Severe malaria-bacterial sepsis co-infection among critically ill children admitted to pediatric emergency departments in Kinshasa

Abstract

BACKGROUND

Severe malaria remains a major cause of pediatric critical illness and early mortality in sub-Saharan Africa. In pediatric emergency settings, concomitant bacterial sepsis is often under-recognized because of overlapping clinical manifestations, which may delay appropriate antimicrobial therapy and worsen outcomes. Data on severe malaria-bacterial sepsis co-infection among critically ill children in Kinshasa remain limited.

AIM

To determine the frequency of severe malaria-bacterial sepsis co-infection and identify the associated clinical and biological factors among critically ill children.

METHODS

This prospective multicenter observational study was conducted from September 2014 to September 2015 in four pediatric emergency departments in Kinshasa, Democratic Republic of Congo. Children aged 6 months to 15 years admitted with life-threatening conditions and confirmed severe malaria were included. Bacterial sepsis was defined by a positive blood culture for a pathogenic organism. Clinical and laboratory variables were compared using Fisher’s exact test and the Mann-Whitney U test. Multivariable logistic regression was performed to identify factors independently associated with co-infection.

RESULTS

Among 138 children admitted with life-threatening conditions, 96 (69.6%) had confirmed severe malaria. Severe malaria-bacterial sepsis co-infection was identified in 11 children, representing 11.5% of severe malaria cases. In univariate analysis, respiratory distress, splenomegaly, and hypoglycemia were more frequent among co-infected children. In multivariable analysis, respiratory distress remained independently associated with bacterial co-infection. Early mortality in the emergency department was higher among co-infected children than among those without co-infection. Gram-negative bacteria, predominantly Enterobacteriaceae, accounted for most bloodstream isolates.

CONCLUSION

Severe malaria-bacterial sepsis co-infection is relatively frequent among critically ill children in Kinshasa and is associated with poorer early outcomes. Early and systematic assessment for sepsis at admission may help improve the initial management of children in pediatric emergency settings.

Key Words: Severe malaria; Bacterial sepsis; Co-infection; Critically ill children; Pediatric emergency care; Early mortality; Sub-Saharan Africa

Core Tip: Among children admitted with severe malaria, concomitant bacterial sepsis is not uncommon and may be missed at presentation, contributing to early mortality. In pediatric emergency departments in Kinshasa, simple clinical signs-particularly respiratory distress-may help identify children at higher risk of co-infection, supporting early sepsis assessment and prompt integrated management in resource-limited settings.

INTRODUCTION

Malaria remains a major public health problem in sub-Saharan Africa, particularly in the Democratic Republic of Congo, where it continues to cause substantial morbidity and mortality among children[1,2]. According to the World Health Organization (WHO), an estimated 282 million malaria cases and 610000 malaria-related deaths occurred worldwide in 2024, with approximately 95% of cases and deaths occurring in the WHO African Region[3]. Children under 5 years of age account for nearly 75% of malaria-related deaths in this region Severe malaria is a medical emergency, and outcomes depend largely on the timeliness and quality of early care. In resource-limited settings, this burden may be further increased by concomitant bacterial infections, which are frequently unrecognized at admission[3].

Severe malaria-bacterial sepsis co-infection represents a major diagnostic and therapeutic challenge because both conditions share overlapping clinical manifestations. Co-infection has been associated with more severe clinical presentations and increased mortality, particularly among African children[4-11]. In addition, empirical antibiotic use in children with severe malaria, often initiated without microbiological confirmation, remains controversial and may contribute to inappropriate treatment and antimicrobial resistance[5,10]. Accurate diagnosis is further hindered by limited access to microbiological investigations, delayed availability of results, financial barriers, and limited laboratory capacity, all of which may lead to under-recognition at admission and suboptimal early treatment decisions.

