Stroke in children

Arterial ischemic stroke.

In the 46 children with arterial ischemic stroke, hematologic or metabolic disorders were identified in 9 (20%), and 7 of these had multiple risk factors. The presence of aCLA (n = 3) was the most frequent prothrombotic condition observed in our series. Hyperhomocysteinemia was found in one child who was homozygous for C677T point mutation of the MTHFR gene. This child also had renal failure, hypercholesterolemia, and folate deficiency.

A cardiac disorder was present in 9 (20%) of 46 patients. Six of these children had cyanotic congenital heart disease, including one child with trisomy 21 and another with pulmonary artery stenosis and a patent foramen ovale. In three children with cyanotic congenital heart disease, the stroke occurred in the context of polycythemia (n = 2) or cardiac surgery (n = 1).

Primary vascular disorders were present in 11 (31%) of the 36 patients with arterial ischemic stroke who underwent conventional or MR angiography. In 6 (17%), a moyamoya pattern was identified. All children with moyamoya were white. Of the six moyamoya patients, one had trisomy 21 with congenital heart disease and polycythemia, another had trisomy 21 and aCLA, and a third had aCLA.

We found nonspecific arteriographic changes in a further 10 (24%) of the 36 patients with arterial ischemic stroke who underwent angiography. Clinical or laboratory findings did not suggest a specific cause in these patients. In particular, a history of varicella was absent. The arteriographic abnormalities were localized to the proximal portion of the large intracranial arteries (n = 7) or to the distal portion (n = 3). These abnormalities consisted of single stenosis (n = 3), multiple unilateral stenoses (n = 5), and multiple bilateral stenoses (n = 2). At the site of the stenoses, no specific signs of dissection were noted, including no intimal flap, double lumen, or string signs. MRI showed no sign of methemoglobin within the arterial wall. Nonspecific changes on arteriography were present in all three patients with arterial ischemic stroke who had had a nonspecific upper respiratory infection within the preceding 4 weeks.

In ischemic stroke patients, three had extrinsic arterial compression caused by underlying expanding processes, including subarachnoid hemorrhage with a large intracranial hematoma (n = 1), subdural hematoma (n = 1), and diffuse brain edema secondary to acute liver failure with transtentorial herniation and compression on posterior cerebral artery (n = 1).

Sinovenous thrombosis.

Risk factors were found in all five patients with sinovenous thrombosis. Two patients had a nonspecific infection. Three children had a hematologic or metabolic disorder, including one with nephrotic syndrome and dehydration.

Hemorrhagic stroke.

Among the 21 patients with hemorrhagic stroke, risk factors included vascular abnormalities in 14 patients (67%), hematologic disorder in 2 patients (10%), and bleeding into an intracranial tumor in 2 patients (10%). In 3 (14%) of the 21 patients, no risk factor was identified. No patient with hemorrhagic stroke had multiple risk factors.

Death.

Death was more frequent in children with hemorrhagic stroke (29%) than in ischemic stroke (16%). Death occurred in 4 (40%) of the 10 children with recurrent hemorrhagic or ischemic stroke and in 10 (16%) of the 62 children with a single episode of stroke. Deaths resulted from complications of stroke (n = 11), underlying cardiac disorders (n = 2), and acute respiratory distress syndrome (n = 1).

Hematologic and metabolic disorders.

