Predictors of Mortality in Burn Patients at Selected Tertiary Public Hospitals, Addis Ababa, Ethiopia: A Two-Year Retrospective Study

1. Introduction

Burn injury is a type of tissue damage caused by exposure to thermal, electrical, chemical, or radiation energy and remains a major cause of preventable injury worldwide [1]. Burns represent a significant global public health challenge, particularly in low- and middle-income countries (LMICs), where access to timely and specialized burn care is often limited [2]. According to the World Health Organization, an estimated 180 000 deaths occur annually due to burn injuries, with the vast majority occurring in LMICs [1]. Beyond mortality, burns are associated with prolonged hospitalization, permanent disability, psychological trauma, and substantial socioeconomic consequences, disproportionately affecting children and economically productive age groups [3,4,5].

In Africa, burn injuries account for a substantial proportion of trauma-related morbidity and mortality. Systematic reviews from sub-Saharan Africa demonstrate a persistently high burden of severe burns, with hospital-based case fatality rates averaging around 17%, substantially higher than in high-income countries [6,7]. This elevated mortality is largely attributable to delayed presentation, limited intensive care capacity, inadequate burn units, and shortages of trained personnel and essential supplies [6,7,8]. In Ethiopia, burn injuries are increasingly recognized as a major public health problem and a significant contributor to trauma-related mortality [7]. A systematic review and meta-analysis reported a pooled burn mortality prevalence of 6.99% (95% CI: 4.8–9.41) among hospitalized burn patients in Ethiopian hospitals, highlighting the substantial fatality burden associated with burn injuries in the country [9]. Contributing factors include reliance on open fire cooking, unsafe electrical infrastructure, crowded living conditions, delayed referral, and limited public awareness of burn prevention [10]. Established predictors of burn-related mortality include total body surface area (TBSA) burned, presence of inhalation injury, burn depth, patient age, comorbidities, and timing of definitive care [12,13].

Although advances in surgical techniques, resuscitation protocols, and infection control have improved outcomes in high-income settings, burn care in many African countries continues to face challenges, including inadequate staffing, limited ICU capacity, and insufficient access to wound care materials and reconstructive surgery [6,7]. Despite the documented burden of burn injuries in Ethiopia and pooled mortality estimates from systematic reviews, hospital-specific, contemporary data on burn-related outcomes remain limited [10,11]. National burn registries are absent [6], and burn care services are concentrated in a few urban centers, with AaBET and Yekatit 12 Hospitals serving as the main referral centers for moderate to severe burns [11]. Consequently, policymakers and clinicians face challenges in implementing evidence-based interventions tailored to local needs [10]. Evaluating the magnitude of in-hospital mortality and identifying clinical and demographic predictors at these major referral centers is critical for improving survival, optimizing resource allocation, and informing the development of targeted burn care strategies in resource-limited settings [14,15,16]. Therefore, this study aims to determine in-hospital mortality among burn patients admitted to AaBET and Yekatit 12 Hospitals and to identify the clinical and demographic factors associated with mortality.

2. Materials and Methods

A facility-based cross-sectional study was conducted at two public referral centers providing specialized burn care in Addis Ababa, Ethiopia: AaBET and Yekatit 12 Hospitals.

Yekatit 12 Hospital, operating under the Addis Ababa City Administration Health Bureau, is a long-established burn care center with a dedicated 19-bed burn unit (12 adult and 7 pediatric beds). AaBET Hospital, affiliated with St. Paul's Hospital Millennium Medical College, was established in 2016 to address gaps in emergency and surgical care. Its burn unit, under the Department of Plastic and Reconstructive Surgery, also has a 19-bed capacity (12 adult and 7 pediatric beds).

Although no formal national burn referral guidelines exist, patients are typically referred to these centers for moderate to severe burns, inhalation injuries, burns involving critical areas (face, joints), chemical or electrical burns, and burn-related complications such as infection or sepsis. Yekatit 12 Hospital primarily serves patients from Addis Ababa and nearby regions, while AaBET functions as a national referral center receiving severe burn cases from across the country.

The study included all patients of any age and sex with acute burn injuries admitted to the selected hospitals between September 1, 2021, and November 1, 2023.Exclusion criteria were: Missing or irretrievable medical charts, Incomplete documentation of key study variables, Patients documented as leaving against medical advice

At both institutions, burn patients are initially assessed and resuscitated in the Emergency Department by emergency physicians. Fluid resuscitation for major burns is guided by the Parkland formula, a well-established protocol. Patients requiring ongoing care are admitted to the burn unit following initial stabilization. Patients enrolled in Community-Based Health Insurance (CBHI) receive inpatient burn care free of charge, whereas uninsured patients cover treatment costs out-of-pocket.

