Hospitalize or discharge the emergency department patient with syncope? A systematic review and meta-analysis of direct evidence for SAEM GRACE.

Abstract

Syncope is a frequent reason for hospitalization from the emergency department (ED), but the benefit of hospitalization is unclear. This systematic review and meta-analysis (SRMA) aims to cohere and synthesize the best current evidence regarding the potential benefit of hospitalization for ED syncope patients for developing an evidence-based ED syncope management guideline. We conducted a SRMA according to the patient-intervention-control-outcome (PICO) framework: In patients 16 years of age or older who present to the ED with syncope (population), does hospitalization (intervention) or direct ED discharge (comparison) improve short-term outcomes (outcome)? The primary outcome was a composite of all adverse events as defined by individual studies, up to 30 days. Two reviewers independently assessed articles for inclusion and methodological quality. We measured heterogeneity among included studies with I-squared statistic and used GRADE criteria to assess the quality of evidence. Our search strategy identified 2140 publications and included 18 publications (510,545 participants) in the analysis. All studies reported higher rates of adverse events in hospitalized patients (0.7%-43.8%) compared to discharged patients (0%-3.7%). Our meta-analysis detected considerable statistical heterogeneity. The GRADE assessment for all adverse events and all-cause mortality revealed risk ratios of >5 favoring ED discharge for both outcomes at a median follow-up of 30 days. However, point estimates are limited by serious risk of bias, inconsistency, imprecision, indirectness, and publication bias. Due to the uncertainty of the available evidence, this SRMA's findings do not support a recommendation for or against hospitalizing patients presenting to ED with syncope. However, discharging low-risk patients with syncope from the ED is associated with a low risk of short-term adverse events.