Initial experience with TCD as the sole determinant for shunting during CEA under general anesthesia.

Shung Lee MD, Jim Dooner MD, Violet Hung, Victoria BC

Purpose: To describe our experience with TCD in our initial 100 patients and to summarize the literature regarding accuracy for predicting the need for shunting during CEA under GA (sensitivity, specificity, PPV, NPV and ROC characteristics).

Background:  Although many surgeons perform CEA under locoregional anesthesia, a significant number of CEA are performed under GA. To prevent cerebral ischemia during carotid clamping, routine shunting is an acceptable technique in this circumstance. However, shunts will be unnecessary in up to 90% of patients. In addition to adding time and cost to the procedure, shunts have been implicated as a cause of carotid dissection and intraoperative air and particulate emboli. Thus, selective shunting is desirable when performing CEA under GA. The three modalities that are commonly utilized are distal ICA stump pressures, EEG, and TCD. The literature suggests that none of these modalities offers 100% accuracy in determining the use of shunt during carotid occlusion. Fortunately, the risk of a false negative with either modality is extremely small and in fact may be acceptable when weighed against the potential morbidity of a policy of routine shunting.

Results: We reviewed our first 100 patients utilizing a policy of shunting if the MCA velocity dropped to <50% of baseline upon clamping of the carotid bifurcation. Indications for surgery were symptomatic carotid disease in 65% and asymptomatic carotid disease in 35%. The MCA velocity profile pre and post anesthesia, during dissection, at cross-clamp, and with or without shunting are reported. The percentage of patients shunted is reported. The percentage of ipsilateral perioperative neurologic events was <2% with none of the events occurring during the immediate post-operative period suggesting that intraoperative cerebral ischemia was not responsible for these cerebrovascular events

Conclusion: In our experience, we believe that TCD for determining selective shunting under general anesthesia is safe and accurate resulting in lower utilization of shunts. Given the low percentage of patients who require shunting when CEA is performed under locoregional anesthesia and the low incidence of stroke during CEA, a large number of procedures would need to be completed in order to validate TCD as a sole determinant for the need to shunt during CEA. In this regard, the literature is not conclusive. Further study of TCD in awake patients undergoing CEA is warranted.