Background-The cholesteryl ester transfer protein (CETP) plays a central role in reverse cholesterol transport. Currently, it remains unresolved whether circulating CETP is causally associated with coronary heart disease (CHD). We aimed to investigate this causal association using CETP gene rs708272 polymorphism as an instrument in a Mendelian randomization meta-analysis. Methods and Results-We searched PubMed and EMBASE before May 2014. Data and study quality were assessed in duplicate. Thirty-four articles (17 813 CHD patients and 22 203 controls) were qualified. Overall analyses revealed a significant association of rs708272-B1 allele with a reduced CHD risk compared with B2 allele under allelic (odds ratio and 95% confidence interval: 0.87 and 0.82-0.92; P<0.001), homozygous genotypic (0.74 and 0.66-0.83; P<0.001), and dominant (0.87 and 0.80-0.94; P<0.001) models. Carriers of rs708272-B1B1 genotype (weighted mean difference and 95% confidence interval: -0.21 and -0.41 to 0.00 mu g/dL; P=0.052) or B1 allele (-0.15 and -0.30 to 0.00 mu g/dL; P=0.056) had a marginally lower circulating CETP level compared with B2B2 genotype carriers. In Mendelian randomization analysis, there was a 25% (odds ratio and 95% confidence interval: 0.75 and 0.19-0.91) and a 17% (0.83 and 0.41-0.96) significantly reduced risk of CHD by a reduction of 0.2 mu g/mL in circulating CETP for the comparison of B1B1 genotype and B1 allele with B2B2 genotype, respectively. There were low probabilities of publication bias. Conclusions-Our findings demonstrate that the long-term genetically reduced circulating CETP might be causally associated with the low risk of CHD.