https://orcid.org/0000-0002-8463-8794
Figures
Abstract
Shock is considered one of the most important mechanisms of critical illness in children. However, data on paediatric shock in sub-Saharan Africa is limited, which constrains development of effective treatment strategies. We aimed to describe the prevalence, mortality, and aetiology of paediatric shock in a tertiary hospital in Malawi. Children aged two months to 16 years presenting with shock (FEAST criteria; respiratory distress and/or impaired consciousness, and at least one sign of impaired circulation; capillary refill>3 seconds, cold extremities, weak pulse, or severe tachycardia) to the emergency department were included and followed-up prospectively using routinely collected data between February 2019 and January 2020. Prevalence, mortality and aetiology of shock were reported for both the FEAST criteria and World Health Organization (WHO) definition. The association between aetiology and mortality was assessed with univariable analysis. Of all screened admissions (N = 12,840), 679 (5.3%) children presented with shock using FEAST criteria and the mortality was 79/663 (11.9%). WHO-defined shock applied to 16/12,840 (0.1%) and the mortality was 9/15 (60.0%). Main diagnoses were viral/reactive airway diseases (40.4%), severe pneumonia (14.3%), gastroenteritis (11.3%) and presumed sepsis (5.7%). Children diagnosed with presumed sepsis and gastroenteritis had the highest odds of dying (OR 11.3; 95%-CI:4.9–25.8 and OR 4.4; 95%-CI:2.4–8.2). Considering the high mortality, prevalence of paediatric shock (FEAST and WHO definitions) in Malawi is high. Sepsis and gastroenteritis are diagnoses associated with poor outcome in these children. Consensus on a clinical meaningful definition for paediatric shock is essential to boost future studies.
Citation: Kumwenda M, Assies R, Chathima G, Khofi H, van Woensel JBM, Chimalizeni Y, et al. (2024) Prevalence, mortality, and aetiology of paediatric shock in a tertiary hospital in Malawi: A cohort study. PLOS Glob Public Health 4(1): e0002282. https://doi.org/10.1371/journal.pgph.0002282
Editor: Devan Jaganath, University of California San Francisco, UNITED STATES
Received: March 29, 2023; Accepted: November 18, 2023; Published: January 8, 2024
Copyright: © 2024 Kumwenda et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The ethical approval granted by the Research Ethics Committee (COMREC) did not include permission to share (pseudonymised/anonymised) data for research purposes other than covered in our approved ethical protocol. Depending upon the research questions of new investigators we will need to ask for additional approval from the Research Ethics Committee in Blantyre (COMREC). Therefore, we suggest that the main applicant of this study for the Research Ethics Committee (COMREC), Dr. Calis, assesses if additional use of the data is covered under the ethical protocol. Alternatively, COMREC could contacted directly for approval to access the data (job.calis@gmail.com or comrec@medcol.mw).
Funding: This work was supported by Kamuzu University of Health Sciences, Department of Paediatrics and Child Health (MK). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Globally, 5 million deaths of under-five year old children occurred during 2020, of which more than half occurred in sub-Saharan Africa [1]. In-hospital mortality in these countries remains high and most of these deaths occur within the first 24 hours of admission [2]. Improving emergency care is essential to increase survival of these critically ill children.
The World Health Organization (WHO) Emergency Triage Assessment and Treatment (ETAT) provides guidelines for emergency care of critically ill children in low-resource settings, focussing on respiratory distress, circulatory failure and seizures [3]. Treatment of circulatory failure, or shock, in children in sub-Saharan Africa, however, has been an issue of debate since the largest trial on fluid bolus treatment in paediatric shock was published in 2011 (the Fluid Expansion As Supportive Therapy or FEAST trial). This trial showed that administration of fluid boluses could have detrimental effects in children with febrile illness and impaired perfusion [4]. The WHO has subsequently adapted the ETAT guidelines to include these results, distinguishing between different definitions for shock and treatment algorithms [5].
The publication of the results of the FEAST trial initiated a still unresolved debate regarding the optimal treatment for shock in children in sub-Saharan Africa and other LMIC. Since then, at the time of writing, only one study by Mbevi et al. has been published focusing on shock in African children [6]. Furthermore, the lack of a universal definition for shock hampers the interpretation of study results. The FEAST trial applied criteria based on international guidelines that are different from the more stringent definition for paediatric shock in the WHO ETAT guideline [4, 5]. These definitions mostly consist of combinations of clinical signs, including capillary refill time, cold peripheries, heart rate and blood pressure, which also has its limitations. For example, increased heart may also be seen in several other conditions in which circulation is not necessary compromised such as fever, severe respiratory distress, or pain. Few studies have reported these different definitions and subsequent differences in interpretation of study results in a low-resource setting other than the ones included in the FEAST trial [6, 7].
