Evaluation of confounding in epidemiologic studies assessing alcohol consumption on the risk of ischemic heart disease