or parathyroid disease; or adrenal gland problems. Other exclusion criteria included past or present smoking (more than 3 years) and heavy alcohol use (>2 drinks per day). Also, individuals who participated in a strength training program 2 or more days per week for more than 3 months a year were not eligible to participate. Although nonathletes were excluded if they reported training and competing in any sport in the past 10 years, recreational exercisers were allowed to participate. The nonathletes were matched to the master cyclists by age (± 2 years) and body weight (± 2 kg). In 2008 (7 years later), all former participants were invited to return to the study's laboratory for follow-up testing. Of the 27 cyclists and 24 nonathletes assessed at baseline, 22 cyclists (81.4%) and 19 nonathletes (79.2%) completed follow-up measures.
Bone Density Measurements
Bone mineral density (g/cm2) of the lumbar spine (L1-L4), proximal femur, and total body was assessed by DXA using a Lunar DPX-NT densitometer (GE/Lunar Corp, Software Version 4.0, Madison, WI). Soft tissue mass also was obtained from the total body scan to determine body composition. All baseline and follow-up scans were conducted on the same machine and by the same technologist. Quality assurance (QA) tests were performed each morning of use, using a standard with tissue-equivalent material with 3 bone-simulating chambers of known bone mineral content. In vivo BMD precision, which was determined by scanning 30 subjects twice with repositioning between scans and calculated as the coefficient of variation (% CV), was 1.2% for the lumbar spine, 1.7% for total hip, and 0.8% for the total body.
To determine the percentage of cyclists and nonathletes who presented with osteopenia or osteoporosis, we used criteria published by the International Society of Clinical Densitometry (ISCD) (12) for men older than 50 years of age. Osteopenia is defined as having a BMD value between 1 and 2.5 standard deviations (SD) below the peak reference value of a young adult male of the same ethnicity (i.e., T-score between −1 and −2.5 at the spine, femoral neck, or total hip), whereas osteoporosis is defined as a T-score ≤ −2.5 SD.
At baseline and follow-up, the cyclists recorded the number of days per week, hours per day, and months per year spent cycling. Many cyclists kept training diaries and referred to them for these data. The cyclists also reported any new sport or physical activity, including weight training, and the number of years (or months if <1 year) they had engaged in each activity since their baseline visit to the lab. The nonathletes were asked to report, for a typical week, the number of days and hours they engaged in impact exercise or weight training; they also reported the number of months per year they performed these activities. All subjects were asked whether their exercise participation had increased, decreased, or remained the same since their initial visit to the study's laboratory.
At baseline, participants completed a 3-day dietary recall, from which mean daily calcium intake was determined, including calcium from supplements. At follow-up, a semi-quantitative food frequency for calcium was administered (6), which included a question on calcium supplement use. Changes in total calcium intake were estimated from these measures.
At baseline, all participants received a printout of their BMD scans and a detailed explanation of their results following completion of data collection and analysis. For the present study, we were interested in learning whether there had been any differential changes made in bone-health behaviors among the cyclists and nonathletes. Therefore, participants were asked if they had made changes in their exercise habits since baseline and, if so, the reason for any change. Possible responses included the following: “to improve my bone density”; “to improve my health unrelated to my bone density”; “to help control my body weight”; “to maintain or increase muscle mass”; or “for reasons unrelated to my health, weight, or muscle mass.” They also were asked if they “made any changes in their dietary habits related to calcium.” Responses included “None,” “I increased intake of dairy products,” “I began taking and continue to take a calcium supplement,” or “I increased dairy intake and am taking a calcium supplement.” At follow-up, participants were also asked to report any medication they were taking for their bone health and for how many months or years since baseline had they been taking the medication(s).
Data were analyzed using the Statistical Package for the Social Sciences software (version 15.0; SPSS, Inc., Chicago, IL, USA). Descriptive data are presented as group mean ± SD. Paired t-tests were used to compare within-group changes in physical characteristics, calcium intake, and training and physical activity in cyclists and nonathletes, respectively. Student t-tests for independent samples were used to determine whether the group differences in BMD we had observed at baseline were still evident at follow-up. Bone sites of interest included the lumbar spine (L1-L4), total hip, femoral neck, and total body. We conducted analysis of covariance (ANCOVA) with repeated measures to determine possible differential changes in BMD between cyclists and nonathletes. We adjusted for changes in variables known to affect BMD and that were significantly different between cyclists and nonathletes, including body mass index (BMI), lean tissue mass, and calcium intake. We also adjusted for the number of years subjects reported engaging in any new impact exercise or weight training since baseline because both of these modes of exercise may positively influence BMD. We covaried for this change in exercise rather than exclude those subjects completely because in the initial study cyclists were excluded if they reported any regular impact exercise or weight training. Independent t-tests were used to compare BMD at follow-up between study participants who reported engaging in weight training or impact exercise since the baseline assessment. Chi-square analysis was used to compare the difference in proportion of cyclists and proportion of nonathletes who met BMD T-score criteria for osteoporosis or osteopenia/low bone mass. The alpha level was set at p ≤ 0.05 for significance.
