Reduced Training Is Associated With Increased Loss of BMD*

Abstract

This 8‐year controlled, follow‐up study in 66 Swedish soccer women evaluated the effect of training and reduced training on BMD. The players who retired during the follow‐up lost BMD in the femoral neck, whereas the controls did not.

Introduction: Physical activity during adolescence increases BMD, but whether the benefits are retained with reduced activity is controversial.

Materials and Methods: At baseline, DXA evaluated BMD in 48 active female soccer players with a mean age of 18.2 ± 4.4 (SD) years, in 18 former female soccer players with a mean age of 43.2 ± 6.2 years and retired for a mean of 9.4 ± 5.3 years, and in 64 age‐ and sex‐matched controls. The soccer women were remeasured after a mean of 8.0 ± 0.3 years, when 35 of the players active at baseline had been retired for a mean of 5.3 ± 1.6 years.

Results and Conclusions: The players still active at follow‐up had a higher BMD at baseline than the matched controls in the femoral neck (FN; 1.13 ± 0.19 versus 1.00 ± 0.13 g/cm²; p = 0.02). The yearly gain in BMD during follow‐up was higher in the active players than in the controls in the leg (0.015 ± 0.006 versus 0.007 ± 0.012 g/cm², p = 0.04). The soccer players who retired during follow‐up had a higher BMD at baseline than the matched controls in the FN (1.13 ± 0.13 versus 1.04 ± 0.13 g/cm²; p = 0.005). The players that retired during follow‐up lost BMD, whereas the controls gained BMD during the study period in the FN (−0.007 ± 0.01 versus 0.003 ± 0.02 g/cm² yearly; p = 0.01). The soccer players already retired at baseline had higher BMD at study start than the matched controls in the leg (1.26 ± 0.09 versus 1.18 ± 0.10 g/cm²; p = 0.01). The former players who were retired at study start lost BMD, whereas the controls gained BMD during the study period in the trochanter (−0.006 ± 0.01 versus 0.004 ± 0.014 g/cm² yearly; p = 0.01). This study shows that, in girls, intense exercise after puberty is associated with higher accrual of BMD, and decreased physical activity in both the short‐term and long‐term perspective is associated with higher BMD loss than in controls.

soccer women, physical activity, BMD, follow‐up study, femoral neck

INTRODUCTION

Cross‐sectional and prospective uncontrolled and controlled studies suggest that weight‐bearing physical activity increases the accrual of BMD during growth and that the skeleton seems to be most responsive to mechanical load during the pre‐ and peripubertal years.^(1–6) Furthermore, activities with a high strain and a strain applied at a high rate and at different angles are the most effective load for increasing BMD, whereas the duration of the activity, calculated as minutes of physical activity per day, seems to be of minor importance.^(7–12) Soccer exercise has most of the beneficial effects described above, and individuals subjected to soccer exercise during growth and adolescence, both men and women, are reported to have 1‐2 SD higher BMD in weight‐loaded regions in comparison with controls.^(13–17) This benefit would likely more than halve the fracture risk if retained into old age.^(18,19) However, before vigorous exercise during adolescence can be recommended as a prevention strategy for osteoporosis, data must be presented to corroborate that the beneficial effects of exercise are retained with reduced activity level. Most cross‐sectional studies suggest that BMD benefits of 0.5‐1.0 SD are retained for one to two decades of retirement, a benefit no more than one‐half that observed in active athletes, but that virtually no benefits are retained after three to four decades of retirement, the years when the incidence of fragility fractures rises exponentially.^{(1,15,20–26)} However, even if higher BMD does not exist in former athletes compared with controls after decades of retirement, residual benefits in bone size and bone quality may exist, benefits that are not captured by DXA measurements but that could increase bone strength and reduce fracture incidence. Furthermore, cross‐sectional observations may be influenced by secular changes in the duration and intensity of the exercise. Elderly retired soccer players could train less vigorously in their youth than young soccer players today. Thus, they may have attained a lower peak BMD, producing the apparent accelerated diminution seen in this cross‐sectional study. We cannot exclude this possibility. Neither can we exclude that differences in lifestyle between former athletes and controls could influence the outcome. Thus, prospective data with retirement from exercise are needed before we can draw inferences with a higher level of evidence. At present, no conclusive answers can be obtained from published longitudinal data, because studies report both that BMD benefits are maintained with cessation of activity^(27–32) and that BMD benefits are lost with retirement.^(33–38) Because all prospective studies are short‐term reports, catching only a short period of retirement, we must wonder if the retirement period might have been too short to capture a faster BMD loss compared with controls. Until now, there are no prospective controlled studies following young girls from intense activity into a retirement period exceeding 5 years. Furthermore, no prospective controlled studies have thus far been published that follow BMD changes in women retired for decades.

