Subtle cognitive decline following CEA occurs and persists for at least several weeks after surgery. This decline was absent in a control group.

OBJECTIVE-Carotid endarterectomy (CEA) is an effective means of stroke prevention among appropriately selected patients; however, neuropsychometric testing has revealed subtle cognitive injuries in the early postoperative period. The purpose of this study was to establish whether serum levels of two biochemical markers of cerebral injury were correlated with postoperative declines in neuropsychometric test performance after CEA. METHODS-Fifty-five consecutive patients underwent a battery of neuropsychometric tests 24hours before and 24 hours after elective CEA. Two patients were excluded because of postoperative strokes. The pre-and postoperative serum levels of S100B protein and neuronspecific enolase for injured patients, defined as those who exhibited significant declines in neuropsychometric test performance (n = 12), were compared with the levels for uninjured patients (n = 41).RESULTS-There were no significant differences in the baseline S100B levels for the two groups. Injured patients exhibited significantly higher S100B levels, compared with uninjured patients, at 24, 48, and 72 hours after surgery (P < 0.05). There were no significant differences in neuron-specific enolase levels for injured and uninjured patients at any time point.CONCLUSION-These data suggest that subtle cerebral injuries after CEA, even in the absence of overt strokes, are associated with significant increases in serum S100B but not neuron-specific enolase levels. Analyses of earlier time points in future studies of subtle cognitive injuries and biochemical markers of cerebral injury after CEA may be revealing. KeywordsCarotid endarterectomy; Cerebral ischemia; Neuron-specific enolase; Neuropsychological tests; S100B Carotid endarterectomy (CEA) is an effective means to prevent stroke among patients at high risk. Although the risk of major stroke after CEA is relatively low, the incidence of subtle cognitive injuries, as revealed by neuropsychometric tests (NPMTs), is quite high (4,6,11). NPMTs are sensitive measures of subtle neurological dysfunctions that are not revealed in routine neurological examinations. We recently demonstrated that as many as A variety of biochemical markers of severe neurological injury have been described. Neuron-specific enolase (NSE) is a glycolytic enzyme found in neurons and neuroendocrine cells (19). The calcium-binding protein S100B is an important participant in neuronal and glial cell-cell communication and intracellular signal transduction and exhibits its highest levels of expression in the nervous system (33). These molecules have been demonstrated to be markers of cerebral injury in animal models of stroke (1,9) and among human stroke patients (3,5,21,23,27).Among patients with cerebral infarctions, high S100B levels were correlated with mildly poorer performance on NPMTs, but these results did not reach significance (32). Another study demonstrated no correlation between increased NSE and S100B levels and a decline in performance on NPMTs after CEA (25). In that study, however, the au...

These data suggest that subtle cerebral injuries after CEA, even in the absence of overt strokes, are associated with significant increases in serum S100B but not neuron-specific enolase levels. Analyses of earlier time points in future studies of subtle cognitive injuries and biochemical markers of cerebral injury after CEA may be revealing.

Rightward deviation on bisection of a horizontal line is well described in patients with right brain injury and left hemineglect. Because of the observation that hemineglect patients may bisect very short lines to the left of the true midpoint (the so-called crossover effect), additional models have been proposed to incorporate this finding into existing theories of hemineglect. We investigated a line-length effect in six patients with left hemineglect. When presented with any set of lines of uniform (reference) length, percentage rightward deviation on line bisection remained constant across different line lengths. When lines of a second length were mixed into any uniform set of lines, bisection performance on the reference lines changed. A rightward shift in the perceived midpoint of the reference line occurred if the added lines were shorter than the reference lines; a leftward shift occurred if the added lines were longer. Leftward shifts included shifts across the true midpoint, reproducing the crossover effect. Shifts in the perceived midpoint occurred both on a manual line bisection task and on a line bisection discrimination task in which no manual response was required. We propose that the crossover effect may be part of a more general stimulus-context effect in which the perceived midpoint of a line is related not to absolute length, but to the line's length relative to other lines with which it is presented. Such a context effect has not hitherto been described in the neglect syndrome. A possible mechanism for the effect is a generalization of length estimation produced by the combined influence of the focal stimulus and all stimuli that precede it.