We performed a GWAS of trochanter and intertrochanter bone mineral density (BMD) in the Framingham Heart Study and replicated in three independent studies. Our results identified one novel locus around the associated variations at chromosomal region 3q13.32 and replicated two loci at chromosomal regions 3p21 and 8q24. Our findings provide useful insights that enhance our understanding of bone development, osteoporosis, and fracture pathogenesis. Introduction Hip trochanter (TRO) and intertrochanter (INT) subregions have important clinical relevance to subtrochanteric and intertrochanteric fractures but have rarely been studied by genome-wide association studies (GWASs). Methods Aiming to identify genomic loci associated with BMD variation at TRO and INT regions, we performed a GWAS utilizing the Framingham Heart Study (FHS, N = 6, 912) as discovery sample and utilized the Women's Health Initiative (WHI) African-American subsample (N = 845), WHI Hispanic subsample (N = 446), and Omaha osteoporosis study (N = 971), for replication. Results Combining the evidence from both the discovery and the replication samples, we identified one novel locus around the associated variations at chromosomal region 3q13.32 (rs1949542, discovery p = 6.16 x 10(-8), replication p = 2.86 x 10(-4) for INT-BMD; discovery p = 1.35 x 10(-7), replication p = 4.16 x 10(-4) for TRO-BMD, closest gene RP11-384F7.1). We also replicated two loci at chromosomal regions 3p21 (rs148725943, discovery p = 6.61 x 10(-7), replication p = 5.22 x 10(-4) for TRO-BMD, closest gene CTNNB1) and 8q24 (rs7839059, discovery p = 2.28 x 10(-7), replication p = 1.55 x 10(-3) for TRO-BMD, closest gene TNFRSF11B) that were reported previously. We demonstrated that the effects at both 3q13.32 and 3p21 were specific to the TRO, but not to the femoral neck and spine. In contrast, the effect at 8q24 was common to all the sites. Conclusion Our findings provide useful insights that enhance our understanding of bone development, osteoporosis, and fracture pathogenesis.