Short-term outcomes of acute heart failure hospitalizations in Ethiopia: A multicenter prospective study

Abstract

Objectives

To determine the incidence and predictors of unplanned 30-day readmission and in-hospital mortality among adults hospitalized with AHF in Ethiopia.

Methods

A multicenter prospective observational study was conducted in six referral hospitals in the Amhara region between December 2023 and April 2024. Adults (≥18 years) with AHF were consecutively enrolled and followed up until discharge and 30 days post-discharge. Kaplan–Meier survival analysis and Cox regression were used to estimate outcomes and predictors.

Results

Of the 1131 patients, 275 (24.4 %) were readmitted within 30 days and 121 (10.7 %) died in the hospital. Independent predictors of readmission included hyponatremia (AHR = 10.5; 95 % CI: 3.1–36.2), thrombocytopenia (AHR = 16.7; 95 % CI: 4.8–58.3), ischemic heart disease (AHR = 6.9; 95 % CI: 1.8–27.0), Charlson Comorbidity Index ≥4 (AHR = 6.5; 95 % CI: 1.7–24.6), poor physician adherence to guideline-directed therapy (AHR = 8.2; 95 % CI: 2.3–30.1), and low patient adherence (AHR = 4.8; 95 % CI: 1.6–14.5). Prescription of ACE inhibitors, beta-blockers, and SGLT2 inhibitors at discharge significantly reduced the readmission risk (AHR range: 0.09–0.30). The predictors of in-hospital mortality included reduced ejection fraction, tachycardia, hypoxemia, left bundle branch block, pulmonary hypertension, elevated creatinine, severe hypertension, and pneumonia.

Conclusion

AHF patients in Ethiopia experience high short-term readmission and mortality. Correcting electrolyte imbalances, improving comorbidity management, strengthening physician adherence to guideline-directed therapy, and promoting patient adherence are essential for improving outcomes.

1. Introduction

Heart failure (HF) is a major global public health problem, affecting more than 64 million people and accounting for substantial morbidity, hospitalizations, and premature deaths worldwide [1,2]. Despite advances in pharmacological and device-based therapies, patients with acute heart failure (AHF) remain at a high risk of early readmission and in-hospital mortality [3,4]. These outcomes reflect both the underlying severity of the disease and persistent deficiencies in inpatient and post-discharge care.

In high-income countries (HICs), 30-day readmission rates after AHF hospitalization generally range between 15 % and 25 %, and in-hospital mortality between 4 % and 10 %, supported by well-organized care transitions and broader use of guideline-directed medical therapy (GDMT) [5,6]. In contrast, low- and middle-income countries (LMICs) report substantially higher rates of both outcomes, largely due to delayed presentation, limited diagnostic and therapeutic resources, and fragmented post-discharge follow-up [[7], [8], [9]].

Recent studies from sub-Saharan Africa highlight the disproportionate burden of HF in younger populations, often driven by hypertension, ischemic heart disease, cardiomyopathies, and rheumatic valvular disease [[10], [11], [12]]. In Ethiopia, these conditions are the leading causes of HF; however, the true burden, clinical profiles, and outcomes of AHF remain poorly characterized. Prior research has been mainly single-center and retrospective, limiting external validity and failing to capture the diversity of patients across different regions and care settings [8,13,14].

Unplanned 30-day readmission is increasingly recognized as a quality-of-care metric, reflecting incomplete stabilization, premature discharge, or lack of continuity of care [15,16]. Similarly, in-hospital mortality remains an important marker of acute clinical instability, often associated with renal dysfunction, electrolyte imbalance, pneumonia, and comorbid conditions [17,18]. Understanding the predictors of these adverse outcomes is critical for guiding targeted interventions to reduce avoidable hospitalizations, improve survival, and strengthen HF care systems in resource-limited environments.

To address these gaps, we established the first multicenter prospective acute heart failure (AHF) registry in Ethiopia. This design enables systematic and standardized data collection across multiple tertiary hospitals, reducing recall and selection bias while capturing regional variations in clinical characteristics, management, and outcomes. The multicenter, prospective approach enhances data quality, representativeness, and generalizability, providing robust evidence to inform clinical management strategies, healthcare resource allocation, and the development of national heart failure care guidelines for Ethiopia and comparable LMIC settings [9,19,20].

Accordingly, this study aimed to determine the incidence of unplanned 30-day readmission and in-hospital mortality and to identify their clinical and healthcare-related predictors among adults hospitalized with acute heart failure in Ethiopia.

