Nonlinear effect of body mass index on postoperative survival following isolated heart transplantation

Abstract

Background

Guidelines regarding recipient's body mass index (BMI) for heart transplant are evolving with variable cutoffs depending on the country and institution. It is imperative to provide updated nonlinear estimates of postoperative risk attributable to a recipient’s BMI to evaluate the relevance of existing cutoffs.

Methods

A total of 30,787 patients were analyzed from the United Network for Organ Sharing (UNOS) database. Patients receiving an isolated heart transplant ages 18 and older since 2010 were included. Overall survival was the primary outcome. A multivariate Cox proportional hazards model was applied and included a penalized smoothing spline term for recipient BMI and risk factors such as diabetes. We assessed the overall significance of the nonlinear penalized spline terms using an asymptotic Wald test.

Results

The cohort consisted of 662 (2.2%) BMI <18.5, 9,359 (30%) BMI 18.5 to 24.9, 10,997 (36%) BMI 25 to 29.9, 9,550 (31%) BMI 30 to 39.9, and 206 (0.7%) BMI ≥40 patients. The nonlinear spline terms for recipient BMI were statistically significant (p < 0.01). The hazard ratio (HR) appeared to grow linearly in BMI at an inflection point of BMI = 26. No inflection point was observed at either of the International Society for Heart and Lung Transplantation recommended cutoffs of BMI = 30 (HR 1.11, confidence interval [CI] 1.07-1.15) or BMI = 35 (HR 1.29, CI 1.24-1.37).

Conclusions

After multivariable adjustment, there is no sharp cutoff in survival risk at either BMI = 30 or BMI = 35. Unlike previously reported, postoperative survival risk grows approximately linearly in the BMI range from 26 to 40.