MP23: The yield of computed tomography of the head in patients presenting with syncope: a systematic review

Abstract

Introduction: Syncope accounts for 1-3% of Emergency Department (ED) visits. Previous studies have reported overuse of computed tomography (CT) of the head among syncope patients. Professional organizations including Choosing Wisely have recommended against its use in the absence of high-risk features. However, a review of CT head use among syncope patients and its diagnostic yield has not been previously reported. Methods: We conducted a systematic review using EMBASE, Medline, and Cochrane databases from inception to August 2016. We included studies involving adult syncope patients that reported CT head use and its diagnostic yield during acute management by a two-step process: first title/abstract review and then full-text review of selected articles. We excluded case reports, narrative reviews and those involving children. We collected the proportion of patients who had CT head performed, and its diagnostic yield. Outcomes included identification of acute intracranial conditions (hemorrhage, mass or infarct) that require further management. Two reviewers independently abstracted the data and discrepancies were resolved by consensus. We calculated inter-observer reliability for inclusion in the systematic review using kappa values. We performed meta-analysis for diagnostic yield of the CT head. Results: Fifteen studies with 2,802 syncope patients in four sub-groups (proportion of patients among whom CT head was performed and its yield in ED and inpatient settings; studies that reported only the yield among those with CT head performed; and the use and yield among syncope patients ≥65 years old) were included. The inter-observer agreement for inclusion of final articles for meta-analysis was κ=0.925 [95% CI: 0.861-0.990]. Seven ED studies (n=1,261) reported 55.7% patients (95% CI: 32.1-78.0%) had head CT performed with a yield of 4.0% (95% CI: 2.7-5.6%); 5 studies with 1138 hospitalized patients reported that 38.6% (95% CI: 20.4-58.6%) had head CT with a yield of 1.1% (95% CI: 0.4-2.2%). The yield among studies that report only outcomes for CT head was 2.3% and the yield among patients’ ≥65 years was 7.7%. Conclusion: Our review found that a very high proportion of syncope patients had CT head performed during acute management with a very low diagnostic yield. The yield is higher among patients ≥65 years old. A robust tool to identify patients who require a CT head will reduce unnecessary testing.