Technique and clinical results of carotid stump back-pressure to determine selective shunting during carotid endarterectomy

Technique and clinical results of carotid stump back-pressure to determine selective shunting during carotid endarterectomy Joseph P. Archie, Jr., PhD, MD, Raleigh, N.C. A method of confirming carotid back pressure accuracy, variability during carotid damping, and the clinical results with a modified back pressure shunt criterion were evaluated in 665 carotid endarterectomies. Mean arterial pressure, back pressure, and internal jugular vein pressure were measured. Cerebral perfusion pressure (back pressure - jugular vein pressure) and the collateral to hemisphere vascular resistance ratio, (ratio = [arterial pressure- back pressure]/[baek pressure- jugular vein pressure]) were calculated. A shunt was used when cerebral perfusion pressure < 18 mm Hg. Back pressure accuracy was confirmed by test occlusion of the internal carotid artery distal to the plaque. Initial back pressure values were falsely high in 83 (12.5%) carotid endarterectomies. The mean SD (n = 665, nun Hg) were arterial pressure = 84. -+ 9.6, back pressure = 41. -+ 15.9, jugular vein pressure = 6.2 -+ 3.9, cerebral perfusion pressure = 35.1 ± 5.7, and resistance ratio = 1.85 -+ 1.44. Perfusion pressure was <18 mm Hg in 82 (12.3%), of which 74 (11.1%) were shunted, and 8 (1.2%) had perfusion pressure increased -> 18 mm Hg during carotid endarterectomy with phenylephrine. Back pressure was <25 mm Hg in 17 (16.1%), -<25 in 114 (17.1%), and <5 mm Hg in 481 (72.3%). Pressures were continuously monitored during 28 carotid endarterectomies, and all had a positive linear relationship between arterial pressure and back pressure, and minimal variability in the back pressure/arterial pressure and resistance ratios. Only two patients (.3%) had a new nettrologic deficit in the first 12 hours after carotid endarterectomy. Unless confirmation techniques are used, erroneously high carotid stump back pressure measurements may occur in 1% to 15% of carotid endarterectomies, resulting in failure to use a shunt in some depending on the pressure criterion used. Patients with cerebral perfusion pressure ~> 18 rnm Hg and back pressure ->25 mm Hg can safely undergo carotid endarterectomy without a shunt. (J VASC SURG 1991;13:319-27.) A shunt is the standard method of protection from cerebral ischemia during carotid endarterectomy (CEA). Since a shunt is necessary to prevent irreversible neurologic cell damage in less than 2% and perhaps only 1% ofceas, 1'2 most routine shunting and most selective shunting may be unnecessary. How best to identify the small number of patients who need a shunt remains undetermined. With improved CEA results and the current gold standard of less than 2% perioperative permanent neurologic deficits, it is clearly important to place a shunt in the few patients at risk. One of the most popular and simplest techniques of selective shunt determination From Wake Medical Center, Raleigh, and The University of North Carolina, Chapel Hill, N.C. Presented at the Thirty-eighth Scientific Meeting of the North American Chapter, International Society for Cardiovascular Surgery, Los Angeles, Calif., June 4-6, i99. Reprint requests: Joseph P. Archie, Jr., MD, 3417 Williamsborough Ct., Raleigh, NC 2769. 24/6/25766 is the intraoperative Carotid back or stump pressure method initially described by Moore and Hall 3 and Moore et al. 4 However, this method continues to be controversial and nonstandardized 2 years later. Many conflicting reports exist concerning the correlation between the three commonly used shunt determination methods: carotid back pressure, electroencephalogram, and neurologic status during regional anesthesia, s-~ Further, carotid back pressure is reported to be widely variable in some patients independent of changes in systemic arterial pressure. 11 This is a report of the clinical resuks of a large consecutive series of CEAs performed with selective shunting determined by a modified carotid back pressure criterion. It also identifies errors in back pressure measurement technique that may produce falsely elevated values, and evaluates the within patient variability of carotid back pressure during the period of occlusion for CEA. 319

