Prognostic value of the emergency heart failure mortality risk grade score in patients admitted to the emergency department for cardiogenic acute pulmonary edema: a prospective cohort study

¹Mongi Slim Academic hospital, Emergency department, Faculty of Medicine of Tunis, University of Tunis El Manar, Tunis, Tunisia

^&Corresponding author
Wiem Demni, Mongi Slim Academic hospital, Emergency department, Faculty of Medicine of Tunis, University of Tunis El Manar, Tunis, Tunisia

Introduction    

Cardiogenic acute pulmonary edema is a common clinical manifestation of acute heart failure. It is one of the most frequent causes of emergency department visits and intensive care unit admissions. It results from a sudden increase in pulmonary capillary wedge pressure, secondary to left ventricular systolic and/or diastolic dysfunction, severe valvular disease, or cardiac rhythm or conduction disturbances [1,2]. Over the past few decades, the prevalence of cardiogenic pulmonary edema has steadily increased. According to the European Society of Cardiology Heart Failure III Registry, published in 2026, acute heart failure affects more than 64 million people worldwide and is becoming increasingly prevalent [3]. Despite therapeutic advances, the prognosis for patients with cardiogenic pulmonary edema remains poor, with an estimated in-hospital mortality rate of 15 to 20% and a one-year survival rate of approximately 50% [4].

In a Tunisian study published in 2021, involving 2 040 patients with acute heart failure from a multicenter registry, 20% presented with acute heart failure, and pulmonary congestion emerged as a predictor of poor prognosis [5]. To date, few scoring systems have been proposed to study and predict the prognosis of patients admitted to the emergency department for management of cardiogenic pulmonary edema. The EHMRG is a clinico-biological score proposed previously to evaluate the prognosis of patients with acute heart failure by assessing short-term mortality risk. Researchers studied this prognostic score in 86 emergency departments in Canada among patients admitted for congestive heart failure. It has been shown to predict early mortality in these patients on day 7, to stratify them based on their risk of death, and to predict their short-term outcomes, thus allowing for patient prioritization, guiding therapeutic decisions, and providing more appropriate and targeted care [6].

In Tunisia, and particularly in North Africa, few or no studies have evaluated the prognostic value of this score in patients admitted to the emergency department for management of cardiogenic pulmonary edema. To assess its clinical applicability in the North African population, which may differ from the initial cohorts in epidemiological and clinical characteristics, comorbidities, and access to care, external validation of this score is essential to evaluate its potential integration into emergency department risk stratification; hence the interest in this cohort. This study aimed to assess the prognostic value of the EHMRG score for mortality at day 7, day 30, and 6 months in patients with cardiogenic acute pulmonary edema.

Methods     

Study design and framework: this was a longitudinal, prospective, analytical study, which was carried out over a period of 18 months (from July 2022 to December 2023) in the emergency department of the Mongi Slim University Hospital Center in Tunisia, which receives more than 76000 consultations/year and more than 3600 hospitalizations/year, the majority of which came from Tunis.

Study population: in this study, we included patients aged 18 years and older who were admitted to the emergency department for management of cardiogenic pulmonary edema and who had given their consent to participate. We did not include patients admitted to the emergency department for acute respiratory distress unrelated to cardiogenic pulmonary edema, patients being treated for end-stage renal disease, or patients undergoing chronic hemodialysis. We subsequently excluded from our study patients with cardiogenic pulmonary edema in whom nephrogenic overload was suspected or confirmed during hospitalization.

Sampling strategy and patient recruitment: patient recruitment was conducted prospectively and consecutively, 24 hours a day, 7 days a week. It was conducted across the department's various functional units: all eligible patients meeting the inclusion criteria were enrolled prospectively and consecutively. No formal sample size calculation was performed. As this was a prospective external validation study of a prognostic score, we consecutively included eligible patients in the cohort throughout the study period.