2. Methods

2.1. Study setting and periods

A multicenter prospective observational study was conducted from December 1, 2023, to April 30, 2024, at six comprehensive and specialized hospitals in the Amhara region, Ethiopia: Felege Hiwot Comprehensive Specialized Hospital (FHCSH), University of Gondar Comprehensive Specialized Hospital (UoGCSH), Debre Tabor Comprehensive Specialized Hospital (DTCSH), Debre Markos Comprehensive Specialized Hospital (DRCSH), Tibebe Ghion Comprehensive Specialized Hospital (TGCSH), and Debre Berhan Comprehensive Specialized Hospital (DMCSH). These hospitals serve as referral centers for large catchment populations and provide inpatient and outpatient care for cardiovascular diseases, including heart failure.

2.2. Study design

A prospective observational design was employed. Patients were recruited consecutively at each hospital and followed from admission through discharge and for 30 days post-discharge.

2.3. Population and eligibility criteria

Adults (≥18 years) admitted with a primary diagnosis of acute heart failure (AHF), confirmed clinically and supported by echocardiographic or laboratory findings, regardless of ejection fraction, were included. Patients unwilling to provide consent, those transferred from other study hospitals to avoid duplication, and those without reachable contact information for follow-up were excluded.

2.4. Sample size, sampling technique, and site allocation

This study aimed to establish a multicenter prospective registry of acute heart failure patients in Ethiopia using a census approach. The total sample size of 1131 participants was derived from all consecutive, eligible patients admitted to six tertiary hospitals. This approach ensured representativeness and external validity and allowed for meaningful subgroup analyses and improved generalizability to other Ethiopian and sub-Saharan African hospital settings. The study population was selected using a conventional sampling technique to enhance the representativeness. Data collection was conducted across multiple sites, with site allocation proportional to patient volume, ensuring balanced and comprehensive coverage as depicted in Fig. 1.

Fig. 1.

Diagram illustrating patient recruitment, eligibility assessment, enrollment, and 30-day follow-up across six comprehensive specialized hospitals in the Amhara region, Ethiopia.

Footnotes: Abbreviations: FHCSH, Felege Hiwot Comprehensive Specialized Hospital; UoGCSH, University of Gondar Comprehensive Specialized Hospital; DTCSH, Debre Tabor Comprehensive Specialized Hospital; DRCSH, Debre Markos Comprehensive Specialized Hospital; TGCSH, Tibebe Ghion Comprehensive Specialized Hospital; JUMC, Jimma University Medical Center.

2.5. Study variables

The study analyzed in-hospital mortality and 30-day unplanned readmission as dependent variables, with the independent variables categorized into patient-related factors, clinical-related factors, and healthcare-related factors. Patient-related factors included socio-demographic and health history, clinical-related factors included presenting signs and symptoms, and healthcare-related factors included care coordination and follow-up.

2.6. Data collection instruments and procedures

The data collection instruments were developed in English after an intensive literature review [3,5,15,[21], [22], [23], [24]]. The data collection instruments were developed in English following an extensive review of the existing literature to ensure comprehensiveness and relevance to the study objectives. These instruments included structured questionnaires and data abstraction forms designed to capture sociodemographic information, clinical variables, treatment adherence, and outcome measures.

Data were collected directly from patients, caregivers, and medical records by two trained nurse professionals who underwent a rigorous orientation and training session to ensure uniform understanding and application of the data collection tools. To minimize observer bias and enhance reliability, inter-rater reliability assessments were conducted during the training phase by having both nurses independently collect data from a pilot sample, with discrepancies discussed and resolved.

Post-discharge follow-up data were obtained through multiple sources, including telephone interviews, follow-up clinic visits, and review of patients' medical charts, ensuring comprehensive capture of unplanned 30-day readmissions and mortality. To overcome language barriers and improve patient comprehension, the questionnaires were translated from English into Afan Oromo and Amharic—the predominant local languages—and back-translated to English to verify accuracy.

Adherence to prescribed treatments was assessed through a combination of physician reports, pharmacy dispensing records, and patient self-reports, providing a multidimensional evaluation of medication adherence. Regular supervision and daily monitoring of data collection activities by the principal investigator ensured the completeness, consistency, and timely resolution of any issues encountered during the process. The study used the 5-item Medication Adherence Report Scale (MARS-5) to measure medication adherence. A score below 20 was considered poor, indicating missed doses or inadequate regimen. A score above 20 indicated good adherence, indicating consistent use. This helps identify patients at risk of adverse clinical outcomes.

In addition, periodic data quality checks were performed, including cross-verification of the entered data against the source documents and random re-interviews of a subset of participants to confirm data accuracy. These measures aimed to maintain high data integrity and minimize missing or inconsistent information throughout the study.

2.7. Data quality assurance

Prior to the main study, the data collection instruments were pretested on 5 % of the planned sample size in a similar patient population at one of the participating hospitals. This pretest aimed to evaluate the clarity, relevance, and applicability of the questionnaire and data abstraction forms. Feedback from the pretest was used to refine the instruments and address any ambiguities or logistical challenges.

