Case Fatality of Subarachnoid Hemorrhage by Aneurysm Location

BACKGROUND AND OBJECTIVES: According to region-specific hospital-based studies, the location of ruptured intracranial aneurysm (RIA) influences the case fatality rate (CFR) of aneurysmal subarachnoid hemorrhage (SAH). However, little is known about whether CFRs vary by RIA location in population-based studies that include prehospital SAH deaths. We assessed whether CFRs and CFR-trends differ by RIA location using whole population data from Finland and New Zealand. METHODS: We used externally validated administrative databases to identify all nonhospitalized and hospitalized SAH cases in Finland and New Zealand from 2001 to 2017. Using the ICD-10, we categorized RIAs into anterior communicating artery (Acom) (ICD-10 I60.2), internal carotid artery (ICA) (I60.0/I60.3), middle cerebral artery (MCA) (I60.1), and vertebrobasilar artery (VBA) (I60.4/I60.5) locations. To validate the location-specific SAH diagnoses, we used external hospital- and population-based datasets and autopsy data. We calculated sudden death rates (those occurring before admission to a ward) and overall 30-day CFRs, and computed age-, sex, and country-adjusted risk ratios using a Poisson regression model with 95% CIs. RESULTS: Among 13,470 SAH cases (5,056 from New Zealand; median age 58 years; 61.3% women), 26.6% had Acom, 18.4% ICA, 29.5% MCA, 11.5% VBA, and 14.0% other/unspecified RIAs. The overall 30-day CFRs were the greatest for VBA (54.1%), followed by MCA (40.5%), Acom (29.1%), and ICA (28.5%) RIAs. Location-specific sudden death rates were 33.0%, 21.6%, 11.9%, and 9.9%, respectively. Between 2001-2003 and 2015-2017, overall 30-day CFRs declined significantly for VBA (24%, 95% CI 13%-34%) and MCA (15%, 95% CI 5%-24%) RIAs. Location-specific differences in CFRs were similar between countries, but temporal decreases were observed only in Finland. Between 2001-2003 and 2014-2017, the proportion of VBA RIAs increased by 35.9% (from 10.3% to 14.0%). DISCUSSION: SAH CFRs vary significantly by RIA location, with VBA and MCA RIAs having the greatest CFRs, mainly due to high sudden death rates. It is unclear whether aneurysms with these high-risk locations could benefit from improved primary/secondary prevention or prehospital management strategies. These findings cannot be directly applied to the preventive treatment of unruptured intracranial aneurysms.