BACKGROUND: Most patients with acute coronary syndrome quit smoking when hospitalized, although several have been found to relapse and resume smoking within 3 months. AIM: This paper reports a study to identify factors that can predict who will resume smoking after hospitalization for an acute coronary syndrome. METHODS: Patients (n = 1320) below the age of 75 years, admitted to a Swedish university hospital coronary care unit with acute coronary syndromes, between September 1995 and September 1999, were consecutively included. Data were collected from hospital medical records and included information on previous clinical history, former illnesses and smoking. During their hospitalization, an experienced nurse interviewed the patients by using a structured questionnaire to obtain additional information. Patients were followed up 3 months after the discharge. Those who continued to smoke (non-quitters) were compared with those who had stopped (quitters) with regard to age, sex, medical history, clinical course, and intention to quit. To identify factors independently related to continued smoking, a logistical regression in a formal forward stepwise mode was used. RESULTS: Of the patients admitted, 33% were current smokers. Three months after discharge, 51% of these patients were still smoking. There were no significant differences in age, gender or marital status between non-quitters and quitters. In a multivariate analysis, independent predictors of continued smoking were: non-participation in the heart rehabilitation programme (P = 0.0008); use of sedatives/antidepressants at time of admission (P = 0.001); history of cerebral vascular disease (P = 0.002), history of previous cardiac event (P = 0.01); history of smoking-related pulmonary disease (P = 0.03) and cigarette consumption at index (P = 0.03). CONCLUSIONS: Smoking patients who do not participate in a heart rehabilitation programme may need extra help with smoking cessation. The findings may provide means of identifying patients in need of special intervention.