The present study sought to resolve the contradictory evidence as to whether the G-->A substitution at position -75 of the apoA-I gene promoter raises HDL cholesterol (HDL-C) levels by examining the effect of this polymorphism in French Canadians, a relatively genetically homogeneous population. Among 308 women, carriers of the A allele displayed 12% and 10% higher mean plasma HDL-C and apoA-I concentrations, respectively, than did noncarriers. Among 345 men, no effect of the A allele was noted. The frequency distribution of HDL-C levels in women carrying the A but not the G allele appeared bimodal, with one peak corresponding to the mean of the noncarriers and a second to higher HDL-C. Thus it appears that only a subset of A alleles confers high HDL-C levels. This hypothesis was supported by data from four kindreds within which some but not all A alleles segregated with hyperalphalipoproteinemia. The data suggest that the A substitution in the apoA-I gene promoter does not directly confer high HDL-C levels but may be in linkage disequilibrium with other sequence polymorphism(s) at this locus in a subset of alleles that raise HDL-C levels.