Without improved understanding of the impact of the FEAST trial and other shock definitions on the prevalence, outcome and aetiology of shock it is difficult to design adequate treatment protocols for paediatric shock in general and specifically for our context. Therefore, we performed this cohort study to assess the prevalence, mortality and potential aetiology of shock in children in Malawi, applying both the FEAST and WHO definition. First, we aimed to determine the prevalence of shock in children upon admission to hospital and their clinical outcome. Second, we aimed to describe clinical characteristics and aetiology of children with shock, and the association between aetiology and mortality.
Methods
Study setting
This study was conducted at Queen Elizabeth Central Hospital (QECH), a large tertiary hospital in Blantyre, Malawi, that has a catchment area of over six million people, of whom more than half are children under 15 years old [8, 9]. All children are admitted through the paediatric accident and emergency department to one of the five paediatric wards. Local guidelines for shock follow recommendations in WHO ETAT and clinicians at QECH are trained in WHO ETAT guidelines [3]. Reporting was done using the STROBE recommendations [10].
Study design and participants
Children aged two months to 16 years with shock upon admission were included and followed up prospectively until discharge or death using routinely collected data over a 12-month period (1 February 2019–31 January 2020). Shock was defined using a modified FEAST definition, to be able to compare our results with the FEAST trial, the landmark paper on shock in African children (S1 Table). The definition is modified by applying it to all children presenting to the emergency department, not only to children presenting with febrile illness. All paper hospital admission files were screened daily. For eligible children, informed consent was subsequently obtained from parents or guardians. Therefore, two study populations were considered, an eligible population consisting of children that met the inclusion criteria and an included population consisting of children whose guardian or parents consented to use further clinical details for this study. For children who died before parents or guardians could give informed consent, we received ethical approval to include them for further analysis in this study. By including these children, we attempted to avoid selection bias in further assessment of the aetiology of shock. Children were prospectively followed up by retrieving their paper admission files daily during weekdays or the next working day if they were admitted on weekends and public holidays. Data were manually entered from the admission file into an electronic case record form in Open Data Kit (2020 Get ODK Inc.) after discharge or death. Fig 1 shows a flow diagram of screened, eligible and included children.
The ‘eligible population’ consists of all children that meet the inclusion criteria. The ‘included population’ are children for whom consent was given for analysis of clinical characteristics and follow-up of these children. The ‘aetiology population’ consists of all children with known outcome and discharge/death diagnosis.
Data collection
The primary outcome was in-hospital mortality. Prevalence and mortality were derived from the eligible population (all children who met the inclusion criteria), for which the total screened children was used as the denominator to represent all admissions through the emergency department. None of the eligible children were discharged directly home or transferred to another hospital. A subset of the eligible population also classified as WHO shock, or as “shocked” by the clinician (S1 Table) [5]. If one of these variables was missing, it was considered normal. The paediatric department uses a predefined and concise paper hospital admission form that is completed by the attending physician in the emergency department, which we used for the case record form (S1 Fig). Vital signs are therefore the first upon presentation before initial treatment, recorded by the clinician. All additional tests were independently done as indicated by the clinician or routinely as part of the services offered at QECH (S2 Table). Aetiological diagnosis was the clinical diagnosis recorded by the attending clinician upon discharge. To assess the aetiology of shock, we grouped the main discharge diagnosis of the included children into diagnostic categories, which were considered as exposure variables (S3 Table).
Missing data
Patient outcome was extracted from hospital logbooks and additional information on the discharge or death diagnosis was retrieved from the paper admissions files. Lost to follow-up of outcome and/or main discharge diagnosis was reported. Main reasons for missing data were not recorded clinical data by attending physicians or missing paper files.
Statistical analysis
Data were analysed using IBM SPSS version 23. The prevalence of shock was calculated by dividing the total number of eligible children with shock on admission by the total number of screened paediatric admissions. Mortality was calculated by dividing the total number of deaths by the total eligible children with known outcome. Diagnostic categories were compared between the two definitions for shock using Chi square or Fisher’s exact test for categorical data. A p-value of <0.05 was considered statistically significant. In an explorative analysis we assessed associations between diagnosis and mortality using univariable analysis and presenting odds ratios and 95% confidence intervals.
Ethics statement
For this observational study, we obtained required approval by the College of Medicine Research and Ethics Committee (COMREC –P.12/18/2557). Parents or guardians of eligible children provided written informed consent. The ethical committee provided a waiver to include the data of children who died prior to consent. During data collection, MK, RA and GC had access to individual patient information. The dataset includes anonymized data. Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the S1 Checklist.
Results
Of the 13,995 admitted children during the 12 months, 12,840 (91.7%) hospital files could be retrieved for screening. Of these 12,840 children, 679 fulfilled the eligibility criteria for shock based on the FEAST criteria (eligible population) from which prevalence and mortality could be derived. Of these children, 505 could be included for descriptive analysis of clinical characteristics and 470 children for further analysis of aetiology and mortality (Fig 1).
Prevalence
Prevalence of shock on admission amongst all screened children was 5.3% (679/12,840, Fig 1). The WHO shock definition applied to 0.1% (16/12,840) of all screened children. Clinicians reported shock as an admission diagnosis in 0.3% (36/12,840) of all screened children.
