25227 A Preoperative Risk-Stratification Model for Medical and Surgical Complications in Complex Head and Neck Microvascular Reconstruction

Saturday, October 11, 2014: 1:15 PM
Patrick Gerety, MD , Plastic Surgery, University of Pennsylvania, Philadelphia, PA
Marten N Basta, BA , Plastic Surgery, University of Pennsylvania, Philadelphia, PA
Megan Fracol, BA , Plastic Surgery, University of Pennsylvania, Philadelphia, PA
John P Fischer, MD , Plastic Surgery, University of Pennsylvania, Philadelphia, PA
Joseph M. Serletti, MD , Plastic Surgery, University of Pennsylvania, Philadelphia, PA
Suhail Kanchwala, MD , Division of Plastic Surgery, University of Pennsylvania, Philadelphia, PA
David W. Low, MD , Division of Plastic Surgery, University of Pennsylvania, Philadelphia, PA
Liza C. Wu, MD , Division of Plastic Surgery, University of Pennsylvania, Philadelphia, PA

Background

Free tissue transfer has become the standard of care for reconstructing head/neck defects. Medical and flap complications have been associated with tremendous hospital costs in these patients, and a better understanding of predisposing factors is essential to improve cost-efficacy and patient satisfaction.  This project aims to create the first preoperative risk-stratification tool in head/neck reconstruction based upon our free flap experience.

Methods

An institutional retrospective chart review identified all free flap head/neck reconstructions performed from 2005-2013.  Demographic information and operative characteristics were collected as preoperative risk factors. Intra/post-operative complications and total hospital cost for were detailed.  Primary outcomes were flap complication (anastomotic revision, arterial/venous flap thrombosis, or partial/full flap loss) and medical complication (heart failure, MI, acute renal failure, VTE, pneumonia, respiratory failure, septicemia, or stroke).  Preoperative factors were analyzed via step-wise multivariate logistic regression, with significant variables entered into a bootstrap model. Final adjusted beta-coefficients were utilized to generate weighted risk scores for each variable.  Each patient was assigned an aggregate risk score for each complication, yielding the risk-assessment tool.

Results

438 free flap head/neck reconstructions were included.  Medical complication incidence=15.8% and significant predictors included BMI≥35 (OR=3.09), cardiovascular comorbidity (OR=2.72), floor-of-mouth defect (OR=3.03), fibula flap (OR=2.62), and prolonged operative time (OR=6.31).  Patients were risk-stratified into 3 groups: low (complication=6.8%), moderate (complication=36.5%), and high (complication=57%). Flap complication incidence=13%, with predictors including anemia (OR=3.05), radiation (OR=3.95), and fibula flap (OR=3.80).  Patients were risk-stratified into 3 groups: low (complication=3%), moderate (complication=19%), and high (complication=46.7%). Patients at intermediate/high risk of medical or flap complications had significantly longer length of stay and more reoperations.  Total hospital costs for patients at high-risk for medical complications were $80,368 (vs. $37,606 for low-risk patients), while patients at high-risk for flap complications had total costs=$81,890 (vs. $36,188 for low-risk patients). The risk models demonstrated high accuracy for discriminating complication risks with C-statistic=0.78 (medical) and C-statistic=0.77 (flap).

Conclusion

This study identifies important independent preoperative risk factors for medical and flap complications in head/neck microsurgical reconstruction with high accuracy. Importantly, the model stratifies patients according to risk, demonstrating that high-risk groups suffer significantly more morbidity and incur substantially higher costs.  As the first such model in head and neck microvascular reconstruction, it can serve as a simple yet accurate tool for perioperative decision-making.