Definitions and diagnostic criteria: since 2005, the various nosological frameworks of "heart failure" or "acute heart failure" have been integrated into a single clinical entity called "acute heart failure syndrome" [7]. Cardiogenic pulmonary edema constitutes a sub-entity of acute heart failure. The diagnosis of cardiogenic pulmonary edema can be made based on several anamnestic, clinical, biological, radiological, and echocardiographic criteria [8]. In this study, the diagnosis of acute cardiogenic pulmonary edema was based on: anamnestic criteria (presence of cardiovascular risk factors; history and/or treatments suggestive of underlying heart disease; recent onset or worsening of dyspnea; and presence of orthopnea or paroxysmal nocturnal dyspnea); clinical criteria (presence of crackles and/or wheezing on lung auscultation; signs of right-sided heart failure); Radiological criteria (presence of alveolar syndrome: bilateral, symmetrical perihilar opacities with indistinct margins, vascular redistribution toward the lung apices, ± cardiomegaly, ± presence of pleural effusion); echocardiographic criteria (elevated left ventricular filling pressures (LVFP), impaired left ventricular ejection fraction (LVEF), presence of segmental wall motion abnormalities, and presence of underlying heart disease or heart valve disease) and biological criteria (B-type natriuretic peptides (BNP) and NT-proBNP). Chest X-ray and transthoracic echocardiography were performed routinely in all patients, whereas natriuretic peptide levels were measured only in cases of diagnostic uncertainty. The protocol for managing patients admitted to the emergency department with cardiogenic pulmonary edema is based on the latest 2021 recommendations of the European Society of Cardiology [9].

Outcome criteria: the primary outcome in this study was mortality accrued within 7 days of hospitalization. The secondary outcomes were 30-day and 6-month mortality, as determined by the study investigators.

Emergency heart failure mortality risk grade score calculation and prognostic assessment: the EHMRG score is easily calculated in the emergency department. It is composed of 10 simple parameters: age, systolic blood pressure, heart rate, pulse oxygen saturation, serum creatinine level, serum potassium level, troponin level, active neoplasm, metolazone use, and pre-hospital medical transport. This score was first studied in Canada in 2012 by Lee et al. [6] and was validated in 2018 by the same team in a second study, the "ACUTE study" [10]. It predicts mortality on day 7 and day 30 of care for patients admitted to the emergency department for acute heart failure. Based on the EHMRG score, the patients can be classified into five subgroups according to their seven-day and 30-day mortality risk: i) category 1: "very low risk": EHMRG score ≤ -49.1; ii) category 2: "low risk": -49 ≤ EHMRG score ≤ -15.9; iii) category 3: "intermediate risk": -15.8 ≤ EHMRG score ≤ 17.9; iv) category 4: "high risk": 18 ≤ EHMRG score ≤ 56.5; v) category 5: "very high risk": EHMRG score > 56.5. So, beyond its ability to predict mortality, this score allows patients to be stratified by risk level into "very low- or low-risk patients", who can be managed in less intensive care units and discharged after a shorter hospital stay, and "high- or very high-risk patients," who are classified as having a poorer prognosis and require more intensive care. The original EHMRG score was calculated for each patient using the variables included in the original derivation model. The EHMRG score at admission was calculated using an online calculator (MEDCalc) [11]. Missing data were minimal (<5%) and were managed using complete-case analysis. No imputation method was applied.

Data collection: initial data were collected for each patient by the emergency physician who treated them. We collected sociodemographic and medical history data, initial clinical examination data, additional test data, EHMRG score items, data related to therapeutic management, and clinical course data. Mortality on day 7, day 30, and at 6 months was investigated by the study investigators.

Statistical analysis: to better describe the baseline characteristics of our cohort population, we calculated absolute and relative frequencies (percentages) for categorical variables (medical histories, previous hospitalization, presence of right heart failure signs, ultrasound data and use of different types of ventilation) and means, medians, standard deviations (SD), and ranges (minimum and maximum extreme values) for quantitative variables (age, sex ratio, left ventricular ejection fraction, and EHMRG score). The EHMRG score was calculated for each patient using the variables included in the original model. To assess the prognostic value of the EHMRG score in patients with cardiogenic pulmonary edema, we established receiver operating characteristic (ROC) curves according to mortality at day 7, day 30, and 6 months.

