Purpose

Access site complications following peripheral vascular intervention (PVI) are associated with prolonged hospitalization and increased mortality. Prediction of access site complication risk may optimize PVI care; however, there is no tool designed for this. We aimed to create a clinical scoring tool to stratify patients according to their risk of developing access site complications after PVI.

Methods

The Society for Vascular Surgery’s Vascular Quality Initiative database yielded 27,997 patients who had undergone PVI at 131 North American centers. Clinically and statistically significant preprocedural risk factors associated with in-hospital, post-PVI access site complications were included in a multivariate logistic regression model, with access site complications as the outcome variable. A predictive model was developed with a random sample of 19,683 (70%) PVI procedures and validated in 8,314 (30%).

Results

Access site complications occurred in 939 (3.4%) patients. The risk tool predictors are female gender, age > 70 years, white race, bedridden ambulatory status, insulin-treated diabetes mellitus, prior minor amputation, procedural indication of claudication, and nonfemoral arterial access site (model c-statistic = 0.638). Of these predictors, insulin-treated diabetes mellitus and prior minor amputation were protective of access site complications. The discriminatory power of the risk model was confirmed by the validation dataset (c-statistic = 0.6139). Higher risk scores correlated with increased frequency of access site complications: 1.9% for low risk, 3.4% for moderate risk and 5.1% for high risk.

Conclusions

The proposed clinical risk score based on eight preprocedural characteristics is a tool to stratify patients at risk for post-PVI access site complications. The risk score may assist physicians in identifying patients at risk for access site complications and selection of patients who may benefit from bleeding avoidance strategies.