Previous studies have reported bacteremia in 5% to 23% of children with severe malaria, involving a wide range of bacterial pathogens[5,6,11]. However, data on severe malaria-bacterial sepsis co-infection in children in Kinshasa remain scarce. The Democratic Republic of the Congo has experienced recurrent malaria outbreaks, including episodes of malaria co-infection with typhoid fever. In the Panzi health zone of Kwango Province, southeast of Kinshasa, successive outbreaks of typhoid fever in 2023 and malaria in 2024 were reported, with case fatality rates of 6% to 7%[12,13]. At both the national level and in Kinshasa, data on severe malaria-bacterial sepsis co-infection in children remain scarce and fragmented. At the same time, the Integrated Management of Childhood Illness strategy has been implemented for children younger than 5 years, and national recommendations on sepsis management and rational antibiotic use have been adopted; however, their practical impact in routine care remains limited.

The objective of this study was therefore to determine the frequency of severe malaria-bacterial sepsis co-infection and to identify the associated clinical and biological factors among critically ill children admitted to pediatric emergency departments in Kinshasa.

MATERIALS AND METHODS

Study design and setting

This was a multicenter, prospective, hospital-based observational study conducted from September 2014 to September 2015 in four referral hospitals in Kinshasa, Democratic Republic of Congo. The study was conducted in the pediatric emergency departments of these hospitals and included children admitted with life-threatening conditions.

Study population and inclusion criteria

Children aged 6 months to 15 years admitted to the pediatric emergency departments with at least one sign of a life-threatening condition were eligible for inclusion. Severe malaria was defined according to WHO criteria[2]. Among children with microscopy-confirmed severe malaria, bacterial sepsis co-infection was assessed.

Operational definitions

Severe malaria was defined according to WHO criteria as microscopy-confirmed Plasmodium falciparum infection (positive thick blood smear) associated with at least one clinical or biological criterion of severity[2].

Bacterial sepsis was defined as an acute infectious syndrome with signs of severity, confirmed by a positive blood culture for a pathogenic organism. Blood culture contaminants were defined as commensal organisms from the skin or mucosal flora and were excluded from the analysis.

Severe malaria-bacterial sepsis co-infection was defined as the concomitant presence of confirmed severe malaria and a positive blood culture for a pathogenic organism.

Life-threatening conditions were defined in accordance with the WHO Emergency Triage Assessment and Treatment recommendations[14]. Hypoglycemia was defined as a blood glucose level below 45 mg/dL.

Therapeutic management

All children with confirmed severe malaria received parenteral antimalarial treatment, mainly artesunate, together with standard emergency supportive care. Antibiotics were not administered systematically at admission; empirical antibiotic therapy was initiated only in children with marked clinical deterioration, whereas in other cases, antibiotics were deferred pending blood culture results when clinically feasible.

Data collection and paraclinical investigations

Sociodemographic and clinical data were prospectively collected at admission using standardized case report forms. Laboratory investigations were performed in specialized facilities according to standard procedures. Blood cultures and antimicrobial susceptibility testing were conducted at the National Institute of Biomedical Research. Malaria diagnosis was established by microscopy using thick and thin blood smears at the Department of Parasitology. Biochemical and hematological analyses, including blood glucose, alkaline reserve, C-reactive protein, and complete blood count, were performed at the laboratory of the Faculty of Pharmaceutical Sciences. Microorganisms considered to be blood culture contaminants were classified as negative results, in accordance with international recommendations.

Study variables

The variables analyzed included sociodemographic, clinical, and laboratory parameters. Sociodemographic variables included age and sex. Clinical variables comprised prostration, convulsions, respiratory distress, coma, dark urine (hemoglobinuria), bleeding, jaundice, shock, mucocutaneous pallor, and oligo-anuria, which was defined as urine output < 0.5 mL/kg/hour in a febrile child. Laboratory variables included hematological, biochemical, and microbiological parameters.

Sample size

The sample size was estimated based on the prevalence of bacterial co-infection reported in the literature among children with severe malaria in endemic settings[8]. A minimum sample size of 128 patients was considered sufficient to identify approximately 10 cases of severe malaria-bacterial sepsis co-infection, thereby allowing an exploratory multivariable analysis limited to one predictor variable, in accordance with commonly accepted recommendations aimed at reducing the risk of overfitting in regression models.

