The indications for free flaps have been more or less clarified; however, the course of reconstruction after the failure of a free flap remains undetermined. Is it better to insist on one's initial choice, or should surgeons downgrade their reconstructive goals? To establish a preliminary guideline, this study was designed to retrospectively analyze the outcome of failed free-tissue transfers performed in the authors hospital. Over the past 8 years (1990 through 1997), 3361 head and neck and extremity reconstructions were performed by free-tissue transfers, excluding toe transplantations. Among these reconstructions, 1235 flaps (36.7 percent) were transferred to the head and neck region, and 2126 flaps (63.3 percent) to the extremities. A total of 101 failures (3.0 percent total plus the partial failure rate) were encountered. Forty-two failures occurred in the head and neck region, and 59 in the extremities. Evaluation of the cases revealed that one of three following approaches to handling the failure was taken(1) a second free-tissue transfer; (2) a regional flap transfer; or (3) conservative management with debridement, wound care, and subsequent closure by secondary intention, whether by local flaps or skin grafting. In the head and neck region, 17 second free flaps (40 percent) and 15 regional flaps (36 percent) were transferred to salvage the reconstruction, whereas conservative management was undertaken in the remaining 10 cases (24 percent). In the extremities, 37 failures were treated conservatively (63 percent) in addition to 17 second free flaps (29 percent) and three regional flaps (5 percent) used to salvage the failed reconstruction. Two cases underwent amputation (3 percent). The average time elapsed between the failure and second free-tissue transfer was 12 days (range, 2 to 60 days) in the head and neck region and 18 days (range, 2 to 56 days) in the extremities. In a total of 34 second free-tissue transfers at both localizations, there were only three failures (9 percent). However, in the head and neck region, seven of the regional flaps transferred (47 percent) and four cases that were conservatively treated (40 percent) either failed or developed complications that lengthened the reconstruction period because of additional procedures. Six other free-tissue transfers had to be performed to manage these complicated cases. Conservative management was quite successful in the extremities; most patients' wounds healed, although more than one skin-graft procedure was required in 10 patients (27 percent). In conclusion, a second free-tissue transfer is, in general, a relatively more reliable and more effective procedure for the treatment of flap failure in the head and neck region, as well as failed vascularized bone flaps in the reconstruction of the extremities. Conservative treatment may be a simple and valid alternative to second (free) flaps for soft-tissue coverage in extremities with partial and even total losses.