Breast reconstruction with tissue expansion and subsequent implant insertion frequently leads to acceptable outcomes, but can be associated with complications such as asymmetry, capsular contracture and implant extrusion. Many of these complications can be remedied by local revision, but in severe cases conversion to an autologous reconstruction may be required for salvage. This group of patients has not been adequately studied in the past, specifically as to the impact of the explanted implant milieu on the performance of a variety autologous reconstructions, especially the Free TRAM Flap. Therefore, this study retrospectively reviewed patients whose failed implant reconstructions were converted to autologous tissue. During the 10-year period 1992-2002, forty patients who met theses criteria were identified. Seventeen patients had saline implants (25 implants) and 23 had silicone gel implants (30 implants). The leading reason for conversion was capsular contraction in 20 pts (50%). In the remainder of patients, other concerns led to explantation and autologous salvage, including asymmetry, unsatisfactory cosmetic result, implant migration, silicone concern and rupture of implant. All patients underwent anterior capsulectomy and in 28 (70%) posterior capsulectomy, either partial or complete, was performed as well. The reconstructive procedures included 26 free TRAM flaps (18 unilateral, 8 bilateral), 12 pedicle TRAM flaps (5 unilateral, 7 bilateral) and 2 Latissimus Dorsi mypcutaneous flap with implant. In the free TRAM group (34 flaps), the recipient vessels were the thoracodorsals in 20 pts (59%), internal mammary in 13 flaps (38 %), and subscapular vessels in one patient (3%). Drains were used in all patients to reduce seroma and left in place an average of 8.8 days (range 4-15). Despite this, 2 pts (5%) required interventional treatment for persistent seroma. In 4 patients, free TRAM reconstruction could not be performed because excessive scaring around recipient vessels prevented safe isolation and pedicled reconstructions were performed instead. In the free TRAM group, there were 4 (15%) instances of microvascular compromise; 2 arterial thromboses, 1 venous thrombosis and 1 venous compression, leading to complete flap loss in 2 pts (8%). These data suggest that autologous salvage after failed implant breast reconstruction can be safely accomplished, but may be associated with slightly higher conversion to pedicled reconstruction and complete flap loss compared with historical non-implanted delayed reconstruction controls. Seroma formation does not appear to be increased, as would have been expected after extensive capsulectomy.