32 Archie Journal of VASCULAR SURGERY METHODS AND PATIENTS During the period 1982 to 1989, 672 consecutive CEAs were performed by the author at two community hospitals. All but three were done with the patient under general anesthesia. Six hundred sixtyfive (98.9%) of these were submitted to a selective shunt criterion based on carotid back pressure and are the subject of this report. In the other seven CEA cases the pressure criterion was not applicable for the following reasons: the inability to confirm back pressure accuracy in three, an occluded internal carotid artery at CEA in two (thrombosis after arteriogram but reopened at operation), an open distal internal carotid artery not communicating directly with the common carotid artery in one, and severe symptomatic fibromuscular dysplasia and atherosclerosis precluding the confirmation technique in one. No deaths or morbidity occurred in these seven patients, but they are not included in the statistical analysis of the results. The 665 CEAs (365 men, 3 women) were performed on 575 patients (318 men, 257 women). The age was 65.9 -+- 8.8 (mean + 1 SD) years (65.1 men, 67. women) with a range of 39 to 9 years. Carotid endarterectomy was done concomitantly with coronary artery bypass in 37 patients and with abdominal aortic surgery in three. There were 648 primary and 17 reoperations for recurrent stenosis. Patients with proximal common carotid artery occlusion or stenosis requiring bypass were not included, nor were patients undergoing only common carotid and/or external CEA. Associated diseases and risk factors were coronary artery disease in 398 (69%) patients, hypertension in 36 (63%), and diabetes in 81 (14%). Standard or arterial digital two-view arch arteriography was performed in all patients with selective carotid views in most. The percent diameter stenosis was calculated by use of the equation: (1 - minimum internal or common carotid diameter/normal distal internal carotid diameter) 1. On the operated side this was 72.9% -+ 29.5% (mean + 1 SD) (74.6% men, 7.9% women), and on the nonoperated side it was 3.7% +_ 38.2% (3.7% men, 28.9% women). Fifty-eight (8.7%) patients had contralateral internal carotid artery occlusion. Indications for CEA were transient ischemic attack in i88 (28.3%), amaurosis fugax in 96 (14.4%), reversible ischemic neurologic deficit in 2 (3.%), completed stroke in 98 (14.7%), nonlateralizing global ischemia in 7 (1.5%), and asymptomatic (greater than 75% diameter stenosis) in 193 (29.%). Of the asymptomatic subset, 86 were staged operations after contralateral CEA for symptomatic disease in 71 and asymptomatic stenosis in 15. The global ischemic subset was older (7.5 years, p <.5) than the other indications for CEA subsets by analysis of variance (ANOVA). This subset also had a higher percent contralateral carotid stenosis (55.8%, p <.1). The global ischemic and asymptomatic subsets had higher percent operated side carotid stenosis (88.2% and 86.5%, respectively, p <.1) than the other subsets. All patients had either a radial or a brachial artery cannula placed for continuous arterial pressure monitoring during the operation and early postoperative period. Neurologic examination was performed by the author on all patients in the recovery room and subsequently in the intensive care unit. Systemic arterial blood pressure was maintained in each patient's normal range when necessary during and after operation by continuous infusion of nitroglycerine and/or nitroprusside for hypertension or with phenylephrine (Neosynephrine) for hypotension. Standard surgical techniques were used including systemic heparin, careful carotid dissection before occlusion and during pressure measurements, optical magnification, meticulous end point management, and liberal use of patch reconstruction. End point tacking sutures were rarely necessary. Doppler ultrasound confirmation of the adequacy of reconstruction was performed during operation in all patients. Patch reconstruction with autogenous greater saphenous vein was used in 55 (75.9%) operations (484 full patches, 21 partial patches). When adequate saphenous vein was not available, polytetrafluoroethylene patches (Gore-Tex, registered trademark of W.L. Gore & Assoc., Elkton, Md.) were used in 18 (2.7%) and Dacron in 16 (2.4%) operations. Interposition reversed autogenous greater saphenous vein bypass from the common carotid bulb to the internal carotid artery was used for reconstruction in eight (1.2%) patients because of the excessive length of disease in the internal carotid artery or poor conclifton of the internal carotid artery after endarterectomy. After exposure of the carotid arteries, a 22-gauge needle connected to a calibrated strain-gauge transducer system (with on-line monitoring of the pressure waveform and digital pressure readout) was used to measure mean internal jugular vein pressure. The needle was then placed in the common carotid artery proximal to the bulb and the significantly diseased segment, and the mean systemic arterial pressure was measured. Elastic loops were routinely used to oc-

