Comparison of Hemodynamic Cerebral Ischemia and Microembolic Signals Detected During Carotid Endarterectomy and Carotid Angioplasty
Background and Purpose There has been concern about carotid percutaneous transluminal angioplasty (PTA) carrying a greater risk of cerebral ischemia than carotid endarterectomy. We set out to compare cerebral hemodynamics and microembolization during carotid PTA and CEA.
Methods We used transcranial Doppler to monitor the middle cerebral artery of 28 patients undergoing carotid PTA (n=14) or carotid endarterectomy (CEA) with a shunt (n=14). Each period during which the internal carotid artery was occluded by PTA balloon or by clamp when the shunt was not in place was timed. Individual periods were summated to give a total occlusion time. Ischemic time was defined as the period for which mean middle cerebral artery velocity fell to a third or less of baseline. Microembolic signals were counted during each procedure.
Results CEA resulted in significantly longer individual and total occlusion time than PTA (mean individual occlusion time, seconds), CEA, 168±51; PTA, 20±7; P<.001; mean total occlusion time; CEA, 337±70; PTA, 26±10; P<.001. Ischemic time was also significantly longer during CEA than during PTA (CEA, 165±40; PTA, 17±5; P=.001). There were significantly more microembolic signals during PTA than during CEA (mean number of microembolic signals during CEA, 52±64; during PTA, 202±119; P=.001). There was no correlation between any of the parameters measured and periprocedural stroke, which occurred in one patient in each group.
Conclusion PTA results in less hemodynamic ischemia but more cerebral microembolism than CEA. In this small series, however, it is not possible to comment on the relations between ischemic time, microembolism, and stroke.
There has been considerable interest recently in the role of carotid PTA in the management of carotid stenosis.1 2 3 There is no data available from randomized controlled trials, but published results from series of patients treated with carotid PTA look promising with an overall rate of stroke or death similar to that of surgery.4
One criticism of carotid PTA is that atherosclerotic material dislodged either by the PTA catheter or by the PTA balloon may embolize to the cerebral circulation and result in cerebral ischemia.2 During CEA the ICA is clamped above and below the stenosis. This operative field is flushed several times before releasing the clamps to remove atherosclerotic debris that might embolize to the cerebral circulation. The external carotid artery clamp is usually released before the ICA clamp so that any remaining debris will embolize to the branches of this vessel instead of to vessels distal to the ICA. These precautions are not possible during carotid PTA with a single angioplasty balloon. Furthermore, whereas the atherosclerotic material is removed during CEA, it is cracked and split during PTA and may continue to produce emboli. Another concern is that balloon inflation results in occlusion of the carotid artery with the risk of cerebral ischemia. During CEA, however, temporary ischemia during clamping of the carotid artery can be reversed by the use of a shunt. This option is not possible during carotid PTA.
Earlier studies have shown that microembolic signals characteristic of particulate or gaseous cerebral emboli have been detected during both CEA5 and carotid PTA.6 Several studies have been reported in which TCD has been used to study hemodynamic changes during CEA5 7 8 9 10 ; one study has reported hemodynamic changes during carotid PTA,11 but a direct comparison in similar patients by the same observers has not yet been reported. We therefore set out to compare cerebral hemodynamics and microembolism during carotid surgery with a shunt and during carotid PTA in a randomized study of the procedures.
Materials and Methods
Periprocedural monitoring of MCA blood flow with TCD was attempted in 38 consecutive patients randomized between carotid PTA or CEA. Three patients were excluded because of failure to obtain a good signal with TCD. An additional 7 patients were excluded because a shunt was not used during the operations. Fourteen patients were monitored during carotid PTA and 14 during CEA. All patients had severe (70% to 99%) stenosis as measured by the common carotid method.12 All surgery was performed by one of two consultant vascular surgeons (T.L., R.S.T.) and all carotid PTA by two interventional radiologists (T.B., A.C.) assisting one another. The MCA distal to the stenosis was insonated at a depth of 46 to 56 mm with a sample volume of 10 mm with TCD (EME 2000 or EME Pioneer 4100). The transducer was attached with a head strap throughout the procedure and the Doppler signal recorded onto digital audiotape for subsequent analysis, which was done with the EME Pioneer 4100. All recordings were performed by a single individual (F.C.). Mean MCA velocity at the start of either procedure was considered to be the baseline velocity. Patients were monitored for 20 minutes after CEA or PTA (recovery). This was timed from the end of suturing the skin (CEA) or from removal of the catheter (PTA). A twenty-minute period was selected as patients generally remain in the theater (CEA) or in the angiography room (PTA) for this period of time after the procedures. It was impossible to monitor patients for longer periods because the recording would have had to have been interrupted while the patient was being transferred back to the ward as the TCD equipment requires mains electricity supply.