Assessment of classical performance indicators of the EHMRG score: after verifying that the area under the ROC curve is greater than 0.5, we chose as "cut-off" the value of the EHMRG score that corresponded to the best "sensitivity - specificity" pair while calculating the positive and negative predictive values. The optimal cut-off value was determined using the Youden Index. We categorized patients with cardiogenic pulmonary edema into two groups: those who died within day 7 ("deceased on day 7") versus those who survived ("survivors on day 7"). To compare two means (age, sex ratio, left ventricular ejection fraction, and EHMRG score), we used student's T-test for independent samples and, if invalid, the non-parametric Mann-Whitney test. To compare two percentages (medical histories, previous hospitalization, presence of right heart failure signs, ultrasound data, and use of different types of ventilation), Pearson's Chi-squared test and, if invalid, Fisher's exact test were used.

To evaluate the prognostic performance of the EHMRG score regarding in-hospital evolution, we subsequently divided the included patients into two groups according to the calculated cut-off value: patients with EHMRG score ≥ cut-off versus patients with EHMRG score < cut-off. To compare two percentages (In-hospital mortality, use of different types of ventilation, Use of vasoactive drugs, Transfer to intensive care), we used Pearson's chi-squared test or, if invalid, Fisher's exact test. We used student's T-test to compare two means (duration of oxygen therapy). In all statistical tests, the significance level (p-value) was set at 0.05, and risk was calculated using the odds ratio (OR) with a 95% confidence interval.

Ethical considerations: this study was conducted in accordance with the ethical principles of research and the Declaration of Helsinki. All patients were informed in advance of the aim of our observational study, and informed consent was obtained by the physician in charge of the patient. Initially, the study was approved by the thesis committee of the Tunis Faculty of Medicine. Local ethics committee of the Mongi Slim University Hospital Center of Tunis (Tunisia) approval was secondarily obtained before manuscript submission (approval No. 19/2026).

Results     

Population characteristics: during the study period, 119,148 patients visited the emergency department for all pathologies. The eligibility criteria were not met for 118,912 patients. Initially, 236 patients were eligible. Twelve patients were subsequently excluded due to an association with nephrogenic pulmonary edema. A total of 224 patients hospitalized in the emergency department for management of cardiogenic pulmonary edema were included. At 30 days and 6 months of follow-up, the number of unreachable patients was 3 (1.3%) and 36 (16%), respectively. Ultimately, the number of patients followed till the end of the study was 188 (84%) (Figure 1). The mean age of the included patients was 71 ± 11 years, with a range from 32 to 98 years. The male-to-female ratio was 1.07. The main comorbidities observed in the patients were hypertension (66.5%), diabetes (48.7%), coronary artery disease (34.8%), and heart failure (29.9%). Thirty-four patients (15.2%) had previously experienced at least one episode of acute cardiogenic pulmonary edema. The overall characteristics and mortality at day 7 of the study population are presented in Table 1. In-hospital mortality was 12.5%. Cumulative mortality on day 7, day 30, and six months was 17%, 22%, and 28%, respectively.

EHMRG score calculation and its prognostic value in terms of 7-day, 30-day, and 6-month mortality: the median EHMRG score was 45 [3-84], with a range from -125 to 351. Table 2 illustrates the EHMRG score variables for the included population. Based on the EHMRG score mortality risk categories, patients were classified as "very high risk" and "high risk" in 44.1% and 22.8% of cases, respectively (Figure 2). No deaths were reported in the "very low" and "low risk" groups at day 7 and day 30, and only one death was reported for the "very low" risk group at six months. When examining the prognostic value of the EHMRG score, the 7-day mortality analysis revealed satisfactory discriminatory ability for mortality, with an area under the ROC curve (AUC) of 0.84 (p<0.001, 95% CI= [0.77-0.91]) and a cutoff of 50.2 (sensitivity=87%, specificity=71%, positive predictive value=38%, and negative predictive value=96%). For 30-day mortality, the AUC of the EHMRG score was 0.87 (p<0.001; 95% CI= [0.81-0.92]), with a cutoff of 61.3 (sensitivity=82%, specificity=76%, positive predictive value=49%, and negative predictive value=94%). At six months, the AUC of the EHMRG score was 0.85 (p<0.001; 95% CI= [0.79-0.90]), with a cutoff of 59.2 (Se=80%, Sp=74%, PPV=49% and NPV=94%) (Figure 3 and Table 3).