The purpose of this study was to assess the effect of training and short‐term and long‐term detraining on BMD in young women subjected to intense physical activity during growth and adolescence in a nonrandomized, prospective, controlled study.

MATERIALS AND METHODS

BMC and BMD (g/cm²) were evaluated by DXA at baseline (1994) in 96 active competing female soccer players with a mean age of 18.3 ± 4.0 years and two to five training sessions per week, which was predominantly soccer training and supported by impact‐loaded and aerobic training. BMD was also evaluated in 25 former soccer players with a mean age of 40.0 ± 4.5 years and retired for a mean of 9.7 years (range, 5‐20 years).⁽¹⁵⁾ All women were white.

Eight years later, 14 women had moved out of the region and 5 lived abroad, leaving 102 individuals to be invited to a follow‐up evaluation. Four women were pregnant at this time and thus excluded from further measurements. Of the 98 who could be included, 70 consented to participate, 50 players active at baseline and 20 players retired at baseline, an attendance rate of 71%. Before any calculations were made, we further excluded one woman who was active at baseline and two women who were retired because of medications or a disease known to affect the bone metabolism. We also excluded an outlier: one women who was active at baseline but 6 years older than the second oldest active soccer player. This resulted in the inclusion of 13 women who at follow‐up were still active soccer players training predominantly with soccer, which was supported by impact‐loaded and aerobic training, 35 women who were active players at baseline but had retired from competitive training during the study period, so that at follow‐up they had a mean retirement period of 5.3 ± 1.6 years, and 18 women who had already retired at baseline, so that at follow‐up they had a mean retirement period of 17.3 ± 5.4 years.

The active and retired soccer players were compared with 64 women living in the same region, none exercising at a competitive level but some at a recreational level, predominantly with aerobics, distance running, and weight training. These women were measured at baseline during 1992‐1994 as a part of a larger prospective followed normative database with the same DXA machines as the soccer players.⁽³⁹⁾ From this cohort, we retrospectively matched the controls to the soccer players by sex and age (Table 1). No attempt was made to match for body weight. None of these women had a disease or was taking medication known to affect bone metabolism, and as among the cases, all were white. From this cohort, we created three separate control groups matched by age to the soccer players still active at follow‐up, the players who retired during the study period, and the former players who had retired already at baseline (i.e., one individual could be matched to more than one control cohort).

Table 1.

Age, Height, Weight, BMD (g/cm²), and BMC (g) Evaluated by DXA at Baseline and at 8‐Year Follow‐up in Soccer Women Who Were Active Both at Baseline and at Follow‐up and in Matched Controls, in Soccer Women Who Were Active at Baseline but Who Retired During the Follow‐up, and in Matched Controls and in Soccer Women Who Were Retired Both at Baseline and at Follow‐up and in Matched Controls

Two DXA devices (Lunar DPX‐L, software version 1.3y; Lunar, Waukesha, WI, USA) were used during the study. The apparatus was calibrated each day using a standardized phantom. The precision of the DXA measurements in vivo was as follows: total body, 0.4%; spine, 0.5%; femoral neck, 1.6%. Additionally, the cross‐calibrations of the machines found a difference of <0.4%. The evaluation included BMD in the total body, the head, the spine, the lumbar spine, the hip (femoral neck, Ward's triangle, trochanter), and the leg. From the hip scans, the narrowest width of the femoral neck (FN; cm) was measured by a ruler included as software in the DXA equipment. BMC data are presented from total body and head, so that the effect of dental work could be subtracted. At baseline, 15 of the active and retired soccer players were measured only in the hip region; total body, head, spine, lumbar spine, and leg data were achieved from 12 players who were active throughout the study and 21 active players who retired during the follow‐up.⁽¹⁵⁾ Body weight was measured by an electric scale, and body height was measured by a wall‐tapered height meter. A questionnaire, previously used in several studies,^{(14,15,25,40)} evaluated lifestyle factors such as smoking habits, alcohol intake, coffee consumption, illnesses, medications, use of contraceptives, blue collar workers, individuals excluding anything in food intake, age at menarche, age at menopause, current and past exercise history, years of active exercise career, type and level of activity, and years of retirement both at baseline and at follow‐up.

Statistical calculations were performed with Statistica, version 5.3 (StatWin). Data are presented as mean ± SD. Student's t‐test between means was used to compare BMD in soccer players and controls. A χ² test or a Fisher's exact test evaluated differences in lifestyle factors between the groups. A p value of <0.05 was considered as a statistically significant difference. The ethics committee at the University of Lund approved the study.

