Purpose: Free flap loss, while uncommon, can be a devastating experience for patient and surgeon. After flap failure, a second free flap may be problematic on a number of levels, yet traditional (non- microsurgical) reconstructive approaches may yield inferior results and the morbidity of failing to complete the reconstruction must be taken into account. The approach to the management of a failed initial free flap appears to be primarily influenced by functional outcome. In the head and neck second free flaps are often required to cover vital structures or to restore critical function. Alternatively, as breast reconstruction may be considered quality of life issue, if the first free flap fails, it is unlikely that a second will be performed. While clinically endorsed, this tenet has not received objective study and analysis.

Methods: We retrospectively examined all initial free flaps performed at 2 institutions for reconstruction of the breast, head and neck (H&N) and lower extremity (LE) over a 12 year period. Data analyzed included patient demographics, co-morbidities, complications, number/ type of initial free flaps, and operative procedure(s) for wound closure.

Results: 3789 free flaps in the 3 study categories were performed 7/1/1992 - 6/30/2004. Operative mortality rate (60 days) was 0.4%(N=16)-all H&N pts(rate=0.9%). Case distribution: Breast = 1,945 with total and partial failure rates of 1.4% (N=28) & 1.6% (N=32; H&N: 1,741 with total and partial failure rates of 2.9% (N=51)& 4.2% (N=73) respectively; LE: 103 with total and partial failure rates of7% (N=7) and 3% (N=3). The etiology of flap failure (arterial, venous or combined) was equally distributed and similar in all categories. Flap failure rates were not statistically different in the first 7 versus last 5 years of study. Co- morbidities, age and complications were highest in the H&N population, yet not statistically different between the failed and successful groups in any category. In breast reconstruction of the initial 28 total flap failures 46% (n=13) did not complete their reconstruction. Rationale cited for deferring reconstruction included: Oncologic indications N=4; Surgeon recommendation, N= 2;Pt preference N=6,unknown N=1. 54% (N=15)completed their reconstruction utilizing: Expander/implant (80%, N=12); Latissimus dorsi/implant (13.5%, N=2), a second free flap for breast was used only once(7%).An average of 2.6 (range(R)0-5) additional procedures (exclusive of NAR) performed an average 2.9 mos (R 2 days-14mos) after the initial flap were required for reconstruction. 98% (N=30) of the partial losses completed their breast reconstructions primarily with local/balancing procedures necessitating 0-7 additional operations. Of the 51 initial flap failures in H&N, 82.4% (N=42) completed their reconstruction, 17.6% (N=9) did not. Initial flap losses were managed as follows: Second free flap 44% (N=54); pedicled flap 26%; (N=14 primary closure 20% (N=11);skin graft 9%(N=5); requiring an average of 2.6 (R 1-7) additional procedures. In 4 cases a 3rd free flap was performed after failure of the first 2 flaps. Our LE experience more limited, salvage of flap failures, utilized a second free flap 57% ( N=4); 3rd free flap (N=1); Debridement/ primary closure (N=2-all thigh wounds)requiring an average of 1.8(R 1-3) additional procedures.

Conclusions: This preliminary investigation suggests that practice patterns result in a differential approach to the management of a failed initial free flap based on anatomic region. Unknown, is the outcome of wounds when initial or second free flaps are not employed for closure/ reconstruction, especially if the ultimate outcome is a quality of life issue. Additional prospective outcome studies will be required to answer the deeper questions: Are our indications for reconstruction and/ or closure of defects using free flaps too liberal? Do alternative techniques provide comparable results to free tissue transfer in terms of form, function, and aesthetics?