Systematic Review/Meta-analysis
Prediction of Early Adverse Events in Emergency Department Patients With Acute Heart Failure: A Systematic Review

https://doi.org/10.1016/j.cjca.2017.09.004Get rights and content

Abstract

Background

Acute heart failure (AHF) accounts for a substantial proportion of Emergency Department (ED) visits and hospitalizations. Previous studies have shown that emergency physicians' clinical gestalt is not sufficient to stratify patients with AHF into severe and requiring hospitalization vs nonsevere and safe to be discharged. Various prognostic algorithms have been developed to risk-stratify patients with AHF, however there is no consensus as to the best-performing risk assessment tool in the ED.

Methods

A systematic review of Medline, PubMed, and Embase up to May 2016 was conducted using established methods. Major cardiology and emergency medicine conference proceedings from 2010 to 2016 were also screened. Two independent reviewers identified studies that evaluated clinical risk scores in adult (ED) patients with AHF, with risk prognostication for mortality or significant morbidity within 7-30 days. Studies included patients who were discharged or admitted.

Results

The systematic review search generated 2950 titles that were screened according to title and abstract. Nine articles, describing 6 risk prediction tools met full inclusion criteria, however, prognostic performance and ease of bedside application is limited for most. Because of clinical heterogeneity in the prognostic tools and study outcomes, a meta-analysis was not performed.

Conclusions

Several risk scores exist for predicting short-term mortality or morbidity in ED patients with AHF. No single risk tool is clearly superior, however, the Emergency Heart Failure Mortality Risk Grade might aid in prognostication of mortality and the Ottawa Heart Failure Risk Score might provide useful prognostic information in patients suitable for ED discharge.

Résumé

Introduction

L’insuffisance cardiaque aiguë (ICA) constitue une part importante des visites au service des urgences (SU) et des hospitalisations. Les études antérieures ont démontré que l’expérience clinique des médecins d’urgence n’est pas suffisante pour répartir les patients atteints d’une ICA en un groupe de patients atteints gravement et nécessitant une hospitalisation vs en un groupe de patients non atteints gravement et aptes à obtenir un congé de l’hôpital en toute sécurité. Divers algorithmes de pronostics ont été élaborés pour stratifier le risque des patients atteints d’ICA, toutefois, il n’y a aucun consensus sur le meilleur outil d’évaluation du risque au SU.

Méthodes

Nous avons mené une revue systématique de Medline, PubMed et Embase jusqu’en mai 2016 à l’aide de méthodes établies. Nous avons également examiné les comptes rendus majeurs des conférences de 2010 à 2016. Deux examinateurs indépendants ont relevé les études qui permettaient d’évaluer les scores de risque clinique chez les patients (SU) adultes atteints d’ICA, qui avaient un pronostic de risque de mortalité ou de morbidité importante entre 7 et 30 jours. Les études portaient sur les patients qui avaient obtenu leur congé d’hôpital ou avaient été admis.

Résultats

La recherche de revues systématiques a permis de générer 2950 titres que nous avons examinés en fonction du titre et du résumé. Neuf articles qui décrivaient 6 outils de prédiction du risque répondaient entièrement aux critères d’inclusion. Toutefois, la performance pronostique et la facilité de mise en œuvre au chevet du patient sont limitées pour la plupart. En raison de l’hétérogénéité clinique des outils de pronostics et des résultats des études, nous n’avons pas réalisé de méta-analyses.

Conclusions

De nombreux scores de risque existent pour prédire la mortalité à court terme ou la morbidité des patients du SU atteints d’ICA. Aucun outil de risque n’est nettement supérieur. Toutefois, le Emergency Heart Failure Mortality Risk Grade aiderait au pronostic de mortalité et le Ottawa Heart Failure Risk Score fournirait des renseignements utiles sur le pronostic des patients admissibles au congé du SU.

Section snippets

Protocol and registration

The reporting of the search methods and results for this review follow the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines for reporting of systematic reviews.5 The protocol was published on the University of York, Centres for Reviews and Dissemination Web site (registration number CRD42017067290).

Study eligibility criteria

Full-text randomized controlled trials and observational studies were considered for inclusion. Studies must have included ED patients with the clinical

Study selection

The results of the screening process are outlined in the PRISMA flow diagram shown in Figure 1. The original search identified a total of 1921 unique citations after removal of duplicates. Initial screening of titles resulted in 1651 records being excluded. Of the 270 articles screened according to abstract, 250 were excluded, leaving 20 articles for full-text review. Of these, 11 articles were excluded, leaving 9 articles describing 6 risk prediction tools included in the evidence synthesis.

