EEG changes during awake carotid endarterectomy

Cost-Effectiveness of Two Decision Strategies for Shunt Use During Carotid Endarterectomy

Background

Arterial shunting during carotid endarterectomy (CEA) is essential in some patients because of insufficient cerebral perfusion during cross-clamping. However, the optimal diagnostic modality identifying these patients is still debated. None of the currently used modalities has been proved superior to another. The aim of this study was to assess the cost-effectiveness of two modalities, stump pressure measurement (SPM) versus electroencephalography (EEG) combined with transcranial Doppler (TCD) during CEA.

Methods

Two retrospective cohorts of consecutive patients undergoing CEA with different intraoperative neuromonitoring strategies (SPM vs. EEG/TCD) were analyzed. Clinical data were collected from patient hospital records. Primary clinical outcome was in-hospital stroke or death. Total admission costs were calculated based on volumes of healthcare resources. Analyses of effects and costs were adjusted for clinical differences between patients by means of a propensity score, and cost-effectiveness was estimated.

Results

A total of 503 (239 SPM; 264 EEG/TCD) patients were included, of whom 19 sustained a stroke or died during admission (3.3 vs. 4.2%, respectively, adjusted risk difference 1.3% (95% CI −2.3–4.8%)). Median total costs were €4946 (IQR 4424–6173) in the SPM group versus €7447 (IQR 6890–8675) in the EEG/TCD group. Costs for neurophysiologic assessments were the main determinant for the difference.

Conclusions

Given the evidence provided by this small retrospective study, SPM would be the favored strategy for intraoperative neuromonitoring if cost-effectiveness was taken into account when deciding which strategy to adopt.

Somatosensory Evoked Potentials and Electroencephalography during Carotid Endarterectomy Predict Late Stroke but not Death

BACKGROND

Late stroke and death rates are anticipated to be higher in patients undergoing carotid endarterectomy (CEA) compared with healthy counterparts. However, little is known regarding predictors other than the baseline comorbidities. We have recently shown that dual intraoperative somatosensory evoked potentials (SSEPs) and electroencephalography (EEG) monitoring improves the ability to predict perioperative strokes. We seek to determine if dual intraoperative monitoring (IOM) can further predict long-term strokes and death.

METHODS

Consecutive patients who underwent CEA under dual SSEP and EEG IOM between January 1, 2000 and December 31, 2010 were analyzed. Patients were divided in 2 groups, those with and those without IOM changes. IOM changes were classified as either occurring during carotid cross-clamp placement or at any time during the operation. End points were time to stroke and death. Log-rank tests and Cox regression analysis were used to identify predictors of postoperative stroke and death.

RESULTS

A total of 853 CEAs (mean age 70.6 ± 9.5 years, 58.7% male, 38.9% symptomatic) were performed during the study period with a mean clinical follow-up of 48 ± 38 months. One hundred seven patients (13.6%) had significant SSEP or EEG changes at the time of clamping, while considerably more patients (217, 25.4%) had SSEP and/or EEG changes recorded at any point during the procedure, including clamping. Baseline characteristics including rates of bilateral disease, statin use, and antiplatelet use were similar between groups. Female gender, symptomatic disease, and significant contralateral carotid stenosis were more frequent in the group with IOM changes. The overall stroke-free survival rate at 5 years was significantly higher in patients without IOM changes (94.7% vs. 88.2%, P < 0.05) and at 10 years (86.1% vs. 78.0%, P < 0.05). Despite differences in stroke-free survival, overall survival at 10 years was not different between groups (44.0% in patients with IOM changes vs. 42.8% in patients without, P = 0.7). Renal insufficiency (hazards ratio [HR] 2.13, P = 0.03), diabetes (HR 1.84, P = 005), and age > 80 at the time of operation (HR 3.24, P = 0.001) were significant predictors of late stroke, while statins were significantly protective (HR 0.55, P = 0.05). Controlling for these factors, IOM changes (HR 2.5, P = 0.004) were a strong predictor of long-term risk of stroke after CEA.

CONCLUSION

Intraoperative SSEP and/or EEG changes are predictive of late stroke but not death following CEA indicating a need for further elucidation and management of the underlying risk factors driving the elevated stroke risk in this subset of CEA patients.

Selective shunt during carotid endarterectomy using routine awake test with respect to a lower shunt rate

PURPOSE

To evaluate shunt rate and discuss the resultsrelated to selective shunt placement during carotid endarterectomy (CEA) using routine awake test.

