Yield of further diagnostic work-up of cryptogenic subarachnoid hemorrhage based on bleeding patterns on computed tomographic scans

BACKGROUND: Subarachnoid hemorrhage (SAH) is cryptogenic in 15% of cases. Despite reports of proven recurrence, additional diagnostic studies are not often recommended when no abnormalities were identified on the initial study with digital subtraction angiography (DSA). In our retrospective review of outcomes after cryptogenic SAH, we identify diagnostic strategies that most often yielded the source of bleeding. METHODS: Of 719 patients admitted with SAH from 1998 to 2003,92 (12.8%) patients had findings negative for a bleeding source on initial four-vessel DSA. Based on computed tomographic scans, SAH was categorized as perimesencephalic in 45 patients (mean age, 48 yr) and nonperimesencephalic in 47 patients (mean age, 53 yr). All underwent cerebral magnetic resonance imaging and magnetic resonance angiography; select patients underwent additional studies. Multiple variables were analyzed. Outcomes at the time of discharge were categorized according to the modified Rankin Scale. RESULTS: After perimesencephalic SAH, 44 (97.8%) patients had good scores (0-2) on the modified Rankin Scale, and one patient (2.2%) was deceased. Six (13.3%) patients experienced complications, one (2.2%) experienced vasospasm, and two (4.4%) had hydrocephalus. Further studies in perimesencephalic SAH yielded a diagnosis in 13.9% of patients. After nonperimesencephalic SAH (mean Hunt and Hess score of 2.2), hospital and intensive care unit stays averaged 12 and 8.3 days, respectively. Outcomes were good in 30 (63.8%) patients, poor (modified Rankin Scale 3-5) in 11 (23.4%), and six (12.8%) died. Further studies in nonperimesencephalic SAH exhibited positive findings in 21.3% of patients. Eighteen (38.3%) patients had complications, nine (19.1 %) experienced vasospasm, four (8.5%) had recurrent SAH, and 12 (25.5%) had hydrocephalus. CONCLUSION: Compared with perimesencephalic SAH, nonperimesencephalic SAH was associated with significantly (P < 0.01) longer hospital and intensive care unit stays, greater complication rates, and worse outcomes. Positive findings after further work-up after initial negative DSA in 16% of our patient population confirms that cryptogenic SAH is not necessarily nonaneurysmal, but that a bleeding risk exists. Therefore, we advocate repeat DSA and/or computed tomographic angiography after cryptogenic SAH.