Volume 13 Number 2 February 1991 Carotid stump back-pressure and carotid endarterectomy 321 clude the superior thyroid and external carotid arteries, and either a loop or vascular forceps were used to occlude the common carotid artery proximal to the needle to produce the back or stump pressure. Confirmation of the back pressure accuracy was then performed by test occlusion of the internal carotid artery distal to the plaque. Confirmed back pressure waveform patterns are characterized by complete damping of the waveform without a rise in mean pressure, whereas falsely elevated back pressures are readily identified by incomplete or absence of damping of the back pressure waveform and/or an increase in mean pressure as illustrated in Fig. 1. If the back pressure was not confirmed, a search was made to determine if the cause was incomplete occlusion of the common carotid artery, usually because of plaque, or of the external carotid artery, because of,,~ early or posterior branch not included in the vessel loop. After correction accurate back pressure was confirmed by repeating the above steps. The observed numeric errors in mean back pressure measurements were not routinely recorded but ranged from 5 to 5 mm Hg above true values. Ipsilateral collateral cerebral perfusion pressure was calculated by subtracting jugular venous pressure from carotid back pressure. 12,13 In some patients with low cerebral perfusion pressure and low arterial pressure, phenylephrine was used to increase arterial pressure (and thus cerebral perfusion pressure) to or above the 18 mm Hg shunt threshold criterion? 4 If cerebral perfusion pressure was less than 18 mrn Hg a shunt was used. When patients with marginal perfusion pressures (18 < cerebral perfusion pressure < 25 mm Hg) had a decrease in arterial pressure,uring CEA, phenylephrine was used to keep cerebral perfusion pressure >18 mm Hg. The ratio of ipsilateral collateral to hemisphere cerebral vascular resistance was calculated, R = (Pa - Pc) / (Pc - Pv) (where Pa = mean arterial pressure; Pc = mean carotid back pressure; Pv = mean internal jugular vein pressure). 1~ In 28 patients in which a shunt was not used (cerebral perfusion pressure -> 18 mm Hg) the carotid back pressure was continuously measured during the 15- to 3-minute occlusion required for CEA by inserting the small end of a Javid shunt into the internal carotid artery and connecting it to an online pressure monitor system. Mean systemic arterial pressure and mean back pressure were simultaneously recorded at 2-minute intervals. Since carotid back pressure is known to vary with arterial pressure, the variability of cartoid back pressure and cerebral per- A B D : I. i i m a Time Fig. 1. Schematic of carotid back pressure waveforms. The dashed vertical time line is the onset of confirmation test occlusion of the internal carotid artery. A confirmed accurate back pressure waveform is given by A. Three commonly identified waveforms associated with incomplete occlusion of the common (B) or external carotid arteries (C, D) and a falsely high back pressure are also shown. fusion pressure independent of arterial pressure during CEA was analyzed by calculating the ratio of carotid back pressure to arterial pressure and the resistance ratio. All patient and numeric data were stored prospectively in a computer registry. Analysis of variance (A_NOVA) and paired and unpaired t tests were performed by use of commercial software programs (BMDP, Statistical Software Inc., Los Angeles, Calif.). RESULTS Of the 665 CEAs, 82 (12.3%) had a confirmed collateral cerebral perfusion pressure less than 18 mrn Hg. Eight of these had cerebral perfusion pressure increased to ->18 mm Hg with phenylephrine and were not shunted. The remaining 74 (11.1%) had a shunt placed during CEA. The cerebral perfusion pressure < 18 mm Hg shunt protocol was violated twice (2/665,.3%) (cerebral perfusion pressure = 32 and 25 mm Hg) when the patient moved with seizure-type activity under general anesthesia 2 and 4 minutes after carotid occlusion, and a shunt was placed. Prior experience with this observation sug-

322 Archie Journal of VASCULAR SURGERY Table I. The number and frequency of 665 CEAs meeting various shunt pressure criteria Cerebral perfusion pressure Carotid back pressure Carotid back pressure >25 mm Hg and cerebral perfusion pressure <18 mmrtg Carotid back pressure ->25 mm Hg and cerebral perfusion pressure <18 mm Hg Shunt criteria No. (%) <18 mm Hg 82 (12.3%) ~" <25 mm Hg 172 (26.2%) <25 mm Hg 17 (16.1%) --<25 mm Hg 114 (17.1%) <3 mm Fig 183 (27.5%) <5 mm Hg 481 (72.3%) 4 (.6%) 6 (.9%) ~74 (11.1%) after increasing systemic arterial pressure and cerebral perfusion pressure. gests that cerebral ischemia could be the cause. 