METHODS

Patients with CEA from 2007 to 2011 were retrospectively reviewed from prospectively collected data. The need for shunt placement was determined by the awake test, based on the alteration in the neurologic examination. We collected data by using the clinical records and imaging studies, and investigated factors related to selective shunt such as collateral circulation and contralateral internal carotid artery (ICA) stenosis.

RESULTS

There were 45 CEAs under regional anesthesia with the awake test in 44 patients. The mean age was 61.8 ± 7.1 years old. There were 82.2% (37/45) of males, and 68.9% (31/45) of symptomatic patients. Selective shunt placement had been performed in only two (4.4%) patients. Among them fewer cases (4%) had severe (stenosis >70%) contralateral ICA lesions, and more cases (91%) of complete morphology of the anterior or posterior circulation in the circle of Willis. There was no perioperative stroke, myocardial infarctionor death, and asymptomatic new brain lesions were detected in 4 patients (9%), including 2 cases of selective shunt placement.

CONCLUSION

CEA under routine awake test could besafe and feasible method with low shunt placement rate in selected patients.

Neurophysiologic monitoring in neurosurgery

This article focuses on the application of neurophysiologic monitoring in uniquely neurosurgical procedures. Neurophysiologic monitoring provides functional testing and mapping to identify neural structures. Once identified, the functionality of the central and peripheral nervous system areas at risk for neurosurgical injury can be monitored. It discusses the use of motor-evoked potentials, sensory evoked potentials, electromyography and electroencephalography to assess neurologic change.

Prospective randomized trial of routine versus selective shunting in carotid endarterectomy based on stump pressure

BACKGROUND

The use of shunting in carotid endarterectomy (CEA) is controversial. This randomized trial compared the results of routine (RS) vs selective shunting (SS) based on stump pressure (SP).

METHODS

Two-hundred CEA patients under general anesthesia were randomized into RS (98 patients) or SS (102 patients), where shunting was used only if systolic SP (SSP) was <40 mm Hg. Clinical and demographic characteristics were comparable in both groups. Patients underwent immediate and 30-day postoperative duplex ultrasound follow-up. Analysis was by intention-to-treat.

RESULTS

Of 102 SS patients, 29 (28%) received shunting. Indications for CEA were similar (42% symptomatic for RS; 47% for SS, P = .458). The mean internal carotid artery diameter was comparable (5.5 vs 5.5 mm, P = .685). Mean preoperative ipsilateral and contralateral stenosis was 76% and 38% for RS (P = .268) vs 78% and 40% for SS (P = .528). Mean preoperative ipsilateral and contralateral stenosis was 79% and 56% in the shunted (P = .634) vs 78% and 34% in the nonshunted subgroup of SS patients (P = .002). The mean SSP was 55.9 mm Hg in RS vs 56.2 for SS (P = .915). The mean SSP was 33 mm Hg in the shunted vs 65 in the nonshunted subgroup (P < .0001). Mean clamp time in the nonshunted subgroup of SS was 32 minutes. Mean shunt time was 35 minutes in RS and 33 in SS (P = .354). Mean operative time was 113 minutes for RS and 109 for SS (P = .252), and 111 minutes in shunted and 108 in the nonshunted subgroup (P = .586). Mean arteriotomy length was 4.4 cm for RS and 4.2 for SS (P = .213). Perioperative stroke rate was 0% for RS vs 2% for SS (one major and one minor stroke, both related to carotid thrombosis; P = .498). No patients died perioperatively. Combined perioperative transient ischemic attack (TIA) and stroke rates were 2% in RS vs 2.9% in SS (P > .99). The overall perioperative complication rates were 8.3% in RS (2 TIA, 3 hemorrhage, 1 myocardial infarction [MI], and 1 asymptomatic carotid thrombosis) vs 7.8% in SS (2 strokes, 1 TIA, 3 hemorrhage, 1 MI, and 1 congestive heart failure; P = .917).

CONCLUSIONS

RS and SS were associated with a low stroke rate. Both methods are acceptable, and surgeons should select the method with which they are more comfortable.

