Bone mineral density (BMD) is under strong genetic control. Recent work has suggested that a polymorphism affecting an Sp1 binding site in the collagen I (COLI) A1 gene is associated with BMD and vertebral fracture in postmenopausal women. We analyzed this polymorphism in relation to BMD and bone turnover in 220 healthy premenopausal women aged 31-57 years. There were 61% SS homozygotes, 35% Ss heterozygotes, and 4% ss homozygotes, genotype frequencies similar to those previously reported in other Caucasian populations. Women in the three genotype groups were matched for age, body weight, physical activity, smoking habits, and oral contraceptive use, but height was greatest in the SS group and lowest in the ss group (p = 0.03). Between-group comparisons by analysis of variance (ANOVA) showed that COLI A1 genotype was significantly associated with spine BMD (p = 0.05), total body BMD (p = 0.046), and total body bone mineral content (BMC) (p = 0.02), but the differences between extreme genotypes were small (4, 5, and 10%, for spine BMD, total body BMD, and total body BMC, respectively). After adjustment for height, the differences between genotypes decreased and were no longer significant by ANOVA (p = 0.08, 0.17, and 0.33 for spine BMD, total body BMD, and total body BMC). Furthermore, no significant difference between genotypes was observed for femoral neck, trochanter, Ward's triangle, or forearm BMD. COLI A1 genotype was associated with serum C-terminal extension propeptide of type I collagen (p = 0.04), with lowest levels in ss individuals, but not with any other marker of bone formation (osteocalcin, alkaline phosphatase, and type I collagen N-terminal extension propeptide) or bone resorption (urinary excretion of type I collagen C and N telopeptide breakdown products). The COLI A1 Sp1 polymorphism is associated with height, peak total body BMD and BMC, and spine BMD. The genotype-specific differences account for only a small proportion of variance in BMD at these sites and are not significant after adjustment for height, suggesting that part of the effect on bone mass may be due to differences in body size. Our data support the view that COLI A1 may be a candidate gene for regulation of bone mass, but our results must be treated with caution, in view of the small number of ss individuals, and will require confirmation in larger studies.