Vasospasm and Thrombus Formation as Possible Mechanisms of Stroke Related to Alkaloidal Cocaine
Background “Crack” cocaine (alkaloidal cocaine) induces ischemic stroke. However, the mechanisms by which this occurs are not well documented in humans. We present pertinent information on three patients whose ischemic strokes involved the territory of the internal carotid artery and were associated with crack use.
Case Descriptions These patients were investigated clinically, radiologically, intraoperatively, and/or histopathologically at the same institution, and the diagnostic evaluations did not reveal a definite cardiac or hematologic cause for stroke. Large filling defects were noted on conventional carotid angiography in two of these patients; in the third patient, the histopathological changes were compatible with vasospasm. To our knowledge, these changes have not been previously documented in human arteries.
Conclusions We suggest that some brain infarcts among crack cocaine users may result from vasospasm of large arteries and secondary intravascular thrombosis.
Crack cocaine induces ischemic stroke.1 The mechanisms by which this occurs are not clearly defined and probably are multifactorial.2 We present three patients with ischemic strokes associated with exposure to crack cocaine and show evidence that large-vessel vasospasm and clot formation may have occurred.
A 35-year-old African-American woman had an abrupt onset of left frontal supraorbital sharp, piercing headache approximately 1 hour after smoking crack cocaine. Minutes after the headache onset she noted heaviness of her left limbs; the feeling was more marked in the arm. On arrival to the hospital she was dysarthric and had left lower facial weakness and left hemiparesis. Cardiac auscultation was normal. Blood pressure was 132/84 mm Hg. Position and vibration senses and stereognosis were impaired on the left. She was mildly anemic. Electrolytes, platelet count, prothrombin and partial thromboplastin times, hepatitis B surface antigen, renal function, electrocardiogram (ECG), and admission head CT scan were normal. Conventional cerebral angiography performed 30 hours after cocaine use showed a large filling defect in the proximal right internal carotid artery (ICA) (Fig 1⇓). A carotid thrombectomy performed 43 hours after cocaine use showed a white fibrin clot obstructing the carotid artery. The carotid endothelial surface was normal by direct visualization. Subsequently, two-dimensional and M-mode echocardiography showed no thrombus, valve, or wall motion abnormalities, and 24-hour Holter monitor revealed a normal sinus rhythm. Histological study of the clot showed neither calcification nor microorganisms. Her neurological function improved dramatically after surgery, and she was subsequently discharged from the hospital.
A 32-year-old African-American woman who had had a convulsion several months earlier was brought to the emergency department from a “crack house” because of lethargy, dysarthria, and right-sided weakness after smoking crack. Cardiac rate and rhythm were regular. A 2/6 systolic ejection murmur was heard at the left sternal border. Blood pressure was 110/70 mm Hg. The patient was lethargic and mute and did not appear to understand questioning. Her left eye skewed down and out with some correction on lateral gaze. She was able to fix on objects. Right upper motor neuron facial weakness was present, and there was marked right hemiparesis. Clonus and extensor plantar responses were elicited bilaterally. Hemoglobin, serum electrolytes, arterial blood gases on room air, coagulation profile, chest x-ray film, and transthoracic echocardiogram were all normal. The urine drug screen was positive for cocaine metabolite and nicotine. Cultures were negative on three blood samples. The ECG revealed sinus bradycardia. The electroencephalogram showed a left hemispheric disturbance with suppression of background rhythm; rhythmic delta activity was observed in both hemispheres. CT of the head with and without contrast showed a nonenhancing hypodensity in the left frontal and temporal lobes anteriorly, associated with moderate mass effect and compression of the lateral ventricle, with left-to-right shift of the midline structures (Fig 2⇓, left panel). CT scan 1 day later showed no change in infarct size, but we noted effacement of the basal cisterns thought to be the result of brain swelling and uncinate gyrus herniation. She lapsed into coma and died 72 hours after admission.
At autopsy, the intradural arteries on the left had a smaller caliber than those on the right. There was no atherosclerosis or inflammation of the blood vessels. The meninges at the base of the brain were mildly fibrotic. The left cerebral hemisphere was pale and swollen, and its appearance was consistent with a recent infarct in the distribution of the left anterior and middle cerebral arteries. We noted bilateral transtentorial herniation, more marked on the left, and several fresh hemorrhages in the medial aspect of the left frontal lobe and the left striatum (Fig 2⇑, top right panel). There was softening of the entire left frontal lobe and of the anterior portion of the left parietal lobe. We noted vascular engorgement but no hemorrhage, infarct, or other lesion in the brain stem. Histopathologically, multiple sections from the left hemisphere and occasional sections from the right hemisphere showed cerebral vessels with abnormal internal elastic lamina infolding and tunica media degeneration (Fig 2⇑, bottom left panel). There was no evidence of vessel inflammation or infection. The heart was normal. Fig 3⇓ provides comparative (“control”) normal human anterior cerebral arteries taken from two other patients.