Implementation of an artificial neuronal network to predict shunt necessity in carotid surgery

In carotid surgery, it could be useful to know which patient will tolerate carotid cross-clamping in order to minimize the risks of perioperative strokes. In this clinical study, an artificial neuronal network (ANN) was applied and compared with conventional statistical methods to assess the value of various parameters to predict shunt necessity. Eight hundred and fifty patients undergoing carotid endarterectomy for a high-grade internal carotid artery stenosis under local anesthesia were analyzed regarding shunt necessity using a standard feed-forward, backpropagation ANN (NeuroSolutions); NeuroDimensions, Gainesville, FL) with three layers (one input layer, one hidden layer, one output layer). Among the input neurons, preoperative clinical (n = 9) and intraoperative hemodynamic (n = 3) parameters were examined separately. The accuracy of prediction was compared to the results of a regression analysis using the same variables. In 173 patients (20%) a shunt was used because hemispheric deficits or unconsciousness occurred during cross-clamping. With the ANN, not needing a shunt was predicted by preoperative and intraoperative parameters with an accuracy of 96% and 91%, respectively, where the regression analysis showed an accuracy of 98% and 96%, respectively. Those patients who needed a shunt were identified by preoperative parameters in 9% and by intraoperative parameters in 56% when the ANN was used. Regression analysis predicted shunt use correctly in 10% using preoperative parameters and 41% using intraoperative parameters. Intraoperative hemodynamic parameters are more suitable than preoperative parameters to indicate shunt necessity where the application of an ANN provides slightly better results compared to regression analysis. However, the overall accuracy is too low to renounce perioperative neuromonitoring methods like local anesthesia.

Evaluation of Changes of Systemic Blood Pressure and Shunt Incidence in CEA

BACKGROUND

Induced hypertension is widely recommended as a protective measure in carotid endarterectomy (CEA) to prevent shunt insertion. In this study changes of systemic blood pressure were evaluated in relation to the shunt rate when CEA was performed under local anaesthesia.

MATERIALS AND METHODS

In 930 CEAs performed for a high-grade (>70%) ICA stenosis under local anaesthesia the mean systemic blood pressure was measured preoperatively (RR1) and directly before carotid cross-clamping (RR2). A ratio was calculated from these values (RRR=RR2/RR1). A shunt was only inserted for clinical signs of cerebral ischemia. If that became necessary later after cross-clamping had been tolerated primarily, the blood pressure during this period was also recorded (RR3). Also the presence of a contralateral ICA occlusion and baseline blood pressure levels were considered as factors with potential impact on shunt necessity.

RESULTS

Among the 638 male (69%) and 292 female (31%) patients with a median age of 70 years (ranging from 52 to 91 years) 82 (9%) had a contralateral ICA occlusion. A shunt was used in 177 operations (19%) and significantly more frequent in patients with a contralateral ICA occlusion (39/82=48% vs. 138/848=16%, p<0,001). RRR was significantly reduced in patients who needed a shunt (0.95 (0.41-1.53) vs. 1.0 (0.54-1.9), p=0.002) which was only true for patients with a patent contralateral ICA. The shunt rate did not differ when contrasting RRR thresholds (<0.7 vs. >1.3) or preoperative blood pressure levels (<100 mmHg vs. >120 mmHg) were compared. RRR did not differ between directly or delayed shunted patients. RR3 did not differ significantly from RR2. A regression analysis identified the presence of a contralateral ICA occlusion as the only independent parameter influencing shunt insertion.

CONCLUSIONS

Changes in systemic blood pressure during CEA under local anaesthesia seem to influence shunting rather marginally. The value of induced hypertension to prevent cerebral ischemia should be newly discussed.

Electrophysiologic monitoring in neurosurgery

Electrophysiologic techniques have become common in the neurosurgical operating room. This article reviews the methods used for mapping neural structures or monitoring during surgery. Mapping methods allow identification of target structures for surgery, or for identifying structures to allow avoidance or plot safe pathways to deeper structures. Monitoring methods allow for surgery on nearby structures to warn of encroachment, thereby reducing unwanted injury.

Prospective evaluation of electroencephalography, carotid artery stump pressure, and neurologic changes during 314 consecutive carotid endarterectomies performed in awake patients

OBJECTIVE

This study attempted to correlate neurologic changes in awake patients undergoing carotid endarterectomy (CEA) under cervical block anesthesia (CBA) with electroencephalography (EEG) and measurement of carotid artery stump pressure (SP).

METHODS

Continuous EEG and SP monitoring was measured prospectively in 314 consecutive patients undergoing CEA between April 1, 2003, and July 30, 2006, under CBA. Indications for CEA were asymptomatic 70% to 99% internal carotid artery stenosis in 242 (77.1%), transient ischemic attacks (including transient monocular blindness) in 45 (14.3%), and prior stroke in 27 (8.6%). Mean common carotid artery pressure before clamping, mean SP after carotid clamping, and intraarterial pressure were continuously monitored in all patients. An indwelling shunt was placed when neurologic events (contralateral motor weakness, aphasia, loss of consciousness, or seizures) occurred, regardless of SP or EEG changes.

