Background: We aim to investigate whether the association between circulating interleukin 6 (IL-6) levels and the risk for coronary artery disease (CAD) is robust and perhaps even causal by a meta-analysis implementing mendelian randomization approach with IL-6 gene G-174C polymorphism as an instrument. Methods: Data were available from 19 articles encompassing 9417 CAD patients and 15982 controls. A random effects model was applied irrespectively of between-study heterogeneity, and publication bias was examined using a funnel plot and the corresponding statistics. Results: Overall, comparison of IL-6 gene alleles -174C with -174G had 4% increased risk for CAD (95% confidence interval [95% CI]: 0.97-1.10; P=0.285), accompanying marginal heterogeneity (I-2=38.3%; P=0.033). This association was potentiated in dominant model as odds ratio (OR) reached 1.08 (95% CI: 0.96-1.22; P=0.204) and heterogeneity was significant (I-2=58.4%; P<0.0005). Subgroup analysis by ethnicity indicated that carriers of -174C allele were associated with a 12% increased risk for CAD in prospective studies involving White populations (OR=1.12; 95% CI: 0.95-1.33; P=0.184), whereas the association in East Asians was remarkably reversed with 37-46% reduced risk. Relative to -174GG homozygotes, carriers of -174C allele had an overall 0.24 pg/ml high circulating IL-6 levels (P=0.047). The predicted OR for 1 pg/ml elevation in IL-6 levels was 1.60 (95% CI: 1.44-1.72; P<0.01) in prospective studies involving White populations. Publication biases were absent for all comparisons (P>0.1). Conclusion: Our findings provided strong evidence on the causal association of circulating IL-6 levels with the development of CAD in White populations. (C) 2011 Elsevier Ireland Ltd. All rights reserved.