The Yield of Computed Tomography of the Head Among Patients Presenting With Syncope: A Systematic Review

Background: Overuse of head computed tomography (CT) for syncope has been reported. However, there is no literature synthesis on this overuse. We undertook a systematic review to determine the use and yield of head CT and risk factors for serious intracranial conditions among syncope patients. Methods: We searched Embase, Medline, and Cochrane databases from inception until June 2017. Studies including adult syncope patients with part or all of patients undergoing CT head were included. We excluded case reports, reviews, letters, and pediatric studies. Two independent reviewers screened the articles and collected data on CT head use, diagnostic yield (proportion with acute hemorrhage, tumors or infarct), and risk of bias. We report pooled percentages, I², and Cochran's Q-test. Results: Seventeen articles with 3,361 syncope patients were included. In eight ED studies (n = 1,669), 54.4% (95% confidence interval [CI] = 34.9%–73.2%) received head CT with a 3.8% (95% CI = 2.6%–5.1%) diagnostic yield and considerable heterogeneity. In six in-hospital studies (n = 1,289), 44.8% (95% CI = 26.4%–64.1%) received head CT with a 1.2% (95% CI = 0.5%–2.2%) yield and no heterogeneity. In two articles, all patients had CT (yield 2.3%) and the third enrolled patients ≥ 65 years old (yield 7.7%). Abnormal neurologic findings, age ≥ 65 years, trauma, warfarin use, and seizure/stroke history were identified as risk factors. The quality of all articles referenced was strong. Conclusion: More than half of patients with syncope underwent CT head with a diagnostic yield of 1.1% to 3.8%. A future large prospective study is needed to develop a robust risk tool.