Among the cyclists were regional, national, and international competitors, several of whom were age-group national (n = 5) and world champions (n = 1). The overall retention rate of the study was 80.4% (22 cyclists and 19 nonathletes). Reasons for attrition included the following: unable to contact (n = 5), moved out of area (n = 2), refusal due to time constraints (n = 2), and death (n = 1). Further, 3 cyclists and 1 nonathlete were excluded from the longitudinal data analysis because of their use of medications known to improve BMD, which they began taking at least 3 years before follow-up measures were conducted. Thus, in the current report, the final sample size was n = 37 (19 cyclists and 18 nonathletes). We compared baseline data of subjects who dropped out or who were excluded from the study to those who were included in the study and found no significant differences in any variables, including physical characteristics and BMD at all measurement sites (p > 0.05).
Exercise patterns and calcium intake of participants are reported in Table 2. The nonathletes decreased their frequency of exercise from nearly 5 to 3 days per week (p = 0.001), and their duration from 4.5 to 3.2 hours per week (p = 0.02). The cyclists reported minimal changes in their frequency of training. Although the group mean cycling duration decreased by 1 hour per week, this difference was not statistically significant. Calcium intake at baseline and follow-up was significantly greater among cyclists than nonathletes (p < 0.005). Although the cyclists showed a trend (p = 0.053) toward a decrease in calcium intake, their group mean value remained close to the recommended amount of 1,200 mg/day for men older than 50 years of age. Conversely, the mean calcium intake of the nonathletes was considerably lower than recommended (8).
Group mean (± SD) BMD values for regional and total body BMD at baseline and follow-up are reported in Table 3. Independent t-tests indicated significant group differences at baseline for total hip and femoral neck BMD (p < 0.05) and a trend (p < 0.10) at the total body and spine. At follow-up, the cyclists' BMD was significantly lower than that of nonathletes at all sites (p < 0.05) except the spine, which showed only a trend (p < 0.10). Group-by-time repeated-measures ANCOVA indicated a significant group-by-time interaction for total body BMD (p = 0.03).
Figure 1 illustrates the percentage of cyclists and nonathletes classified as having osteopenia (BMD T-score −1 to −2.5) or osteoporosis (BMD T-score of ≤−2.5) at baseline and follow-up. A significantly greater percentage of cyclists than nonathletes met the ISCD criteria for osteopenia or osteoporosis at baseline (84.2% vs. 50.0%, p = 0.026) and at follow-up (89.5% vs. 61.1%, p = 0.034).
Changes in self-reported behaviors associated with BMD since the initial study in 2001 are reported in Table 4. Nearly 74% and 63% of the cyclists and 72% and 78% of the nonathletes reported not making any changes in calcium intake or exercise habits, respectively. Among the cyclists who did change their behaviors, 4 began taking calcium supplements and 1 reported increasing dairy consumption, 4 began weight-training, 1 began doing impact exercise, and 2 reported both weight training and impact exercise. Among the nonathletes, 2 reported increasing dairy consumption and 3 others reported taking calcium supplements. Three nonathletes reported engaging in both weight training and impact exercise since baseline, and 1 began weight training only. None reported engaging in impact exercise alone. Of the cyclists who reported making a change, 50% did so to increase or improve their bone health. None of the 4 nonathletes made a change for the purpose of improving their bone health (Table 4).
Figure 2 compares the changes in BMD in study participants who reported participation vs. no participation in weight training or impact exercise since the baseline assessment. Subjects who reported engaging in either weight training or impact exercise, or both (n = 11), lost significantly less BMD at the spine and femoral neck compared to those who did not participate in these modes of exercise (n = 26).
