Catechol-O-methyltransferase (COMT) has been suggested to be involved in the pathogenesis and pharmacological treatment of affective disorders. The nonsynonymous single nucleotide polymorphism (SNP) in exon 4 (Val108/158Met; rs4680) influences the COMT enzyme activity. Inconsistent results were found between Val158Met polymorphism (rs4680) and treatment response phenotypes in genetic association studies. However, the haplotype combinations of alleles at the Val108/158Met SNP with the other synonymous SNPs in the COMT gene region have shown association between enzyme activity/amount and COMT-dependent phenotypes. We carried out this study to define the functional impact of COMT genotypes/haplotypes on susceptibility and on treatment response phenotypes of major depressive disorder (MDD). Three hundred and ninety-six patients with MDD diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition [(DSM)-IV] and 295 healthy controls were recruited for this study and genotyped for the seven COMT SNPs (rs2075507, rs737865, rs6269, rs4633, rs4818, rs4680, and rs165599). This is the first study with all these SNPs to investigate for MDD and treatment response phenotypes. Our results show that none of the seven SNPs, including the rs4680, was significantly associated with MDD after permutation correction in single SNP analyses. Although several haplotype combinations showed significance, the combinations of G-T-G-G haplotype for rs6269, rs4633, rs4818 and rs4680 were only present in the MDD group (G-T 4.5%, corrected sim P=0.0001; G-T-G 3.87%, corrected sim P=0.001; G-T-G-G 3.3% corrected sim P=0.0025). In the treatment response phenotypes, the GG genotype of the rs2075507 SNP (located in the promoter region of MB-COMT) was less common in resistant patients in a single SNP analysis with low corrected sim P=0.052 and power=0.086. However, in the haplotype analysis, the haplotypes of exonic SNPs, rs4633, rs4818, and rs4680, were related to the treatment response phenotypes investigated, especially the phenotype of the response to antidepressant treatment. The C-C-A haplotype of these SNPs was overrepresented (almost four-and eight-fold) in the responders compared with the nonresponders and controls, respectively, after Bonferroni correction (corrected sim P=0.048, 0.0001, respectively). Both nonsynonymous and synonymous SNPs within haplotypes may be more relevant than the single SNP in conferring MDD susceptibility and treatment response phenotypes. Despite the limited power of our analysis, this finding suggests that the polymorphic COMT gene that influences catecholaminergic neurotransmission may play a role in the individual response to antidepressants.