Transcranial Doppler monitoring during carotid endarterectomy: is it appropriate for selecting patients in need of a shunt?

PURPOSE

This report summarizes our experience in evaluating a series of 168 patients who underwent a total of 175 carotid endarterectomy procedures under local anesthesia. Patients were monitored by stump pressure (SP) measurement and transcranial Doppler scanning (TCD). The need for shunting was compared between SP/TCD flow velocity reduction and the awake response (gold standard).

METHODS

The study cohort represented 56% of all the carotid patients treated during the study period. Clamping ischemia was defined as the appearance of focal deficit (focal ischemia) or unconsciousness (global deficit) on carotid clamping. In the case of clamping ischemia, a shunt was inserted. To define the optimal value of SP and TCD flow velocity that is able to discriminate patients with clamping ischemia, a receiver operator characteristic (ROC) curve was constructed. Sensitivity and specificity tests, together with negative and positive predictive values (NPV and PPV), were calculated. Cutoff values were defined as the ROC curve values that correlated the highest sensitivity with the highest specificity for both SP and TCD.

RESULTS

Clamping ischemia was present in 18 procedures (10%) in which a shunt was used. No perioperative deaths were recorded. Major perioperative morbidity occurred in one patient (0.6%). Two nondisabling strokes were also recorded (1.8% overall rate of neurologic morbidity). Cutoff values for both SP and TCD, using the ROC curve, were < or = 50 mm Hg and > or = 70% flow velocity reduction from baseline, respectively. SP values of < or = 50 mm Hg or less showed a sensitivity of 100%, a specificity of 83%, a PPV of 40%, and an NPV of 100%. TCD flow monitoring (> or = 70% flow reduction) revealed a lower sensitivity (83%) but a greater ability to avoid false positive results (96% specificity), resulting in increased PPV (71%) and NPV (98%). Combining SP and TCD failed to provide better results in terms of specificity (81%) and PPV (38%).

CONCLUSIONS

SP measurement using a 50 mm Hg cutoff appears to be a reliable predictor of clamping ischemia but requires the use of a shunt in 17% of the patients who would otherwise not require this procedure. In contrast, TCD has greater specificity but is associated with a lower sensitivity, with 17% false negative results. In our experience, both SP and TCD show limitations, as they overestimate or underestimate carotid endarterectomy procedures in need of a shunt. We believe that sensitivity is more important than specificity in carotid endarterectomy, and thus conclude that TCD flow velocity measurement is not an optimal method for detecting clamping ischemia.