Total and regional bone mineral and tissue composition in female adolescent athletes: comparison between volleyball players and swimmers

As life expectancy continues to rise, osteoporosis becomes an increasingly important healthcare concern due its economic impact and harmful effects on human health, especially among women [1]. Adult bone structure is largely determined during the two first decades of life; thus factors that stimulate bone formation during childhood and adolescence have a major role in the prevention of osteoporosis later in life [2]. Particularly, sex, ethnicity, hormones, alcohol consumption, tobacco, nutrition, and exercise are among the most significant contributing factors that can influence bone acquisition during early life [3].

During the period of growth, physical activity (mostly weight-bearing activities) is a particularly relevant factor for achieving an optimum peak bone mass level, due to the positive osteogenic response [4, 5]. In a cross-sectional study, Ginty et al. [6] demonstrated that high-impact loading activities (such as jogging, playing tennis, football, rugby, basketball, and exercising with weights) for 1 h or more a day was associated with greater size-adjusted whole body (+3.4%) and total hip (+8.5%) bone mineral content (BMC) among male adolescents, compared to those with a median of 7 min∙day of participation in high-impact loading activities. Although weight-bearing physical activity during childhood and adolescence has been widely recognized to be beneficial for bone health [7], previous studies have focused predominantly on assessing the combined effect of different high-impact loading sports (mixing participants from different sports into the same bone-loading category) [8, 9]. Thus, the osteogenic effect of specific sports is less well understood.

Muscle and bone are inextricably linked not only mechanically but also genetically and molecularly. Recent studies demonstrated a molecular “cross talk” between muscle and bone, as both tissues release endocrine, paracrine, and autocrine factors that may mediate intercellular communication [10]. Although non-impact loading sports, such as swimming, are widely recognized to have no substantial positive effect on bone health [11, 12], they stimulate muscle contraction in a variety of muscles that can induce hypertrophy and also may potentially stimulate the molecular “cross talk” between muscle and bone [13]. Thus, exploring the effect of sports with clearly different bone-loading mechanisms, such as volleyball and swimming, may help to better understand the interaction between skeletal muscle and bone during growth.

Sex differences in bone quality and strength are well described [14]. By late puberty, boys have higher bone strength than girls, which is due mainly to the larger bone size in boys than in girls [15]. During aging, men have a greater periosteal apposition that increases bone size and offsets bone loss more than in women, yielding fewer males than females at risk for fracture in old age [16]. Thus, the relationship between physical activity and bone health is particularly relevant among females compared to males because they are at increased risk for osteoporosis and fracture later in life. At present the most effective sport modality for bone health promotion in girls is unknown, largely because of the confounding effects of biological maturation on bone development during adolescence [17, 18].

The aim of the current study was to compare total and regional bone (mineral content and density) and soft-tissue composition (fat and lean mass) between female adolescent swimmers and volleyball players. As the positive impact of sports participation on bone mass can be tempered by nutritional factors (such as calcium, protein and total caloric intake), differences in dietary intake were controlled for. It was hypothesized that volleyball players would have higher whole-body and regional BMC and bone mineral density (BMD) than swimmers; no differences in fat mass and lean mass were expected between sports.

Current age, age at menarche, training history, anthropometry, and dietary intake of swimmers and volleyball players are presented in Table 1. Swimmers and volleyball players did not differ significantly in chronological age but differed moderately in age at menarche [d=0.65 (t=2.121, p<0.05)]; volleyball players experienced the first menstruation 0.72 years earlier than swimmers. Swimmers had significantly more years of training [d=1.69 (t=-3.836, p<0.01)] and annual number of training sessions (i.e., training volume) [d=6.30 (t=20.814, p<0.01)] compared to volleyball players. Athletes had mean statures and mean body masses that approximate, respectively, the 53th and 68th age-specific percentiles for U.S. girls [23]. Mean BMI-for-age exceeded the ≥50th percentile in both groups (swimmers: 60th percentile; volleyball players: 73rd percentiles) [23]. Volleyball players were heavier (body mass: d=-0.79 (t=-2.596, p<0.01), with more subcutaneous adipose tissue than swimmers (suprailiac, abdominal and anterior thigh skinfolds: d=-0.67 to -1.06 (t=-2.233 to -3.556, p=0.031 to p<0.01). Volleyball players also had higher levels of cholesterol intake [d=0.87 (t=-2.485, p<0.018)] than swimmers.

Table 2 and Fig. 1 comprises the descriptive statistics for DXA whole body and regional body composition of swimmers and volleyball players. Overall, swimmers (BMC: 2328±338 g) and volleyball players (BMC: 2656±470 g) exceeded in 2.1 and 2.8 standard deviation scores, respectively, the average of international standards for whole body BMC of healthy adolescents of the same race, gender, stature and body mass [28]. Volleyball players had significantly greater BMC and BMD in the whole body [+12.4%: d=-0.80 (t=-2.637, p<0.05) and +5.5%: d=-0.80 (t=-2.574, p<0.05), respectively], subhead [+13.8%: d=-0.83 (t=-2.724, p<0.01) and +6.5%: d=-0.89 (t=-2.893, p<0.01), respectively], trunk [+15.1%: d=-0.76 (t=-2.491, p<0.05) and +7.6%: d=-0.98 (t=-3.233, p<0.01, respectively] and lower limbs [15.8%: d=-1.06 (t=-3.476, p<0.01) and +6.4%: d=-0.71 (t=-2.344, p<0.05), respectively] than swimmers. In addition, volleyball players had significantly greater lean soft tissue in the lower limbs [+5.9%: d=-0.72 (t=-2.217, p<0.05)], whole body [+29.4%: d=-0.87 (t=-2.853, p<0.01)], and trunk fat tissue [+28.2%: d=-0.73 (t=-2.456, p<0.05)].