16 Both of these patients had had a prior ipsilateral stroke, and neither developed a new neurologic deficit. Contralateral internal carotid occlusion or a previous ipsilateral stroke did not influence shunt use. Seventeen patients with initially confirmed 25 > cerebral perfusion pressure - 18 mm Hg had arterial pressure maintained or increased with phenylephrine. The number of patients meeting the shunt criterion used in this report, cerebral perfusion pressure < 18 mm Hg, as well as the number meeting other frequently used back pressure shunt criteria are given in Table I. Initial confirmation of an accurate carotid back pressure was noted in 582 of the 665 CEAs (87.5%). The causes of falsely elevated back pressure, when identified by pressure waveforms (Fig. I) are given in Table II. In the 28 patients with arterial pressure and carotid back pressure continuously monitored during CEA, the absolute percent variability (mean + 1 SD) between 2-minute sequential measurements (n = 317) for the ratio carotid back pressure/arterial pressure was 3.2% + 3.4%, and for the resistance ratio was 6.6% + 6.9%. During CEA carotid back pressure and cerebral perfusion pressure change very little independent of arterial pressure. For the 71 pairs of arterial pressure and carotid back pressure with the same arterial pressure value the absolute variability of carotid back pressure was 2.9% + 3.5%. Finally, the variability between the initial and the final carotid back pressure/arterial pressure and resistance ratios (m 1 SD, n = 28) were -1.7% _+ 3.7% and 5.4% + 9.6%, respectively, not significantly different from the null hypothesis Table II. Confirmation of accuracy of carotid back pressure measurements before CEA No. (%) Initial confirmation 582 (87.5%) Initial falsely elevated pressure 83 (12.5%) Etiology of error Incomplete common carotid 31 (4.7%) occlusion Incomplete external carotid oc- 47 (7.1%) clusion Internal carotid branch 2 (.3%) Undetermined but reconfirmed 3 (.4% ) by paired t testing. Fig. 2 illustrates in four patients the temporal variability of arterial pressure and carotid back pressure during CEA. These four were chosen because of the wide range of arterial pressm~" during CEA and a between patient range of carotid back pressure values. The mean systemic arterial pressure was 84. 14.1 mm Hg (mean _+ 1 SD, n -- 665), the mean carotid back pressure 4.1 _ 15.8 nun Hg, the mean jugular venous pressure 6.2 _ 3.9 mm Hg, the mean cerebral perfusion pressure 35.1 + 15.7 mm Hg, and the resistance ratio value 1.85 1.94. The frequency distribution of cerebral perfusion pressure and carotid back pressure is given in Fig. 3. Mean carotid back pressure was 42.6 mm Hg for men and 39. mm Hg for women (p <.5), and mean cerebral perfusion pressure for men was 36.9 mm Hg and for women 32.8 mm Hg (p <.1). Thus resistance rates was higher in women than men, 2.9 versus 1.65 (p <.5). The arterial pressure, ca-.... ~ rond back pressure, internal jugular vein pressure'; and cerebral perfusion pressure values were similar and not significantly different for the six indications for CEA by ANOVA. However, the resistance ratio value was higher in the subset of 7 CEAs performed for nonlateralizing global ischemic symptoms (mean resistance ratio = 2.58,p <.5 by ANOVA). This may be expected because of the significantly higher degree of contralateral carotid stenosis in this group. The 58 patients with contralateral internal carotid occlusion had a lower carotid back pressure and cerebral perfusion pressure (mean 3.2 and 23.9 mm Hg, p <.1) than the others. This group also had higher probability of shunt use by our criteria, 21 of 58 (36.2%) versus 53 of 67 (8.7%). Also, in the subset of 58 patients with contralateral internal carotid occlusion there was a significant difference in the resistance ratio, for the 16 asymptomatic and 42 symptomatic patients (2.33 vs 3.56, p <.1), a

Voktme 13 Number 2 February 1991 Carotid stump back-pressure and carotid endarterectomy 323 u 4 8 1216 2 24 28 32 36 4 44 48 52 56 6 64 68 72 76 8 84 88 92 961 " 4 8 12 16 2 24 28 32 36 4 44 48 52 56 6 64 68 72 76 8 84 88 92 961 Pressure, mmhg Fig. 2. The frequency distribution of carotid back pressure and cerebral perfusion pressure by 4 mm Hg intervals for 665 CEAs. finding similar to that previously reported, is The mean arterial pressure, carotid back pressure, internal jugular vein pressure, and cerebral perfusion pressure values (mm Hg) for the 76 shunted and 589 nonshunted CEAs were 81.6 and 84.3, 18., and 43.9, 7.2 and 6.1, and 11.8 and 38., respectively. There were significant differences in carotid occlusion times for the patch and primary closure groups and for the shunt and no shunt groups (Table III). The occlusion time (rain:see) for shunt placement was 1 : 34 +.39 and for shunt removal was 