The required sample size was initially calculated using a single population proportion formula, yielding 345 patients. However, to enhance representativeness and statistical power, all eligible patient charts within the study period were included. Consequently, a total of 832 patient records were analyzed.

Data were extracted from secondary sources, specifically the Emergency Health Management Information System (HMIS) logbooks and patient medical records.

The data extraction tool included: Sociodemographic variables: age, sex, residence (urban/rural); Clinical variables: time to presentation, cause of burn, total body surface area burned (TBSA%), burn depth, anatomical site, inhalation injury, comorbidities; Treatment and outcome variables: length of hospital stay, ICU admission, and in-hospital mortality.

Data were collected using a structured, paper-based checklist by two trained nurses. Training was provided on data collection procedures, ethical considerations, and tool utilization.

Data quality was ensured through daily supervision by the principal investigator, including checks for completeness, consistency, and accuracy. Data was entered into Epi Info version 7.2.6.0 and exported to SPSS version 27 for cleaning and analysis.

Descriptive statistics (frequency, percentage, median, and range) were used to summarize patient characteristics. Binary logistic regression analysis was performed to identify predictors of in-hospital mortality. Variables with a p-value < 0.2 in bivariate analysis were included in the multivariable model. Statistical significance was set at p < 0.05, and associations were reported using adjusted odds ratios (AORs) with 95% confidence intervals.

The dataset used and/or analyzed during the current study is available from the corresponding author upon reasonable request. There are no restrictions on data accessibility; however, patient confidentiality is strictly maintained through de-identification of all records.

Ethical approval was obtained from the Institutional Review Board of Addis Continental Institute of Public Health (protocol number: AC_IPHRO/0014/2024), the Research Directorate of St. Paul's Hospital Millennium Medical College, and the Addis Ababa Public Health Research and Emergency Management Directorate (reference number: A/A/9881/227).As the study utilized retrospective patient records, informed consent was waived. All data were anonymized to ensure confidentiality. No generative artificial intelligence tools were used in study design, data collection, analysis, or interpretation. AI-assisted tools were used only for language editing and formatting, which does not require formal disclosure under journal guidelines.

3. Results

Following the exclusion of 32 patients due to incomplete data, a total of 800 patients with new burn injuries were included in the analysis. These patients were admitted to two burn centers in Addis Ababa during the study period.

3.1. Sociodemographic Characteristics

Of the total study population, 459 patients (57.4%) were admitted to Yekatit 12 Hospital, while 341 (42.6%) were admitted to AaBET Hospital. The majority of patients (86.8%) were from urban areas. The median age was 18 years (range: 9 months–90 years), with children under 10 years representing the largest proportion (36.0%). Females accounted for 57.0% of the cohort, resulting in a male-to-female ratio of 1:1.3 (See Table 1).

3.2. Clinical Findings and Burn Characteristics

The median Total Body Surface Area (TBSA) burned was 15% (IQR: 8–22). Inhalation injury was identified in 112 patients (14.0%).Scald burns were the most common mechanism (49.1%), followed by flame burns (24.5%) and electrical burns (24.4%). Most patients (80.6%) presented within 8 hours of injury, and 57.5% initially received care at other health centers before referral. Pre-existing comorbidities were reported in 56 patients, with epilepsy and hypertension being the most common (each 3.0%). The most frequently affected anatomical sites among burn patients were the upper extremities (28.1%), followed by the head and neck combined (16.9%), the head alone (15.3%), and the trunk (12.6%). Other affected regions included the lower extremities (7.8%), the perineum (2.1%), and multiple body regions (13.0%), while isolated neck burns (4.3%) were the least common (See Table 2).

3.3. Patient Outcomes

A total of 45 patients (5.6%) required admission to the Intensive Care Unit (ICU). The overall in-hospital mortality rate was 8.5% (95% CI: 6.5–10.4), corresponding to 68 deaths. Of these deaths, 39 occurred at Yekatit 12 Hospital and 29 at AaBET Hospital. The majority of deaths occurred among female patients (n = 42). The leading causes of death documented in the medical records were septic shock (n = 30), followed by hospital-acquired or ventilator-associated pneumonia (n = 23), and multi-organ failure (n = 15).

3.4. Factors Associated with Burn-Related Mortality

3.4.1. Bivariable Analysis

In the bivariable logistic regression analysis, several variables met the inclusion criteria (p < 0.2) for inclusion in the multivariable model. These factors included burn depth, inhalation injury, mechanism of burn, total body surface area (TBSA) burned, place of hospital stay, head and neck involvement, and source of referral.