During the data collection phase, the principal investigator conducted daily reviews of the completed data forms to ensure completeness, accuracy, and consistency. Any missing or unclear data were promptly identified and rectified through direct communication with the data collectors or by revisiting the source documents.

To further enhance the data quality, the following procedures were implemented:

• ▪Training and Standardization: Data collectors received comprehensive training on the study protocols, data collection tools, and ethical considerations to standardize procedures and minimize variability.
• ▪Supervision: Regular supervisory visits and spot checks were conducted by the principal investigator and study supervisors to monitor adherence to protocols and provide immediate feedback.
• ▪Double Data Entry: Data were entered independently by two trained data clerks into the Epidata software to minimize entry errors. Discrepancies between entries were reconciled by cross-checking the original data sources.
• ▪Consistency Checks: The dataset was examined for logical consistency, range checks, and outliers using the STATA software. Queries arising from these checks were resolved by consulting the original records.
• ▪Data Security: All collected data were securely stored with password protection, and access was limited to authorized personnel to maintain confidentiality and prevent data loss.

2.8. Data processing and statistical analysis

Data entry was performed using Epidata version 4.6, and statistical analyses were conducted using STATA version 17. The distribution of continuous variables was assessed for normality using the Shapiro-Wilk test to determine whether parametric or non-parametric statistical methods were appropriate.

Kaplan–Meier survival curves were generated to visualize the time-to-event data, specifically the risk of unplanned 30-day readmission. Differences between the groups were compared using the log-rank test, which evaluates whether there are statistically significant differences in survival distributions.

To identify the factors associated with unplanned readmission, Cox proportional hazards regression models were used. Before fitting these models, the proportional hazards assumption, which states that the relative risk between groups remains constant over time, was tested using Schoenfeld residuals. Schonfeld residuals are a diagnostic tool in survival analysis that helps verify whether the assumption holds for each explanatory variable. Variables that met this criterion and had a p-value less than 0.25 in the bivariate (unadjusted) Cox regression were selected as candidates for the multivariate (adjusted) Cox regression.

In the multivariate analysis, statistical significance was defined as a p-value of 0.05. Adjusted hazard ratios (AHR) with 95 % confidence intervals quantified the strength and direction of the associations between the predictors and the risk of unplanned readmission.

2.9. Operational definition and definition of terms

• ▪Short-term clinical outcomes: Unplanned hospital readmission within 30 days after discharge and in-hospital mortality.
• ▪Comorbidity: The presence of one or more additional medical conditions in a patient, which may be related or unrelated to the primary diagnosis [21].
• ▪Index admission: The initial hospital admission for acute heart failure.
• ▪Unplanned hospital readmission: Any hospital admission of a patient with heart failure within 30 days after discharge, excluding scheduled follow-up visits or emergency visits unrelated to heart failure.
• ▪Follow-up time: The duration from the date of discharge until the occurrence of unplanned readmission or censoring.
• ▪Hazard risk status: The patient's condition at the end of the follow-up period, classified as either censored (no readmission) or having experienced an unplanned readmission.
• ▪Physician adherence: Prescription of guideline-directed medical therapy (GDMT), including ACE inhibitors (ACEI), angiotensin receptor blockers (ARB), angiotensin receptor–neprilysin inhibitors (ARNI), β-blockers, mineralocorticoid receptor antagonists (MRA), and SGLT2 inhibitors (SGLT2i).
• ▪Patient adherence: Measured using the Medication Adherence Report Scale (MARS-5); a score < 20 indicates non-adherence.
• ▪ Laboratory cut-offs:
• •Hyponatremia: serum sodium <135 mmol/L
• •Thrombocytopenia: platelet count <150 × 10⁹/L
• •Elevated creatinine: > 1.2 mg/dL
• •Additional laboratory thresholds defined according to the standard clinical guidelines.

3. Results

3.1. Study cohort derivation and hospital distribution

The study assessed 1556 patients, excluding 425 due to in-hospital mortality. A total of 1131 discharged alive were included in the final 30-day readmission analysis, with the University of Gondar Comprehensive Specialized Hospital contributing the most (Fig. 1).

3.2. Baseline characteristics

A total of 1131 adults hospitalized with acute heart failure were included in the analysis. The mean age was 56 ± 12 years, and 57 % were men. Most participants (77 %) resided in rural areas and were admitted to cardiac wards. Chronic decompensated heart failure accounted for nearly three-quarters of the cases, while 27 % presented with de novo AHF. More than one-third of the patients had a preserved ejection fraction (HFpEF), and most were in NYHA functional class II–IV. Hyponatremia, thrombocytopenia, and elevated serum creatinine were frequent laboratory abnormalities. Pulmonary edema and pneumonia were the most common radiographic findings, and ischemic heart disease was the predominant echocardiographic abnormality. Common comorbidities included hypertension (16 %), chronic kidney disease (15 %), anemia (18 %), and ischemic heart disease (38 %). Nearly three-quarters had prior HF hospitalizations, with ischemic and rheumatic valvular diseases being the leading etiologies (Table 1).