A 45-year-old African-American man developed left-sided weakness, throbbing bitemporal headache nausea, vomiting, dysphagia, slurred speech, and urinary incontinence at an unclear time after drinking approximately a fifth of a gallon of vodka. He said that he had smoked “20 rocks” of crack cocaine approximately 4 days before. He also said that he had used crack 2 weeks earlier and at that time had experienced left-sided incoordination and gait imbalance with resolution in 2 days. He had no history of diabetes mellitus, hypertension, or seizures. He had smoked half of a pack of cigarettes per day during the previous 3 to 4 years.
The patient was afebrile and without pertinent general physical findings. Blood pressure was 130/90 mm Hg. He was oriented but lethargic and tended to drift off to sleep unless constantly given verbal stimulation. The patient exhibited questionable left visual field defect, a ptotic left upper eyelid, decreased pinprick sensation in the left side of the face, left lower facial weakness, dysarthric speech, tongue deviation to the left, and marked left hemiparesis with arm involvement much greater than leg.
Urinalysis, complete blood count, platelet count, prothrombin and partial thromboplastin times, blood cultures, serum protein electrophoresis, cholesterol level, and chest x-ray film were unremarkable. Ethanol was not detected in the plasma. A urine toxicology screen was positive for cocaine metabolite and nicotine. The ECG showed normal sinus rhythm with a nonspecific T-wave abnormality. CT of the head showed a hypodensity in the territory supplied by the right middle cerebral artery (MCA) and a hyperdensity in the M-1 segment of the right MCA. Conventional cerebral angiography showed a thrombus in the proximal right ICA with associated right MCA occlusion (Fig 4⇓).
Follow-up MRI showed changes consistent with an infarct in the distribution of the right MCA with hemorrhagic transformation and a decreased left temporal lobe signal on T2-weighted imaging. Routine electroencephalography showed a right subcortical frontotemporal disturbance. Transthoracic echocardiography showed global left ventricular hypokinesis with an ejection fraction of 41%. Transesophageal echocardiography did not show thrombus. Carotid duplex Doppler performed 8 days after cerebral angiography showed a 10% to 15% left ICA bulb stenosis and a 15% stenosis in the right ICA bulb, without evidence of the intraluminal clot. The patient’s level of consciousness became normal within 36 hours, and his hemiparesis partially improved. The patient was lost to follow-up.
We present three patients whose angiographic and pathological data suggest that vasospasm and/or thrombus formation may be important mechanisms of crack cocaine–induced cerebral infarction. Patients 1 and 3 had intra-arterial clot without angiographic evidence of vasospasm, and patient 2 had vasospasm without evidence of clot. The vasospasm changes were more prominent in the left- than the right-sided cerebral arteries, which, with the smaller size of the left-sided vessels at autopsy, is consistent in our opinion with more severe cerebral vasoconstriction on the left side.
Multiple mechanisms for the neurovascular complications of crack cocaine use have been previously proposed.1 2 3 4 5 Cocaine and its metabolites (norcocaine and benzoylecgonine) are potent vasoconstrictors, and the chronic use of crack cocaine may lead to prolonged vasoconstriction due to the more protracted half-lives of active metabolites,6 which are perhaps even longer in association with ethanol,7 as in our third patient. Our observations in patient 2 suggest that large cerebral vessels may develop tunica media and elastic lamina damage presumably related to cocaine-induced vasoconstriction. These pathological changes have been reported in cardiac vessels associated with cocaine use.8 The changes were seen in multiple sections and territories and bilaterally, making artifactual changes unlikely. Special staining of this patient’s vessels did not disclose other potential causes for these changes (ie, atherosclerosis). To our knowledge, there is no other known diagnostic entity associated with these cerebral vascular pathological changes. Several previously described patients had stereotypical transient ischemic attacks every time they smoked crack and then suffered a cerebral infarct in the same territory as the transient ischemic attack.1 This clinical observation is also compatible with vasospasm as an underlying mechanism. One could also speculate that vasospasm led to thrombus formation on occasion in these patients (and that transient vasospasm did the same in our patients 1 and 3).
Cocaine-induced vasoconstriction has been suggested by cerebral angiography in several patients with cocaine-associated ischemic strokes.1 2 3 4 5 This vasoconstrictive response to cocaine has also been reported in several animal studies and is summarized in the Table⇓ to lend experimental support for our clinically developed hypothesis of vasospasm.9 10 11 12 13 14 Rabbits and other animals that have had large cranial arteries exposed to cocaine also develop similar changes (range, 20% to 100% reduction in vessel diameter),10 11 12 13 including angiographic evidence of vasospasm that is severe enough to result in vascular occlusion. However, vasodilation has also been described14 15 (Table⇓). The effect of cocaine on vessel diameter may depend on the location of the vessel14 and the dose of cocaine used. It is known that cocaine prevents the uptake of catecholamine from nerve terminals, resulting in the availability of epinephrine to interact with postsynaptic receptors. This in turn can potentiate adrenergic transmission and can result in vasoconstriction, thereby leading to ischemia.1 6 Cocaine is also known to block the uptake of serotonin.6 As with catecholamine, the presence of elevated levels of serotonin could potentiate vasoconstriction of the large- and medium-sized arteries of the cerebral circulation.