To our knowledge, this is the first longitudinal study of BMD in older master cyclists. However, a recent study that tracked BMD changes over 1 year of training and racing in male road cyclists (mean age ∼35 years) reported approximately 1-1.5% mean decrease in BMD at the proximal femur from preseason to the end of racing season, with only partial recovery of bone mass during 3 months of off-season (2). In the present study, all cyclists had been training and racing for an average of 27 years and had engaged in little to no weight-bearing activity for most of their adult life. We found that cyclists and nonathletes had similar losses in regional BMD over the 7 years of follow-up, as demonstrated by the lack of an interaction effect between the cyclists and nonathletes at the spine and proximal femur sites. However, after adjusting for changes in calcium intake, lean tissue mass, BMI, and impact exercise or weight training, there was a significantly greater decline in total body BMD in cyclists compared to nonathletes. We also found that among all subjects, those who reported participating in weight training or impact exercise since the baseline assessment lost significantly less bone mass at the spine and femoral neck than those subjects who reported no participation in these modes of exercise. Although the total hip and total body sites were not significantly different between these groups, Figure 2 illustrates a pattern of less bone loss at all bone sites in subjects that participated in weight training or impact exercise than those that did not. Given that nearly 37% of the cyclists, compared to 22% of the nonathletes, began weight training and/or impact exercise following baseline assessment, this change in their exercise habits may have resulted in less of a group mean decline than otherwise expected in the cyclists' regional BMD because many studies have shown the positive effects of high impact or resistance/weight training on bone mass in adult men and women (1,7,15,19,20). We had hypothesized that BMD would decrease more in cyclists at all bone sites if they continued to engage only in cycling and not perform weight-bearing recreational or occupational physical activity.
Another possible reason for the lack of differential change in regional BMD between the cyclists and nonathletes is that 4 subjects were excluded from data analysis because they reported taking bone-building medications since the baseline assessment. Excluding these subjects may have biased the change scores because the excluded individuals had low baseline BMD (thus the reason for beginning medication use) and may have otherwise been losing bone mineral at an accelerated rate. Because 3 of the 4 excluded subjects were cyclists, if we would have retained them in our final analysis, we may have observed a significant difference in regional BMD between the groups.
As we had observed in the previous cross-sectional study (16), there was a consistent pattern of lower BMD in cyclists than nonathletes. At the 7-year assessment, all bone sites were significantly lower in cyclists than controls, except the spine, which showed a trend (p = 0.09). Close observation of the lumbar spine scans indicated that some subjects developed osteophytes and possible arthritic changes that may have artificially increased their BMD values. In those subjects, lumbar BMD appeared to have increased, which resulted in large standard deviations and a smaller decrease in mean values at follow-up. Others have shown that the spine may not be a reliable site for longitudinal assessment of BMD in middle-aged or older adults as a result of age-related changes that confound interpretation of BMD at this bone site (11).
To compare BMD values of study participants to their age-matched peers, we examined BMD Z-scores, which express BMD values relative to a reference population matched for age, weight, height, and ethnicity. At the initial assessment, BMD Z-scores of the master cyclists were approximately 10% lower than that of both the reference population and the nonathletes at the hip and lumbar spine, whereas the nonathletes were representative of their age-matched peers, with BMD Z-scores ranging from 99% to 103% of the reference population (16). In the current study, the BMD Z-scores of both cyclists and nonathletes remained at similar percentages of the reference population, which suggests that the changes in absolute BMD in both groups tracked as predicted.
The large differences in BMD between cyclists and nonathletes remain unexplained. In the baseline study we matched cyclists to nonathletes by age and body weight, and included only nonsmokers and those with no medical conditions or taking medications known to affect bone mass, to eliminate these known confounders. We therefore concluded that the group differences we observed may have been a result of the cyclists spending very little time doing any weight-bearing activity over many years of cycle training, even though they engaged in a high volume of aerobic training. This interpretation of our data was supported by other studies in adult athletes showing that nonimpact sports, even if performed vigorously, do not provide an osteogenic stimulus compared to impact sports in which mechanical loading is applied unevenly and at a very high rate, such as in gymnastics, volleyball, judo, and soccer (1,5,15,21). It is also possible that the cyclists self-selected into their sport, in part because of their low body mass, which is important for success in cycling, particularly on hilly terrain. Alternatively, 7 years may not have been sufficient time to discern small group differences in the rate of bone loss in our small sample because the expected rate of decline in men in their 50s is less than 0.5% per year (9). Thus, additional follow-up assessments are needed to plot changes over a longer time period.
