Association of VDR and Estrogen Receptor Genotypes with Bone Mass in Postmenopausal Caucasian Women: Different Conclusions with Different Analyses and the Implications
Hong-Wen Deng,Jian Li,Jin-Long Li,M. Johnson,G. Gong,R. R. Recker
DOI: https://doi.org/10.1007/s001980050177
1999-01-01
Osteoporosis International
Abstract:Much work has been done on the association between vitamin D receptor (VDR) genotypes and bone mineral density (BMD). Despite considerable effort, the results are inconsistent. While the VDR association remains unresolved, studies have expanded to other candidate genes (i.e., estrogen receptor (ER) genotypes), also yielding inconsistent results. A few studies have suggested that interaction effects between VDR and ER genotypes significantly affect BMD. We assessed associations of BMD with VDR BsmI genotypes, and ER XbaI and PvuII polymorphisms (denoted as ERX and ERP respectively) with spine, femoral neck, distal radius BMD, and with total body bone mineral content (tbBMC) in 108 US Mid-western postmenopausal Caucasian women. We statistically controlled for confounding factors such as height, weight, etc., in the analysis. No significant association was detected for ER genotypes with spine and radius BMD, or for VDR genotypes with femoral neck and radius BMD and tbBMC. No significant interaction between VDR and ER genotypes was detected in our sample. However, the VDR genotypes are significantly (p = 0.004) associated with *5.8% spine BMD variation. Both ERX and ERP genotypes are significantly (p = 0.02) associated with *3.5% femoral neck BMD variation. ERX genotypes are significantly (p = 0.03) associated with *2.4% tbBMC variation. However, if the data were analyzed by simple ANOVA as in some previous studies, without adjusting statistically for confounding factors, all the significant results we found here would have gone undetected. Our findings suggest that: (1) VDR and ER genotypes may have different effects on BMD at different sites and on tbBMC; and (2) if significant factors influencing bone are not appropriately controlled, true significant associations can easily be missed. These findings may offer a partial explanation for some of the earlier inconsistent results of association studies on BMD with VDR and ER genotypes.