The sensibility and specificity of cerebral oximetry, measured by INVOS - 4100, in patients undergoing carotid endarterectomy compared with awake testing.

BACKGROUND: Selective shunting during carotid endarterectomy (CEA) is advocated to reduce shunt-related stroke. Cerebral monitoring is essential for temporary carotid shunting. Many techniques are available for cerebral monitoring; however, none is superior to monitoring the patient's neurological status (awake testing) while performing the procedure under local anesthesia (LA). Cerebral oximetry (CO) has previously been used to show the adequacy of cerebral circulation in patients undergoing CEA. This investigation was designed to compare the performance of the INVOS-4100 cerebral oximeter and the neurologic functions, by means of detecting cerebral ischemia induced by carotid cross-clamping, in patients undergoing CEA under LA, namely cervical plexus block.
METHODS: Patients scheduled for CEA under LA were included. Patients converted to general anesthesia (GA) or other types of operations other than CEA were excluded from this study. We enrolled 100 consecutive patients from January 2009 to December 2010. Bilateral regional cerebrovascular oxygen saturation (rSO(2)) was monitored in all patients, in addition to the awake testing. Changes in rSO(2) following carotid artery clamping were recorded. A drop greater than 20% was considered as an indicator of cerebral ischemia that might predict the need for carotid shunting. Patients were only shunted based on the awake testing.
RESULTS: Of the 100 patients undergoing CEA under LA, 9 showed a significant drop in rSO(2) (range: 22.6-32.8%, mean: 26.4%): only three of them required shunting, while the remaining 6 had no changes in consciousness after internal carotid artery (ICA) cross-clamping and it was not necessary to place a shunt (false positive). Compared to the preclamping values, a significant decrease in rSO(2) was found on the hemisphere of the operated side, while no significant change was observed contralaterally. Ninety-one patients had no significant changes of CO values: in 89 of them there was no consciousness deterioration, so we didn't place a shunt (true negative), but 2 patients showing a non-significant post-clamping decline in CO saturation (1.5% and 18.2%) required shunting based on the awake testing (2 false negative). In the current study, the median drop in rSO(2) was 19% (range: 1.5-26.4%) in the 5 patients that required shunting. This represents a sensitivity of 60% and a specificity of 25% for CO in comparison to the awake testing.
CONCLUSION: The results of this study suggest that the usefulness of CO in predicting cerebral ischemia is modest. Cerebral monitoring with INVOS-4100 has a high negative predictive value, but the positive predictive value is low.