Statistical analysis

Data were analyzed using appropriate statistical software. Categorical variables were summarized as frequencies and percentages, whereas continuous variables were presented as means ± SD or medians with ranges, as appropriate according to their distribution. Comparisons between children with and without severe malaria-bacterial sepsis co-infection were performed using Fisher’s exact test for categorical variables and the Mann-Whitney U test for continuous variables. The strength of associations between clinical or biological variables and bacterial co-infection was estimated using odds ratios (ORs) with 95% confidence intervals (95%CIs). An exploratory multivariable logistic regression model was then constructed to identify factors independently associated with bacterial co-infection. Variables with a P value < 0.20 in univariable analysis, together with variables considered clinically relevant, were entered into the multivariable model. A two-sided P value < 0.05 was considered statistically significant.

RESULTS

Study population and prevalence of severe malaria-bacterial sepsis co-infection

During the study period, a total of 1230 pediatric admissions were recorded across the four emergency departments. Among these, 138 children were admitted with life-threatening conditions. Of these 138 children, 96 (69.6%) had microscopy-confirmed severe malaria. Pathogenic organisms were isolated from blood cultures in 20 patients (14.5%), including 11 children with severe malaria-bacterial sepsis co-infection. Therefore, severe malaria–bacterial sepsis co-infection accounted for 11.5% (11/96) of children with confirmed severe malaria and 0.89% of all pediatric admissions (11/1230; 95%CI: 0.50-1.59).

Baseline characteristics of the study population

The baseline characteristics of the study population are summarized in Table 1. The mean age was 4.5 ± 3.4 years, and 57.9% of the children were male. Children younger than 5 years accounted for 60.8% of admissions. Among the included children, 96 (69.6%) had confirmed severe malaria, whereas blood cultures were negative or sterile in 85.5% of cases.

Table 1 Baseline characteristics of children admitted with life-threatening conditions (n = 138).

Variables	n (%)
Age (years)
< 5	84 (60.8)
≥ 5	54 (39.2)
Sex
Female	58 (42.1)
Male	80 (57.9)
Microscopy-confirmed malaria
Positive	96 (69.6)
Blood culture
Positive for pathogenic organisms	20 (14.5)
Negative (sterile or contaminants)	118 (85.5)

Clinical and biological factors associated with co-infection

The clinical and biological factors associated with severe malaria-bacterial sepsis co-infection are presented in Table 2. In univariable analysis, respiratory distress (OR = 10.6; 95%CI: 2.8-39.9) and splenomegaly (OR = 9.9; 95%CI: 2.7-37.3) were significantly associated with co-infection. Hypoglycemia was also significantly more frequent among co-infected children (P = 0.038). No significant between-group differences were observed for the other clinical or laboratory variables. In exploratory multivariable logistic regression, respiratory distress remained the only factor independently associated with severe malaria-bacterial sepsis co-infection (adjusted OR = 6.5; 95%CI: 1.6-26.0).

Table 2 Factors associated with severe malaria-bacterial sepsis co-infection in children (n = 138), n (%)/mean ± SD.

