Reexamining BSA as a Preoperative Predictor of Risk of Prosthetic Joint Infection

BACKGROUND: Prosthetic joint infection (PJI), a dreaded complication of arthroplasty, has been found to correlate with increasing body mass index (BMI) and body surface area (BSA). Recent data suggest that BSA may be a better predictive tool for assessing infection risk. We further evaluated this association in an orthopedic specialty hospital arthroplasty cohort to evaluate whether BSA is a predictor of prosthetic joint infection. METHODS: A retrospective cohort of hip and knee arthroplasty patients between 2009 and 2014 was identified through administrative hospital data using ICD-9 codes. Patients with a BMI <14 or >60 kg/m(2), height <142 or >200 cm, and weight <36 or >226 kg were excluded. BSA was calculated using the DuBois formula, and assessed both as a continuous and as a categorical variable. Univariate analyses were done using χ (2) tests and adjusted models were assessed using logistic regression. RESULTS: 17,859 knee and 18,128 hip patients were identified. 1.1% of knees and 0.74% of hips were infected.]Mean BSA was 1.9 m(2) (±0.2 m(2)). BSA was significantly associated with PJI in hips (P = 0.004), but not knees, when analyzed as a continuous variable in unadjusted models. However, this association lost its significance after adjusting for PJI risk factors. Additionally, when assessed as a categorical variable in a multivariate model, BSA in the highest quartile (>2.11) was not associated with PJI. CONCLUSION: After evaluating BSA as a continuous and categorical variable, we failed to find an association between BSA and infection risk in THA or TKA. The impact of BSA decreased after multivariate adjustment. BSA may not be optimal as a predictor of preoperative risk. DISCLOSURES: All authors: No reported disclosures.