The proper approach to assessing the impact of the fact that e-cigarettes were not available before 2007.

Abstract

In a cross-sectional study, one computes the association between exposures (in this case e-cigarette and cigarette use) and an outcome (in this case having had an acute myocardial infarction) among data observed at a fixed point in time. In this case the specific issue is whether the myocardial infarctions may have occurred before e-cigarette use. Rodu & Plurphanswat provided an alternative analysis of the Population Assessment of Tobacco and Health (PATH) data set to one we published, that found an association between having had a myocardial infarction and e-cigarette use. Their analysis selectively changes data for e-cigarette users who started using e-cigarettes before their myocardial infarction, which is inappropriate in a cross-sectional study. Their inappropriate analysis led to the biologically implausible result that non-daily e-cigarette use was protective against myocardial infarction. A standard cross-sectional analysis that treats all exposure groups the same is the appropriate way to do such studies.