Occlusion time was defined as the time during CEA that the common carotid artery was clamped and there was no shunt in situ and as the time during carotid PTA that the angioplasty balloon was inflated. Individual occlusion times were summated to obtain a total occlusion time. Ischemic time was defined as the period for which the mean MCA flow velocity was reduced to less than a third of the baseline, either by clamp without shunt, shunt malfunction, excess pressure on the carotid artery, or cardiovascular changes (CEA) or by catheter obstructing the ICA lumen, balloon inflation, or cardiovascular changes (PTA).
Microembolic signals were identified according to previously defined criteria13 by short and usually unidirectional signals visible in the Doppler spectrum with an intensity of 8 db or more above the background signal and accompanied by a characteristic “click.” During PTA, microembolic signals may arise from the catheter, balloon inflation, or contrast injection. Contrast injection results in a shower of microembolic signals that cannot be counted individually.14 Therefore, only microembolic signals that were detected in the MCA while the catheter was being manipulated or during or immediately after balloon deflation were included. It has been suggested that the multiple microembolic signals detected at the time of balloon deflation may be particulate.6 Those associated with movement of the catheter in the absence of recent contrast injection are also likely to be particulate, and those occurring immediately after contrast injection are likely to be air. During CEA, microembolic signals detected during carotid dissection must be particulate as the carotid artery has not been opened. Subsequently, microembolic signals may be particulate or air.
To attempt to compare the number of solid microembolic signals during carotid PTA with the number during CEA, the number of microembolic signals during carotid dissection (CEA) was compared (1) with the number of microembolic signals during and immediately after balloon deflation,(2) with the number of microembolic signals at other times during catheter manipulation, and (3) with the total number of microembolic signals during catheter and manipulation and balloon deflation. The total number of microembolic signals during CEA was compared with the total number of microembolic signals during any period of catheter manipulation and balloon deflation. The number of microembolic signals detected during recovery from CEA or PTA were compared.
Tapes were analyzed blind to the name of the patient and stage of the procedure. It was not possible to be blinded to whether the procedure was carotid PTA or CEA because the TCD recording showed frequent contrast injection during carotid PTA, which did not occur during CEA.
The first 10 recordings of each procedure were reviewed by the same observer on two separate occasions 1 month apart and blind to the name of the patient. The purpose of this procedure was to obtain an estimate of intraobserver variability in tape interpretation. One 2-hour tape that included over 100 potential microembolic signals was reviewed independently by an experienced transcranial Doppler user from another center to confirm that our criteria for the detection of microembolic signals was concordant with that of a second center. All potential microembolic signals from all procedures were reviewed by a second observer from St. George’s Hospital Medical School and recorded as microembolic signals only if both observers from this single center agreed.
Fisher’s exact test was used to compare the presence of vascular risk factors between the two groups. Student’s t test was used to compare the degree of stenosis (treated and contralateral ICA), occlusion times, ischemic times, and microembolic signals detected in the CEA group versus those in the carotid PTA group.
There was no significant difference in the degree of stenosis between the two groups (PTA mean stenosis treated 79%; range, 70% to 90%, median, 80%; CEA mean stenosis treated 81%; range, 70% to 95%, median, 80%) or between the degree of stenosis of the contralateral ICA (PTA mean contralateral stenosis 48%; range, 0% to 100%, median, 50%; CEA mean contralateral stenosis 46%; range, 0% to 100%, median, 55%). The contralateral ICA was occluded in four patients in each group. There was no significant difference between vascular risk factors in each group (Table 1⇓).
Detection of Microembolic Signals
Interobserver agreement in the detection of microembolic signals was 93% and intraobserver agreement was 95%.
The average number of microembolic signals detected during carotid dissection was 15 (range, 0 to 123; SD, 32). This was significantly fewer than the average number of 46 microembolic signals detected during balloon deflation after carotid PTA (range, 12 to 148; SD, 39; t=−2.3; P=.03; Table 2⇓). The average number of microembolic signals detected during carotid dissection was also significantly less than either the number detected during other periods of radiological manipulation (average number of microembolic signals, 156; range, 2 to 325; SD, 93; t=−5.34; P<.001) or the total detected during any period of catheter or balloon manipulation, including balloon deflation (average number of microembolic signals, 202; range 18 to 426; SD, 119; t=−5.67; P<.001).
There were significantly fewer microembolic signals detected during CEA than during carotid PTA (t=−4.15, P=.001) with the total number for CEA being 52 (range, 0 to 269; SD, 64) compared with 202 (range, 18 to 426; SD, 119) for PTA.
The average number of microembolic signals detected during recovery from CEA was 19 (range, 0 to 100; SD, 29). Fewer microembolic signals were detected during recovery from PTA (average, 5; SD, 6; range, 0 to 17), but this difference was not significant.