Table 1.

Baseline characteristics of patients hospitalized with acute heart failure in Ethiopia (N = 1131).

Characteristic	n (%) or mean ± SD
Sociodemographic characteristics
Age (years), mean ± SD	56 ± 12
Male sex	648 (57.3)
Rural residence	866 (76.6)
Admitted to the cardiac ward	795 (70.3)
Clinical presentation
Chronic AHF	825 (72.9)
De novo AHF	306 (27.1)
Ejection fraction ≥50 % (HFpEF)	418 (37.0)
NYHA class II–IV	942 (83.3)
Median length of hospital stays (days, IQR)	10 (6–15)
Key laboratory and imaging findings
Hyponatremia (low serum sodium)	198 (17.5)
Thrombocytopenia	147 (13.0)
Elevated serum creatinine levels	231 (20.4)
Pulmonary edema (X-ray)	978 (86.5)
Pneumonia (X-ray)	259 (22.9)
Electrocardiography and echocardiography
Atrial fibrillation	177 (15.7)
Left bundle branch block	122 (10.8)
Ischemic heart disease	407 (36.0)
Pulmonary hypertension	153 (13.5)
Comorbidities and precipitation factors
Hypertension	183 (16.2)
Chronic kidney disease/AKI	170 (15.0)
Anemia	206 (18.2)
Ischemic heart disease (comorbidity)	424 (37.5)
Severe pneumonia	143 (12.6)
Drug discontinuation	130 (11.5)
Charlson Comorbidity Index ≥4	189 (16.7)
Admission history and etiology
Previous HF admission	838 (74.1)
Major etiology: Ischemic heart disease	494 (43.7)
Rheumatic valvular disease	242 (21.4)
Hypertensive heart disease	100 (8.8)
Degenerative valvular disease	105 (9.3)

Abbreviations: AHF = acute heart failure; HFpEF = heart failure with preserved ejection fraction; NYHA = New York Heart Association; AKI = acute kidney injury; CCI = Charlson Comorbidity Index.

Note: Detailed sociodemographic, clinical, imaging, and comorbidity data are available in Supplementary Tables S1–S5 (appendix B).

3.3. Unplanned hospital readmission rate in heart failure patients

This study followed 1131 patients with heart failure who were discharged alive. A total of 275 patients were readmitted within 30 days, with a 24.35 % incidence. The time at risk was 52,352 days, with a readmission incidence rate of 0.9 % per day. The median readmission was 30 days. The highest readmission rate was 25–28 days (Fig. 2).

Fig. 2.

Kaplan–Meier curve depicting the cumulative incidence of unplanned 30-day readmission among adult patients hospitalized with acute heart failure across six specialized hospitals in the Amhara region, Ethiopia.

Footnote: Supplementary Appendix B includes additional Kaplan–Meier survival analyses stratified by physician adherence to guideline-directed medical therapy, platelet count, serum sodium, pneumonia status, and other clinical variables (Figs. S1–S5), with corresponding log-rank test p-values (< 0.05 for all comparisons).

3.4. Unplanned readmission and its predictors in heart failure

The Cox regression analysis revealed that several predictors significantly impact the risk of unplanned readmission in patients. Low serum sodium, low platelet count, ischemic heart disease, a Charlson Comorbidity Index score, poor guideline adherence, and discharge medications have all been found to have a higher risk of unplanned readmission. Patients discharged on ACEIs, BBs, and SGLT2 inhibitors had a 70 % lower risk of unplanned readmission compared with those not taking these medications (Table 2).

Table 2.

Multivariate Cox proportional hazards regression analysis identifying predictors of unplanned 30-day readmission among patients with acute heart failure admitted to six specialized hospitals.

Predictor	Category	CHR (95 % CI)	p-value	AHR (95 % CI)	p-value
Serum sodium	High	1.30 (0.55–3.10)	0.55	2.68 (0.64–11.24)	0.177
	Normal	1	–	1	–
	Low	2.38 (1.28–4.42)	0.006	10.52 (3.06–36.18)	0.001
Platelet count	Normal	1	–	1	–
	Low	23.00 (9.10–58.30)	0.001	16.73 (4.80–58.30)	0.001
Heart block	Yes	3.67 (1.04–10.85)	0.042	8.60 (1.10–70.00)	0.0431
	No	1	–	1	–
CCI (≥4)	Yes	20.26 (8.00–51.36)	0.001	6.46 (1.72–24.57)	0.006
	<4	1	–	1	–
Discharge beta-blocker	Yes	0.43 (0.24–0.77)	0.004	0.17 (0.04–0.74)	0.018
	No	1	–	1	–
SGLT2 inhibitor at discharge	Yes	0.32 (0.10–1.04)	0.059	0.09 (0.02–0.52)	0.007
	No	1	–	1	–
MARS-5 Score (<20)	Yes	11.46 (6.30–20.84)	0.001	4.82 (1.60–14.47)	0.005
	≥20	1	–	1	–
Physician adherence to GDMT (Poor)	Poor	7.00 (2.47–8.87)	0.001	8.23 (2.25–30.10)	0.001
	Moderate	1	–	1	–
	Good	0.46 (0.14–1.49)	0.194	0.14 (0.01–1.55)	0.109

Footnotes.

