Invited commentary: should arteriographic case-control studies be used to identify causes of atherosclerotic coronary artery disease?

Arteriographic case-control studies still have a role in defining associations and dose-response relations between newly purported risk factors and atherosclerotic coronary artery disease. They remain one of the better ways to sort out whether a risk factor acts to cause clinical coronary disease via effects on coronary atherosclerosis, infarction independent of atherosclerosis, or both. It remains unclear if the selection processes for coronary arteriography infuse bias into the selection of cases, though the report by Reed and Yano suggests it does not. Control group selection remains problematic, as it does for all case-control studies, and considerable effort should be exercised to avoid the many potential biases in selection of both cases and controls. In this latter statement there is agreement with authors Reed and Yano. However, the avoidance of arteriographic studies threatens to curb the initiation of that sequence of studies that can, in a relatively short duration of time, provide useful information on the potential for a risk factor intervention to inhibit coronary atherosclerosis and, possibly, clinical coronary events. This, then, raises the final point for emphasis. Arteriographic studies, be they case-control, prospective, or experimental, may be most powerful when paired with those using clinical endpoints. In addition to the chance for replication of findings with another set of endpoints, the issue of mechanism of action via the atherosclerotic process can be addressed. Interestingly, the Honolulu Heart Program, with its laudable efforts to collect both autopsy and arteriographic quantitations of atherosclerosis, provides perhaps the best illustration in the current literature of the power of using multiple endpoints for coronary artery disease to more completely elucidate the role of risk factors in the natural history of that disease.