Does magnetic resonance imaging produce further benefit for detecting a bleeding source in subarachnoid hemorrhage of unknown origin?

Background: Spontaneous subarachnoid hemorrhage (SAH) without evidence of a bleeding source on the first digital subtraction angiogram (DSA) - also called SAH of unknown origin - is observed in up to 27% of all cases. Depending on the bleeding pattern on CT scanning, SAH can be differentiated into perimesencephalic (PM-SAH) and non-perimesencephalic SAH (NON-PM-SAH). The aim of our study was to investigate the effectiveness of magnetic resonance imaging (MRI) for detecting a bleeding source in SAH of unknown origin.

Methods: We retrospectively reviewed 1,226 patients with spontaneous SAH between January 1991 and December 2008 in our department. DSA was performed in 1,068 patients, with negative results in 179 patients.

Results: Forty-seven patients were categorized as having PM-SAH and 132 patients as having NON-PM-SAH. MRI of the brain and the craniocervical region was performed within 72 h after diagnosis of SAH and demonstrated no bleeding sources in any of the PM-SAH and NON-PM-SAH patients (100% negative).

Conclusions: In our experience MRI did not produce any additional benefit for detecting a bleeding source after SAH with a negative angiogram. The costs of this examination exceeded the clinical value. Despite our results MRI should be discussed on a case-by-case basis because rare bleeding sources are periodically diagnosed in cases of NON-PM-SAH.