¹

Wide CI for heart block reflects small subgroup size and statistical uncertainty; interpret with caution.

• ▪
  CHR: crude hazard ratio; AHR: adjusted hazard ratio; CI: confidence interval; p-value: statistical significance.
• ▪
  SGLT2: Sodium-glucose cotransporter-2 inhibitor; CCI: Charlson Comorbidity Index; MARS-5: Medication adherence rating scale; GDMT: Guideline-directed medical therapy.
• ▪
  Analysis: Multivariate Cox proportional hazard regression; AHRs adjusted for all variables listed.
• ▪
  Detailed sociodemographic, clinical, imaging, comorbidity, and heart failure admission pattern data, including full multivariate Cox proportional hazards regression results, are provided in Supplementary Appendix A (Sections S1–S5, Tables 4–9).

3.5. Predictors of in-hospital mortality

This study found that the key predictors of mortality risk among patients included pulse rate, reduced oxygen saturation, dyspnea during exertion, left bundle branch block, pulmonary hypertension, elevated serum creatinine, severe hypertension, and severe pneumonia. These factors increase the mortality risk, necessitating urgent care and prompt treatment for high-risk patients (Table 3).

Table 3.

Predictors of in-hospital mortality among acute heart failure patients admitted to six specialized hospitals in the Amhara region.

Predictors	Category	AHR (95%CI)t1	P-value
HFEF		0.84(0.73–0.96)	0.012
Diastolic blood pressure		0.93(0.84–1.02)	0.131
Pulse rate		1.90(0.85–2.76)	0.001
Respiratory rate		1.03(0.88–1.22)	0.702
dyspnea		1.21(1.02–1.44)	0.031
Pleural effusion	Yes	1.31(0.28–6.16)	0.735
	No	1
Dyspnea on exertion	Yes	2.07(2.04–2.95)	0.027
	No	1
LBBB	Yes	19.70(1.54–25.20)	0.022
	No	1
IHD	Yes	5.31(0.39–72.97)	0.212
	No	1
Pulmonary HTN	Yes	21.88(1.87–56.15)	0.014
	No	1
Serum creatinine increased		2.76(1.44–5.27)	0.002
Sever hypertension	Yes	5.47(1.17–21.85)	0.041
	No	1
Severe pneumonia	Yes	8.14(2.42–36.25)	0.005
	No

Footnotes.

¹Reference, AHR reflects per-unit change, e.g., pulse rate (1 beat/min), oxygen saturation (1%). AHR: adjusted hazard ratio; CI: confidence interval; p-value: statistical significance. HFEF: Heart failure with preserved ejection fraction; IHD: Ischemic heart disease; LBBB: Left bundle branch block; Pulmonary HTN: Pulmonary hypertension. Analysis: Multivariate Cox proportional hazard regression; AHRs were adjusted for all listed predictors. Interpretation caution: Wide CIs for LBBB, IHD, and pulmonary HTN reflect small subgroup sizes.

4. Discussion

This study assessed the magnitude and determinants of 30-day unplanned hospital readmissions among patients admitted with acute heart failure (AHF) across six specialized hospitals in the Amhara region. The overall 30-day unplanned readmission rate was 24.35 %, indicating a substantial post-discharge burden and reflecting both clinical and system-level challenges in heart failure management. Multivariable Cox regression analysis identified several independent predictors of early readmission, including low serum sodium, low platelet count, ischemic heart disease, heart block, Charlson Comorbidity Index (CCI) ≥4, and poor medication adherence (MARS-5 score < 20). In contrast, the prescription of beta-blockers (BBs), angiotensin-converting enzyme inhibitors (ACEIs)/angiotensin receptor blockers (ARBs), and SGLT2 inhibitors at discharge was associated with a significantly lower risk of readmission. Furthermore, poor physician adherence to GDMT markedly increased the risk of unplanned readmission. Collectively, these findings underscore that both clinical factors, reflecting disease severity and comorbidity burden, and health system factors, such as provider adherence and continuity of care, play crucial roles in determining early rehospitalization among patients with AHF.

