Repeat Catheter Angiography in Patients with Aneurysmal-Pattern Angiographically Negative Subarachnoid Hemorrhage.

BACKGROUND: A significant proportion of patients with subarachnoid hemorrhage have a normal cerebral angiogram. Patients with angiographically negative subarachnoid hemorrhage (anSAH) with either perimesencephalic- (panSAH) or aneurysmal-pattern hemorrhage (aanSAH, also known as diffuse anSAH) have an excellent prognosis, but only if underlying vascular abnormalities are completely excluded. The rate of occult aneurysms in patients with aanSAH varies widely across studies. The purpose of this study was to quantify the value of repeat DSA in these patients. METHODS: We reviewed the records of all patients initially diagnosed with aanSAH after a screening DSA at a single tertiary neurovascular referral center from January 2006-April 2018. Patients with panSAH and traumatic SAH were excluded. We also performed a systematic review and meta-analysis of positive second DSAs in previously published case series of patients with aanSAH who underwent two serial DSAs. For meta-analysis, PubMed Central, MEDLINE and Cochrane Library databases were searched for pertinent studies up to November 2019. The rate of aneurysm detection on repeat angiography was extracted from each study. Pooled rates for positive second angiogram were calculated as untransformed proportions in a binary random-effects model meta-analysis. Inter-study heterogeneity was calculated using an I2 statistic. RESULTS: Three of 27 patients (11.1%) with aanSAH and at least two DSAs were subsequently found to have a cerebral aneurysm in our institutional dataset. Twenty-six studies in our systematic review met inclusion criteria, and the pooled rate of positive second angiogram was 10.4% (95% CI 7.3%-13.5%, P < 0.001). Substantial inter-study heterogeneity was observed in the meta-analysis (I2 = 61.7%, P < 0.001). CONCLUSIONS: One in 10 patients with aanSAH has an occult ruptured aneurysm. A second-look DSA should be strongly considered in these cases.