Delayed Cerebral Ischemia and Vasospasm After Spontaneous Angiogram-Negative Subarachnoid Hemorrhage: An Updated Meta-Analysis.

OBJECTIVE: To report clinical outcomes of delayed cerebral ischemia (DCI) and vasospasm in angiogram-negative subarachnoid hemorrhage (SAH) according to hemorrhage patterns, perimesencephalic hemorrhage (PMH) and non-PMH.

METHODS: Online databases from January 1990 to December 2017 were systematically reviewed. A fixed-effects model was used to control heterogeneity. To resolve publication bias, the trim and fill method was used to estimate number of missing studies and adjusted odds ratio (OR). Subgroup analysis of data from studies that defined angiogram-negative subarachnoid hemorrhage by angiography repeated at least twice or computed tomography angiography was performed.

RESULTS: Among 24 studies including 2083 patients, 23/985 patients (2.3%) with PMH and 144/1098 patients (13.1%) with non-PMH had DCI, indicating that patients with PMH experienced significantly lower DCI than patients with non-PMH (OR = 0.219; 95% confidence interval [CI], 0.144-0.334). Regarding vasospasm, 99/773 patients (12.8%) with PMH and 231/922 patients (25.1%) with non-PMH exhibited vasospasm, indicating that patients with PMH experienced significantly lower vasospasm than patients with non-PMH (OR = 0.445; 95% CI, 0.337-0.589). Funnel plots show asymmetry indicating possible publication bias. After trimming 10 studies for DCI and 7 for vasospasm, the adjusted ORs remained significant between PMH and lower risks of DCI and vasospasm. Subgroup analysis of 789 patients in 8 studies showed a lower risk of DCI (OR = 0.268; 95% CI, 0.151-0.473) and vasospasm (OR = 0.346; 95% CI, 0.221-0.538) in patients with PMH.

CONCLUSIONS: PMH showed a significantly lower risk of DCI and vasospasm than non-PMH. Clinical outcomes of angiogram-negative subarachnoid hemorrhage, based on meta-analysis of individual patient data, need to be investigated.