Survival prediction in patients with chronic limb-threatening ischemia who undergo infrainguinal revascularization

Objective: Accurate survival prediction critically influences decision-making in caring for patients with chronic limb-threatening ischemia (CLTI). The Bypass versus Angioplasty in Severe Ischaemia of the Leg (BASIL) trial demonstrated that in patients who survived >2 years, there was a significant advantage to infrainguinal bypass compared with endovascular intervention, which increased with time. Validated survival models for patients with CLTI are lacking. Methods: The Vascular Quality Initiative was interrogated for patients who underwent infrainguinal bypass or endovascular intervention for CLTI (January 2003-February 2017). Cox survival models were generated using only preoperative variables. Survival at 30 days, 2 years, and 5 years was modeled separately. Patients were defined as low risk (30-day survival >97% and 2-year survival >70%), medium risk (30-day survival 95%-97% or 2-year survival 50%-70%), and high-risk (30-day survival <95% or 2-year survival <50%). Results: Among 38,470 unique CLTI patients, 63% (n = 24,214) underwent endovascular intervention and 37% (n = 14,256) underwent infrainguinal bypass. Kaplan-Meier estimates of overall survival at 30 days, 2 years, and 5 years were 98%, 81%, and 69%, respectively. The proportion of patients in the low-, medium-, and high-risk groups was 84%, 10%, and 6.5%, respectively. Patients in the low-risk group were significantly less likely to undergo endovascular intervention compared with those in the high-risk group (low risk, 59% endovascular; high risk, 75% endovascular; P < .0001). Independent predictors of death were similar in all three models, with greatest magnitude of effect associated with age >80 years, oxygen-dependent chronic obstructive pulmonary disease, stage 5 chronic kidney disease, and bedbound status. The C index for the 30-day model, 2-year model, and 5-year model was 0.76, 0.72, and 0.71, respectively. Procedure type (open or endovascular) was not significant in any models and did not have an impact on C indices. Conclusions: These survival prediction models, derived from a large U.S. cohort of patients who underwent revascularization for CLTI, demonstrated good performance and should be validated. Most CLTI patients considered candidates for limb salvage were of average perioperative risk and were predicted to survive beyond 2 years. These models can differentiate patients into low-, medium-, and high-risk groups to facilitate evidence-based revascularization recommendations that are consistent with current treatment guidelines.