This suggests a relatively high rate of readmissions, concerns with the initial hospitalization's quality of treatment, or challenges with the shift from inpatient to outpatient care. A similar readmission rate was reported in Ethiopia [25], complying with the current study, and a higher unplanned readmission rate was reported [26,27]. On the other hand, a lower readmission rate was reported in other studies [5,28]. This disparity could be partially explained by the differences in the healthcare settings examined. While the current study was conducted exclusively among patients admitted to medical wards, the other studies referenced included a mix of patients from critical care units and other inpatient subunits as well. Additionally, there may have been differences in the baseline characteristics of the patient populations included across the studies. Factors such as disease severity, comorbidities, and access to certain medications or specialized care could have varied between the study cohorts.

Additionally, this study reveals that the unplanned 30-day readmission and in-hospital mortality rates among patients hospitalized with acute heart failure (AHF) are similar to those reported in several prominent registries worldwide. These differences are due to variations in healthcare infrastructure, socioeconomic factors, patient demographics, comorbid conditions, and adherence to clinical management guidelines. The THESUS-HF registry reported high rates of early readmission and in-hospital mortality, highlighting challenges such as limited access to diagnostic tools, shortages of evidence-based therapies, and delayed presentation of patients with advanced disease [24]. The INTER-CHF registry also identified ischemic heart disease, renal dysfunction, and poor adherence to medication as major predictors of readmission and mortality [19]. The ADHERE registry showed lower rates of 30-day readmission and in-hospital mortality, highlighting gaps in care transitions, patient education, and medication access [29]. The European Society of Cardiology Heart Failure Long-Term registry reported intermediate rates of readmission and mortality, emphasizing the importance of evidence-based treatment protocols [3]. The Brazilian Heart Failure Registry also reported similar rates, highlighting the need for targeted interventions focusing on comorbidity management and adherence support [30].

Risk factors that contribute to 30-day unplanned readmissions are low serum sodium, low platelet counts, heart block, ischemic heart disease, Charlsol comorbidity index above 4, and MARS-5 score below 20. On the other hand, medications prescribed at discharge, such as ACEI, BB, and SGLT2 inhibitors, were associated with lower readmission risks.

The current study has shown that low serum sodium levels or hyponatremia can significantly increase the risk of unplanned hospital readmissions. This implies a 2.4-fold higher risk for patients with hyponatremia than those with normal sodium levels. This association has been observed in previous studies [23,31]. In addition, in the present study, the risk of unplanned readmission was found to increase by more than twelvefold among patients with a low platelet level at admission as compared to a normal platelet level. In line with this study, other studies have shown a decrease in platelet levels and an increase in unplanned readmissions [25].

In the current study findings, ischemic heart disease increased the risk of unplanned readmissions by sevenfold compared with those without ischemic heart disease. The consistent findings across these current and previous studies demonstrate a strong and significant association between ischemic heart disease and an elevated risk of unplanned, frequent readmissions in patients with heart failure [27,32]. Likewise, the study indicated that the risk of unplanned readmissions was found to be around seven times higher for the higher comorbidity Charlson index than for the lower comorbidity Charlson index. This finding agrees with other study findings, which found an increased risk of unplanned readmission compared with the lower comorbidity index [14]. The present findings showed that patients with heart failure who had heart block had a higher chance of needing to be readmitted. This implies that patients with heart block had an adjusted hazard ratio (AHR) of readmission of approximately 9 times. Contrary to this finding, a lower hazard ratio was reported [33]. A discrepancy could occur due to the sample size, statistical power, methodology, and specific patient characteristics.

This study found that patients who received BBs at discharge were less likely to be readmitted than those who did not receive BBs. This implies that BB prescription at discharge has a protective effect against unplanned readmission. This finding was consistent with the study conducted in Ethiopia [7]. Additionally, the present study found that ACEI/ARB prescription at discharge had a protective effect on unplanned readmission compared with non-prescribed at discharge. Therefore, the findings show that there is a potential opportunity to improve patient outcomes and reduce healthcare use by focusing on evidence-based medication management practices during care transitions, which could have meaningful clinical and economic implications. The findings of this study align with the results reported in previous related research [34,35]. In addition, this study showed that SGLTI2 prescribed at discharge decreased unplanned readmissions. This finding agrees with the study done in South Korea [17]; in addition, this finding is supported by the report's findings from multinational nations [20,36,37]. The discrepancy may be due to the strict follow-up of the patients and adherence to the treatment protocol in the previous study.

In this study, patients with a MARS-5 score below 20 had nearly 5 times higher risk of unplanned hospital readmission compared with those with a MARS-5 score of 20 or higher. This suggests that the MARS-5 score itself is an independent predictor of unplanned readmissions. This result is consistent with previous studies showing that patients who do not adhere to their prescribed medication have a significant increase in readmission [38]. Previous studies have shown that poor medication adherence is associated with an increased risk of unplanned readmissions in patients with heart failure [26]. Improving medication adherence in patients with heart failure can help optimize treatment outcomes, reduce the risk of exacerbations, and decrease the likelihood of unplanned readmissions. The current study found a concerning discrepancy compared with prior research. Poor physician adherence to clinical guidelines was associated with an 8-fold increased risk of unplanned hospital readmission. This is in contrast to other studies that documented a lower hazard ratio for this relationship [3]. This discrepancy between the current findings and the results from previous research is related to the ejection fraction and the study population. This resulted in a significant link between medication adherence and the risk of unplanned 30-day readmission in patients with acute heart failure. Patients with a MARS-5 score below 20 had a five-fold higher risk of readmission. This highlights the importance of adherence in preventing hospital readmissions and improving patient outcomes.

