We tested the effectiveness of an individually delivered behavioral multicomponent smoking intervention (SI) against offering advice only (AO) to 267 patients after coronary arteriography. After 6 months, 51% of AO patients and 62% of SI patients reported abstinence. Validated rates were 34% and 45% for AO and SI patients, respectively. Logistic regression analyses, controlling for severity of illness, stage of change, and self-efficacy, among other variables, showed that, at 6 months, the SI had the most effect for patients with more severe coronary artery disease (CAD) who had been admitted with a myocardial infarction (95% confidence interval = 2.05, 124.85). At 12 months, only severity of disease mediated SI effects (95% confidence interval = 3.10, 58.00). Similar results were seen for cotinine-validated cessation. This study confirms the effectiveness of individually administered SI for more seriously ill patients with CAD and raises questions as to how to better intervene with those individuals with less severe disease.