The Vascular Quality Initiative 30-day stroke/death risk score calculator after transfemoral carotid artery stenting.

OBJECTIVE

Carotid artery stenting (CAS) was introduced as an alternative carotid revascularization procedure in patients deemed to be at high risk for carotid endarterectomy. Although techniques and selection criteria for patients have dramatically improved, CAS continues to have higher risk of stroke and death in comparison to carotid endarterectomy. Several risk factors are known to be associated with worse outcomes. Whereas knowledge of these independent factors is helpful, clinical decision-making is further refined when these are considered in aggregate. This study aimed to develop a score to predict the risk of stroke/death after transfemoral CAS (TFCAS).

METHODS

We analyzed the Vascular Quality Initiative CAS data set from 2010 to 2018. Lesions due to trauma, dissection, or transcarotid artery stenting and cases performed without an embolic protection device were excluded. Univariable and multivariable logistic regression methods with bootstrapping (1000 repetitions) were used to identify predictors associated with 30-day stroke/death. Stepwise backward selection for variables was used to achieve model parsimony. A risk score was made by converting regression coefficients for each predictor to integers from which probability was calculated. Scores were grouped into simplified categories.

RESULTS

We identified 10,753 patients undergoing TFCAS during the study period with a combined 30-day stroke/death rate of 4.1%. On multivariable adjustment, independent predictors of 30-day stroke/death included age (odds ratio [OR], 1.05; 95% confidence interval [CI], 1.03-1.06; P < .001), nonwhite race (OR, 1.42; 95% CI, 1.16-1.74; P = .001), diabetes (OR,1.34; 95% CI, 1.08-1.67; P = .01), coronary artery disease (OR, 1.40; 95% CI, 1.13-1.73; P = .001), congestive heart failure (OR, 1.41; 95% CI, 1.07-1.85; P = .02), symptomatic status (OR, 2.11; 95% CI, 1.64-2.72; P < .001), and contralateral occlusion (OR, 1.64; 95% CI, 1.22-2.19; P = .001). On the other hand, preoperative use of statins (OR, 0.074; 95% CI, 0.59-0.93; P = .02) and dual antiplatelet therapy (P2Y12 inhibitors and aspirin; OR, 0.46; 95% CI, 0.32-0.66; P < .001) were associated with a significant reduction in stroke/death after TFCAS. The model had a C statistic of 69.0%. The coefficients of these predictors were used to develop a risk score calculator that estimates the probability of 30-day stroke/death after TFCAS.

CONCLUSIONS

In an analysis of 10,753 patients undergoing TFCAS between 2010 and 2018, significant predictors of perioperative stroke or death included old age, nonwhite race, symptomatic status, diabetes, coronary artery disease, congestive heart failure, and contralateral occlusion in addition to perioperative dual antiplatelet therapy and statin use. These variables were used to develop a risk score calculator that estimates the probability of 30-day stroke/death after TFCAS. External validation of this tool in different populations of patients and data sets is warranted to evaluate its predictive performance.