The study found large adjusted hazard ratios for heart block, pulmonary hypertension, and thrombocytopenia, which are plausible due to advanced heart failure. However, wide confidence intervals suggest statistical imprecision due to smaller subgroup sizes and unmeasured confounding factors. These findings highlight the complexity of predicting readmission in real-world acute heart failure populations, requiring future multicenter studies with larger sample sizes and stratified analyses. Taken together, these results reflect the multifactorial nature of early readmissions and highlight the interplay between patient severity and healthcare delivery factors.

The study reveals a strong link between poor physician adherence to guideline-directed medical therapy (GDMT) and unplanned readmission in patients with heart failure with reduced ejection fraction. This is due to the clinical and system-level dynamics, such as inconsistent drug availability and limited follow-up infrastructure. These findings underscore the clinical importance of GDMT adherence and the broader systemic challenges in low- and middle-income contexts.

4.1. In-hospital mortality and its predicators

The findings of this study provide crucial insights into the predictors of in-hospital mortality among heart failure patients in the rural Gondar district of Ethiopia. Understanding these predictors is essential for improving clinical outcomes, as early identification of high-risk patients can facilitate timely and targeted interventions. Left ventricular Ejection Fraction (LVEF) emerged as a significant predictor of mortality [39]. Higher ejection fractions correlate with better cardiac function, suggesting that maintaining or improving LVEF is critical in managing heart failure [40].

Interventions aimed at improving quality of life, functional capacity and reducing HF hospitalization such as optimizing medical therapy with ACEI/ARB/ARNI, SGLT2i, beta-blockers, and mineralocorticoid receptor antagonists, should be targeted [41,42]. Regular echocardiographic evaluations should be part of routine care, enabling timely adjustments in management [43].

Another significant predictor was heart rate, where greater rates were linked to a higher chance of death (AHR 1.90). Elevated heart rate can be the result of compensatory mechanisms in decompensated heart failure and frequently reflect higher myocardial oxygen demand and lead to increased adverse outcomes [[44], [45], [46]] Clinicians can identify patients at risk of worsening by constantly monitoring heart rates during hospital stays [47]. Patients with high heart rate can be stabilized with therapeutic measures such as inotropic drug administration or diuretic adjustment. Furthermore, knowing how heart rate and mortality are related might guide clinical judgment and result in more individualized treatment plans.

Similarly, dyspnea (AHR 1.21) levels play a critical role in assessing patient status. Low oxygen saturation can exacerbate HF by increased pulmonary hypertensions. Currently, pulmonary hypertension in these patients a consequence of elevated ventricular diastolic pressure, and the primary approach to treatment should focus on optimizing GDMT for HF [[48], [49], [50]].

The presence of cardiomegaly significantly increased the risk of mortality (AHR 9.31). Cardiomegaly often indicates advanced heart failure and increased risk of mortality rates and highlights the need for comprehensive cardiac evaluation [51]. Patients exhibiting this feature may require hospitalization for optimization of their heart failure management, including the use of diuretics to relieve fluid overload and monitoring for complications [52,53] The identification of cardiomegaly can also signal the need for more aggressive treatment strategies, including consideration for advanced heart failure therapies [54].

Dyspnea on exertion (AHR 2.07) is a common symptom of heart failure that serves as an important clinical marker. Its presence indicates functional impairment, can signal worsening heart failure, and its increased mortality by more than two times [55,56]. Effective management of dyspnea is crucial, as alleviating this symptom can significantly enhance quality of life which can improve cardiac efficacy and enhance exercise tolerance in HF patients. Clinicians should prioritize the use of diuretics and other heart failure therapies to manage congestion effectively and improve patient comfort [57].

The association of left bundle branch block (LBBB) with increased mortality risk (AHR 19.70) is particularly concerning. LBBB is often seen in patients with advanced heart disease increases mortality risk by ten times and may necessitate specialized management approaches [58,59]. The risk of mortality is high compared to the previous findings, may the study population is hospitalized and in acute phase HF. Cardiac resynchronization therapy is a promising intervention for patients with LBBB, improving both functional status and survival [60]. This highlights the importance of identifying patients who may benefit from advanced interventions early in their hospitalization.

