Risk prediction of 30-day readmission after infrainguinal bypass for critical limb ischemia

Abstract

Hospital readmission after lower extremity bypass is a large cost burden and has become a focal point for policy change directed at disease-specific bundling strategies. The purpose of this study was to evaluate rates and predictors of 30-day readmission from a large, multicenter trial data set. We analyzed the PRoject of Ex-Vivo vein graft ENgineering via Transfection III (PREVENT III) data set of 1404 critical limb ischemia (CLI) patients undergoing lower extremity vein graft bypass at 83 North American centers. The primary end point was readmission ≤30 days of discharge. Secondary end points included graft patency and limb salvage evaluated in the context of readmission. The data set was split into a two-thirds derivation set and a one-third validation set for the purposes of creating a risk prediction model. A whole number integer risk score was assigned to independent predictors of readmission. Summary risk scores were collapsed into categories and defined as low (0-1 points), medium (2-5 points), and high (>5 points). We analyzed 1356 vein graft bypass patients, of which 23 (1.7%) died in-hospital and were excluded from the readmission analyses. In the derivation data set of 866 patients, 211 (24.4%) were readmitted ≤30 days of discharge. The most common reasons for readmission were wound infection in index leg (39.8%), an additional procedure in the index leg (20.8%), and nonvascular reasons (19%). By multivariable analysis, factors associated with 30-day hospital readmission (odds ratio [95% confidence limits]) included female gender (1.5 [1.0, 2.1]), current smoking (1.6 [1.1, 2.4]), in-hospital loss of graft patency (1.8 [1.0, 3.2]), dialysis (2.0 [1.2, 3.2]), and tissue loss (1.7 [1.1, 2.5]). In the derivation set, rates of readmission correlated to risk category. The 30-day readmission rates were 15.6% for low-risk patients (0-1 points), 24.1% for moderate-risk (2-5 points) patients, and 38.0% for high-risk (>5 points) patients. Similarly, in the validation set, the rates were 16.5%, 25.4%, and 38.1% for low-, medium-, and high-risk groups, respectively. Thirty-day readmission was not associated with loss of long-term graft patency but was associated with long-term limb loss (hazard ratio, 2.1; 95% confidence interval, 1.4-3.1; P= .0002). Readmission after lower extremity bypass for CLI is common (24%). Certain characteristics, such as female gender, current smoking, dialysis-dependence, tissue loss, and in-hospital graft-related events, are associated with increased risk. Readmission is associated with long-term limb loss. These data provide benchmark values for this complex patient population and may prove useful when hospital readmission is used as a quality metric for hospital performance.