C-reactive protein (CRP) is a commonly used inflammatory marker and elevated CRP levels are shown to increase the risk of coronary artery disease (CAD). Sequence variations in the CRP gene believed to influence the protein levels have been extensively investigated in CAD community. Most of the published studies, however, have reported mixed findings. The objective of the present study was to examine the associations of CRP variants (+942G>C, -717A>G, +1444C>T) with genetic risk of CAD by use of a meta-analysis.The human case-control studies were identified through online search, hand search, and contacting the authors of original articles. We performed both random-effect and fixed-effect meta-analysis to estimate CAD risk (odds ratios, OR). This analysis combined 16 studies in total. We found +942G>C was not associated with CAD risk when all data were pooled together, nor did we find a significant association in subgroup analyses. Meta-analysis of +1444C>T studies showed a similar trend. However, a borderline association with CAD risk was revealed for -717A>G (random-effect: OR = 0.53, 95% CI = 0.28-1.00 for the homozygous model; random-effect: OR = 0.51, 95% CI = 0.26-1.00 for the recessive model).These data suggest that the CRP gene variants examined may not modulate CAD risk.