Anzeige
Erschienen in:
01.01.2008 | Original Article
Effects of the sample size of reference population on determining BMD reference curve and peak BMD and diagnosing osteoporosis
Erschienen in: Osteoporosis International | Ausgabe 1/2008
Einloggen, um Zugang zu erhaltenAbstract
Summary
Establishing reference databases generally requires a large sample size to achieve reliable results. Our study revealed that the varying sample size from hundreds to thousands of individuals has no decisive effect on the bone mineral density (BMD) reference curve, peak BMD, and diagnosing osteoporosis. It provides a reference point for determining the sample size while establishing local BMD reference databases.
Introduction
This study attempts to determine a suitable sample size for establishing bone mineral density (BMD) reference databases in a local laboratory.
Methods
The total reference population consisted of 3,662 Chinese females aged 6–85 years. BMDs were measured with a dual-energy X-ray absorptiometry densitometer. The subjects were randomly divided into four different sample groups, that is, total number (Tn) = 3,662, 1/2n = 1,831, 1/4n = 916, and 1/8n = 458. We used the best regression model to determine BMD reference curve and peak BMD.
Results
There was no significant difference in the full curves between the four sample groups at each skeletal site, although some discrepancy at the end of the curves was observed at the spine. Peak BMDs were very similar in the four sample groups. According to the Chinese diagnostic criteria (BMD >25% below the peak BMD as osteoporosis), no difference was observed in the osteoporosis detection rate using the reference values determined by the four different sample groups.
Conclusions
Varying the sample size from hundreds to thousands has no decisive effect on establishing BMD reference curve and determining peak BMD. It should be practical for determining the reference population while establishing local BMD databases.