Peroxisome proliferator-activated receptors (PPARs) alpha, delta and gamma are nuclear transcription factors that regulate fatty acid biosynthesis. Our objectives were to determine the effects of PPAR haplotypes on biochemical, angiographic, clinical phenotypes and their responses to treatment with fluvastatin. We genotyped 372 Lipoprotein and Coronary Atherosclerosis Study subjects for seven single nucleotide polymorphisms (SNPs) in PPARalpha (-35 089A>C, 484C>G), delta (-4401C>T, 294T>C) and gamma (34C>G, 25 506C>T, 161C>T) by restriction mapping or 5' exonuclease assay. We reconstructed and estimated haplotypes frequencies using four algorithms. Linkage disequilibrium (LD) was calculated by D' and haplotype effects by permutation and regression analyses. The PPARD and PPARG SNPs were in LD. The baseline plasma triglyceride levels and their responses to treatment with fluvastatin were associated with PPARD haplotypes (P = 0.01). Triglyceride levels were lowest and highest in homozygotes with diplotypes 3 and 4 (130.1 +/- 40.8 and 194.2 +/- 44.6 mg/dl, P < 0.001), respectively. PPARD haplotype 3 was also an independent determinant of plasma apolipoprotein (apo)B (P = 0.021) and apoC-III (P = 0.001) levels, mean number of coronary lesions (P = 0.046) and changes in triglyceride (P = 0.01) and apoC-III (P = 0.047) levels in response to fluvastatin. Plasma triglyceride levels (P = 0.044), the mean number of coronary lesions (P = 0.026) and changes in minimum lumen diameter in response to fluvastatin (P = 0.022) were also associated with PPARG haplotypes. No significant associations between PPARA haplotypes and the phenotypes or significant interactions between PPAR haplotypes and the occurrence of new clinical events were detected. PPARD and PPARG haplotypes are independent determinants of plasma levels of lipids, severity of coronary atherosclerosis and its response to therapy.