1:48-1:33. The clinical results are given in Table IV. Of the four strokes, two reversible ischemic neurologic deficits and transient ischemic attacks that occurred in the hospital, only two were present in the first 12 hours after operation, and both were identified in the recovery room after the patient awoke from general anesthesia. One was total contralateral hemiplegia, the patient immediately returned to the operating room, and a thrombosed internal carotid artery with a primarily closed long arteriotomy was reopened and ~atched. The patient was not improved and died 2 weeks later of the stroke. The other early neurologic deficit was contralateral hand muscle weakness that cleared. This patient had a normal postoperative duplex scan. The other three hospital strokes occurred 12 to 72 hours after CEA. One was due to internal carotid artery thrombosis after a reoperation for recurrent symptomatic high-grade stenoses. The other two patients had normal duplex scans. One patient undergoing a combined CEA and coronary artery bypass with mild disease in the contralateral internal carotid artery had a contralateral cerebral ischemic stroke 4 days after surgery (not included in Table IV). DISCUSSION The currently used methods of determining when to selectively use a shunt during CEA include neurologic examination after carotid artery occlusion under regional anesthesia, electroencephalographic (EEG) monitoring, carotid stump back pressure measurement, and occasionally regional cerebral blood flow measurement, s6,17-2 Contradictory and confusing information exists between both shunt se-

324 Archie Jottmal of VASCULAR SURGERY 15 I,i @ 1-5- ++...~+ ~+' +'8., " +,,+,,,.~nl.. ii.b [] [] O ~8 I I I I 5 7 9 11 13 15 Systemic Arterial Pressure, Pa (mmhg) Fig. 3. Sequentially measured mean systemic arterial pressure and carotid back pressure in four patients during CEA. Table III. Carotid endarterectomy occlusion and shunt times for primary closure and patch reconstruction (min: sec, mean - 1SD) Total Primary closure Patch (n = 665) (n = 118) (n = 547) No shunt 27:1 _+ 6:53 21:5 --- 6:2 28:32 --- 6:46 589 11 488 Shunt 35:12 8:48 26:27 --- 4:8 36:3 --- 8:37 76 17 59 p <.1 for all combinations of shunt, no shunt, patch, and primary closure except shunt-primary closure versus no shunt patch by ANOVA. Interposition operations are included in patch group. Shunt times include ischemic times for placement and removal of shunt. lection methods and the shunt criterion tbr each method. For example, the frequency with which a shunt was used with the neurologic evaluation under regional anesthesia method was 9% to 21%, s-: with EEG monitoring it was 1% to 29%, :7 and with carotid back pressure or cerebral perfusion pressure it was 1% to 5%),+,:3,2o When carotid back pressure and neurologic examination under regional anesthesia were compared in the same patients, the resuits were conflictings: 9 more frequently than they were in agreement, a~'9'm Granader et al. s placed shunts in 21% of their CEAs based on neurologic examination under regional anesthesia and found a high correlation with back pressure for shunt use when carotid back pressure < 25 mm Hg and shunt nonuse when carotid back pressure > 5 mm Hg, but conflicting results when 5 > carotid back pres- sure > 25 mm Hg. Others found conflicting results for carotid back pressure > 5 mm Hg with neurologic examination 9 and EEG.S One study suggested that all patients with carotid back pressure values less than 7 mm Hg be shunted. 6 Some of these studies point out high arterial pressure values in patients who have new neurologic deficits after CEA, an invalid argument unless the source of the ischemia is identiffed. Even more confusing is the report of Ferguson 7 who performed 12 CEAs with patients under general anesthesia without a shunt and one hat,, a stroke. He found that 34% had EEG changes as well as a poor correlation with carotid back pressure and EEG when compared to regional cerebral blood flow measured with 133-xenon 133. Clearly, much of the conflict may be due to liberal shunt criterion by all methods. A shunt may be unnecessary in 9% or more of patients selected for shunting by all selective shunt methods and criterion. The stroke rate of a pooled series of 2964 CEAs performed without a shunt was 1%, perhaps only half of which were interoperative. 2 No method or criterion has been established to identify the very small percentage of patients who need a shunt to prevent permanent neurologic damage during CEA. However, it may not be advisable to search for a highly specific method such as lowering the carotid back pressure or perfusion pressure criterion to 1 to 12 mm Hg. First, if such a criterion were used some patients would probably have significant transient cerebral ischemia during carotid occlusion that could produce subtle nondetectable permanent neurologic damage. Sec-

