Predictors of early graft failure after infrainguinal bypass surgery: a risk-adjusted analysis from the NSQIP

Abstract

Introduction and objectives: Infrainguinal bypass surgery (BPG) is accompanied by significant 30-day mortality and morbidity, including early graft failure. The goal of this study is to identify patient- and procedure-specific factors which predict the rate of early graft failure in contemporary practice.

Methods: Data was obtained from the private sector National Surgical Quality Improvement Program, a prospective, validated database collected between 2005 and 2008 from 211 hospitals, using primary and modifier Current Procedural Terminology codes for BPG. The primary endpoint was graft failure at 30 days. Procedural parameters, patient demographics and clinical variables were analyzed by univariate and multivariate methods.

Results: There were 9217 BPG procedures (limb salvage, 49%; infrapopliteal distal anastomosis, 43%; prosthetic 32%) with patient variables: age 67 ± 12 years, male 64%, diabetes 44%, dialysis 7.4%. Mortality was 2.4%, major morbidity was 17.3%, and graft failure rate was 6.3% at 30 days. Multivariate predictors of graft failure demonstrated correlation (p-value, OR) with female gender (p = 0.0054, 1.29), limb salvage indication (p < 0.0001, 1.60), infrapopliteal anastomosis (p < 0.0001, 2.15), composite graft (p = 0.0436, 1.82), current smoking (p = 0.0007, 1.36), impaired sensorium (p = 0.0075, 2.13), emergency procedure (p < 0.0001, 2.03), previous vascular procedure (p = 0.0005, 1.39), and platelets >400K (p = 0.0019, 1.49). High-risk composite constructs utilizing these significant predictive factors can identify cohorts of patients with up to a 98-fold increase in odds of early graft failure.

Conclusions: These results describe common risk factors that correlate with early graft thrombosis including the unique description of its association with thrombocytosis. Additional risk factors thus identify a subset of patients who are at highest risk for early BPG failure. This data may be used to refine patient selection.