To the Editor:
We read with interest the recent case report by McCabe et al,1 in which they describe a patient with fatal cerebral reperfusion hemorrhage after internal carotid artery (ICA) stenting. In their interesting discussion, the authors argue that this disastrous complication most likely occurred because cerebral perfusion pressure overwhelmed the vasoconstrictive capacity of the arteriolar circulation. Because a brain CT scan performed before carotid stenting disclosed diffuse patchy leukoaraiosis in the deep white matter, the authors suggest that leukoaraiosis could be a risk factor for reperfusion hemorrhage after carotid stenting. We provide an additional case report of reperfusion injury after carotid stenting that supports the main conclusion of McCabe and colleagues and provides some further understanding of this dreadful complication.
On November, 10, 1999, a 43-year-old right handed man was admitted to our stroke unit, 2 days after a sudden onset of left-sided facial and brachial paresis. His past medical record was relevant for a history of hypertension since the age of 20, with no other atherosclerotic risk factors or past cerebrovascular events. On admission, his blood pressure was 150/80 mm Hg and his neurological exam was normal. A brain CT scan showed a 3.5-cm ischemic infarction in the right centrum ovale consistent with acute embolism. An old ischemic lacunar infarction 10 mm in size was also observed in the genu of the right internal capsule. Periventricular or brain stem leukoaraiosis and brain stem lacunar infarctions were absent. A Doppler ultrasound of the cervical vessels suggested >90% stenosis of the right ICA and normal findings on both vertebral arteries and left ICA. The following day a selective injection on the right common carotid artery disclosed 95% stenosis of the ICA, with very slow and faint intracranial filling. A deep ulceration of the stenotic segment was also noted. A selective injection on the left common carotid artery showed marked cross-flow through the circle of Willis, a fetal origin of the posterior communicating artery and patent intracranial vessels bilaterally. The vertebrobasilar circulation lacked abnormal findings, including atherosclerotic changes. During the following week the patient remained asymptomatic, his blood pressure remained within normal values, and 1000-IU/h unfractionated heparin IV was started. One week after the onset of symptoms, a right carotid stenting was planned via a perfemoral approach under local anesthesia. In the early morning of the procedure, the patient was given aspirin (300 mg) and clopidogrel (225 mg). Our antithrombotic protocol includes, during the month after the procedure, the combination of aspirin 300 mg daily and clopidogrel 75 mg daily. An intravenous heparin bolus (5000 IU) was given and heparin was then maintained at 1000 IU/h. In agreement with Guterman and colleagues,2 urokinase (200 000 IU) was infused through a microcatheter proximal to the ulcerated lesion before introduction of the balloon. The stenosis was crossed with a Hannibal 0.014″ J angioplasty wire and 0.7 mg atropine IV was given. The stenosis was predilated with a Bijou balloon (3.5 × 20 mm) and stented with a Carotid Wallstent Monorail (8 × 29 mm long). The stent was further dilated using a 5 × 20-mm RxViaTrac 14 balloon. Hemodynamic and neurological functions were constantly monitored during the procedure by an anesthesiologist. The blood pressure varied between 180/80 mm Hg and 135/75 mm Hg during the procedure. After carotid stenting an angiogram showed the complete recanalization of the right ICA without local complications. An aortic arch injection disclosed a much faster and pronounced filling of the right carotid circulation. A stroke neurologist confirmed the normal neurological state of the patient at the end of the procedure. The patient was transferred to the neurointensive care unit for observation. Over the next 6 hours, the patient was treated with 1000 IU/h unfractionated heparin IV. His blood pressure was recorded every 15 minutes, with values ranging between 100/60 mm Hg and 125/60 mm Hg. He was free of headache and his neurological exam remained completely normal. Suddenly, the patient became confused, and several minutes later he lost consciousness, disclosing no seizure activity. His blood pressure increased to 180/75 mm Hg but returned to normal in the ensuing 5 minutes without hypotensive therapy. He was immediately transferred to the CT scan room, where a large left thalamic hemorrhage was detected that extended into the ventricules and the midbrain. No radiological signs of hemispheral edema were detected. An urgent full blood count was normal, the activated partial thromboplastine time was 66 seconds (equivalent to 0.3 to 0.5 U/L heparin levels), and fibrinogen was 3.7 g/L (normal 1.5 to 4.5 g/L). Seven days later the patient died in the neurointensive care unit. Autopsy was not authorized.