1. Occlusion Time
The mean duration of individual periods of ICA occlusion in all patients monitored during CEA was 168 seconds (range, 77 to 268; SD, 51; Table 3⇓), giving a total average occlusion time of 337 seconds (range, 223 to 492; SD, 70).
The mean duration of individual balloon inflation during PTA was 20 seconds (range, 14 to 41; SD, 7) and on average the balloon was inflated 1.3 times, giving a mean total occlusion time of 26 seconds (range, 14 to 42; SD, 10).
Mean individual occlusion times during PTA were significantly shorter than mean individual times in patients treated with CEA (t=15.3, P<.001). Mean total occlusion times during PTA were significantly shorter than mean total occlusion times in patients treated with CEA (t=16.6, P<.001).
2. Ischemic Time
During CEA, MCA flow velocity was reduced to less than a third of baseline in 5 patients. The average individual ischemic period was 165 seconds (range, 112 to 209; SD, 40; Table 4⇓). During carotid PTA, MCA flow velocity was reduced to less than a third of baseline in 9 patients for an average of 17 seconds (range, 10 to 27; SD 5). The mean individual ischemic time was significantly less for carotid PTA than for CEA (t=−8.2, P=.001).
One patient in the CEA group (patient C4) had a major stroke 30 minutes after the procedure. The mechanism of this was thrombosis and occlusion of the endarterectomy site. MCA flow velocity did not fall to an ischemic level in this patient and individual and total occlusion time were both close to the mean for the group of patients treated with CEA The number of microembolic signals detected during dissection, and the overall total number was below the mean number for the CEA group. This patient had just above the average number of microembolic signals detected during recovery. The carotid artery was reexplored and patency restored; however, the neurological symptoms did not resolve. A second patient (C5) was found to be dysphasic and with a weak right arm immediately after reversal of anesthesia; the carotid artery was reexplored and was found to be occluded with a fresh thrombus. Similarly, the number of microembolic signals and occlusion periods in this patient were not above average. However, more than an average number of microembolic signals were detected during recovery. The thrombus was removed and the patient had no residual neurological deficit.
One patient (A3) in the PTA group had a major stroke during the procedure. This was caused by rupture of the angioplasty balloon while inflated and subsequent carotid artery spasm. The patient was asystolic for 14 seconds, less than the mean ischemic time in the PTA group. The mean individual occlusion time in this patient was just above the mean for the group as a whole (21 versus 20 seconds), but the total occlusion time was longer (42 versus 26 seconds). The mean MCA blood flow velocity after restoration of sinus rhythm was 23 cm/s, which was not reduced sufficiently from the baseline value (48 cm/s) to be considered ischemic. The mean number of microembolic signals from catheter or balloon manipulation during this PTA was less than the average (137 versus 156) as was the mean number of microembolic signals from balloon deflation (12 versus 46).
This randomized comparison of 28 patients demonstrates that carotid PTA results in significantly less hemodynamic ischemia than CEA, both with respect to occlusion time and ischemic time. However, carotid PTA results in significantly more microembolic signals than CEA during the procedure, though not in the 20 minutes immediately after surgery. This has not been reported previously.
In this small series, there was no correlation between hemodynamic ischemia or microembolic signals detected during either procedure and neurological outcome because only one patient in the group treated with PTA and two patients in the group treated with CEA had a periprocedure stroke or TIA. In contrast to these findings, Eckert et al11 monitored 22 patients undergoing carotid PTA and showed a significant association of neurological disturbance (either TIA or stroke) in patients in whom mean middle cerebral artery blood velocity fell by >50% during balloon inflation. The period of reduced mean blood flow velocity is not stated, but balloon inflation was repeated up to three times for 10 to 40 seconds, and therefore the period of reduced mean blood flow velocity would not have exceeded 120 seconds. None of our PTA patients had occlusion or ischemic times approaching this length. In a series reported by Spencer5 in which 500 CEA operations were monitored with TCD, MCA mean velocity fell to <30% of baseline in 81 patients, all but 3 of whom were shunted. Hypoperfusion was not significantly associated with cerebrovascular complications.
Convincing evidence that microembolic signals detected by TCD predict stroke risk is lacking, although Spencer5 reported a significantly higher number during CEA in patients who suffered cerebrovascular complications. Gaunt et al15 have reported a relation between the detection of persistent microembolic signals during the immediate postoperative period after CEA and both carotid artery thrombosis and the development of major neurological deficits. In their series, three patients developed carotid artery thrombosis in the early postoperative period. During the 30 minutes after CEA, 672 microembolic signals were detected in one of these patients and 157 microembolic signals in another. In our series, two patients developed carotid artery thrombosis associated with neurological deficit after CEA. Both had more microembolic signals detected during recovery than the average for the group treated with CEA but not as many as were detected in the patients studied by Gaunt. Four other patients(C6, C7, C9, C11) also had similar rates of microembolic signals after CEA and two (A6, A9) after PTA. None of these patients developed any neurological deficit, and all carotid arteries were shown to be patent after PTA or CEA. It is our policy to observe patients in whom we continue to detect microembolic signals in the 20 minutes after CEA or PTA and to reintervene only if neurological symptoms develop.