Volume 13 Number 2 February 1991 Carotid stump back-pressure and carotid endarterectomy 325 Table IV. Mortality and morbidity for 665 CEAs Posthospital Hospital (3 day) Total Death Stroke 1 1 Cardiac 5 4 9 Total 6 (.9%) 4 (.6%) 1 (1.5%) Stroke (survival) 3 (.4%) I (.2%) 4 (.6%) Total death and stroke 9 (1.4%) 5 (.8%) 14 (2.1%) (95% confidence) (.5-2.4%) (.1-1.4%) (1.1-3.2%) RIND 2 2 4 TIA 2 1 3 Myocardial infarction 3 1 4 (survival) Patch rupture ~ 2 1 3 Reoperation for hemorrhage 4 1 5 RIND, Reversible ischernic neurologic deficit; TIA, transient ischemic attack. ~AII female veins, one immediate in OR, the two others also included in reoperation for hemorrhage, all survived with one RIND included in RIND total. ond, even very experienced surgeons would have little oppommity to maintain their skill and judgment in shunt use. Some of the problems with the carotid back pressure method is related to methodology and criteria. Some authors use systolic back pressure 2 rather than mean, 3,4,12'1s but many do not report which is used in their studies. Mean back pressures below 3 mm Hg have severely damped waveforms mad systolic and mean values are within 1 to 5 mm Hg of each other. When mean pressure is >3 mm Hg there may be a 5 to 25 mm Hg difference between systolic and mean back pressure. Further, the zero reference level for pressure measurement relative to the patient's position on the operating table is rarely mentioned. This is very important since an inaccurate zero pressure _)ference may produce up to an 18 mm Hg error, particularly if the patient is not in the supine position) 2'13 One advantage of using the cerebral perfusion pressure modification is that jugular venous pressure is subtracted from carotid back pressure thereby nulling any zero reference error. For example, if the zero reference error is "X" mm Hg then cerebral perfusion pressure = back pressure plus "X" minus internal jugular vein pressure plus "X" = back pressure minus internal jugular vein pressure provided the same needle-manometer system is used to measure both pressures. Pressure criterion for shunt selection is a source of confusion. The original recommendation by Moore and HalP to use a shunt when carotid back pressure -< 25 mm Hg is strongly supported by our study and others. 4"1 '21 Some authors recommend shunt use when carotid back pressure < 5 mm H A 892 ~,,, or carotid back pressure > 7 mm t-ig, 6 ~r carotid back pressure is completely unreliable, s This study illustrates the potential for error in carotid back pressure measurements. The 12.5% rate of falsely elevated carotid back pressure values on initial unconfirmed measurement will probably vary with the technique of vessel occlusion. Careful and gentle handling and occlusion of the carotid segments so as not to produce emboli may account for some incomplete occlusions, particulary those caused by common carotid artery plaque. However, unless confirmation methods such as those described herein are used, one cannot be confident that the measured value is accurate. This may explain in part the higher average carotid back pressure values reported in other studies, as well as some of the conflicting correlations between carotid back pressure and other methods of selective shunt determination. Another criticism of the selective shunt method is within patient variability of repeated carotid back pressure measurements. 11 In a previous study we found repeated carotid back pressure values to vary with mean systemic arterial pressure in a predictable and linear manner. 14 The minimal within-patient variability in the carotid back pressure/arterial pressure and resistance ratios of the 28 patients in whom arterial pressure and carotid back pressure were continuously measured during CEA should put to rest the question of carotid back pressure variability independent of arterial pressure variability. Further, pre CEA carotid back pressure and cerebral perfusion pressure values arc reliable predictors of arterial pressure and cerebral perfusion pressure during CEA. The clinical results of this study stro@y support the use of low carotid back pressure and cerebral perfusion pressure values for selective shunt criterion. Only one of the 489 (.2%) CEA done without a shunt (cerebral perfusion pressure--- 18 mm Hg)