Despite the absence of systematic screening in our cohort, we found at least one hematologic or metabolic disorder in about 25% of ischemic stroke patients. Prothrombotic disorders also were frequent in a recent series,3 being associated with 35 (38%) of 92 pediatric ischemic strokes. We have shown that hematologic and metabolic disorders are frequently found in combination with other risk factors in patients with ischemic stroke. The presence of multiple risk factors has been previously reported to increase the risk of thrombosis considerably.9 Based on these observations, we believe that an extensive hematologic and metabolic screening must be part of the workup of pediatric ischemic strokes, even when a cause is known. Arterial ischemic strokes should be investigated as thoroughly as sinovenous thromboses, ruling out the presence of aPL antibodies and de-ficiencies in protein C, protein S, and antithrombin, as well as aPCR (or factor V Leiden) and hyperhomocysteinemia (or C677T MTHFR gene mutation). Although a recent study reports that 20% of patients with sinovenous thrombosis were heterozygous carriers of the G20210A prothrombin gene point mutation,10 the relative importance of this mutation in patients with arterial ischemic stroke remains to be defined.11 Sickle cell disease must be ruled out in black patients with stroke because a transfusion program prevents stroke recurrence.12 Mitochondrial disorders also are important to exclude in patients presenting with sudden neurologic deficit.

Cardiac disorders.

Although echocardiography was not done for all patients in our series, we found a cardiac risk factor in about 20% of patients with ischemic stroke. In previous reports, up to 15% of ischemic strokes were attributed to cardiac disorders,7,13 most frequently to congenital heart disease.14 Therefore, cardiac evaluation remains essential in the investigation of any ischemic stroke, especially in younger children. Congenital heart disease tends to cause stroke during the first 4 years of life, often in the setting of cardiac surgery or catheterization.14 Among our six patients with congenital heart disease, only one had a cardioembolic stroke associated with cardiac surgery. As found in our series, cyanotic congenital heart disease has been associated with moyamoya disease and polycythemia.15 In children with congenital heart disease and ischemic stroke, these two associated conditions must be ruled out because such findings can modify therapeutic decisions.

Vascular abnormalities.

A recent large series7 found specific vasculopathies in 18% of pediatric ischemic strokes. We identified recognizable vasculopathies in almost 25% of our patients with arterial ischemic stroke. This percentage may underestimate their prevalence because not all patients in our series underwent vascular imaging. These findings underline the importance of vascular investigation in all arterial ischemic strokes in children. Moyamoya was the most frequent recognizable vasculopathy in our cases of ischemic stroke. Although the exact pathophysiologic mechanism of moyamoya remains unknown, it has been associated with many conditions, including trisomy 21 and congenital heart disease.15,16 The association of moyamoya with lupus anticoagulant17 or with a significantly elevated blood level of IgG aCLA18,19 also has been reported, albeit rarely. We report here two more cases of moyamoya associated with aCLA and arterial ischemic stroke. Surprisingly, in our series, aCLA was found in 2 of 6 patients with moyamoya but in only 1 of 31 without moyamoya (p = 0.06). Despite the few patients in our series, this finding may suggest an association between aPL antibodies and moyamoya in children with arterial ischemic stroke, rather than only a rise in IgG aCLA as an epiphenomenon in response to an acute cerebral event. Although the link between aPL antibodies and thrombosis also remains unknown, vascular endothelium is believed to be a major target for the antibodies.20 A direct cause–effect relationship cannot be established. However, a thrombogenic effect of aPL antibodies may induce chronic progressive occlusion of cerebral vessels, characteristically of the supraclinoid portion of the internal carotid arteries with telangiectatic vessel proliferation, a combination that constitutes the moyamoya pattern seen on angiography.21 Alternatively, moyamoya may cause the remodeling of endothelial membrane phospholipids into a more immunogenic target, thereby triggering the synthesis of aPL antibodies. A third possibility is that both moyamoya and aPL antibodies result from a common underlying disorder and independently induce arterial ischemic stroke. Therefore, we believe that testing for aPL antibodies should be done in all children with ischemic stroke, and we agree with other investigators19 that it is especially important in those with moyamoya.

