Bone mineral density in diagnosis of osteoporosis: reference population, definition of peak bone mass, and measured site determine prevalence.

A population-based study was performed in order to compare different definitions of peak bone mass, and to apply the corresponding T-scores for different skeletal sites to a cohort of 70-yr-old women for studying the prevalence of osteoporosis. Bone mineral density (BMD) of the hip, lumbar spine, and forearm was measured by dual X-ray absorptiometry (Hologic 4500) in 296 women ages 16-31 yr and 210 women age 70 yr. Peak bone mass occurred in women in their early 20s at the proximal femur and at 28 and 31 yr at the spine and forearm, respectively. BMD cutoff levels were compared to machine-specific cutoff values for the different sites. When applied to our cohort of 70-yr-old women, the prevalence of osteoporosis at the total hip was 9-25%, depending on which peak bone mass the T-score of -2.5 was based. The prevalence in the spine was 28-33% and in the forearm 45-67%. Osteoporosis in at least one of the three measured sites was documented in 49-72% of the population sample. Our results show that the use of T-score to define osteoporosis results in a highly different prevalence rate in a given population depending on the reference population and the skeletal sites chosen for measurement.