Pulmonary hypertension (AHR 21.88) was strongly associated with increased mortality. Elevated pulmonary pressures can exacerbate right heart failure and lead to death [61]. The current finding is in line the previous study findings. Clinicians should evaluate for pulmonary hypertension in heart failure patients and consider therapies such as pulmonary vasodilators for those with significant elevation in pulmonary pressures [62]. This approach can potentially improve outcomes and reduce mortality risk, as well as enhance patients' quality of life by alleviating symptoms associated with pulmonary hypertension [48,49,63].

Serum creatinine levels (AHR 2.76) serve as a vital indicator of renal function, which is frequently compromised in heart failure patients. The relationship between renal impairment and mortality emphasizes the need for vigilant monitoring of kidney function. Management strategies should focus on optimizing fluid status and avoiding acute kidney injury, which can be achieved through careful diuretic management and consideration of nephrology consultation when necessary. This dual focus on heart and kidney health is crucial, as renal dysfunction can lead to a vicious cycle of worsening heart failure.

The impact of severe hypertension (AHR 5.47) on mortality risk further emphasizes the need for comprehensive care that addresses comorbid conditions. The current study findings are in line with the previous study findings [64]. So effective management of blood pressure can alleviate the burden on the heart and improve overall prognosis. Strategies such as lifestyle modifications, medication adherence, and regular follow-up are essential components of hypertension management in heart failure patients [65]. The interplay between hypertension and heart failure underscores the importance of an integrated approach to patient care, where multiple comorbidities are managed concurrently [66].

Lastly, the presence of SCAP (AHR 8.14) indicates the complexity of heart failure management. Some studies have found that severe CAP can increase mortality risk in patients with heart failure by a hazard ratio ranging from 1.5 to 3.0 or even higher, depending on factors such as the severity of heart failure, comorbid conditions, and the specific characteristics of pneumonia [67]. It highlights the need for a systemic approach that considers potential complications associated with heart failure. Identifying patients with SCAP allows for tailored management strategies that address both cardiac and systemic health. This comprehensive approach can improve patient outcomes, as it ensures that all aspects of a patient's health are being taken into consideration.

The implications of these findings are profound. Clinicians must remain vigilant in monitoring these predictors to identify high-risk patients early in their hospitalization. Implementing standardized protocols for assessing these parameters can help streamline care and ensure that interventions are initiated promptly.

5. Conclusion

Acute heart failure significantly affects patients and healthcare systems in Ethiopia, exhibiting a high in-hospital mortality rate of 27.3 % and a 30-day unplanned readmission rate of 24.4 %. Key predictors of in-hospital mortality include low oxygen saturation, high pulse rate, dyspnea, left bundle branch block, pulmonary hypertension, high serum creatinine, severe hypertension, and preserved versus reduced ejection fraction. Early readmissions are linked to low serum sodium, thrombocytopenia, heart block, ischemic heart disease, high comorbidity burden, and poor adherence to both medications and guidelines. Conversely, evidence-based discharge medications, such as beta-blockers and ACEIs/ARBs, can lower the risk of readmission. The findings emphasize the need for early risk assessment, adherence to guideline-directed therapies, close monitoring of at-risk patients, and system performance improvement to mitigate mortality and readmissions in acute heart failure.

CRediT authorship contribution statement

Getachew Yitayew Tarekegn: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Legesse Chekleba: Writing – review & editing, Writing – original draft, Investigation, Funding acquisition. Tilaye Arega Moges: Writing – review & editing, Writing – original draft, Visualization, Project administration, Methodology. Fisseha Nigussie Dagnew: Writing – review & editing, Writing – original draft, Software, Resources, Conceptualization. Samuel Berihun Dagnew: Writing – review & editing, Writing – original draft, Visualization, Validation, Data curation. Sisay Stiotaw Anberbr: Writing – review & editing, Writing – original draft, Validation, Supervision, Software, Resources. Behailu Terefe Tesfaye: Writing – review & editing, Writing – original draft, Validation, Supervision, Software, Resources, Project administration, Methodology, Data curation, Conceptualization.

Ethical considerations

The study received ethical approval from the Institutional Review Board of Debre Tabor University (Reference No: DTU/674/24). Prior to data collection, official permission was obtained from the hospital directors and heads of the internal medicine departments of all participating institutions. Written informed consent was obtained from each participant after providing full information about the study's purpose, procedures, and potential risks and benefits. Participants were informed of their right to withdraw from the study at any stage without any consequences. Confidentiality and anonymity of all patient information were strictly maintained throughout the study. All procedures were conducted in accordance with the ethical principles outlined in the Declaration of Helsinki.

Funding

The authors declare that they did not receive any funding for this research work.

Declaration of competing interest

There were no financial or commercial relationships that could be interpreted as creating a conflict of interest in the research.

Appendix A. Supplementary data

Supplementary material

Data availability

The raw data will be available upon the reasonable request of the corresponding authors.