![[ Visit Client Website ]](images/banner.gif)
Methods: Patients undergoing breast reconstruction since 2003 with follow up in our institution’s medical record system were reviewed. Chi-squared and independent t-tests were used to identify variables associated with infection; associated variables (p<0.05) were used to build a logistic regression.
Results: Four hundred thirty-three patients were included, of whom 88 had infections. Patients who had infections were similar age (p=0.810) and BMI (p=0.072) to those without but had more comorbidities (1.23±1.21 versus 0.92±1.29, p=0.042). A binary logistic regression controlling for number of comorbidities, chemotherapy, radiation, Caucasian ethnicity, dehiscence, seroma, and implant exposure was constructed. Implant exposure (OR 6.71, 2.29-19.7) and Caucasian ethnicity (OR 1.72, 1.03-2.88) were the only factors predictive of infection. Patients with infections were more likely to decline further reconstructive procedures (OR 2.10, 1.21-3.64) and require more procedures overall (5.08±2.35 versus 3.74±1.75, p<0.0001), largely driven by more implant exchanges (1.89±1.74 versus 1.29±1.13, p=0.017). Infected patients were not any less likely to finish their reconstructions, as indicated by nipple reconstruction or tattooing (OR 0.90, 0.58-1.49).
Conclusions: Contrary to previous studies, we did not find that radiotherapy or chemotherapy were associated with infection in a regression model. The relationship between implant exposure and implant infection is fairly straightforward but the reason Caucasian patients were more likely to have infections is less clear. While it was reassuring to find that infected patients were not less likely to complete breast reconstruction, the increased number of procedures these patients undergo is concerning from a cost and risk-exposure standpoint.