3.4.2. Multivariable Logistic Regression Analysis

In the multivariable model, five variables demonstrated a statistically significant association with burn-related mortality. Patients with deep burns (AOR = 1.97; 95% CI: 1.01–3.84), inhalation injury (AOR = 6.53; 95% CI: 2.14–19.95), TBSA ≥15% (AOR = 3.33; 95% CI: 1.29–8.61), and ICU admission (AOR = 14.42; 95% CI: 5.87–35.40) had significantly higher odds of death. Conversely, referral from another health facility (AOR = 0.18; 95% CI: 0.053–0.61) was associated with reduced odds of mortality (See Table 3).

4. Discussion

This study analyzed burn mortality and associated factors at two major burn referral hospitals in Addis Ababa—AaBET and Yekatit 12—examining 800 admitted patients over a two-year period (2021–2023). We found an in-hospital mortality rate of 8.5%, which is comparable to the Tanzanian pediatric burn study reporting 7.1% [17] and the South Gondar pediatric study in Ethiopia (8.5%) [16]. Our mortality rate is lower than the 19.6% and 30% reported in Nigerian studies [18,19] and the 17% average across sub-Saharan Africa [7]. Similarly, higher mortality rates have been observed in South Africa (23%) [20] and in high-risk units in India (47.3%) [21]. In the Islamic Republic of Iran, a study reported a 35.8% mortality rate among burn patients [22]. In contrast, mortality rates are substantially lower in high-income countries such as Singapore (4.6%) and Spain (4.3%) [23,24].

Overall, the comparatively lower mortality observed in our study is likely attributable to a younger, mixed-age population and lower average burn severity (mean TBSA ≈18%). In contrast, studies reporting substantially higher mortality involved older patient populations, including high-risk cohorts in India and Nigeria [21,25]; larger burn sizes (TBSA ≥30–40%) in Nigerian, Iranian, and Indian ICU-based studies [21,25,26]; and ICU-restricted cohorts with severe burns [21,26]. Additionally, a higher burden of comorbidities and associated conditions—including sepsis, inhalation injury, HIV infection, and chronic medical illnesses—has consistently been identified as a major contributor to mortality in these higher-mortality settings [27,28].

In this study, inhalation injury was identified as a significant predictor of mortality, with patients who had inhalation injury being 6.53 times more likely to experience in-hospital mortality than those without. This finding is consistent with previous reports, including Güldoğan et al., Queiroz et al., and Ahuja et al., who similarly identified inhalation injury as an important determinant of burn mortality [27,28,29]. This association is likely due to the increased risk of respiratory failure and sepsis in patients with inhalation injuries [13,28]. Also, we found those with TBSA of ≥15% were 3.33 times more likely to experience burn related hospital mortality than their counter partner. This aligns with previous studies by Ryan et al. and others [29,30]. Patients with TBSA ≥15% are at substantially higher risk of burn-related mortality because this threshold marks the transition from a localized injury to a systemic pathophysiological response [30,31,32]. Burns exceeding 15–20% TBSA trigger a systemic inflammatory response syndrome (SIRS) characterized by massive cytokine release, capillary leak, and endothelial dysfunction, leading to intravascular volume depletion, burn shock, and early organ hypoperfusion [33,34]. Increasing TBSA is also associated with profound immunosuppression, impairing both innate and adaptive immune responses and markedly increasing susceptibility to sepsis and multiorgan failure, which are the leading causes of late burn mortality [33,35,36]. In addition, extensive burns induce a hypermetabolic and catabolic state, with elevated energy expenditure, muscle protein breakdown, and insulin resistance, predisposing patients to delayed wound healing, infectious complications, and nutritional depletion [33,37,38]. Furthermore, as TBSA increases beyond 15%, there is a proportional rise in the need for aggressive fluid resuscitation, repeated surgical interventions, blood transfusions, and intensive monitoring, which may be particularly challenging in resource-limited settings, thereby compounding mortality risk [31,34]. Collectively, these mechanisms explain why TBSA thresholds as low as 15% have been consistently identified as independent predictors of mortality in both adult and pediatric burn populations, including the findings reported by Ryan et al. and other large cohort studies [13,29].

