This meta-analysis of tea consumption in relation to stroke, myocardial infarction, and all coronary heart disease is based on 10 cohort studies and seven case-control studies. The study-specific effect estimates for stroke and coronary heart disease were too heterogeneous to be summarized (homogeneity p < 0.02 for stroke, p < 0.001 for coronary heart disease). Only the relative risk estimates for myocardial infarction (seven studies) appeared reasonably homogeneous (homogeneity p = 0.20). The incidence rate of myocardial infarction is estimated to decrease by 11% with an increase in tea consumption of 3 cups per day (fixed-effects relative risk estimate = 0.89, 95% confidence interval: 0.79, 1.01) (1 cup = 237 ml). However, evidence of bias toward preferential publication of smaller studies that suggest protective effects urges caution in interpreting this result. The geographic region where the studies were conducted appeared to explain much of the heterogeneity among coronary heart disease, myocardial infarction, and probably stroke results. With increasing tea consumption, the risk increased for coronary heart disease in the United Kingdom and for stroke in Australia, whereas the risk decreased in other regions, particularly in continental Europe.