Variables	Co-infection (n = 11)	No co-infection (n = 127)	P value
Age (years)
< 5	7 (63.6)	77 (60.6)	1.000
≥ 5	4 (36.4)	50 (39.4)	-
Sex
Female	7 (63.6)	51 (40.2)	0.199
Male	4 (36.4)	76 (59.8)	-
Clinical signs
Pallor	4 (36.4)	41 (32.3)	0.553
Respiratory distress	7 (63.6)	18 (14.2)	0.0012
Splenomegaly	7 (63.6)	19 (15.0)	0.0012
Hepatomegaly	2 (18.2)	23 (18.1)	0.243
Convulsions	1 (9.1)	9 (7.1)	1.000
Jaundice	2 (18.2)	7 (5.5)	0.203
Dark urine (hemoglobinuria)	0 (0.0)	6 (4.7)	1.000
Coma	0 (0.0)	3 (2.4)	1.000
Shock	0 (0.0)	3 (2.4)	1.000
Biological parameters
White blood cells (cells/mm3), median (min-max)	13700 (5500-55000)	9200 (3300-85000)	0.245
Platelets (× 103/μL), median (min-max)	130 (118-174)	144 (50-454)	0.160
C-reactive protein (mg/L), median (min-max)	48 (0-48)	48 (0-96)	0.691
Hematocrit (%)	25.0 ± 7.5	25.2 ± 7.3	0.554
Blood glucose (mg/dL)	36.4 ± 12.1	67.8 ± 41.0	0.0381
Alkaline reserve (mmol/L)	17.4 ± 7.8	18.6 ± 5.6	0.637
Clinical outcome
Deaths in the emergency department	2 (18.2)	3 (2.4)	0.060

¹Mann-Whitney U test.

²Fisher’s exact test.

Microbiological findings

Figure 1A presents the distribution of pathogenic organisms isolated from blood cultures among children with severe malaria-bacterial sepsis co-infection. Gram-negative organisms predominate, particularly Enterobacteriaceae, including Enterobacter spp. and Salmonella Typhi. The antimicrobial susceptibility patterns of the isolated pathogens are shown in Figure 1B.

Figure 1 Microbiological findings. A: Distribution of pathogenic organisms isolated from blood cultures among children with severe malaria-bacterial sepsis co-infection (n = 11); B: Antimicrobial resistance profiles of pathogenic organisms isolated from blood cultures among children with severe malaria-bacterial sepsis co-infection.

Early mortality in the emergency department

Overall, five deaths (3.6%) occurred in the emergency departments during the study period. Early mortality was higher among children with severe malaria-bacterial sepsis co-infection than among those without co-infection [2/11 (18.2%) vs 3/127 (2.4%)], although the difference did not reach statistical significance.

DISCUSSION

In this multicenter prospective study conducted in pediatric emergency departments in Kinshasa, severe malaria-bacterial sepsis co-infection was identified in 11.5% of children with microscopy-confirmed severe malaria. In multivariable analysis, respiratory distress was the only factor independently associated with co-infection. No detailed comparative analysis was performed across the four referral centers. Although data collection was standardized, the limited number of patients enrolled at each center, particularly the small number of co-infected cases, precluded robust statistical comparisons between centers. Nevertheless, the overall clinical characteristics and prevalence of co-infection appeared broadly similar across sites, suggesting a relative homogeneity in the clinical profiles of critically ill children admitted to these hospitals. Taken together, these findings underscore the diagnostic difficulty of distinguishing severe malaria from bacterial sepsis at admission in resource-limited emergency settings.

Prevalence of severe malaria-bacterial sepsis co-infection

The prevalence of severe malaria-bacterial sepsis co-infection observed in this study is consistent with that reported in several African series[7-11]. Hospital-based studies conducted in sub-Saharan Africa have shown considerable variability in the prevalence of bacterial co-infection among children with severe malaria[15]. A meta-analysis by Church and Maitland reported a pooled prevalence of approximately 6.4%[16], whereas individual studies from sub-Saharan Africa have reported rates ranging from 4.6% to 7.8%. In contrast, higher prevalences, ranging from 20% to 25%, have been described in some settings, particularly in Nigeria[7-10,15].

These variations may be explained by differences in the diagnostic criteria used for severe malaria and bacterial sepsis, variations in microbiological capacity, and local epidemiological patterns. In the present study, inclusion was restricted to children admitted with life-threatening conditions, which may partly explain the higher prevalence observed compared with that reported in broader or less severely ill pediatric populations. Moreover, the reclassification of contaminated blood cultures as negative results, in accordance with international recommendations, likely reduced the risk of overestimating bacterial co-infection and improved comparability with methodologically rigorous studies[8-15], including recent meta-analytic evidence[16]. Length of hospital stay could not be reliably assessed because this variable was not collected in a standardized manner across all participating centers.