RESULTS

Shunt placement was necessary because of neurologic changes in 10% (32/314) of all CEAs performed under CBA. Only 3 patients (1.4%) of 216 required shunt placement if SP was 50 mm Hg or more, vs 29 (29.6%) of 98 if SP was less than 50 mm Hg (P < .00001; sensitivity, 29.8%; specificity, 98.6%). In patients with SP of 40 mm Hg or more, 7 (2.6%) of 270 required shunt placement, vs 25 (56.8%) of 44 if SP was less than 40 mm Hg (P < .00001; sensitivity, 56.8%; specificity, 97.4%). Ischemic EEG changes were observed in 19 (59.4%) of 32 patients (false-negative rate, 40.6%) requiring shunt placement under CBA. Three patients had false-positive EEG results and did not require shunt placement (false-positive rate, 1.0%). The perioperative stroke/death rate was 4 (1.2%) in 314. All strokes occurred after surgery and were unrelated to cerebral ischemia or lack of shunt placement.

CONCLUSIONS

Ten percent of patients required a shunt placement during CEA under CBA. Shunt placement was necessary in 56.8% of patients with SP less than 40 mm Hg. EEG identified cerebral ischemia in only 59.4% of patients needing shunt placement, with a false-positive rate of 1.0% and a false-negative rate of 40.6%. Both SP and EEG as a guide to shunt placement have poor sensitivity. Intraoperative monitoring of the awake patients under regional anesthesia (CBA) is the most sensitive and specific method to identify patients requiring shunt placement.

Regional or general anesthesia for carotid endarterectomy? Evidence from published prospective and retrospective studies

OBJECTIVES

The aim of this study was to compare outcomes of patients undergoing carotid endarterectomy under regional or general anesthesia for any new neurologic impairment, stroke, stroke and/or death, death, and myocardial infarction.

DESIGN

Meta-analysis.

SETTING

A search of the National Library of Medicine of the United States PUBMED from 1966 up to June 11, 2005, with the following key words: "carotid surgery or endarterectomy and regional anesthesia."

MEASUREMENTS AND MAIN RESULTS

Forty-eight studies (14 prospective and 34 retrospective) were analyzed. The optimal information size was achieved only when all studies were analyzed together (prospective and retrospective). Regional anesthesia reduced the rate of any new neurologic impairment (odds ratio = 0.60; 95% confidence interval, 0.48-0.75; p < 0.00001), stroke (0.54 [0.43-0.68], p < 0.00001), stroke and/or death (0.62 [0.49-0.78], p < 0.0001), death (0.65 [0.48-0.87], p = 0.004), and myocardial infarction (0.50 [0.36-0.70], p < 0.0001).

CONCLUSION

The number of patients included in randomized controlled trials or even in prospective studies is too low to allow any conclusions on the differences in outcome between the 2 anesthetic techniques. Better outcomes are suggested when results from retrospective studies are added.

Patient selection for carotid stenting versus endarterectomy: A systematic review

Carotid artery stenting has emerged as an alternative to carotid endarterectomy for the treatment of severe extracranial carotid stenosis in patients with anatomic or clinical factors that increase their risk of complications with surgery, yet there remains a substantial amount of variability and uncertainty in clinical practice in the referral of patients for stenting vs endarterectomy. By undertaking a thorough review of the literature, we sought to better define which subsets of patients with "high-risk" features would be likely to preferentially benefit from carotid stenting or carotid endarterectomy. Although only a single randomized trial comparing the outcomes of carotid stenting with distal protection and endarterectomy has been completed, a wealth of observational data was reviewed. Relative to endarterectomy, the results of carotid stenting seem favorable in the setting of several anatomic conditions that render surgery technically difficult, such as restenosis after prior endarterectomy, prior radical neck surgery, and previous radiation therapy involving the neck. The results of stenting are also favorable among patients with severe concomitant cardiac disease. Carotid endarterectomy, alternatively, seems to represent the procedure of choice among patients 80 years of age or older in the absence of other high-risk features. Overall, existing data support the concept that carotid stenting and endarterectomy represent complementary rather than competing modes of therapy. Pending the availability of randomized trial data to help guide procedural selection, which is likely many years away, an objective understanding of existing data is valuable to help select the optimal mode of revascularization therapy for patients with severe carotid artery disease who are at heightened surgical risk.

