Interrater agreement for high grade carotid artery stenosis measurement and treatment decision.

ABSTRACT

OBJECTIVES: Randomized trials in North America (NASCET, ACAS) and Europe (ECST) have shown a beneficial effect of endarterectomy for patients with high grade carotid artery stenosis. The results of the NASCET and the ECST further suggest that the effect of endarterectomy differed by degree of stenosis, supporting the importance of stenosis measurement as a factor in the decision process regarding surgery. We investigated the interrater agreement for carotid artery stenosis measurements and treatment decision in a post hoc study on patients undergoing carotid surgery. METHODS: In a one-year series, 45 consecutive patients underwent preoperative conventional cerebral angiography followed by endarterectomy. Using a magnifying eyepiece and applying the two different measurement criteria of the randomized trials, angiograms were re-evaluated post hoc by three masked raters. Intra-class correlation coefficients (ICCs) with one-sided 95% confidence intervals (CIs) were calculated for the estimation of interrater agreement for degree of stenosis. Conger s kappa (k) statistics were used for the estimation of interrater agreement for a dichotomized stenosis evaluation, i.e. therapeutic decision on surgery (cut-off point for symptomatic stenosis 70%, cut-off point for asymptomatic stenosis 60%). RESULTS: ICCs were.74 (CI.63) for NASCET/ACAS criteria and.72 (CI. 59) for ECST criteria. k values were.55 (CI.42) for NASCET/ACAS criteria and.57 (CI.44) for ECST criteria. Disagreement for a therapeutic decision was seen in 6 of 23 symptomatic patients by NASCET criteria, in 2 of 23 symptomatic patients by ECST and in 4 of 22 asymptomatic patients by ACAS criteria. CONCLUSIONS: Overall, the interrater agreement for stenosis measurements was good. Agreement for therapeutic decisions on carotid surgery, however, was less strong. These findings suggest that accurate stenosis measurement may not suffice for reliable treatment decisions in patients with high grade carotid artery stenosis.