In addition to its direct effects on the cerebral circulation, cocaine may concomitantly potentiate thrombus/embolism formation through its known cardiomyopathic effects.16 17 This remains a possibility in our third patient. Platelet function and protein C activity can also be altered by cocaine and contribute to cerebral ischemia.1 We do not have long-term clinical or laboratory follow-up data on patients 1 and 3 to confirm or refute that a persistent hypercoagulable state might have contributed to the strokes. None of our cases had clinical evidence of prior clotting events.
Vasospasm not associated with cocaine has also been implicated in recurrent, stereotypical transient cerebral or ocular ischemia.18
We suggest that brain infarcts among crack cocaine users may result secondary to large cerebral artery vasospasm with secondary intravascular thrombosis (with or without distal embolization). Hemorrhage into an infarct may reflect reperfusion after resolution of vasoconstriction or recanalization from dissolution of the thrombus (see “Case 3”).
Although cocaine may cause acute hypertension associated with stroke, none of our patients were documented to have markedly elevated blood pressures on admission. This does not exclude a brief episode of severe hypertension that could be associated with cerebral vasoconstriction. Although acute hypertension may play a role in cerebral hemorrhage from cocaine, the hemorrhagic infarcts seen in our patients were not associated with hypertension and are more likely to be associated with reperfusion. In patient 2, the hemorrhage was unlikely to be due to the consequences of herniation based on the location of the hemorrhagic lesion.
This study was supported in part by National Institutes of Health grant NS-23393 and by a Grant-in-Aid from the American Heart Association, Michigan Affiliate.
Reprint requests to Steven R. Levine, MD, Center for Stroke Research, Department of Neurology K-11, Henry Ford Hospital and Health Sciences Center, 2799 W Grand Blvd, Detroit, MI 48202-2689. E-mail firstname.lastname@example.org.
- Received June 20, 1994.
- Revision received March 2, 1995.
- Accepted March 2, 1995.
- Copyright © 1995 by American Heart Association
Levine SR, Brust JCM, Futrell N, Ho K-L, Blake D, Millikan CH, Brass LB, Fayad P, Schultz LR, Selwa JF, Welch KMA. Cerebrovascular complications of the use of the ‘crack’ form of alkaloidal cocaine. N Engl J Med. 1990;323:669-704.
Levine SR, Brust JCM, Futrell N, Brass LB, Blake D, Fayad P, Schultz LR, Millikan CH, Ho K-L, Welch KMA. A comparative study of the cerebrovascular complications of cocaine: alkaloidal versus hydrochloride: a review. Neurology. 1991;41:1173-1177.
Mody CK, Miller BL, McIntyre HB, Cobb SK, Goldberg MA. Neurologic complications of cocaine abuse. Neurology. 1988;38:1189-1193.
Gold MS, Washton AM, Drackis CA. Cocaine abuse: neurochemistry, phenomenology, and treatment. Natl Inst Drug Abuse Res Monogr Ser. 1985;61:130-150.
Altura BM, Altura BT, Gebrewold A. Alcohol-induced spasms of cerebral blood vessels: relation to cerebrovascular accidents and sudden death. Science. 1983;220:331-333.
Powers RH, Madden JA. Vasoconstrictive effects of cocaine, metabolites and structural analogs on cat cerebral arteries. FASEB J. 1990;4:A1095. Abstract.
Langner RO, Bement CI, Perry LE. Arteriosclerotic toxicity of cocaine. Natl Inst Drug Abuse Res Monogr Ser. 1988;88:325-336.
Bement CL, Cohen L, Nielson SW, Langner RO. Arterial injury in rabbits following alternate day injections of cocaine. FASEB J. 1989;3:A297. Abstract.
Wang A-M, Suojanen JN, Colucci VM. Cocaine- and methamphetamine-induced acute cerebral vasospasm: an angiographic study in rabbits. AJNR Am J Neuroradiol. 1990;11:1141-1146.
Diez-Tejedor E, Tejada J, Munoz J. Cerebral arterial changes following cocaine IV administration: an angiographic study in rabbits. J Neurol. 1992;239(suppl 2):S38. Abstract.
Dohi S, Jones MD, Hudak ML, Traystman RJ. Effects of cocaine on pial arterioles in cats. Stroke. 1990;21:1710-1714.
Chokshi SK, Moore R, Pandian NG, Isner JM. Reversible cardiomyopathy associated with cocaine intoxication. Ann Intern Med. 1989;111:1039-1040.