Nontraumatic subarachnoid hemorrhage in the setting of negative cranial computed tomography results: external validation of a clinical and imaging prediction rule.

STUDY OBJECTIVE Clinical variables can reliably exclude a diagnosis of nontraumatic subarachnoid hemorrhage in patients with negative cranial computed tomography (CT) results. We externally validated 2 decision rules with 100% reported sensitivity for a diagnosis of subarachnoid hemorrhage, among patients undergoing lumbar puncture after a negative cranial CT result: (1) clinical rule: presence of any combination of age 40 years and older, neck pain or stiffness, loss of consciousness, or headache onset during exertion; and (2) imaging rule: cranial CT performed within 6 hours of headache onset. METHODS This was a matched case-control study of patients presenting to 21 emergency departments between 2000 and 2011. Patients with a diagnosis of subarachnoid hemorrhage as determined by lumbar puncture after a negative cranial CT result were screened for inclusion. A matched control cohort was selected among patients with a diagnosis of headache after negative cranial CT and lumbar puncture results. RESULTS Fifty-five cases of subarachnoid hemorrhage meeting inclusion criteria were identified, 34 (62%) of which were attributed to cerebral aneurysms. External validation of the clinical rule demonstrated a sensitivity of 97.1% (95% confidence interval [CI] 88.6% to 99.7%), a specificity of 22.7% (95% CI 16.6% to 29.8%), and a negative likelihood ratio of 0.13 (95% CI 0.03 to 0.61) for a diagnosis of subarachnoid hemorrhage. External validation of the imaging rule revealed that 11 of 55 subarachnoid hemorrhage cases (20%) had negative cranial CT results for tests performed within 6 hours of headache onset. CONCLUSION The clinical rule demonstrated useful Bayesian test characteristics when retrospectively validated against this patient cohort. The imaging rule, however, failed to identify 20% of subarachnoid hemorrhage patients with a negative cranial CT result.