The systematic search revealed 558 publications. Excluding duplicates, 354 articles remained for screening. In 77 records, the title or abstract contained the search terms and the information that at least some patients were children, and a further eight publications were also identified from the reference lists. From these 85 publications, 68 publications matched the inclusion criteria, providing data on 90 childhood BAS cases.4,5,s1–s66 Sixty-six publications were case reports or case series5, s1–s52, s54–s66 and two were prospective case studies (registries) .4,s53
Clinical features, management, and outcome of all basilar artery stroke cases
In total, 97 children with BAS (seven SNPSR cases, 90 reported cases) were included in this study. There was a male predominance (73 males, 24 females), particularly in the 5- to 10-year age group. Median age was 9 years (IQR 6–13y). Ethnicity was indicated in 16 cases (eight Caucasian, four African, and four Asian). Presenting signs and symptoms are listed in Figure 1, with impaired consciousness and hemi- or quadriparesis being the most frequent mode of presentation.
The NIHSS was provided in four cases and information to calculate the PedNIHSS score was available in a further 86 patients. The median PedNIHSS score at diagnosis was 15 (IQR 4–27).
Prodromes (Fig. 2), defined as transient signs and symptoms occurring at least 24 hours before the onset of stroke, were reported in 42 patients (43%) with a median time of 22 days before onset of stroke symptoms (range 1–120d; IQR 11–37.5d).
Time from onset of BAS symptoms to diagnosis was up to 88 days. In the 17 children in whom the delay was indicated in hours, the median was 19 hours (range 2–96h), with only six children being diagnosed within 12 hours.
Diagnosis was established by magnetic resonance imaging/magnetic resonance angiograpy in 43 children, catheter angiography in 37, computed tomography in 14, and autopsy in three cases.
BAO localization was indicated in 89 patients, being short in 37 (23 in the distal third, 12 mid-basilar, two proximal) and long in 52 patients (28 whole length of the basilar artery, 22 middle and distal thirds, two proximal and middle thirds).
Aetiology was mostly vasculopathic (n=38: dissection [n=26], vasculitis [n=6], steno-occlusive cerebral arteriopathy [n=2], and other [n=4], such as fibromuscular dysplasia), followed by haematological (n=9), and cardioembolic (n=4). Aetiology was unknown in 40 patients and not stated in six. The cases with known aetiology were published more recently (p=0.016) and more likely to have been investigated with magnetic resonance angiography (p=0.028) than cases with unknown aetiology.
Contributing risk factors were reported in 53 patients and multiple risk factors in nine patients. Risk factors included trauma (n=28), infection (n=10), cardiac disease (n=8), haematological disorders (n=8), elevated risk of arteriopathy (n=3, e.g. post irradiation), positive family history of thromboembolic events (n=4), dehydration (n=1), migraine (n=1), or other entities (n=3).
Blood pressure at diagnosis was only stated in 37 out of 97 children (38%), being above the 95th centile for age in eight patients, low in one, and normal in 28.
Antithrombotic or thrombolytic acute-phase treatment was administered to 47 patients, consisting of acetylsalicylic acid (n=16), heparin (n=20), intra-arterial thrombolysis (n=12), and systemic (intravenous) thrombolysis (n=5), with five children receiving more than one agent (in combination or consecutively). Endovascular mechanical thrombectomy was combined with acetylsalicylic acid in one patient and balloon angioplasty in another. Time from onset of symptoms to antithrombotic treatment was up to 7 days, with a median of 20 hours (range 4–168h), in the 16 children in whom the delay was measured in hours. Six patients were treated within 12 hours.
PedNIHSS score, delay to treatment, and outcome were not significantly different in children who received endovascular therapy. Haemorrhagic transformation was described as minimal in one child. All children undergoing thrombolysis had a complete BAO.
Forty-five patients received long-term antithrombotic treatment (18 acetylsalicylic acid, 15 oral anticoagulants, 12 heparin).
Outcome was indicated in 90 patients, being good in 45 and poor in 45 (eight patients died; Fig. S1, supporting information online). When dichotomizing outcome according to the Basilar Artery International Cooperation Study (BASICS) on BAS in adults,7 defining poor outcome as an mRS score of 4 or more, 31 (35%) children fell into this category, compared with 58 (65%) children with a good to moderate outcome (mRS score 0–3). Nine of the 13 patients presenting with a locked-in syndrome had a poor outcome (mRS >2). Median follow-up lasted 6 months (IQR=3–24mo). Eight patients (8%) had a recurrent stroke, with good outcome in three, poor in two, and no information in three patients. Stroke recurrence was located in the basilar artery in six cases whereas localization was not specified in two cases. Three of these eight children received antithrombotic treatment (one acetylsalicylic acid, one heparin, one oral anticoagulant), one had no antithrombotic treatment, and no information on secondary prevention was provided in four cases.
Follow-up imaging with information on recanalization of the basilar artery was provided in 50 cases, with no recanalization in 16, partial in 18, and complete in 16 patients.
In a univariate analysis, outcome was significantly worse with a high PedNIHSS score (r=0.47; 95% CI 0.29–0.63; p<0.001), coma (r=0.39; 95% CI 0.19–0.56; p<0.001), or quadriplegia (r=0.34; 95% CI 0.13–0.51; p=0.001) as presenting signs, as well as with a long BAO (r=0.24; 95% CI 0.02–0.44; p=0.03); and significantly better after an acute antithrombotic, thrombolytic, or mechanical endovascular treatment (r=−0.24; 95% CI −0.46–0.003; p=0.047). In a multivariate analysis including PedNIHSS score, length of BAO, BAO recanalization, antithrombotic, thrombolytic and mechanical endovascular treatment, quadriplegia, and coma, only PedNIHSS score was significantly associated with outcome (β=0.06; 95% CI 0.009–0.11; r2=0.131; p=0.02). A PedNIHSS score of up to 17 was the cut-off value set by the receiver operator curve, distinguishing (p=0.001) the good- (mRS 0–2, n=45) from the poor-outcome group (mRS 3–6, n=45) with an odds ratio of 5.0, a sensitivity of 71%, and a specificity of 67% (Fig. 3).
Figure 3. Receiver operator curve for the Pediatric National Institutes of Health Stroke Scale as a predictor of good outcome (modified Rankin scale [mRS] 0–2; n=45) versus poor outcome (mRS 3–6; n=45). Area under the curve=0.8201 (95% confidence interval 0.7260–0.9142).
Download figure to PowerPoint