Key points are not available for this paper at this time.
BACKGROUND: The postmastectomy patient faces a plethora of choices when opting for autologous breast reconstruction; however, multi-institutional data comparing the available techniques are lacking. The National Surgical Quality Improvement Program (NSQIP) database provides a robust patient cohort for comparing outcomes and determining independent predictors of complications for each autologous method. STUDY DESIGN: The NSQIP database was retrospectively reviewed from 2006 to 2010, identifying 3,296 autologous breast reconstruction patients. Univariate analyses compared complication and reoperation rates. Multivariable logistic regression analyses of 4 cohorts (free flaps, pedicled transverse rectus abdominis myocutaeous (TRAM) flaps, latissimus, and all flaps in aggregate) determined complication rates and independent risk factors for complications and specific outcomes of interest (surgical site infection SSI, flap failure, reoperation) in all flap types. RESULTS: American Society of Anesthesiologists (ASA) classification ≥ 3, body mass index > 30 kg/m(2), recent surgery, delayed reconstruction, and prolonged operative times are significant predictors of increased complications in autologous reconstructions. Rates of complications, flap failure, and reoperation were highest in the free tissue transfer group (p < 0.001). Latissimus flaps showed significantly lower rates of complications than other autologous methods (p < 0.001). Pedicled TRAM patients had the highest incidences of venous thromboembolic disease and SSI. CONCLUSIONS: This large-scale, multicenter evaluation of outcomes in autologous breast reconstruction found that free flaps have the highest captured 30-day complication and reoperation rates of any autologous reconstructive method; complications in latissimus flaps were surprisingly few. Pedicled TRAM and latissimus flaps remain the most commonly used autologous reconstructive methods. In addition to providing statistically robust outcomes data, this study contributes significantly to patient education and preoperative planning discussions.
Gart et al. (Sun,) studied this question.