Summary of main results

This review has identified 6 risk prediction tools that might be useful in predicting short-term adverse events and guiding disposition decision for ED patients with AHF. The hierarchy of evidence supporting implementation and use of a risk prediction tool should include derivation, internal validation, external validation, and evidence of effect on either patient-oriented or health services outcomes.18, 19 No risk prediction tool for AHF has achieved this level of supporting evidence. However,

Conclusions

Our systematic review of the literature identified several risk prediction scores for predicting mortality and other adverse events in ED AHF patients. Of these, the EHMRG and OHFRS are supported by the most robust bodies of evidence but differ in important ways with respect to their usability and the types of outcomes that the scores are designed to predict. These risk scores might provide useful prognostic information for ED AHF patients when making disposition decisions, but have yet to show

Funding Sources

This work was supported by a systematic review grant from the Alberta Health Services Emergency Strategic Clinical Network.

Disclosures

The authors have no conflicts of interest to disclose. Dr McRae is a coinvestigator with the Ottawa/Canadian Heart Failure Risk Score team.

References (22)

  • J. McGowan et al.

    PRESS Peer Review of Electronic Search Strategies: 2015 guideline statement

    J Clin Epidemiol

    (2016)
  • I.G. Stiell et al.

    Methodologic standards for the development of clinical decision rules in emergency medicine

    Ann Emerg Med

    (1999)
  • D.S. Lee et al.

    Risk stratification in acute heart failure

    Can J Cardiol

    (2014)
  • Heart and Stroke Foundation. 2016 Report on the Health of Canadians: the Burden of Heart Failure. Available at:...
  • M. Gheorghiade et al.

    Acute heart failure syndromes

    Circulation

    (2005)
  • D.S. Lee et al.

    Early deaths in patients with heart failure discharged from the emergency department: a population-based analysis

    Circ Heart Fail

    (2010)
  • J.A. Ezekowitz et al.

    Acute heart failure in the emergency department: short and long-term outcomes of elderly patients with heart failure

    Eur J Heart Fail

    (2008)
  • A. Liberati et al.

    The PRISMA Statement for Reporting Systematic Reviews and Meta-Analyses of studies that evaluate health care interventions: explanation and elaboration

    PLoS Med

    (2009)
  • S.P. Collins et al.

    Early management of patients with acute heart failure: state of the art and future directions–a consensus document from the SAEM/HFSA acute heart failure working group

    Acad Emerg Med

    (2015)
  • S. Garcia-Gutierrez et al.

    Creation and validation of the acute heart failure risk score: AHFRS

    Intern Emerg Med

    (2017)
  • J. Hsiao et al.

    Validating the acute heart failure index for patients presenting to the emergency department with decompensated heart failure

    Emerg Med J

    (2012)
  • Cited by (16)

    • Comparison of machine learning and the regression-based EHMRG model for predicting early mortality in acute heart failure

      2022, International Journal of Cardiology
      Citation Excerpt :

      Using only 10 predictor variables, the model can predict mortality outcomes at two time points, and identifies those who are low risk, with 0% mortality at 7- and 30-days in those in the lowest risk quintile [9]. While other predictive models have been derived for use in patients with HF in the emergent setting, few have been prospectively validated, and none are used routinely in clinical practice [10]. Only the EHMRG model has been compared with and found to be superior to physician estimated risk, with c-indices of 0.71 vs. 0.81 for prediction of 7-day mortality in favor of the mathematical model [9,11].

    • Short-term mortality risk score for de novo acute heart failure (ESSIC-FEHF)

      2020, European Journal of Internal Medicine
      Citation Excerpt :

      In addition, previously, predictive models and risk scores have been developed to predict poor outcomes for all patients who arrive at the ED with symptoms of AHF. Further, only some of them have been developed to stratify by risk of adverse outcomes in either outpatients or inpatients with heart failure [7]. This research team hypothesized that various different variables determine prognosis in the FEHF and that it is possible to combine them in a risk score to assess probability of death during the vulnerable phase after diagnosis of heart failure.

    View all citing articles on Scopus

    See page 178 for disclosure information.

    View full text