Sensitivity of newer-generation computed tomography scanners for subarachnoid hemorrhage: a Bayesian analysis

Subarachnoid hemorrhage (SAH) is a life-threatening condition considered in patients presenting to the emergency department (ED) with acute and severe-onset headache. Currently, the practice pattern for suspected SAH is to perform a non-contrasted computed tomography (CT) scan of the head, followed by lumbar puncture (LP) if the CT is negative. Newer-generation 16-slice CT scanners have been shown in one study to be very sensitive for SAH.

We sought to validate these findings at our institution by retrospectively analyzing the sensitivity of our 16-slice or better CT scanner and performing a bayesian analysis with the results.

We utilized ED electronic medical records and the Department of Neurosurgery research database to search for patients admitted from the ED with a diagnosis of SAH from January 1, 2005 to December 31, 2008. We found a total of 134 patients admitted with SAH during this time frame.

Average age was 53.8 years; 62% were female. Presenting complaint was headache in 57%, paresthesia or weakness in 7%, unresponsive in 10%, confusion or altered mental status in 5%, and "other" in 10%. Sensitivity of 16-slice or better CT scanner in our study was 131/134, or 97.8% (95% confidence interval 93.1-99.4%). No patient with a negative CT had a lesion requiring intervention.

Our study confirms the high sensitivity of 16-slice or better CT scanners for SAH. This calls into question the need for LP after negative head CT when 16-slice CT or better is used.