Table 3 summarizes the interrelationship between age, training experience, and body size descriptors with variables extracted from DXA, by type of sport. In swimmers, BMC, BMD (whole body, subhead, trunk, and upper limbs) and fat tissue (whole body, trunk, and lower limbs) are moderately to largely correlated with CA (r=+0.46 to +0.59, p<0.05). No significant associations were noted within the same variables for volleyball players. With one exception – lower limbs lean soft tissue in volleyball players (r=+0.40, p<0.05) – no correlations were found between age at menarche and total and regional bone mineral and tissue composition. Training experience in swimmers is positively related to the upper limbs BMC (r=+0.49, p<0.05) and fat tissue (r=+0.49, p<0.05), and lower limbs fat tissue (r=+0.52, p<0.05). In volleyball players, years of training only correlated significantly with lower limbs BMD (r=+0.43, p<0.05). Correlations between stature and BMC in swimmers are large to very large (r=+0.64 to +0.82, p<0.01) and are generally higher than those for BMD (r=+0.57 to +0.66, p<0.01) and lean soft tissue (whole body: r=+0.52, p<0.05; lower limbs: r=+0.70, p<0.01). The BMC [with one exception: lower limbs (r=+0.46, p<0.05)] and BMD of volleyball players are not significantly related to stature. In contrast to bone mineral parameters, stature is largely related to lean soft tissue (whole body: r=+0.63, p<0.01; trunk: r=+0.64, p<0.01; lower limbs: r=+0.61, p<0.01). Although the correlation data between body mass and variables extracted from DXA for swimmers and volleyball players are somewhat different, they reflect moderate to very large [BMC, BMD, lean soft tissue (whole body, upper and lower limbs) and fat tissue (whole body and trunk)] and moderate to nearly perfect [BMC, BMD (whole body, subhead and trunk), lean soft tissue (whole body, upper and lower limbs), and fat tissue] positive associations (p<0.05), respectively.

Sport-related variation for total and regional bone and soft-tissue composition, when chronological age, age at menarche, training experience, stature and body mass were statistically controlled by MANCOVA is presented in Table 4. Moderate differences (ES-r=0.32) between swimmers and volleyball players persisted for BMD at the lower limbs (0.098 g∙cm^-2, 7.8%).

This study showed that volleyball players presented greater BMC and BMD, in the whole body, and greater lean soft tissue in the lower limbs with respect to swimmers. Although swimming stimulates muscle hypertrophy [29], our results showed that swimmers had moderately less skinfold thickness (a measure of subcutaneous fat), whole body and trunk fat tissue. Differences between groups in lean mass were mostly small. Thus, the mechanical and non-mechanical stimuli associated with swimming may not be sufficient to trigger the responsiveness of bone cells. After adjustments for potential confounders (i.e., chronological age, age at menarche, years of training, stature, and body mass) the bone content, lean soft tissue and fat tissue were similar between groups. Differences persisted for the lower limbs, with volleyball players presenting higher BMD compared to swimmers.

A secular decline in the age at menarche occurred in the Portuguese population [30]. Findings suggest that the decline is associated, to a large extent, with a reduction in the number of girls who mature late [25]. Allowing for normal variability, training is not related to age at menarche in athletes [31] and variation in mean ages at menarche within a sport is especially evident in swimmers and volleyball players [25]. Mean ages at menarche were 13.31±1.33 years and 12.59±0.95 years for swimmers and volleyball players, respectively. Volleyball players approximates the mean age of menarche for Portuguese school girls calculated using recall methods (i.e., 12.53±1.27 years) [30]. Although there is a difference of 0.72 years between groups’ means, about two-thirds of the present sample attains menarche between 12.0 and 14.0 years. The limited variability may explain the reduced interrelationships between age at menarche with variables extracted from DXA (see Table 3), suggesting that they are somewhat independent.

The current sample has a mean stature and mean body mass which approximate, respectively, the 53th and 68th age-specific percentiles for U.S. girls [23]. The trend for elevated mass-for-stature likely reflects the advanced maturity status of the athletes [31]. Mean BMI-for-age exceeded the ≥50th percentile in both groups (swimmers: 60th percentile; volleyball players: 73rd percentiles) [23]. This was consistent with observations for other samples of youth swimmers and volleyball players [25, 32]. In the current study, dietary data suggested differences only for cholesterol intake with higher amounts being consumed by volleyball players compared to swimmers. Volleyball players were significantly heavier (+6 kg) but not taller than swimmers. Differences between groups in the components assessed by whole-body DXA were moderate for mineral content (328 g) and fat tissue (5200 g) and trivial for lean soft tissue (100 g). Allowing for the limitation of the comparison, swimmers and volleyball players exceeded in 2.1 and 2.8 standard deviation scores, respectively, the average of international standards for whole body BMC of healthy adolescents [28]. This suggests greater BMC of female adolescent swimmers and volleyball players compared to healthy female adolescents, likely with positive effects on bone health later in life.