Several interesting observations were made in behavior change from baseline to follow-up. When asked whether they changed behaviors related to bone health as a result of learning about their BMD following the initial assessment, 63% of the cyclists and 78% of the nonathletes reported no change in their exercise habits. Similarly, 74% of the cyclists and 72% of the nonathletes also reported no change in dietary habits related to consuming dairy products or taking a calcium supplement. Of interest, only 3 of the 13 cyclists classified as osteopenic or osteoporotic at the initial assessment reported a change in their exercise habits for the purpose of improving their BMD. These cyclists began resistance training in an effort to maintain their BMD. The same 3 cyclists also reported that they began taking a calcium supplement, but none reported they increased their dairy intake. Among the 7 nonathletes with osteopenia or osteoporosis in 2001, none reported changing exercise or diet behaviors specifically to improve bone health. These findings underscore the need for behavior change strategies, in addition to education, to effect positive change in health behaviors.
Our data indicate that nearly 90% of the master cyclists and 61% of the nonathletes met the ISCD criteria for either osteopenia or osteoporosis. This finding is alarming, particularly for the cyclists who are at high risk for traumatic fractures from crashes associated with competitive cycling. We were somewhat surprised at the high percentage of nonathletes categorized as osteopenic/osteoporotic, given their mean age of 58 years. However, we suspect their relatively low body weight may be a contributing factor because body weight is strongly associated with BMD (3). At baseline, we recruited nonathletic men matched for body weight to the cyclists, whose weight and BMI ranged from 61 to 86 kg and 20 to 24 kg/m2, respectively. With a mean BMI of approximately 22 kg/m2, both groups of men were within approximately the 20-25th percentile for BMI (14). Thus, our final sample was composed of men with lower weight and BMI compared to the general population.
Although the retention rate of the study was fairly high, our final sample size limited statistical power because we excluded 4 participants (3 of whom were cyclists) in follow-up data analysis because they were taking bone-building medications. Also, although regular weight training was an exclusionary criterion for study participation at baseline, at follow-up we chose to retain subjects that had adopted weight training rather than further reduce the sample size. Thus, the reduced sample size combined with behavioral changes known to influence bone health may have precluded our ability to detect small differential changes in regional BMD.
Another limitation of the study is the absence of hormone data, particularly testosterone, which has an important role in bone mass (4) and has been shown to be suppressed in some endurance athletes (10) and which may be declining at variable rates among study participants. Future follow-up measures should include measurement of sex steroids and other hormones that are affected by endurance exercise and that either directly or indirectly influence bone turnover and resultant bone mass.
A third limitation of the study was the different methods used to assess calcium intake at baseline and follow-up. Although 3-day diet records were obtained in the initial study, at follow-up many participants indicated that they did not want to record dietary data. Therefore, we chose to substitute a measure of calcium using a semi-quantitative food frequency. Doing so may have introduced errors in estimates of calcium intake. However, reliability for repeated measures of calcium intake in the total study sample yielded an intraclass correlation coefficient of 0.60. Thus, it appears that calcium intake remained fairly stable from 2001 to 2008.
Coaches and health professionals interacting with master cyclists need to educate and assist these athletes in developing behavioral strategies to optimize their bone health through proper diet and exercise, as appropriate for their age and health status. Given the findings in this study, coupled with the high risk of fractures resulting from crashes associated with competitive cycling, we recommend that master cyclists (1) be screened periodically to determine their BMD; (2) complement their cycle training with brief bouts of high-impact exercise and/or high-intensity resistance/weight training several times each week; (3) consume a balanced diet with adequate intake of energy, calcium, vitamin D, and other nutrients important for bone health; and 4) eliminate other known modifiable risks of osteoporosis such as smoking and alcohol intake greater than 2 drinks per day. More information regarding bone health in cyclists will enable sports medicine physicians and other health care professionals who monitor the care of these athletes to make recommendations regarding diet, alternative exercise, or possible pharmacologic intervention to preserve bone mass and prevent early onset of osteoporosis.
1. Andreoli, A, Monteleone, M, Van Loan, M, Promenzio, L, Tarantino, U, and De Lorenzo, A. Effects of different sports on bone density and muscle mass in highly trained athletes
. Med Sci Sports Exerc
33: 507-511, 2001.
2. Barry, DW and Kohrt, WM. BMD decreases over the course of a year in competitive male cyclists. J Bone Miner Res
23: 484-491, 2008.
3. Blain, H. Osteoporosis in men: Epidemiology, physiopathology, diagnosis, prevention and treatment. Rev Med Interne
25(Suppl 5): S552-S559, 2004.
