Nader Toossi, Norman A. Johanson

High body mass index (BMI) has been shown to have a strong association with unfavorable outcomes following total hip arthroplasty. An increasing proportion of patients undergoing total hip arthroplasty have a high BMI value. The article by Wagner et al. is a single institution’s retrospective review of prospectively collected data regarding 21,361 primary total hip arthroplasties performed between 1985 and 2012. The relationship between BMI, as an independent variable, and outcomes were reported as hazard ratios, and survival estimates were made by the Kaplan-Meier method. The findings are largely consistent with those of the current literature; however, the treatment of BMI as a continuous variable is unique and represents an advance in understanding obesity in more of a “dose-response” relationship to surgical complications and points to a more refined risk-adjustment strategy. This is the greatest strength of the study.

A limitation of the study is its failure to address the combined effect of diabetes and high BMI on the surgical outcomes. Diabetes mellitus is strongly associated with both high BMI values and certain surgical complications, such as periprosthetic joint infection and wound dehiscence. Although the authors offered a credible explanation for the increased risk of infection, they did not present statistical evidence of controlling for the confounding effect of diabetes in their analysis. There are different ways to control for the confounding effect of diabetes on the outcome, such as case-matching, stratification, or adjusting for diabetic status in a multivariable analysis. The investigators could have stratified the patients on the basis of their diabetic status, or they could have adjusted for diabetes in the Cox model in addition to the other variables, such as age, sex, and preoperative diagnosis.