Angiogram-negative Subarachnoid Hemorrhage: Rebleeding, Complications, Outcomes, and Diagnostic Yield

Introduction: About 15% of subarachnoid hemorrhage (SAH) cases are angiographically negative SAH (anSAH) cases. The diagnostic yield and outcomes in anSAH vary depending on the hemorrhage pattern, while the rebleeding rate remains unclear.

Methods: We retrospectively studied 68 consecutive SAH patients between 2016 and 2021 with a negative initial digital subtraction angiography (DSA) and MRA or CTA. We assessed complications, outcomes, and diagnostic yield of DSA.

Results: Mean age of the 68 patients was 55.5±10.9 years, 30 (44.1%) were female, and 43 (63.2%) exhibited perimesencephalic (PM) anSAH. Hunt and Hess (HH) grade, modified Fisher scale (mFs), blood thickness and Barrow Neurological Institute (BNI) scale, and intraventricular hemorrhage were significantly greater in non-PM-anSAH. Vasospasm on transcranial doppler, delayed cerebral ischemia (DCI), external ventricular drain (EVD) placement, ventriculoperitoneal shunt placement, and seizures were less common in the PM group. Patients with PM-anSAH had a shorter length of stay, lower mRS at discharge, and were more frequently discharged home; however, mRS at 3 months and last follow-up were the same in both bleeding patterns. Three (4.4%) rebleeds occurred (mean follow-up: 632 days) on day 22, 384, and 873 after SAH, 2 (4.7%, p=1.0) in the PM group, with a prior higher mRS at discharge (OR=1.8, p=0.0468). A bleeding source was found in 5 (7.4%) cases, 2 (4.7%, p=0.349) in the PM group, with significantly higher HH grade, more frequent cerebral infarctions and DCI, and need for EVD placement. The mFS and BNI scale could not predict vasospasm, cerebral infarction, or DCI.

Conclusion : Patients with non-PM-anSAH experience more complications. A bleeding source was more frequently found in patients with higher HH grade, cerebral infarction or DCI, and EVDs. Rebleeding was more common after high discharge mRS scores. Repeat imaging in both bleeding patterns is supported by comparable rates of diagnostic yield and rebleeding.