Whether the measured number of microembolic signals recorded during carotid PTA and CEA will lead to significant cerebral damage may depend on the size and nature of the embolic material causing the signal as well as the area of brain to which the material embolizes. It is likely that particulate microembolic signals, particularly those consisting of cholesterol emboli or formed thrombus, may be associated with more cerebral damage than air bubbles or loose platelet aggregates. It is not currently possible to determine the precise nature of individual microembolic signals from parameters such as signal intensity on Doppler because there is wide overlap between different materials.16
It is known that many of the microembolic signals detected during cerebral angiography are gaseous because leaving the contrast to stand after it has been drawn up and injecting it slowly significantly reduces the number of microembolic signals detected.14 In a series of 24 patients undergoing carotid angiography, no correlation was found between microembolic signals detected during the procedure and the risk of periprocedure stroke or new ischemic lesions on magnetic resonance imaging, suggesting that these microembolic signals are not harmful.17 Contrast is injected very frequently during carotid PTA, and many of the microembolic signals detected and included in this series may still have been caused by air instead of particulate matter. It is therefore possible that the finding of significantly more microembolic signals during PTA than CEA may simply reflect microscopic air bubbles introduced during the angiographic sequences. If this is the case, the increased number of microembolic signals associated with carotid PTA may not be clinically significant. To be of value, conclusions about the comparative safety of carotid PTA should be based on randomized studies of clinical outcome. Neurological outcome after carotid PTA has not been compared previously with the detection of microembolic signals, and a larger series of patients would be required to show any association. Neuropsychological outcome has not been reported, but it may prove to be a more subtle test of ischemic insult.
In conclusion, this series has shown that although carotid PTA results in more microembolic signals than CEA, the latter is associated with longer periods of hemodynamic ischemia. The relevance of these findings will only be known when they are compared with neurological and neuropsychological outcome after carotid PTA and CEA. The results of clinical trials addressing these questions are awaited.
Francesca Crawley is funded by a grant from the National Health Service Research and Development Executive. Many thanks to Diana Colquhoun for assistance with Transcranial Doppler.
Selected Abbreviations and Acronyms
|ICA||=||internal carotid artery|
|MCA||=||middle cerebral artery|
|PTA||=||percutaneous transluminal angioplasty|
- Received June 26, 1997.
- Revision received July 29, 1997.
- Accepted September 25, 1997.
- Copyright © 1997 by American Heart Association
McGuiness CL, Burnand KG. Percutaneous transluminal angioplasty of the internal carotid artery. Br J Surg. 1996;1171–1173.
Beebe HG, Archie JP, Baker WH, Barnes RW, Becker GJ, Bernstein EF, Brener B, Clagett GP, Clowes AW, Cooke JP, Creager MA, Cronenwett JL, Dake M, DeWeese JA, Fogarty TJ, Freischlag JA, Goldstone J, Greenfield LJ, Hertzer NR, Hobson RW, Joyce JW, Katzen BT, LoGerfo FW, Mohr JP, Zarins CK. Concern about safety of carotid angioplasty [editorial]. Stroke. 1996;27:197–198.
Brown MM. Carotid and vertebral angioplasty. J Neurol Neurosurg Psychiatry 1997, In press.
Spencer MP. Transcranial Doppler monitoring and causes of stroke from carotid endarterectomy. Stroke. 1997;28:685–691.
Markus HS, Clifton A, Buckenham T, Brown MM. Carotid angioplasty; detection of embolic signals during and after the procedure. Stroke. 1994;25:2403–2406.
Jansen C, Vriens EM, Eikelboom BC, Vermeulen FEE, van Gijn J, Akerstaff RGA. Carotid endarterectomy with transcranial Doppler and electroencephalographic monitoring; a prospective study in 130 operations. Stroke. 1993;24:665–669.
Rothwell PM, Gibson RJ, Slattery J, Warlow CP. Prognostic value and reproducibility of measurements of carotid stenosis. Stroke. 1994;25:2440–2444.
Consensus Committee Criteria of Doppler Microembolic Signals. Basic identification criteria of Doppler microembolic signals. Stoke. 1995;26:1123.
Markus HS, Brown MM. Differentiation between different pathological cerebral embolic materials using transcranial Doppler in an in vitro model. Stroke. 1993;24:1–5.
Gerraty RP, Bowser DN, Infield B, Mitchell PJ, Davis SM. Microemboli during carotid angiography; association with stroke risk factors or subsequent MRI changes. Stroke. 1996;27:1543–1547.