Free pedicled transverse rectus abdominis myocutaneous (TRAM) flap breast reconstruction is often advocated as the procedure of choice for autogenous tissue breast reconstruction in high-risk patients, such as smokers. However, whether use of the free TRAM flap is a desirable option for breast reconstruction in smokers is still unclear. All patients undergoing breast reconstruction with free TRAM flaps at our institution between February of 1989 and May of 1998 were reviewed. Patients were classified as smokers, former smokers (patients who had stopped smoking at least 4 weeks before surgery), and nonsmokers. Flap and donor-site complications in the three groups were compared. Information on demographic characteristics, body mass index, and comorbid medical conditions was used to perform multivariate statistical analysis. A total of 936 breast reconstructions with free TRAM flaps were performed in 718 patients (80.9 percent immediate; 23.3 percent bilateral). There were 478 nonsmokers, 150 former smokers, and 90 smokers. Flap complications occurred in 222 (23.7 percent) of 936 flaps. Smokers had a higher incidence of mastectomy flap necrosis than nonsmokers (18.9 percent versus 9.0 percent; p = 0.005). Smokers who underwent immediate reconstruction had a significantly higher incidence of mastectomy skin flap necrosis than did smokers who underwent delayed reconstruction (21.7 percent versus 0 percent; p = 0.039). Donor-site complications occurred in 106 (14.8 percent) of 718 patients. Donor-site complications were more common in smokers than in former smokers (25.6 percent versus 10.0 percent; p = 0.001) or nonsmokers (25.6 percent versus 14.2 percent; p = 0.007). Compared with nonsmokers, smokers had significantly higher rates of abdominal flap necrosis (4.4 percent versus 0.8 percent; p = 0.025) and hernia (6.7 percent versus 2.1 percent; p = 0.016). No significant difference in complication rates was noted between former smokers and nonsmokers. Among smokers, patients with a smoking history of greater than 10 pack-years had a significantly higher overall complication rate compared with patients with a smoking history of 10 or fewer pack-years (55.8 percent versus 23.8 percent; p = 0.049). In summary, free TRAM flap breast reconstruction in smokers was not associated with a significant increase in the rates of vessel thrombosis, flap loss, or fat necrosis compared with rates in nonsmokers. However, smokers were at significantly higher risk for mastectomy skin flap necrosis, abdominal flap necrosis, and hernia compared with nonsmokers. Patients with a smoking history of greater than 10 pack-years were at especially high risk for perioperative complications, suggesting that this should be considered a relative contraindication for free TRAM flap breast reconstruction. Smoking-related complications were significantly reduced when the reconstruction was delayed or when the patient stopped smoking at least 4 weeks before surgery.