Clinical and biological profiles of co-infected vs non-co-infected children

Respiratory distress was the main clinical factor associated with severe malaria–bacterial sepsis co-infection, in keeping with findings from several African and international studies that have identified respiratory involvement as a marker of severity in children with malaria complicated by bacterial infection[4,7,8]. Studies conducted in Nigeria and East Africa have similarly reported associations between respiratory distress and increased mortality in this population. Splenomegaly, which was associated with co-infection in univariable analysis, has been described as an indirect marker of high parasite burden; however, its value in distinguishing bacterial co-infection remains inconsistent across settings[4-6]. Hypoglycemia, which was more frequent among co-infected children in the present study, has also been reported as a marker of severity in severe malaria, particularly when associated with concomitant bacterial infection[17,18]. Overall, co-infected children appeared to present with greater clinical severity at admission, as reflected by the coexistence of multiple severe manifestations; however, the limited sample size did not allow a more detailed statistical assessment of the combined burden of clinical features.

Microbiological profile and antimicrobial susceptibility

Enterobacteriaceae, particularly Enterobacter spp. and Salmonella Typhi, were the most frequently isolated organisms among co-infected children, consistent with findings from several African studies of severe malaria complicated by bacteremia[7-11]. The blood culture contamination rates observed in the present study were comparable to those reported in other African hospital settings, supporting the decision to classify these isolates as negative results. In addition, the high levels of resistance to commonly used antibiotics highlight the limitations of empirical antibiotic therapy and underscore the need for regular microbiological surveillance to inform antibiotic stewardship in pediatric emergency settings[19,20].

Early mortality in the emergency department

Early mortality in the emergency department was higher among children with severe malaria-bacterial sepsis co-infection than among those without co-infection. Although this difference did not reach statistical significance, probably because of the limited number of deaths, the finding is consistent with previous studies reporting higher mortality among children with severe malaria complicated by bacteremia, with reported death rates ranging from 10% to 30% in several African settings, including Kenya, Malawi, and Nigeria[6-8,15]. In addition, several studies have shown that most deaths occur early during hospitalization, often within the first 7 days, underscoring the critical importance of prompt assessment and early management[6,8,15].

These findings underscore the importance of a thorough clinical assessment at admission, particularly careful evaluation of respiratory distress, to facilitate the early identification of children at high risk of bacterial co-infection. In resource-limited emergency settings, the relatively high frequency of co-infection, together with substantial antimicrobial resistance, challenges the routine use of systematic empirical antibiotic therapy and supports a more targeted, clinically guided approach.

Strengths and limitations

This study has several strengths. It was conducted prospectively across multiple referral hospitals and focused on children admitted with life-threatening conditions, a high-risk population that remains underrepresented in the literature. The use of standardized WHO definitions and microbiological analyses performed in specialized laboratories strengthens the validity of the clinical and biological findings.

Although the data were collected in 2014, the findings remain relevant to current clinical practice in a resource-limited setting such as Kinshasa, where the organization of pediatric emergency care and access to microbiological investigations have changed little over time. The relatively small number of co-infection cases limited the assessment of its impact on mortality and restricted the number of variables that could be included in the multivariable analysis, warranting cautious interpretation. In addition, the hospital-based design and the lack of advanced microbiological investigations may have led to an underestimation of the true burden of bacterial co-infection and may limit the generalizability of the findings.

CONCLUSION

Severe malaria-bacterial sepsis co-infection is relatively common among children admitted with life-threatening conditions in Kinshasa. Respiratory distress appears to be a key clinical marker for the early identification of children at high risk of co-infection. In resource-limited settings, these findings underscore the importance of careful clinical assessment at admission, more rational use of empirical antibiotic therapy, and strengthened microbiological surveillance to improve management and reduce preventable mortality.