RESULTS

Neither at baseline nor at endpoint were any significant differences in anthropometrics or age at menarche found when comparing either the still active soccer players, the players who retired during the study, or the players who were already retired at study start with their matched controls (Tables 1 and 2). All active soccer players and all soccer players who retired during the follow‐up were still on regular menstruation at follow‐up, as were their matched controls. Nine former soccer players who were retired before baseline and 15 matched controls reached menopause by follow‐up (mean age, 49.3 ± 3.3 and 50.5 ± 2.9 years, respectively; p = 0.39). Furthermore, lifestyle factors, except the level of physical activity, differed minimally when the three groups of players were compared with their matched controls (Table 2).

Table 2.

Lifestyle Characteristics at Follow‐up in Soccer Women Who Were Active Both at Baseline and at Follow‐up and in Matched Controls, in Soccer Women Who Were Active at Baseline but Who Retired During the Follow‐up and in Matched Controls, and in Soccer Women Who Were Retired Both at Baseline and at Follow‐up and in Matched Controls>

The soccer players who were active throughout the study had a higher BMD at baseline than the matched controls, reaching significance in the FN, Ward's triangle, and the trochanter region (Table 1). During follow‐up, the yearly gain in BMD was higher in the active players than in the controls, reaching a significant difference in the leg (Table 3); the differences between the groups in BMD were even larger in magnitude at follow‐up (Table 1). Bone size and annual changes in bone size in the soccer players who were active throughout the study were not different at baseline or at follow‐up compared with controls (Table 1).

Table 3.

Annual Changes in BMD During an 8‐Year Follow‐up in Soccer Women Who Were Active Both at Baseline and at Follow‐up and in Matched Controls, in Soccer Women Who Were Active at Baseline but Who Retired During the Follow‐up, and in Matched Controls and in Soccer Women Who Were Retired Both at Baseline and at Follow‐up and in Matched Controls

The soccer players who ended their active career during the follow‐up period had a higher BMD at baseline than the matched controls, reaching significance in the FN, Ward's triangle, the trochanter region, and the leg (Table 1). During follow‐up, the retired players lost BMD, whereas the controls gained BMD in the FN and in the trochanter region; retired players gained less BMD in the spine than the controls (Table 3). Despite the discrepancy in the changes of BMD, the soccer players who ended their active career during the follow‐up period still had a higher BMD, at a lesser magnitude, at follow‐up than the controls in most regions, but this still reached significant difference in the total body and the leg (Table 1). Bone size was larger at baseline but not at follow‐up, and the annual changes in bone size was less in the soccer players who were active at baseline but retired during the study compared with the controls (Table 1).

The soccer players who were retired at baseline had a higher BMD at baseline than the matched controls, reaching a significant difference only in the leg (Table 1). During follow‐up, the soccer players who were retired at baseline lost BMD, whereas the controls gained BMD at a significant level in the trochanter region, whereas a reversed development was seen in the BMD head region (Table 3). Despite the discrepancy in the changes in BMD, the soccer players who were retired at baseline still had higher BMD in the leg at follow‐up than the controls, but at a lesser magnitude (Table 1). Bone size was larger at baseline and at follow‐up, whereas the annual changes in bone size were not different, in soccer players who were retired both at baseline and at follow‐up compared with the controls (Table 1).

DISCUSSION

This report adds the longest prospective observational controlled data thus far that evaluate the skeletal effects of training and detraining in young female former competing athletes. The study indicates that physical activity in the postpubertal period is associated with an increased accrual of BMD and that retirement from physical activity is associated with a faster loss in BMD compared with controls. However, despite this, former soccer players that retired during the follow‐up had higher BMD in the total body and the leg and those retired for a long time had higher BMD in the legs compared with controls. In addition to BMD data, BMC is also presented from total body and the head, so that the effect of dental work could be subtracted. If doing this calculation, all our conclusions remained.

Today there are a variety of reports—cross‐sectional studies, case control studies, observational studies, and prospective controlled trails—that consistently show that physical exercise during growth is associated with increase accrual of BMD.^(3,6,41,42) The data also support the notion that the pre‐ and peripubertal period is the maturational period when exercise may confer the greatest BMD benefits, whereas individuals in the postpubertal years have less response to mechanical load,^(2,5,43) and that exercise during adulthood has been described as at best conferring BMD benefits of a few percentage points.^(32,44) However, the notion is only supported by short‐term prospective controlled studies spanning, at best, 24 months and cross‐sectional studies evaluating the arm‐to‐arm difference in racquet players in relation to starting age of the exercise. This study, providing 8‐year prospective data in exercising young girls, partly opposes the belief common today. The female soccer players in this study who were physically active throughout the study, despite having a higher BMD at baseline, also had a greater accrual of BMD from a mean age of 17‐24 years compared with the controls. In relative changes, this corresponds to a leg BMD increase of 10.6 percentage points in active girls and a 3.5 percentage point increase in the controls, which is a difference of 7.1 percentage points, corresponding to an annual increase of 1.3 and 0.5 percentage points, respectively. In other words, continuous intense exercise for several years in the postpubertal and early adulthood years may also lead to further BMD benefits. The discrepancy with respect to previously published prospective data could be that the girls in this study trained at a higher level and intensity than the women in previously cited studies and that this study, because of the longest follow‐up in young competing girls, captured additive BMD effects over several years of continuous training. However, because this study was not a prospective randomized controlled trial, we cannot draw conclusions regarding causality. It seems unlikely that other differences in lifestyle factors than the level of physical activity could explain the discrepancy in the accrual of BMD, because there were virtually no other differences between the active soccer players and the controls (Table 2). The higher BMD in the active athletes at baseline and the further increased accrual of BMD during the follow‐up period compared with the controls could also be the result of differences in the inherited regulation.

