BACKGROUND: Emerging evidence suggests that endovascular thrombectomy is beneficial for treatment of childhood stroke, but the safety and effectiveness of endovascular thrombectomy has not been compared with best medical treatment. We aimed to prospectively analyse functional outcomes of endovascular thrombectomy versus best medical treatment in children with intracranial arterial occlusion stroke.

METHODS: In this prospective registry study, 45 centres in 12 countries across Asia and Australia, Europe, North America, and South America reported functional outcomes for children aged between 28 days and 18 years presenting with arterial ischaemic stroke caused by a large-vessel or medium-vessel occlusion who received either endovascular thrombectomy plus best medical practice or best medical treatment alone. Intravenous thrombolysis was considered part of best medical treatment and therefore permitted in both groups. The primary outcome was the difference in median modified Rankin Scale (mRS) score between baseline (pre-stroke) and 90 days (±10 days) post-stroke, assessed by the Wilcoxon rank test (α=0·05). Efficacy outcomes in the endovascular thrombectomy and best medical treatment groups were compared in sensitivity analyses using propensity score matching. The Save ChildS Pro study is registered at the German Clinical Trials Registry, DRKS00018960.

FINDINGS: Between Jan 1, 2020, and Aug 31, 2023, of the 241 patients in the Save ChildS Pro registry, 208 were included in the analysis (115 [55%] boys and 93 [45%] girls). 117 patients underwent endovascular thrombectomy (median age 11 years [IQR 6-14]), and 91 patients received best medical treatment (6 years [3-12]; p<0·0001). The median Pediatric National Institutes of Health Stroke Scale (PedNIHSS) score on admission was 14 (IQR 10-19) in the endovascular thrombectomy group and 9 (5-13) in the best medical treatment group (p<0·0001). Both treatment groups had a median pre-stroke mRS score of 0 (IQR 0-0) at baseline. The change in median mRS score between baseline and 90 days was 1 (IQR 0-2) in the endovascular thrombectomy group and 2 (1-3) in the best medical treatment group (p=0·020). One (1%) patient developed a symptomatic intracranial haemorrhage (this patient was in the endovascular thrombectomy group). Six (5%) patients in the endovascular thrombectomy group and four (5%) patients in the best medical treatment group had died by day 90 (p=0·89). After propensity score matching for age, sex, and PedNIHSS score at hospital admission (n=79 from each group), the change in median mRS score between baseline and 90 days was 1 (IQR 0-2) in the endovascular thrombectomy group and 2 (1-3) in the best medical treatment group (p=0·029). Regarding the primary outcome for patients with suspected focal cerebral arteriopathy, endovascular thrombectomy (n=18) and best medical treatment (n=33) showed no difference in 90-day median mRS scores (2 [IQR 1-3] vs 2 [1-4]; p=0·074).

INTERPRETATION: Clinical centres tended to select children with more severe strokes (higher PedNIHSS score) for endovascular thrombectomy. Nevertheless, endovascular thrombectomy was associated with improved functional outcomes in paediatric patients with large-vessel or medium-vessel occlusions compared with best medical treatment. Future studies need to investigate whether the positive effect of endovascular thrombectomy is confined to older and more severely affected children.

FUNDING: None.