Effect of physical activity on bone mineral accretion in adolescent females: A four-year longitudinal study

The weight-bearing of physical activity (PA) is often promoted to increase bone health. One goal of this study was to design a five-year retrospective method of measuring PA across a spectrum of normal activities appropriate for use in an ongoing longitudinal study of bone mineral accretion in adolescent females. Additionally, we modeled the contribution of PA in bone mineral accretion over four years comparing time since menarche and age models. The relationship of athletic menstrual disturbances to bone mineral accretion and the potential role of leptin in predicting these menstrual disturbances were also evaluated.;The modified questionnaire was sent to 314 subjects who had completed the first four years of the study (1992--1996). Test-retest studies were completed on 26 of the 267 respondents to document reliability. Questionnaire Spearman coefficients ranged from r = 0.587 to r = 0.928 with only the 1992--93 values below r = 0.84. Questionnaire validity was measured by correlation with biannual percent body fat and METs/kg/day. Each reported activity was assigned a weight-bearing value using the published body weight impacts of a few activities and all reported activities categorized. The annual and mean osteogenic indicators (Ost) were derived using these values. Annual and average Ost consistently appeared as significant predictors of bone density. The skeletal benefits gained from participating in rigorous sporting events were not linear as Ost was best modeled as a quadratic variable in the total body models. We predicted the additional annual bone mineral accretion for subjects participating in 300, 8000 and 13000 annual Ost to be 1, 26.3 and 42.8 grams of bone mineral, respectively. This conferred a 10% advantage for the 13000 Ost subject over inactive subjects considering normal accretion (40--400 grams). We did not identify bone mineral accretion disadvantages in athletes with menstrual irregularities. Leptin was less predictive than percent body fat when athletic menstrual disruption was modeled with logistic regression.