4. Boonen, S, Vanderschueren, D, Cheng, XG, Verbeke, G, Dequeker, J, Geusens, P, Broos, P, and Bouillon, R. Age-related (type II) femoral neck osteoporosis in men: Biochemical evidence for both hypovitaminosis D- and androgen deficiency-induced bone resorption. J Bone Miner Res
12: 2119-2126, 1997.
5. Calbet, JA, Diaz Herrera, P, and Rodriguez, LP. High bone mineral density in male elite professional volleyball players. Osteoporos Int
10: 468-474, 1999.
6. Cummings, SR, Block, G, McHenry, K, and Baron, RB. Evaluation of two food frequency methods of measuring dietary calcium intake. Am J Epidemiol
126: 796-802, 1987.
7. Daly, RM, Rich, PA, Klein, R, and Bass, S. Effects of high-impact exercise on ultrasonic and biochemical indices of skeletal status: A prospective study in young male gymnasts. J Bone Miner Res
14: 1222-1230, 1999.
8. Food and Nutrition Board. Institute of Medicine. Dietary reference intakes for calcium, phosphorus, magnesium, vitamin D, and fluoride
Washington, DC: National Academy Press, 1997.
9. GE/Lunar Database for Adult Men, Software Version 4.0, 1999.
10. Hackney, AC, Szczepanowska, E, and Viru, AM. Basal testicular testosterone production in endurance-trained men is suppressed. Eur J Appl Physiol
89: 198-201, 2003.
11. Huuskonen, J, Vaisanen, SB, Kroger, H, Jurvelin, JS, Alhava, E, and Rauramaa, R. Regular physical exercise and bone mineral density: A four-year controlled randomized trial in middle-aged men. The DNASCO study. Osteoporos Int
12: 349-355, 2001.
12. International Society for Clinical Densitometry. Diagnosis of osteoporosis in men, premenopausal women, and children. J Clin Densitom
7: 17-26, 2004.
13. Klesges, RC, Ward, KD, Shelton, ML, Applegate, WB, Cantler, ED, Palmieri, GM, Harmon, K, and Davis, J. Changes in bone mineral content in male athletes
. Mechanisms of action and intervention effects. JAMA
276: 226-230, 1996.
14. Kuczmarski, RJ, Carroll, MD, Flegal, KM, and Troiano, RP. Varying body mass index cutoff points to describe overweight prevalence among U.S. adults: NHANES III (1988 to 1994). Obes Res
5: 542-548, 1997.
15. Morel, J, Combe, B, Francisco, J, and Bernard, J. Bone mineral density of 704 amateur sportsmen involved in different physical activities. Osteoporos Int
12: 152-157, 2001.
16. Nichols, JF, Palmer, JE, and Levy, SS. Low bone mineral density in highly trained male master cyclists. Osteoporos Int
14: 644-649, 2003.
17. Rector, RS, Rogers, R, Ruebel, M, and Hinton, PS. Participation in road cycling vs running is associated with lower bone mineral density in men. Metabolism
57: 226-232, 2008.
18. Smathers, AM, Bemben, MG, and Bemben, DA. Bone density comparisons in male competitive road cyclists and untrained controls. Med Sci Sports Exerc
41: 290-296, 2009.
19. Snow, CM, Shaw, JM, Winters, KM, and Witzke, KA. Long-term exercise using weighted vests prevents hip bone loss in postmenopausal women. J Gerontol A Biol Sci Med Sci
55: M489-M491, 2000.
20. Snow-Harter, C, Whalen, R, Myburgh, K, Arnaud, S, and Marcus, R. Bone mineral density, muscle strength, and recreational exercise in men. J Bone Miner Res
7: 1291-1296, 1992.
21. Stewart, AD and Hannan, J. Total and regional bone density in male runners, cyclists, and controls. Med Sci Sports Exerc
32: 1373-1377, 2000.
22. Taaffe, DR, Snow-Harter, C, Connolly, DA, Robinson, TL, Brown, MD, and Marcus, R. Differential effects of swimming versus weight-bearing activity on bone mineral status of eumenorrheic athletes
. J Bone Miner Res
10: 586-593, 1995.
23. USA Cycling membership information. Available at http://www.usacycling.org
. Accessed July 31, 2008.
24. World Health Organization. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: Report of a WHO Study Group. World Health Organization Technical Report 843: 1-129, 1994.