Reduced lean mass constitutes one of the most relevant determinants of risk for low BMD in female adolescent runners [33], while lean mass is related to BMD gains and bone geometry changes in female soccer players [18]. As previously described, part of the osteogenic effect attributed to sport participation may be related to the increase in muscle mass, and subsequent effect on bone cells [12, 34]. Correlations between primary and secondary outcomes in swimmers and volleyball players are summarized in Table 3. Results indicate the complexities involved in attempting to partition out the contribution of age, training and body size, which are often overlooked in comparisons of bone and soft tissue between athletes of different sports.

Bone is a component of body composition that is a focus of attention specifically in the context of preventing osteoporosis later in life [1, 2]. In general, the more mineral accumulated in the skeleton during growth and maturation, the better off the individual will be several decades later when mineral content of the skeleton begins to decline [2, 21]. Evidence from cross-sectional studies suggested that the peak of bone mass acquisition is reached during adolescent years and significantly affects the BMC observed in adulthood [2, 12, 34]. It has been noted that active adolescent males had 8-10% more adjusted BMC at the total body, total hip and femoral neck (p<0.05) in young adulthood and active adolescent females had 9-10% more adjusted BMC at the total hip and femoral neck [21]. In the present study, years of training were positively related to the upper limbs BMC (r=+0.49, p<0.05) in swimmers and with lower limbs BMD (r=+0.43, p<0.05) in volleyball players. All together, this would suggest that the adoption of routines of physical activity including exercise and sport participation during adolescence may itself mediate enhanced skeleton formation, and these benefits are maintained into young adulthood, which may help prevent musculoskeletal diseases, such as osteoporosis, during old age.

The lower values of BMD in swimmers when compared to other sports have been investigated over the last years [29]. Although the mechanisms are not entirely clear, recent evidence point to several possible explanations. It has been theorized that muscle forces produced during sports such as swimming and cycling may not exceed the minimum effective strain stimulus threshold to induce an osteogenic effect [35, 36]. The most relevant aspect of the non-significant effect of swimming on bone gains is due its movement in a “hypogravity” environment [12, 33] for large amount of time per week [18]. The extensive time spent in the water by swimmers may also limit the time they have available to perform other sports, including weight-bearing activities during the day.

The apparent benefits of high-impact loading sport (i.e., volleyball) on BMC and BMD in the present study seemed to be specific to the trunk and lower limbs while differences between volleyball players and swimmers for bone mineral parameters in the upper limbs were trivial. Mechanical loading leads to bone remodelling, adapting the bone structure in response to the mechanical demands [12]. The stress generated by physical exercise on bone stimulates the collagen alignment in the sites directly affected by the activity, leading to higher bone strength [34]. The high quantity of jumps required during a volleyball practice may explain, at least in part, the differences observed for BMD at the lower limbs compared to swimmers. Ferry et al. [18] noted the same effect on lower limbs among girls engaged in soccer practice, in which jumps, kicks, and sprints are commonly performed. In addition, bone cells become desensitized to prolonged mechanical stimulation [37], thus incorporating periods of rest between short vigorous skeletal loading sessions may be a valid strategy to promote osteogenesis. Volleyball practice is characterized by intermittent movement (acceleration and jumps), which may also explain the higher bone mass observed in the female volleyball players compared to swimmers.

A major strength of this study was the inclusion of under-researched late-adolescent female athletes (i.e., volleyball players and swimmers). A further strength of the study was the use of DXA which is considered the safest and most appropriate imaging modality to access body composition and bone status. However, the current investigation is also not without limitations. First, the cross-sectional design prevents comment on causation. The sample size is fairly small, and we were unable to control for participation in other sports, or factors known to affect bone mineral density. Second, no regional site scans at the lumbar spine and proximal femur were performed, and no accurate geometrical properties were captured by DXA. Further research might attempt to control for the influences of biological maturity when seeking to determine the effects of specific physical activities on bone health among young female athletes. This will help inform programs and strategies to enhance bone health during the adolescent period of growth and development, leading to prevention of osteoporosis in later years.

In conclusion, volleyball players had greater BMD at the lower limbs when compared to swimmers. The results support the fact that 3 to 4 times per week of high impact loading activities are associated with higher bone mineral density at regions of interest. The observed skeletal benefits may also translate in positive changes in bone geometry and quality, this providing a substantial increase in bone strength. This observational study provides practical implications for inactive young individuals, young athletes of non-impact loading sports (including swimming) and coaches who can benefit from complementing training routines with osteogenic weight-bearing activities [29], such as resistance, strength and plyometric training.