Frailty as a predictor of outcomes for patients undergoing carotid artery stenting

The concept of frailty has been proposed to capture the vulnerability resulting from aging and has been implemented for the prediction of perioperative outcomes. Carotid artery stenting (CAS) is considered an appropriate minimally invasive procedure for patients considered to high risk to undergo carotid endarterectomy. Recently, the predictive accuracy for perioperative outcomes using the five-item modified frailty index (5mFI) has been reported to be relatively poor for cardiovascular surgery compared with other surgeries. The effects of functional status and the 5mFI on the outcomes after CAS remain unknown. Thus, in the present study, we investigated the relationship between 5mFI, functional status, and perioperative outcomes.

All the patients who had undergone CAS in the Vascular Quality Initiative from November 15, 2016 to December 31, 2018 were included. Good functional status was defined as the ability to perform all predisease activities without restriction using a new variable added to the Vascular Quality Initiative from November 15, 2016 onward. The 5mFI was calculated using functional status and a history of diabetes, chronic obstructive pulmonary disease, congestive heart failure, and hypertension. The perioperative outcomes included in-hospital stroke or death within 30 days after CAS, a prolonged postoperative stay (≥2 days), and nonhome discharge. The associations between functional status, 5mFI, and perioperative outcomes were examined using univariate and multivariable logistic regression, adjusting for sex, age, race, degree of stenosis, symptomatic status, and the usage of preoperative medications. An analysis stratified by functional status was also performed.

Of the 7836 patients, 188 (2.4%) had experienced perioperative stroke or death, 765 (9.8%) had required a nonhome discharge, and 2584 (33.0%) had required a prolonged postoperative stay. A higher (≥0.6 vs <0.6) 5mFI score was associated with greater odds of perioperative stroke or death (adjusted odds ratio [aOR], 2.75; 95% confidence interval [CI], 1.42-5.28; P = .003), non-home discharge (aOR, 2.70; 95% CI, 1.89-3.85; P < .001), and a prolonged postoperative length of stay (aOR, 1.96; 95% CI, 1.56-2.46; P < .001). For the predictive accuracy of the perioperative outcomes, the 5mFI model had an area under the curve for in-hospital stroke or death, nonhome discharge, and prolonged postoperative length of stay of 0.714, 0.767, and 0.668, respectively. The functional status model was not inferior to the 5mFI model for any of these outcomes. In the subgroup analysis, of the asymptomatic patients, a higher 5mFI score was associated with greater odds of perioperative stroke or death (aOR, 7.08; 95% CI, 2.02-24.48; P = .002), nonhome discharge (aOR, 5.87; 95% CI, 2.45-13.90; P < .001), and a prolonged postoperative stay (aOR, 2.60; 95% CI, 1.82-3.71; P < .001).

Frailty, as measured using the 5mFI, and functional status were independent predictors of perioperative stroke or death, nonhome discharge, and an increased length of stay for patients undergoing CAS. These results were greatly pronounced in asymptomatic patients. The results from the present study, thus, caution against the use of CAS for asymptomatic frail patients.