We found nonspecific arteriographic changes in 10 (22%) of 46 children with arterial ischemic stroke. These changes were classified as nonspecific changes because no evidence for specific causes of vasculopathies was found in these patients. Abnormalities were more often multiple, unilateral, and localized proximally on the large intracranial arteries. Similar nonspecific changes have been reported in about 30% of a series6 of 48 Japanese children with ischemic stroke. In that series, mild head trauma with mechanical vascular injury and vasospasm or vasculitis secondary to infection were proposed as causes. In children with arterial ischemic stroke, unexplained arterial stenosis at the circle of Willis (seen on angiograms) has been reported to be frequently spontaneously reversible.22 Among our 46 cases of arterial ischemic stroke, a recent upper respiratory infection was found in 3 (30%) of 10 children with nonspecific arteriographic changes and in none of the 36 children without these changes. This finding suggests that, at least in some children, such changes are secondary to a concurrent or recent infection. Nonspecific infection of the head and neck that produces stimulation of the superior cervical ganglion could predispose cerebral vessels to inflammatory changes and thrombosis. This mechanism has been postulated in postvaricella angiopathy23 and in moyamoya disease.21 Nonspecific arteriographic changes could, in some cases, represent the earliest stages of moyamoya.

Vascular malformation and ruptured intracranial aneurysm were the principal causes of hemorrhagic stroke in our cohort. In one study1 that included both intraparenchymal and extraparenchymal forms of intracranial hemorrhage, these two causes represented about two thirds of hemorrhagic strokes. We agree with other researchers24 that angiography should be done in any spontaneous intracerebral hemorrhage in young patients.

Other risk factors.

In our population, a recent nonspecific infection was found in about 10% of children with ischemic strokes. Another study13 found that a recent or concurrent infection, often affecting the respiratory system, occurred more frequently in children with ischemic stroke than in control subjects (34% versus 11%). Dehydration, secondary hypercoagulability, endothelial damage, altered lipid and prostaglandin metabolism, and secondary vasculitis all can be involved.6,13 Infection is found less often in patients with hemorrhagic stroke.25 We found no ischemic stroke associated with trauma and only one patient with spontaneous arterial dissection. In one study7 of arterial ischemic stroke in the young, arterial dissection was present in none of 81 children younger than 15 years of age but in 3 of 11 children age 15 to 18 years. Our series included only 7 patients with arterial ischemic stroke age 12 to 18 years. This may explain our low frequency of arterial dissections. Recent or concurrent systemic or intracranial infection and mild head and neck trauma should not be overlooked as a stroke risk factor.

Stroke recurrence and death.

Our stroke recurrence rate was similar to the 20% recurrence rate previously reported.2 Recurrence has been linked to migraine,13 metabolic disease,15 moyamoya disease,6 and sickle cell anemia.12 In a series14 of 50 strokes associated with congenital heart disease, the preliminary outcome was death in 8%, no residual symptoms in 40%, and recurrence in only 2%. We have shown that the presence of multiple stroke risk factors is frequent in children with ischemic stroke. We also have shown that recurrent strokes are significantly more likely in patients with multiple risk factors than in those with a single or no identified risk factor. We found a trend toward more frequent death in patients with recurrent stroke, but this association was not statistically significant (p = 0.10). Further studies should re-evaluate this possible association. Our findings provide a rationale for undertaking a vigorous search for every risk factor in every pediatric stroke patient. In particular, risk factors that are frequently found, including prothrombotic conditions, cardiac disorders, and vascular abnormalities, must be ruled out in every ischemic stroke patient, even when a precise cause already has been identified. Some combinations of risk factors may require a multifaceted therapeutic approach. In addition, less frequent risk factors also should be excluded, if clinically suspected, when a stroke remains unexplained and when stroke recurs.

Additional predictors of outcome were not analyzed in our series. Such analyses should be carried out in future studies because prediction of outcome influences the aggressiveness of therapy and rehabilitation and clarifies the prognosis. With our improved understanding of cerebrovascular diseases and the wide availability of the new laboratory and imaging techniques, complete investigations should result in fewer unexplained strokes and more children who can benefit from appropriate, specific treatments designed to prevent stroke recurrence.