Intensive therapies of delayed cerebral ischemia after subarachnoid hemorrhage: a propensity-matched comparison of different center-driven strategies

Intensive therapies of delayed cerebral ischemia (DCI) following aneurysmal subarachnoid hemorrhage (aSAH) have still controversial and unproven benefit. We aimed to compare the overall efficacy of two different center-driven strategies for the treatment of DCI respectively with and without vasospasm angioplasty. Two hundred consecutive patients with aSAH were enrolled in each of two northern European centers. In an interventional center, vasospasm angioplasty was indicated as first line rather than rescue treatment of DCI using distal percutaneous balloon angioplasty technique combined with intravenous milrinone. In non-interventional center, induced hypertension was the only intensive therapy of DCI. Radiological DCI (new cerebral infarcts not visible on immediate post-treatment imaging), death at 1 month, and favorable outcome at 6 months (modified Rankin scale score ≤ 2) were retrospectively analyzed by independent observers and compared between two centers before and after propensity score (PS) matching for baseline characteristics. Baseline characteristics only differed between centers for age and rate of smokers and patients with chronic high blood pressure. In the interventional center, vasospasm angioplasty was performed in 38% of patients with median time from bleeding of 8 days (Q1 = 6.5;Q3 = 10). There was no significant difference of incidence of radiological DCI (9% vs.14%, P = 0.11), death (8% vs. 9%, P = 0.4), and favorable outcome 74% vs. 72% (P = 0.4) between interventional and non-interventional centers before and after PS matching. Our results suggest either that there is no benefit, or might be minimal, of one between two different center-driven strategies for intensive treatment of DCI. Despite potential lack of power or unknown confounders in our study, these results question the use of such intensive therapies in daily practice without further optimization and validation.