Objective assessment of the hemisphere-specific neurological outcome of carotid endarterectomy: a quantitative saccadometric analysis.

ABSTRACT

BACKGROUND: Carotid endarterectomy (CEA) improves the cerebrovascular prognosis of patients with carotid stenosis but carries a risk of causing postoperative neurological deterioration. OBJECTIVE: We assessed hemisphere-specific changes in saccadic eye movements to determine the utility of saccadometry as a quantitative neurosurgical outcome measure. METHODS: Visually evoked saccades were recorded at the bedside before and 2 days after surgery from 30 patients undergoing CEA for symptomatic carotid stenosis. Hemisphere-specific latency distributions were compared using Kolmogorov-Smirnov statistics. Latency distributions were fitted using the Linear Approach to Threshold with Ergodic Rate model and compared with binomial logistic regression. RESULTS: There were 21 males and mean age at surgery was 71±7 years. Following CEA, the distribution of saccades initiated by the cerebral hemisphere distal to the operated artery significantly changed in 25 patients. By contrast, there were 14 significant contralateral-hemisphere saccadic changes (P<.001). Significant contralateral saccadic changes always co-occurred with significant ipsilateral changes and 10 of 14 patients with contralateral saccadic change had contralateral carotid stenosis. There was a significantly greater postoperative reduction in early saccades generated by the ipsilateral hemisphere than by the contralateral hemisphere (P<.02) CONCLUSION: CEA leads to significant hemisphere-specific subclinical changes in saccadic performance and, in particular, differentially affects the proportion of early saccades, a measure of the ability of the frontal cortex to successfully inhibit lower centers, generated by the 2 hemispheres. Saccadometry, a bedside test, provides data that can be statistically compared for individual and groups of patients. It could allow the neurological outcome of carotid surgery to be objectively quantified.