Current issues of ACP Journal Club are published in Annals of Internal Medicine


62% of patients were alive at 1 year after spontaneous subarachnoid hemorrhage

ACP J Club. 1993 Sept-Oct;119:53. doi:10.7326/ACPJC-1993-119-2-053

Source Citation

Longstreth WT Jr, Nelson LM, Koepsell TD, van Belle G. Clinical course of spontaneous subarachnoid hemorrhage: a population-based study in King County, Washington. Neurology. 1993 Apr;43:712-8.



To study the clinical course and outcome of spontaneous subarachnoid hemorrhage in adults.


Inception cohort followed for 1 year.


A county in the state of Washington that has 26 acute care hospitals.


All adults living in King County who had a definite, probable, or possible subarachnoid hemorrhage. Patients were excluded if their bleeding originated from a source other than intracranial aneurysm, including primary intraparenchymal hemorrhages, arteriovenous malformations, trauma, or neoplasms. 171 patients (68% < 65 y old, 139 white, 111 women) were studied. Follow-up was 95%.

Assessment of prognostic factors

Medical records and structured interviews with patients and families provided data on age, smoking status, alcohol consumption, aspirin use, hypertension, physical examination and laboratory evaluation at admission, and surgery for aneurysm.

Main outcome measures

Medical records were used to collect data on mortality, neurologic deficits, physical functioning (Glasgow Outcome Scale), and activities of daily living.

Main results

5 patients died without medical attention; 166 were hospitalized and had a computed tomography (CT) scan of the brain. 162 patients had definite, 7 had probable, and 2 had possible subarachnoid hemorrhage. 103 of 149 (69%) patients had surgery. Of hospitalized patients, 61% were treated with anticonvulsants and 62% with steroids. 19% had symptomatic vasospasm; 24%, hydrocephalus; and 11%, seizures. 30 patients (18%) rebled within 1 year (27 during hospitalization). Mortality at 1 month was 32% and at 1 year, 38%. 52% of the patients who survived had a good outcome (moderate, minimal, or no deficits on the Glasgow Outcome Scale). Multivariate analysis showed that poor outcomes (severe neurologic deficits, vegetative state, or death) were associated with decisions by physicians and family to limit medical support (odds ratio [OR] 46, 95% CI 4.7 to 458); blood on initial CT scan (OR 18.3); and development of hydrocephalus within 3 days (OR 3.4, CI 1.04 to 10.9). Good outcomes were associated with high admission Glasgow Coma Scale scores (OR 0.56) and surgery for aneurysm (OR 0.36, CI 0.12 to 1.07). Time from event to surgery was not an independent factor in predicting prognosis.


At 1 year, approximately one third of patients with subarachnoid hemorrhage had died and one half of the survivors had good outcomes.

Source of funding: National Institute of Neurologic Diseases and Stroke.

For article reprint: Dr. W.T. Longstreth, Jr., Division of Neurology, ZA-95, Harborview Medical Center, 325 Ninth Avenue, Seattle, WA 98104-2499, USA. FAX 206-731-8787.


This excellent population-based study of subarachnoid hemorrhage caused by bleeding aneurysms should not be misconstrued as a population-based study of aneurysms, an important distinction because some neurosurgeons currently advise prophylactic clipping of asymptomatic aneurysms. Moreover, because 10% of aneurysms in this study were obliterated by the bleed, I surmise that some patients, who visited office-based physicians and who had a headache, no loss of consciousness, and a normal neurologic examination, may have had sentinel bleeds. The key to diagnosis in such cases is a careful history that almost always shows a new headache problem, differing from all previous headaches the patient has had. Under such circumstances, a normal CT scan usually indicates the need for a lumbar puncture.

Population-based studies often raise important and sometimes uncomfortable questions about medical care. For example, would the 86 patients in King County who experience an aneurysmal bleed over 1 year do better if their care was regionalized to a few neurosurgic referral centers, rather than being treated in 26 acute care hospitals? An answer is suggested by evidence that 51.5 aneurysm operations were done annually among a population of 1.4 million people, equivalent to a total of 8976 operations for bleeding aneurysms in the United States in 1987 (population of 244 million). Of 3625 neurosurgeons in the United States in 1987, 3005 (83%) reported that they do craniotomies for anterior and posterior communicating artery aneurysms (1). Thus, each neurosurgeon, on average, does only 3 operations for intracranial bleeding aneurysms annually. This assumes a 25% surgical rate for grade V (comatose, moribund) patients, as in this study, for whom aneurysm surgery is, in my opinion, almost always inappropriate (100% poor outcome). How much practice makes perfect is, perhaps, unanswerable at present, but internists and neurologists are well advised to refer their patients with aneurysms to veteran neurosurgeons (2).

Matthew Menken, MD
Robert Wood Johnson Medical SchoolSomerset, New Jersey, USA


1. Pevehouse BC. Comprehensive Neurosurgical Practice Study. Park Ridge, Illinois: American Association of Neurological Surgeons and Congress of Neurological Surgeons; 1988.

2. Menken M. The workload of neurosurgeons: implications of the 1987 practice survey in the USA. J Neurol Neurosurg Psychiatry. 1991;54:921-4.