Background and Purpose-We sought to assess the clinical value of regional cerebral saturation (rSO 2 ) obtained by means of the cerebral oximeter INVOS 3100A (Somanetics) in comparison to monitoring of somatosensory evoked potentials (SEP) for the reliable detection of severe cerebral ischemia requiring shunt placement in the individual patient undergoing carotid surgery under general anesthesia. Methods-In 317 patients undergoing reconstructive surgery on the internal carotid artery, simultaneous recordings of SEP and rSO 2 were obtained throughout the operation. Results-All 287 patients with preserved cortical SEP remained neurologically intact. Shunt placement was performed in 27 patients (9%) after flattening of cortical SEP during cross-clamping of the internal carotid artery. A stable rSO 2 value just before cross-clamping and the lowest value after cross-clamping were registered, and the decrease was calculated. A statistically significant (PϽ0.01) decrease of rSO 2 after cross-clamping could be found in patients without (64.9Ϯ8.3% to 60.9Ϯ9.9%) as well as in patients with consecutive loss of cortical SEP (65.8Ϯ9.1% to 56.1Ϯ13.4%).The difference of the decrease of rSO 2 in both groups was highly significant (6.9Ϯ9.0% versus 15.6Ϯ14.0%; PϽ0.001).However, substantial interindividual variability of rSO 2 and derived change of rSO 2 did not allow the definition of a threshold value indicating need of shunt placement. Conclusions-The

Reversing heparin with protamine reduces postoperative wound drainage after carotid surgery but may predispose to ICA thrombosis and stroke. This is in keeping with a previous retrospective study published during our trial.

Transcranial Doppler ultrasonography (TCD) of the middle cerebral artery (MCA), light-reflective cerebral oximetry and measurement of internal carotid artery stump pressure were compared as methods of monitoring cerebral perfusion during carotid surgery in 33 patients. Median cerebral oxygen saturation was 70 (range 62-85) per cent and TCD-measured mean blood velocity 42 (range 19-91) cm/s before carotid cross-clamping, falling to 68 (53-83) per cent and 16 (0-50) cm/s respectively on application of the clamps (P < 0.001). Stump pressure correlated closely with MCA blood velocity 30 s after the start of cross-clamping (rs = 0.58, P < 0.001), but not with cerebral oxygen saturation. A fall of 5 per cent or more in cerebral oxygen saturation following cross-clamp application was predicted by a decrease in mean MCA blood velocity of at least 60 per cent. Changes in cerebral oxygen saturation correlated significantly with systolic blood pressure throughout the perioperative period (rs = 0.41, P < 0.001). Significant falls in cerebral oxygenation were not predicted by low stump pressure but were associated with large reductions in the mean MCA blood velocity measured by TCD.

A novel instrument using reflected near-infra-red light spectroscopy to measure cerebral oxygen saturation non-invasively was evaluated during carotid endarterectomy; cerebral perfusion was compared with jugular bulb venous oxygen saturation and transcranial Doppler ultrasonographic measurements. Initially, oximetry sensors with light source-detector separation distances of 10 and 27 mm were positioned over the frontal area, while a cannula positioned in the jugular bulb permitted sampling for jugular bulb venous oxygen saturation. To increase cerebral oxygen saturation sensitivity, modified sensors with light source-detector separation distances of 30 and 40 mm were relocated over the middle cerebral artery territory. The changes in cerebral and jugular bulb venous oxygen saturation, and in peak blood flow velocity before and 30 s after carotid clamping and declamping were recorded. The modified cerebral perfusion system achieved improved correlations between cerebral and jugular bulb venous oxygen saturation changes during carotid clamping and declamping (r = 0.92, P < 0.001). The correlation between change in cerebral oxygen saturation and the percentage change in peak flow velocity on both cross-clamping and declamping was equally strong (r = 0.90, P < 0.001). Near-infra-red cerebral spectroscopy reliably detects changes in cerebral oxygen saturation during carotid endarterectomy and may have wide applications in monitoring brain perfusion during neurosurgery and cardiopulmonary bypass surgery, and in closed head injury.

The falls in TCD and CsO(2) were of a similar order of magnitude and must therefore reflect a fall in cerebral perfusion. The ipsilateral ECA contributes significantly to intracranial blood flow and oxygen saturation in severe carotid stenosis.