SHBG transports sex steroid hormones in the blood, and levels in humans are thought to partially be genetically determined. Recently, studies have found a pentanucleotide (TAAAA)n repeat polymorphism in the promoter of the SHBG gene and a missense polymorphism in exon 6 (Asp327Asn) to predict circulating SHBG levels. Based on the potential role of common genetic variation in SHBG to serve as a marker of SHBG levels in the general population, we evaluated the association between the (TAAAA)n repeat polymorphism, Asp327Asn polymorphism, and SHBG levels in a population of African-American, Native Hawaiian, Japanese, Latina, and white healthy postmenopausal women from the Multiethnic Cohort Study (n = 372). Mean SHBG levels were not significantly different between carriers and noncarriers of the Asn327 allele [minor allele frequency range across ethnic groups, 0.02-0.14; Asp/Asn and Asn/Asn genotypes, 33.6 mol/liter; 95% confidence interval (CI), 28.2-40.0; n = 49; Asp/Asp genotype, 30.8 mol/liter (95% CI, 28.7-33.1; n = 296); P = 0.37]. For the repeat polymorphism, we observed six different SHBG repeat alleles segregating in the population (TAAAA6-11), and the distribution of these alleles varied widely across populations. We found suggestive evidence of linkage disequilibrium between the Asn327 allele and the eight-repeat allele in all populations except African-Americans (P > or = 0.08). In analysis of the repeat polymorphism, SHBG levels among carriers of two short alleles (seven or fewer repeats; 31.2 nmol/liter; 95% CI, 27.3-35.6; n = 82) were not statistically different from those of carriers of two long alleles (more than seven repeats; 32.7 nmol/liter; 95% CI, 29.4-36.3; n = 124; P = 0.59). We did, however, observe individual genotypic classes (n = 16) to contribute modestly to the overall prediction of SHBG levels (by analysis of covariance, P = 0.03). Carriers of the six-repeat allele (27.9 nmol/liter; 95% CI, 25.2-30.8; n = 147) were found to have nominally significantly lower SHBG levels than noncarriers (32.4 nmol/liter; 95% CI, 29.7-35.2; n = 202; P = 0.03). This effect was stronger among the subset of women who also carried the Asn327 allele (interaction, P = 0.006). In summary, these results suggest that genetic variation at the SHBG locus may contribute to modest differences in SHBG levels among healthy postmenopausal women, and that much larger studies will be needed to better comprehend the effects of common variations at this locus in predicting circulating SHBG levels.