Correlation of carotid artery stump pressure and neurologic changes during 474 carotid endarterectomies performed in awake patients

PURPOSE

A carotid artery stump pressure (SP) of < 50 mm Hg and abnormal electroencephalography (EEG) changes have been suggested as indications for selective shunting in patients undergoing carotid endarterectomy (CEA) under general anesthesia. We attempted to determine the optimal SP threshold that correlated with neurologic changes in awake patients undergoing CEA using cervical block anesthesia (CBA) and performed a cost comparison with EEG monitoring.

METHODS

Between July 1, 1995, and December 31, 2004, SP was measured during 474 CEAs performed under CBA by inserting a 19-gauge butterfly needle into the common carotid artery. A saline-filled intravenous bag in the patient's contralateral hand was connected to pressure tubing to generate waveforms with hand squeezing that could be visualized on a monitor. Systemic pressure was maintained approximately 10 mm Hg higher than baseline. Accurate SPs were confirmed by the finding of flatline waveforms after internal carotid artery clamping. Selective shunting was performed when neurologic changes occurred (aphasia, inability to squeeze the contralateral hand, decreased consciousness), regardless of SP. During this same period, 142 patients underwent CEA using GA, and SP was also measured.

RESULTS

Shunting was necessary because of neurologic changes in 7.2% (34/474) of all CEAs performed using CBA: 0.9% (3/335) with SPs > or = 50 mm Hg systolic vs 1.0% (4/402) with SPs > or = 40 mm Hg systolic, and 22% (31/139) with SPs < 50 mm Hg systolic vs 42% (30/72) with SPs < 40 mm Hg systolic. If these 474 CEAs had been performed using GA, shunts would have been used in 29% (139/474) of patients for a SP < 50 mm Hg systolic vs 15% (72/474) for a SP < 40 mm Hg systolic. In patients not shunted, the perioperative stroke/death rate was 1.2% in patients (4/332) with SPs > or = 50 mm Hg vs 1.0% (4/398) with SPs > or = 40 mm Hg. Three of the four strokes occurred > 24 hours postoperatively and were unrelated to lack of shunting and ischemia. There was no significant difference in the percentage of patients with SPs > or = 50 mm Hg who underwent CEA using CBA (70%, 335/474) vs GA (67%, 96/142) during this time period. At our hospital, charges for SPe measurement, including anesthesia charges and tubing, were 229 dollars per case vs 3439 dollars per case for EEG monitoring. Use of SP measurements in these 474 patients would have resulted in reduced charges of 1,521,540 dollars compared with EEG monitoring if CEA had been performed under GA.

CONCLUSION

Using 40 mm Hg systolic as a threshold, the need for shunting (15%) and the false-negative rate (1.0%) for SP in our series were equivalent to the results of EEG monitoring during CEA reported in the literature. However, charges for SP measurements are dramatically lower compared with EEG monitoring. Our results suggest that a carotid artery SP > or = 40 mm Hg systolic may be considered as an equally reliable but more cost-effective method to predict the need for carotid shunting during CEA under GA compared with EEG monitoring, but further investigation is warranted.

Bispectral index monitoring may not reliably indicate cerebral ischaemia during awake carotid endarterectomy † †This study was conducted in the Department of Neurosurgery, Newcastle General Hospital, Newcastle Upon Tyne NE4 6BE, UK. Preliminary results were presented at the British Neurosurgery Research Group Meeting, Sheffield, March 29–30, 2001

BACKGROUND

Intraoperative ischaemia during carotid cross-clamping in patients undergoing carotid endarterectomy (CEA) is a major complication and prompt recognition of insufficient collateral blood supply is crucial. Electroencephalogram (EEG) is believed to be one of the useful forms of monitoring cerebrovascular insufficiency during CEA. The aim of this study was to evaluate the utility of bispectral index (BIS) monitoring, a processed EEG parameter, for the reliable detection of intraoperative cerebral ischaemia during awake CEA.

METHODS

We monitored 52 patients continuously with the BIS monitor together with assessment of neurological function (contralateral upper and lower limb strength and the verbal component of the Glasgow Coma Scale for speech) in patients undergoing awake CEA.

RESULTS

Overall mean BIS value in all patients was 96 (SD 2.9). In five patients who showed clinical evidence of cortical ischaemia during carotid cross-clamping, there was no change in the original range of BIS values throughout the procedure (96.7 [3.2]). In one patient BIS values decreased to 38 about 5 min after the incision and recovered within the next 10 min. The mean BIS value in the remaining 46 patients who did not develop clinical signs of ischaemia was 95.4 (2.6). Three cases are presented which demonstrate the inability of the BIS monitor to detect cerebral ischaemia.

