The aim of this study was to examine whether the well-established effect of the common TaqIB polymorphism in intron 1 of the gene for cholesterol ester transfer protein (CETP) on high density lipoprotein cholesterol (HDL-C) concentration and increased risk of myocardial infarction (MI), could be explained by the recently identified -629C>A functional polymorphism in the promoter. Non-fatal MI cases (388 male) and a control group of 794 healthy men were recruited from the 30 year long prospective Reykjavik Study. In the healthy men the frequency of the TaqIB B2 allele was 0.47 (95% CI: 0.44-0.50) and there was a strong allelic association with the -629A allele (D=-0.21, P<0.0001), which had a frequency of 0.52 (95% CI: 0.49-0.56). B2B2 homozygotes displayed 15% higher HDL-C levels than subjects homozygous for the B1 allele (P<0.0001). Homozygotes for the -629A allele displayed 14% higher HDL-C concentrations than subjects homozygous for the -629C allele (P<0.0001). The frequencies of the alleles associated with lower HDL-C were significantly higher in cases compared with controls, 0.59 versus 0.53 (TaqIB B1) and 0.52 versus 0.48 (-629 C) respectively (P<0.05 for both). There was a significantly higher risk for MI in B1B1 homozygotes (OR=1.44, 95% CI: 1.10-1.87, P<0.01), compared to the other genotypes combined. This was not observed for the CC homozygotes (OR=1.16, 95% CI: 0.87-1.54). In addition, homozygotes for the TaqI B2 allele experienced a first MI 2 years later than men with other genotypes, 59 versus 61 years (P<0.05). This effect was not seen for the promoter polymorphism. These results strongly confirm the role of the CETP gene and the TaqIB variant as a risk factor for MI and suggest that another functional polymorphism is yet to be discovered in the CETP gene, that will explain the effect on MI associated with TaqIB observed in this study.