Additionally, patients with deep dermal and full-thickness burns in our study had a significantly higher risk of mortality compared with those with superficial burns (AOR = 1.97). This finding is consistent with reports from multiple settings. A study from Malawi found that full-thickness burns were independently associated with increased mortality, particularly when combined with larger TBSA and inhalation injury [14]. Similarly, Güldoğan et al. reported significantly higher mortality among patients with deep and full-thickness burns, especially in major burn cohorts with TBSA ≥30% [29]. Studies from Iran and India have likewise demonstrated that burn depth is a strong predictor of death, with full-thickness injuries associated with increased risk of sepsis, prolonged hospitalization, and need for intensive care [22,30]. The increased mortality associated with deep dermal and full-thickness burns is biologically plausible. These injuries result in complete destruction of the skin barrier, leading to massive fluid loss, impaired thermoregulation, and a high risk of bacterial colonization and invasive infection [34,36]. Full-thickness burns typically require early excision and skin grafting, and delays in definitive surgical management—particularly common in resource-limited settings—are strongly associated with increased rates of sepsis and multiorgan failure, which remain the leading causes of burn-related mortality [30,39]. Furthermore, deeper burns provoke a more severe systemic inflammatory and hypermetabolic response, characterized by increased energy expenditure, protein catabolism, insulin resistance, and prolonged immune dysfunction, all of which contribute to delayed wound healing, infectious complications, and higher mortality [33,35].

Furthermore, the current study found that patients admitted to the intensive care unit (ICU) were 14.42 times more likely to experience burn-related mortality compared with those managed in the burn ward. This finding is consistent with previous studies that reported disproportionately high mortality among burn patients requiring ICU care [14,28]. An Italian ICU-based study of severe burn patients similarly reported markedly elevated mortality rates, reflecting the concentration of patients with extensive TBSA involvement, inhalation injury, and multiple organ dysfunction within critical care settings [13]. The increased ICU-associated mortality does not reflect harm from ICU admission itself but rather the severity of illness and late referral among patients requiring intensive care. In low- and middle-income settings, ICU admission is frequently preceded by delayed presentation, inadequate early resuscitation, and a high burden of comorbid conditions, which significantly worsen outcomes [14,28]. In addition, limited critical care resources, including shortages of trained personnel, mechanical ventilation capacity, infection control measures, and advanced organ-support modalities, further contribute to the high mortality observed among critically ill burn patients in ICU settings [11,15].

Finally, the current study found that patients referred from health centers and tertiary hospitals were 71% and 82% less likely, respectively, to experience burn-related institutional mortality compared with self-referred patients. This protective effect of formal referral may be explained by earlier recognition of burn severity, initial stabilization prior to transfer, and timely linkage to specialized burn care. Patients referred from health facilities are more likely to have received early fluid resuscitation, analgesia, wound coverage, tetanus prophylaxis, and basic infection prevention measures, which are critical in preventing early burn shock and subsequent complications [34]. In contrast, self-referred patients often present later in the course of injury, frequently without prior resuscitation, and may arrive with advanced hypovolemia, infection, or evolving organ dysfunction [25]. Even basic first aid, fluid management, and airway assessment at a primary level may mitigate early fatal complications. Conversely, self-referred patients, who served as the reference group, had the highest odds of mortality. This likely reflects several disadvantages: delayed presentation, lack of pre-referral resuscitation, and potentially more severe burns in patients who bypass lower-level care because they (or bystanders) perceive the injury as critical.

5. Conclusions

The study found that nearly one in ten patients died after presenting to the two burn centers in Addis Ababa. Factors such as ICU admission, inhalation injury, TBSA ≥15%, and deep dermal/full-thickness burns found to a strong predictive factor for burn related hospital mortality. Whereas, patient referred from health facility was found to be protective factor than that of self-referred burn patients. Therefore, based on the findings of this study, the following recommendations are proposed to guide clinical practice, and public health strategies aimed at reducing burn-related mortality in Ethiopia: Strengthen Critical Care Capacity for Severe Burn Injuries. The markedly increased mortality risk among patients requiring intensive care underscores a major capacity gap. A comprehensive assessment of ICU readiness—including staffing, equipment availability, and common causes of in-ICU mortality is urgently needed. Implement Standardized Early Triage and Risk Stratification Tools specifically for those variables such as TBSA ≥15%, inhalation injury, and deep burn depth—can be translated into a practical triage and risk stratification tool for emergency units. Integrating such a tool into routine assessment can facilitate early identification of high-risk patients, ensure timely resuscitation, expedite specialist consultation, and prioritize allocation of critical care resources. This structured approach would reduce preventable early deterioration. Creating community awareness about the importance of seeking care at nearby health centers and hospitals is crucial, as timely first response, initial intervention, and early resuscitation can prevent delays associated with self-referral and activate referral pathways early, thereby reducing burn-related hospital mortality. Given the high burden of preventable burn injuries, especially among children and households relying on open-flame cooking, prevention efforts remain a cornerstone of mortality reduction. Public health campaigns should address the most common mechanisms identified in this study, focusing on scald prevention in children and flame burn prevention in domestic settings.