This study also suggests that there is an association between retirement from exercise and an increased loss of BMD compared with controls. Thus far, few prospective studies have investigated the effect of reduced training on BMD. The existing studies indicate that, after reduction in activity, there is a decrease in BMD to pretraining levels. However, published studies predominantly include elderly subjects with low exercise‐induced BMD benefits or shorter follow‐up period.^(33–35,45) In this study, the long‐term retired athletes still had a 5.1 percentage point higher BMD in the leg at follow‐up after close to two decades of retirement in comparison with the controls, despite the reduction in the residual high BMD at most sites. Possible explanations for this finding may be that the athletes in this study had reached a much higher BMD during the period of high physical activity, that they had been active longer, that they had started exercising at an earlier age, and that they had exercised at a high level in comparison with the exercisers in previously cited studies. Another possible explanation could be that there is a faster BMD loss just after retirement but that after several years of retirement, this loss reach a lower level not different compared with controls, possibly suggested by data in the spine and the femoral neck (Table 3). However, because there still was a higher BMD loss in the trochanter in the long‐term retired athletes, this contradicts this hypothesis.

Furthermore, it must again be stated that bone strength is not only dependent on BMD; bone size and bone quality are important for skeletal resistance to trauma. In this study, it is known that bone volume is poorly estimated by the DXA method.⁽⁴⁶⁾ In addition, because the higher BMD loss in former athletes occurred at the endosteal surface (as virtually all BMD loss in long bones), a higher BMD loss in former athletes would probably lead to a skeleton that is larger but with a similar BMD compared with controls. This would inevitable lead to a skeleton with higher bone strength, because small changes in bone size are of greater importance for bone strength than small changes in BMD.⁽⁴⁷⁾ Finally, this study includes no evaluation of the bone quality or skeletal architecture, forcing us to draw conclusions as regard to residual benefits in bone strength in former athletes with great caution.

This study could not establish whether retired athletes continue to lose the exercise‐induced benefits after an even longer period of retirement, so there would be no residual higher BMD at the ages when fragility fractures rise exponentially. The obvious decrease in BMD benefits found in active athletes with increased duration of retirement makes us speculate that no benefits will remain in old age, even if this study could not exclude the possibility that a residual high BMD (and larger bone size and residual benefits in bone quality) will persist also in elderly ex‐athletes.

Once more, it seems unlikely that differences in lifestyle factors could explain the discrepancy in the development in BMD, because virtually no differences were found when the active soccer players were compared with the controls (Table 2). The differences in the BMD development when comparing former athletes and controls could theoretically also be caused by differences in the inherited regulation of BMD loss.

Bone size data are conflicting. Finding a larger bone size at baseline but not at follow‐up in the soccer players who retired during the study period could indicate a reversible exercise‐induced effect on bone size. However, this hypothesis is opposed by the findings at baseline and at follow‐up in both the soccer players who were active throughout the study and those who were retired throughout the study. A possible selection bias may explain the conflicting results in the bone size data. In addition, as previously discussed, we know that estimation of bone size from a DXA scan is questionable.⁽⁴⁶⁾

In summary, intense physical activity in the postmenarchal period is associated with a higher accrual of BMD than in controls, indicating additive BMD effects over several years into adulthood. Reduced activity level is associated with a reduction in the residual high BMD, but it cannot be concluded from this study that all benefits will be lost in the long‐term perspective, because there were still residual BMD benefits in the legs after two decades of retirement. In addition, we cannot state if residual benefits in bone size or bone quality remains in former athletes, benefits that possibly could lead to a reduced fracture incidence in older former athletes.

Acknowledgements

This study was supported by grants from the Swedish National Centre for Research in Sports (project 163/02), the Swedish Research Council, and the Region Skane Foundations.

Author notes

The authors have no conflict of interest.