High-Risk Characteristics for Clinical Failure after Isolated Femoropopliteal Peripheral Vascular Interventions

Recent studies suggesting increased late mortality with paclitaxel-coated devices (PCDs) in femoropopliteal peripheral vascular interventions (PVIs) prompted the US Food and Drug Administration to recommend limiting the use of PCDs to "patients at particularly high risk for restenosis". This study's aim is to characterize patients at high risk for restenosis and develop a risk score to guide clinicians in selecting patients for treatment with PCDs.

Patients who underwent isolated femoropopliteal PVI for claudication or rest pain from 2016-2019 in the Vascular Quality Initiative were included. Patients who received intervention with a PCD, hybrid procedures, died within 1 year, or had missing follow-up data were excluded. The primary end point was clinical failure at 1 year defined as > 50% restenosis, loss of patency, reintervention, or major amputation. Data were split randomly into 2/3 for development and 1/3 for validation. A parsimonious multivariable hierarchical logistic regression for clinical failure was developed and a risk score was created using beta-coefficients. The risk score was applied to the validation dataset and tested for goodness-of-fit and discrimination.

Among 4,856 treated patients, 718 (14.8%) experienced clinical failure within 1 year. Clinical failure was associated with age ≤ 50 years, female sex (48.1% vs. 39.5%), insulin-dependent diabetes (29.9% vs. 23.1%), creatinine > 2.0 mg/dL (9.9% vs. 5.7%), prior ipsilateral lower extremity revascularization (48.5% vs. 38.5%), prior ipsilateral minor amputation (5.3% vs. 1.7%), rest pain versus claudication (30.8% vs. 18.7%), occlusion length ≥ 20 cm (18.8% vs. 15.0%), and Trans-Atlantic Inter-Society Consensus II Classification C or D (40.4% vs. 28.0%), all P ≤ 0.01. Risk score development was performed using a multivariable regression. The model demonstrated good fit and discrimination (C-statistic 0.71 in development and 0.72 in validation dataset). Predicted clinical failure was 8.9% for standard-risk (45.9% of interventions), 15.5% for high-risk (44.2% of interventions), and 33.8% for very high-risk patients (9.8% of interventions).

A novel risk score was created with good discrimination for identifying patients at high risk for clinical failure at 1 year after femoropopliteal PVI for claudication and rest pain. Patients at high risk and very high risk for clinical failure may benefit from alternative strategies including PCDs.