326 Archie Journal of VASCULAR SURGERY had a new unexplained neurologic deficit immediately after operation. This criterion yielded a 11.1% shunt rate. If the carotid back pressure < 25 mm Hg criterion had been used, s'4 the shtmt rate would have been 16% (Table I), and only four (.6%) CEAs shunted by the cerebral perfusion pressure < 18 mm Hg criterion would not have shunted by the carotid back pressure < 25 mm Hg criterion. Further, it is interesting that the mean jugular venous pressure was 6.2 mm Hg. Thus a mean carotid back pressure of 25 mm Hg predicts a mean cerebral perfusion pressure of 19 mm Hg, illustrating that the two shunt criteria, carotid back pressure < 25 mm Hg and cerebral perfusion pressure < 18 mm Hg, are quite similar. The advantage of cerebral perfusion pressure over carotid back pressure is that carotid back pressure removes the zero pressure reference level problem, and it identifies patients with high cerebral venous pressure who might otherwise not be shunted by the carotid back pressure criterion. This study supports the use of carotid back pressure and its modification, cerebral perfusion pressure, as safe and effective selective shunt selection techniques to protect patients from inadequate collateral cerebral blood flow that can cause clinically significant neurologic injury during CEA. Patients with carotid back pressure - 25 mm Hg and/or cerebral perfusion pressure -> 18 nun Hg can safely be operated on without the use of a shunt. The use of back pressure alone should be acceptable provided the zero reference is accurate and there is no evidence of high venous pressure. REFERENCES 1. Baker WH, Littooy FN, Hayes AC, Dorner DB, Stubbs D. Carotid endarterectomy without a shunt: the control series. J VASC SURG 1984;1:5-6. 2. Ferguson GG. Intra-operative monitoring and internal shunts: are they necessary in carotid endarterectomy? Stroke 1982;13:287-9. 3. Moore WS, Hall AD. Carotid artery back pressure. A test of cerebral tolerance to temporary carotid occlusion. Arch Surg 1969;99:72-1. 4. Moore WS, Yee JM, Hall Ad. Collateral cerebral blood pressure. An index of tolerance to temporary carotid ocdusion. Arch Surg 1973;16:52-3. 5. Kelly JJ, Callow AD, O'Donnell TF, et al. Failure of carotid stump pressure. Its incidence as a predictor for a temporary shunt during carotid endarterectomy. Arch Surg 1979;114: 1361-6. 6. Hobson RW, Wright CB, Sublett JW, Fedde CW, Rich NM. Carotid artery back pressure and endarterectomy under regional anesthesia. Arch Surg 1974;19:682-7. 7. Ferguson GG. Extracranial carotid artery surgery. Clin Neurosurg 1982;29:573-4. 8. Gnanader DA, Wang N, Comunale FL, Relic DA. Carotid artery stump pressure: how reliable is it in predicting the need for a shunt? Ann Vase Surg 1989;3:313-7. 9. Kwaan HM, Peterson GJ, Cormolly JE. Stump pressure. An unreliable guide for shunting during carotid endarterectomy. Arch Surg 198;115:183-5. 1. Hafner CD, Evans WE. Carotid endarterectomy with local anesthesia: results and advantages. J VASC SURG 1988;7: 232-9. i1. Beebe HG, Start C, Slack D. Carotid artery stump pressure: its variability when measured serially. I Cardiovasc Surg 1989;3:419-23. 12. Archie JP. Hemodynamics of carotid back pressure and cerebral flow during endarterectomy. J Surg Res 1977;23:223-32. 13. Archie JP, Feldtman RW. Determinants of cerebral perfusion pressure during carotid endarterectomy. Arch Surg 1982; 117:319-22. 14. Archie JP, Feldtman RW. Linear response of collateral cerebral perfusion pressure during carotid clamping. J Surg Res 1989;46:253-5. 15. Archie JP, Feldtman RW. Collateral cerebral vascular resistance in patients with significant carotid stenosis. Stroke 1982;13:829-31. 16. Archie JP. Patient movement after carotid clamping: an indication of low cerebral perfusion pressure. A1 J Med Sci 198;I7:1-1. 17. Boysen G. Cerebral blood flow measurement as a safeguard during carotid endarterectomy. Stroke 197I;2:i-1. 18. Sundt TM. The ischemic tolerance of neural tissue and the need for monitoring and selective shunting during carotid endarterectomy. Stroke 1983;14:93-8. I9. Chiappa KH, Burke SR, Young RR. Results of electroen -~ cephalographic monitoring during 367 carotid endarterectomies: use of a dedicated minicomputer. Stroke 1979;1: 381-8. 2. Hayes RJ, Levinson SA, Wylie EJ. Intraoperative measurement of carotid back pressure as a guide to operative management for carotid endarterectomy. Surgery 1972;72:953-6. 21. Smith LL, Jacobson JG, Hinshaw DB. Correlation of neurologic complications and pressure measurements during carotid endarterectomy. Surg Gynecol Obstet 1976;143: 233-6. DISCUSSION Dr. Wesley Moore (Los Angeles, Calif.). Twenty-one years ago, at a meeting of this Society, we introduced the concept that internal carotid artery back pressure measurement was a means of identifying those patients who would require an internal shunt. Since that time, there have been at least as many papers arguing against the validity of this concept as there have been those that have supported it.

