Conventional brain computed tomography for the diagnosis of nontraumatic subarachnoid hemorrahge: A systematic review.

PURPOSE: The aim of this study was to determine the diagnostic accuracy of conventional CT versus any other imaging strategies to diagnose non-traumatic subarachnoid hemorrhage. We included METHODS: cross-sectional, case-control and cohort studies conducted between January 1, 1980 and December 31, 2013. No language restrictions were imposed. Studies should include people of any age and gender group with clinically suspected non-traumatic SAH. Participants must be in the detection phase of the disease entering to the emergency department. There were no preferences in any other demographic characteristic of participants. We intended to compare Conventional CT (index test – including multidetector CT) versus angiography, angiotomography (Computed tomography angiography), digital subtraction angiography (DSA) and Magnetic Resonance (MR). We excluded studies assessing aneurysms with the selected tests. W e designed a search strategy for studies published in Medline via PubMed, CENTRAL, LILACS and EMBASE. No language or publication status restrictions. Other electronic sources were used to find additional studies, such as conference abstracts, Google scholar, DARE and PROSPERO. We looked for additional studies in reference lists of selected articles, contact with authors about knowledge of published or unpublished articles. The results of searches were crosschecked in order to eliminate duplicates. Two investigators independently and blindly screened the titles and abstracts to determine the potential usefulness of the articles. Eligibility criteria were applied to the full text articles during the final selection. The risk of bias was assessed independently by at least two researches using the QUADAS2 tool. No meta-analysis was performed due to lack of data. RESULTS: 588 articles (3630 patients) were found with the search strategies designed, after exclusions, three studies were included in qualitative analyses (Mitchell et al., 2001; Perry et al., 2011; Van dijk, Hupperts, Van der Jagt, Bijvoet, & Hasan, 2001). The overall sensibility and specificity for the computed tomography for Perry et al., 2011 were 92.9% (CI95% 89-95.5%) and 100% (CI95% 99.9-100%) respectively. Van dijk et al., 2001 found a sensibility of 30% for observer 1 and 46% of observer 2. The specificity was 100% in both cases. Mitchell et al., 2001 found the sensibility for CT of 95%. The sensibility for MR T2 phase was 94% and the specificity was 98.5% compared to CT. There was high risk of bias mainly for flow and timing and patient selection. There is a vast clinical and CONCLUSIONS: methodological heterogeneity of studies evaluating CT for NTSAH. There is not a gold standard to diagnose this condition. There was high risk of bias related to these studies.