Abstract Objective: To establish the association between smoking and cognitive decline in patients undergoing coronary artery bypass graft (CABG) surgery. Design: Retrospective review. Setting: Referral center for cardiothoracic surgery at a university hospital. Participants: Four hundred seventeen patients undergoing CABG surgery. Interventions: Based on preoperative data, patients were divided into 2 groups: smokers (n = 185) and nonsmokers (n = 232). Patients who smoked half a pack of cigarettes per day within the last 2 years were identified as smokers, and patients who did not smoke were included in the nonsmoker group. Patients with less than a seventh grade education; an inability to read; or a history of one of the following medical conditions: prior stroke with residual deficit, psychiatric illness, renal disease (creatinine > 2.0 mg/dL), or active liver disease; or patients who quit smoking prior to surgery were excluded from the study. Both groups received similar anesthetic and surgical management. All patients received a battery of neurocognitive tests both preoperatively and 6 weeks after CABG surgery. Neurocognitive test scores were separated into 4 cognitive domains, with a composite cognitive index (the mean of the four domain scores) determined for each patient at every testing period. Measurements and main results: The overall rate of cognitive decline at 6 weeks after surgery in smokers was 36.2%, whereas nonsmokers showed a deficit rate of 36.6%. Nonsmokers were significantly older and presented for surgery on average 6 years later than the smokers. Female sex represented a considerably larger proportion of patients in the nonsmoker group. Smokers had a higher prevalence of myocardial infarction. The univariate analysis of cognitive change at 6 weeks adjusted for age, baseline cognitive index, and education years showed no difference between the 2 groups. Sex, history of myocardial infarction, hypertension, stroke, transient ischemic attack, and duration of cardiopulmonary bypass did not contribute to the multivariate logistic regression model and were dropped from the final analysis. Significant multivariate predictors of neurocognitive dysfunction included age, left ventricular ejection fraction, baseline education level, and baseline cognitive index. Conclusions: This study confirmed previous findings that age, baseline cognitive function, years of education, and impaired left ventricular function are independent predictors of neurocognitive decline at 6 weeks after CABG surgery. Smoking is neither preventive nor causative of cognitive decline after CABG surgery.