Volume 13 Number 2 February 1991 Carotid stump back-pressure and carotid endarterectomy 327 Dr. Archie presented 665 CEAs performed for the range of indications for CEA. He has added his own important modification to back pressure criteria by introducing the concept of true cerebral perfusion pressure as that being equal to internal carotid artery back pressure minus jugular venous pressure, and he used a threshold of 18 mm Hg or less as an indication for shunt use unless the pressure could be artificially raised by vasopressor agents. By use of these methods 74 out of 665 CEAs, or approximately ] 1%, had shunts placed, and the balance were operated on without the encumbrance of an internal shunt. A further important contribution to the study was the monitoring of internal carotid artery back pressure continually in 28 patients and the correlation of internal carotid back pressure with systemic arterial pressure. In these 28, a minimum variability was found between back pressure and systemic arterial pressure, and it introduced an error of no more than 1 to 2 mm Hg. We think that this will ~o a long way to answer the critics of the method, who worry that internal carotid artery back pressure is a onetime only pressure measurement and may vary significantly during the course of carotid clamping mad endarterectomy. Dr. Archie has demonstrated that as long as the systemic mean arterial pressure is monitored and maintained, we can indeed be assured that the internal carotid artery back pressure will be similarly maintained. The bottom line to this study is the fact that there were only two new neurologic deficits within the first 12 hours of reoperation for an incidence of.3%. One of these was easily explained by a perioperative thrombus. The second had an open reconstruction without apparent explanation. Dr. Archie also has an enviable 3-day result with a mortality of 1.5% and a stroke rate of.6% giving him a combined operative morbidity and mortality rate of 2.1%. Dr. Archie indicates an important note of caution, which deserves our attention. His damping method, which includes vessel loop control of the external carotid artery and of the common carotid artery yielded an initial back pressure error rate in aa of 865 operations for an incidence of 12.5%. We believe that this error can be overcome by using arterial clamps on both the common and external carotid arteries as in our original paper. In using this method, there should be no leak and therefore no reason to carry out reconfirmation procedures. Nonetheless, if others use this method, it is important to recognize that it is an important source of error and may lead to erroneous results such as those reported in the literature. In closing, I would like to ask Dr. Archie four questions. Number one, would you consider modifying your damp technique from the vessel loop and forceps to formal vascular damping to eliminate the initial back pressure error? Number two, can you explain why you have obtained such good results in patients who had as their indication for operation prior stroke in the use of selective shunting? In our own series this represents an exception. Number three, if there appears to be direct correlation between cerebral perfijsion pressure and internal carotid artery back pressure, why not simply make one measurement rather than measuring jugular venous pressure to achieve a correction? Finally, in the one patient who awoke with neurologic defidt, what was the indication for operation in that patient? What was the value of the internal carotid back pressure? Was a shunt used? And, in your opinion, what was the mechanism for the complication? Dr. Arehie (Raleigh, N.C.). Thank you Dr. Moore. All of us who use your method are rtuly indebted to you for pioneering it. Dr. Moore asked four questions. The first question regards the incidence of falsely elevated back pressures. This is clearly going to depend on the technique one uses to occlude the common and e:{ternal carotid arteries. If you use a clamp rather than vessel loops or a vascular forceps as I do, you probably have a higher chance of obtaining a complete occlusion. All of you have experienced bleeding from the proximal common carotid artery because of incomplete clamp occlusion because of plaque disease. The same mechanism produces falsely elevated back pressure. The more aggressive one is with the occlusion technique, the lower the inddence of falsely elevated back pressure. In trying to be gentle with the carotid artery and carotid bulb during these measurements I may produce more incomplete occlusions than someone who uses a vascular clamp on the common carotid artery. The external carotid branches that come off posterior are also a problem. I have difficulty finding them, and if they are not occluded during back pressure measurement they will elevate the pressure to some degree. This is why I believe it is important to use a confirmation technique, identify problems, correct them, and then confirm the pressure measurements. Dr. Moore asked about the subset of patients who had preoperative strokes and the good results in this series. I cannot explain why we had such good results with this group of patients. I did not deviate from my shunt protocol except twice, and that was not because of contralateral internal carotid occlusion or previous stroke. The third question was why not just use carotid back pressure rather than cerebral perfusion pressure? I think this depends on the shunt criteria one selects. The few patients that will be missed with a back pressure criteria of 25 mm Hg are the ones with a venous pressure greater than 8 or 9 mm big. Clearly, one can have a patient with a back pressure of 3 mm Hg, a venous pressure of 2, and thus a perfusion pressure of 1 mm Hg, which is far below the adequate level for cerebral protection during the period of carotid damping. Finally, Dr. Moore asked about the mechanism of stroke in the patient who had an unidentified neurologic event. I do not know what the cause was. A duplex scan in the recovery room was normal. I suppose the most likely explanation is intraoperative embolization.