Prediction of mortality according to EHMRG risk categories: according to the risk categories of EHMRG score, the "very high risk" category was significantly associated with a higher mortality risk at day 7, day 30, and 6 months (p<0.001). No deaths were reported in the "very low" and "low" risk groups at day 7 and day 30, and only one death was reported for the "very low" risk group at six months (Figure 4).

Prognostic value of the EHMRG score in patients with pulmonary edema, in terms of in-hospital outcome: according to the EHMRG score cutoff at day 7, an EHMRG score ≥ 50.2 was significantly predictive of in-hospital mortality (p<0.001), the need for mechanical ventilation (p<0.001), and transfer to the intensive care unit (p=0.02). Furthermore, we did not observe a statistically significant relationship between this threshold and the need for non-invasive ventilation, vasoactive drugs, length of hospital stay, or duration of oxygen therapy (Table 4).

Discussion     

Strengths: the use of a score, easy to apply in the emergency department, has previously been validated for assessing short-term mortality in patients with acute heart failure. The lack of Tunisian studies focusing on this score: few, if any, Tunisian studies have been conducted to evaluate this score in patients with cardiogenic pulmonary edema in the emergency department. Telephone follow-up of surviving patients, to assess medium-term (1 month) and long-term (6 months) mortality. To the best of our knowledge, no studies in the literature report 6-month mortality by EHMRG score.

Limitations: single-center design and possible selection bias: this was a single-center study conducted on a population from the northern suburbs of Tunis; this population may not be representative of the Tunisian population and may limit the generalizability of our findings to other healthcare settings, potentially leading to selection bias. The sample size is relatively small: this could reduce statistical power and affect statistical performance. Lack of external validation: although our study provides external validation of the EHMRG score in a limited Tunisian population (compared to the Canadian cohort), multicenter studies involving a larger sample are necessary to confirm the external validity of this score and to allow for its adaptation to the Tunisian context. Absence of multivariable analysis: we addressed the primary aim of this study without resorting to a multivariate analysis; consequently, we were unable to verify the presence of associated confounding factors. Losses to follow-up: although follow-up was carefully conducted, losses to follow-up in 30 days (1.3%) and 6 months (16%) could have influenced the assessment of long-term outcomes and may give rise to attrition bias and potentially affect the precision of long-term prognosis.

Conclusion     

The EHMRG score demonstrates satisfactory prognostic performance on day 7, day 30, and 6 months, with high sensitivity and excellent negative predictive value, in patients with cardiogenic pulmonary edema. These qualities make it a particularly relevant tool in emergencies, allowing, thanks to its ease of use and calculation, the effective exclusion of a high risk of short- and medium-term mortality in patients admitted for cardiogenic pulmonary edema. Its integration into clinical practice could thus contribute to more precise risk stratification and better allocation of therapeutic resources in emergency departments. The present study demonstrates good prognostic performance within this cohort; multicenter studies with a larger sample are needed to confirm its external validity before widespread implementation, especially in other North African populations.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Wiem Demni conceived and designed the study, collected and analyzed data, interpreted the findings, and drafted the manuscript. Yessmine Walha analyzed and interpreted the findings and contributed to the literature review. Youssef Zouaghi collected data and participated in data entry. Dhekra Hamdi and Fadwa Lachtar revised the manuscript and contributed to scientific supervision. Nourelhouda Nouira supervised the study and contributed to methodological validation and critical review of the manuscript. All the authors read and approved the final version of this manuscript.

Tables and figures     

Table 1: general characteristics and mortality on day 7 of the included patients

Table 2: variables of the emergency heart failure mortality risk grade score

Table 3: prognostic performance of the emergency heart failure mortality risk grade score in terms of mortality at day 7, day 30, and 6 months

Table 4: prognostic value of the emergency heart failure mortality risk grade score, in terms of unfavorable in-hospital evolution, of patients with acute pulmonary edema

Figure 1: patient flow diagram

Figure 2: patient distribution across mortality risk categories

Figure 3: receiver operating characteristic curves of the emergency heart failure mortality risk grade score

Figure 4: mortality rates by emergency heart failure mortality risk grade

References