OBJECTIVE: To investigate change in bone mineral density (BMD) in premenopausal women (age, 29-46 years), some of whom were marathon runners with a history of menstrual irregularity. DESIGN: Longitudinal follow-up. SETTING: University medical school. PARTICIPANTS: We investigated 8 sedentary controls (SC) and 19 marathon runners (12 with regular menses (R) and 7 with a history of irregularity (OA) 11.7 +/- 7.9 years before follow-up). MAIN OUTCOME MEASURES: BMD (g/cm2) of lumbar spine (LS) and proximal femur were determined at baseline and follow-up (3-5 years later). We calculated a menstrual history index (MHI) (estimated periods/year since age 13). RESULTS: Body mass, age at menarche, and femoral BMD were not statistically different. Follow-up LS BMD (g/cm2) was lower (p < 0.01) in OA (0.936 +/- 0.060) than in R (1.043 +/- 0.103) and SC (1.094 +/- 0.077), even when covarying for age or both age and mass. No group changed BMD significantly with time. Current MHI was lower (p < 0.001) in OA (9.7 +/- 1.4) than in R (11.3 +/- 0.5) and SC (11.8 +/- 0.4). MHI for the teenage years was lower in OA than in SC but not in R. OA had significantly lower MHI than did R and SC for the third and fourth decades. Only MHI during the third decade correlated significantly with LS BMD for all subjects. CONCLUSIONS: Restoration of LS BMD deficit in women with prior menstrual irregularity aged over 30 is slow and may never reach the same level as age-related controls; secondly, this may be the result of both bone loss in the third decade of life and reduced acquisition during adolescence.
OBJECTIVE: To investigate change in bone mineral density (BMD) in premenopausal women (age, 29-46 years), some of whom were marathon runners with a history of menstrual irregularity. DESIGN: Longitudinal follow-up. SETTING: University medical school. PARTICIPANTS: We investigated 8 sedentary controls (SC) and 19 marathon runners (12 with regular menses (R) and 7 with a history of irregularity (OA) 11.7 +/- 7.9 years before follow-up). MAIN OUTCOME MEASURES: BMD (g/cm2) of lumbar spine (LS) and proximal femur were determined at baseline and follow-up (3-5 years later). We calculated a menstrual history index (MHI) (estimated periods/year since age 13). RESULTS: Body mass, age at menarche, and femoral BMD were not statistically different. Follow-up LS BMD (g/cm2) was lower (p < 0.01) in OA (0.936 +/- 0.060) than in R (1.043 +/- 0.103) and SC (1.094 +/- 0.077), even when covarying for age or both age and mass. No group changed BMD significantly with time. Current MHI was lower (p < 0.001) in OA (9.7 +/- 1.4) than in R (11.3 +/- 0.5) and SC (11.8 +/- 0.4). MHI for the teenage years was lower in OA than in SC but not in R. OA had significantly lower MHI than did R and SC for the third and fourth decades. Only MHI during the third decade correlated significantly with LS BMD for all subjects. CONCLUSIONS: Restoration of LS BMD deficit in women with prior menstrual irregularity aged over 30 is slow and may never reach the same level as age-related controls; secondly, this may be the result of both bone loss in the third decade of life and reduced acquisition during adolescence.