In agreement with McCabe and colleagues, we believe that our patient also illustrates how hyperperfusion hemorrhage can result after carotid stenting in patients with radiological signs of small-vessel disease. The angiogram performed after carotid stenting showed prominent flow increments through previously hypoperfused vessels, which possibly overwhelmed the compensatory capacity of the microcirculation to maintain an adequate control of the cerebral perfusion pressure. The extent to which abrupt flow increments, mechanical stimulation of carotid baroreceptors, or the extent and severity of preceding small-vessel disease played a role in this syndrome is difficult to establish. More careful monitoring of cerebral hemodynamics, through use of methods such as perfusion and difusion MRI or transcranial Doppler, direct measurements of the perfusion pressure, and pathological examination in fatal cases, would be necessary to address these effects. This information could also help in understanding why reperfusion injury after carotid stenting seems to occur at shorter time intervals after the procedure compared with the longer delay (several days) that has been described after carotid endarterectomy. Moreover, larger number of patients with the syndrome would be needed to confirm or refute this clinical presentation.
In the 37 previous carotid stenting procedures performed at our institution, we used the same antithrombotic regime and blood pressure management without hemorrhagic complications. Low-dose urokinase was given in this case because of the ulcerated nature of the stenosis. In agreement with McCabe et al, we excluded the occurrence of bleeding in relation to excessive antithrombotic or fibrinolytic therapy. However, we do not believe that a more aggressive antihypertensive management would have reduced the likelihood of bleeding, as the patient did not disclose postprocedural hypertension. Recently, it was shown that transient hemodynamic changes are extremely frequent during and after carotid stenting.3 However, these changes rarely result in clinical symptoms. On the basis of this small database, it can be argued that small-vessel disease, either as periventricular white matter disease or manifested by lacunar infarction, contributed to the development of hyperperfusion hemorrhage. The effect of chronic hypertension in our patient4 and the effect of aging and tobacco in the patient of McCabe et al5 could have been the most likely predisposing factors. Nevertheless, it remains to be explained why hyperperfusion hemorrhage is not reported more frequently after carotid stenting, because a high prevalence of small-vessel disease can be anticipated in the stented population. Perhaps a retrospective analysis of previous series, such as the one included in CAVATAS, could help to clarify this issue.
Lacking effective treatment for hyperperfusion, hemorrhage prevention is crucial. Although speculative, gradient-echo T2-weighted MR imaging could prove efficacious to detect candidates at greater risk of hyperperfusion complications by providing radiological evidence of previous asymptomatic bleds. The extent of periventricular white matter could also be quantified on MRI. On theoretical grounds but pending further information, a less-aggressive dilatation of stenotic vessels in the presence of concurrent lacunar infarctions or periventricular lucencies could be made hemodynamically safer by allowing a more progressive adaptation of the microcirculation to the new cerebral blood flow state. Meanwhile, we are left with the grief of seeing how preventive therapies turned unexpectedly into devastating complications.
- Copyright © 2000 by American Heart Association
McCabe DJH, Brown MM, Clifton A. Fatal cerebral reperfusion hemorrhage after carotid stenting. Stroke. 1999;30:2483–2486.
Guterman LR, Budny JL, Gibbons KJ, Hopkins LN. Thrombolysis of the cervical internal carotid artery before balloon angioplasty and stent placement: report of two cases. Neurosurgery. 1996;38:620–624.
Quereshi AI, Luft AR, Sharma M, Janardhan V, Lopes DK, Khan J, Guterman LR, Hopkins LN. Frequency and determinants of postprocedural hemodynamic instability after carotid angioplasty and stenting. Stroke. 1999;30:2086–2093.
Chamorro A, Pujol J, Saiz A, Vila N, Vilanova JC, Alday M, Blanc R. Periventricular white matter lucencies in patients with lacunar stroke: a marker of too high or too low blood pressure? Arch Neurol. 1997;54:1284–1288.
Brown MM: Leukoaraiosis. In: Donnan G, Norrving B, Bamford J, Bogousslavsky J, eds. Lacunar and Other Subcortical Infarctions. New York, NY: Oxford University Press; 1995:181–198.
We read with interest the additional case of intracerebral hemorrhage reported by Chamorro et al supporting the suggestion that this is an important complication of carotid stenting. Both our patients shared a number of clinical features, including very severe symptomatic pretreatment stenosis, aspirin therapy, transient hypertension during the procedure, small-vessel disease, and a stable neurological course for 6 to 7 hours after carotid stenting before sudden clinical deterioration. However, their patient was also treated with a high dose of clopidogrel (225 mg), intra-arterial urokinase, and therapeutic heparinization, which may have contributed to the intracranial hemorrhage in their patient. It is also interesting to note that the hemorrhage was in the contralateral thalamus and midbrain in their patient, in contrast to the ipsilateral location in our patient.