CONCLUSIONS

Lack of correlation of BIS with the signs of cerebral ischaemia during CEA makes it unreliable for detection of cerebrovascular insufficiency. We conclude that awake neurological testing is the preferred method of monitoring in these patients.

Intraoperative management: carotid endarterectomies

The management of anesthesia for patients undergoing carotid endarterectomy is challenging and dynamic. Effective management and good outcome requires the anesthesiologist's understanding of cerebral physiology, knowledge of neck anatomy, and understanding of the rapid pathophysiologic changes that occur during carotid artery manipulations. The anesthesiologist must be flexible in the management of patients, who frequently have underlying multiorgan pathology and cardiovascular compromise. Good communication between the anesthetic and surgical teams is needed to avoid irreversible debilitating consequences for the patient.

Intraoperative monitoring during vascular surgery

The principal objectives of intraoperative monitoring are to improve perioperative outcome, facilitate surgery and reduce adverse events, using continuously collected data of cardiopulmonary,neurologic and metabolic function to guide pharmacologic and physiologic therapy. Although sophisticated and reliable apparatus may be used to collect these data they are useless, or even harmful, without proper interpretation. This article provides a comprehensive overview of recent publications on the history,philosophy, and semantics of monitoring.

Prediction of Cerebral Ischemic Tolerance During Carotid Cross-clamping by Angiographic Criteria

OBJECTIVE

To evaluate the prognostic value of angiographic criteria and internal carotid artery (ICA) stump pressure for shunt placement in carotid surgery under local anaesthesia.

DESIGN

Prospective clinical trial at a university hospital.

MATERIALS AND METHODS

In 120 patients a cerebral angiography was initiated before undergoing carotid surgery. Seventy-five percent of the patients were neurologically asymptomatic, 13% had transient ischaemic attacks and 12% suffered from previous strokes. The operation was exclusively performed under local anaesthesia and prior to cross-clamping the ICA stump-pressure was measured. A shunt was inserted only if hemispheric symptoms or unconsciousness occurred independent of the angiographic findings or stump pressure.

RESULTS

In 23% (27/120) a shunt became necessary and significantly (p<0.001) more often when there was a cross-flow towards the contralateral hemisphere (12/20=60%) or if the contralateral ICA was occluded (9/13=69%). The sensitivity for not needing a shunt in case of cross-flow towards the side of operation was 91% (52/57) whereas the specificity was 35% (22/63). ICA stump-pressure was significantly reduced in patients requiring a shunt (31 mmHg) compared to those not needing a shunt (53 mmHg) (p<0.001), but no definitive threshold value was found determining the need for shunting. Intraoperatively, no persistent neurological complication developed.

CONCLUSIONS

Angiographic cross-flow was a good, but not perfect predictor for the need of an intraoperative shunt.

Carotid endarterectomy then and now: outcome and cost-effectiveness of modern practice

BACKGROUND

During the past decade, our practice of performing carotid endarterectomy (CEA) has changed dramatically, most notably by an abrupt shift from routine to selective preoperative angiography, reliance on defined care plans with full-time nurse practitioner oversight, and increasing reliance on eversion endarterectomy and cervical block anesthesia. This study was designed to determine whether these shifts in policy have been associated with lower costs without sacrificing clinical outcome.

METHODS

All patients undergoing CEA from July 1993 to December 2000 were identified, and inpatient and outpatient charts were reviewed. Cost data were obtained from the central hospital accounting system and converted to 2001 dollars. Thirty-day outcomes and costs were quantified each year and compared between each of 2 temporally well-defined groups: those undergoing "routine" versus "selective" angiography and those cared for before and after defined patient care protocols were instituted.

RESULTS

A total of 1168 CEAs were analyzed. Thirty-day combined stroke and death rate was 3.1%, and no trends or significant differences over time were seen. From 1993 to 2000 the cost of CEA fell from $9302 to $6216 (P<.0002), and length of stay was reduced 1 full day (P=.005). Institution of "selective" angiography was associated with an immediate cost savings of approximately $2000 per case (P<.0001), and nurse practitioner oversight along with institution of defined clinical protocols with a $530 (P<.05) decline in nonoperating room-related costs.

CONCLUSIONS

Changes in policy from routine to selective angiography, reliance on defined postoperative care pathways, eversion endarterectomy, and cervical block anesthesia have been associated with significant cost savings, with no compromise in clinical outcome at our institution.