Outcome (mortality)
Outcome was known for 663 of 679 (97.6%) children with shock. Of these 663 children, 79 (11.9%) died. For WHO shock, outcome was known for 15 of 16 children. Of these 15 children, nine (60%) children died.
Clinical characteristics at baseline
Of the 505 children included for further analysis, median age was 17.0 (9.0–36.0) months and 56.6% were male. Fever was reported as a presenting symptom in 77.4% of the children; other main presenting symptoms were respiratory distress (81.4%) and cough (75.1%). On examination, the most frequent clinical signs were severe tachycardia (89.6%) and respiratory distress (82.5%). Fever (>39°C) was measured in 18.0%, pallor in 19.9%, and signs of dehydration in 14.9%. In children with WHO shock, the main presenting symptoms were fever (73.3%) and vomiting (73.3%). The main clinical signs in these children were prostration (93.3%), coma (86.7%), pallor (53.3%), and signs of dehydration (53.3%). Blood cultures were taken on clinical indication in 176/505 (34.9%) children and 8/176 (4.5%) were positive for a micro-organism. Death was more common early during admission, with 27.6% occurring within 24 hours and 55.3% occurring within 48 hours. Table 1 describes clinical characteristics, laboratory results, and management.
Aetiology
Of the 505 included children, 470 (93.1%) children had outcome and diagnosis recorded. Two or more diagnoses were recorded in 190/470 (40.4%). The predominant diagnostic category was viral/reactive respiratory diseases (40.4%). Severe pneumonia was recorded as the main diagnostic category in 14.3%, gastroenteritis in 11.3%, malaria in 12.6%, and sepsis in 5.7% of children with shock according to the FEAST criteria. In children with WHO shock, the main diagnostic categories were gastroenteritis (40.0%), sepsis (26.7%) and malaria (26.7%) (Table 2). In the children that died (n = 75), gastroenteritis (28.0%) and sepsis (22.7%) were the most common diagnoses (Fig 2, S4 Table). Children with HIV had slight increased odds of dying, although this was not statistically significant, while children with poor nutrition had increased odds of dying (OR 1.26; 95%-CI: 0.97–1.63 and OR 1.85 95%-CI: 1.29–2.66, respectively). Children diagnosed with sepsis and gastroenteritis had the highest odds of dying (OR 11.3; 95%-CI: 4.9–25.8 and OR 4.4; 95%-CI: 2.4–8.2, respectively) (Table 2).
(N = 75).
Diagnostic categories where based on the main discharge diagnosis by the clinician.
Discussion
Prevalence of paediatric shock in this cohort study was 5.3% applying the FEAST criteria in children on admission to a tertiary hospital in Malawi. Mortality was 11.9%. Both numbers were highly dependent on the definitions for shock used as prevalence was lower and mortality higher when applying the WHO definition. Aetiologies associated with increased mortality included sepsis and gastroenteritis. Respiratory diseases including bronchiolitis, asthma, and severe pneumonia were most commonly reported as the main underlying diagnosis in shock according to FEAST criteria, however had a low mortality.
Prevalence, mortality and definitions
A uniform bedside definition for paediatric shock that can be used in low-resource settings is lacking [5, 7]. In this study, we reported prevalence and mortality for two definitions for shock. We primarily chose the FEAST criteria for this study as it was used in the largest trial on paediatric shock and its treatment in sub-Saharan Africa to date, and found that it applied to 5.3% of children admitted to QECH. Although the FEAST trial is an important study, neither FEAST nor other studies reported prevalence of shock using the FEAST criteria for shock. Secondly, we applied the WHO shock definition, as it is the most important triage and treatment guideline for low-resource settings worldwide, and found a prevalence of 0.1% of all paediatric admissions. This prevalence is in line with other studies in African children, and corroborates with the FEAST trial reporting that only two percent of these children had WHO defined shock [6, 7].
The different definitions also affected the mortality of shock in this study, which was 11.9% applying the FEAST criteria. This was slightly higher than in the FEAST trial, that reported a 48 hour mortality ranging between 7.3–10.5% [4], which may be explained by better outcomes which are common in randomized controlled trials. We found that mortality in children with WHO defined shock was much higher (60%). This is in line with mortality reported in other studies for children with WHO defined shock, ranging from 41.5% to 100% [7, 11].
Based on the different prevalence and mortality, we believe that neither the FEAST nor the WHO criteria accurately describe paediatric shock. The FEAST criteria may result in overreporting of shock, as nearly half of the children had reactive airway disease which was associated with a mortality of 0%. The WHO definition may result in an underestimation of shock prevalence as it appears to describe children with an advanced stage of shock. These children had extremely high mortality and clinicians reported shock as an admission diagnosis in at least twice the number of children, a concern also raised by other researchers [6, 7]. However, the WHO ETAT guideline and other researchers suggest that the presence of some, but not all, clinical signs of shock used in the WHO definition may be less specific for shock and could also be due to other conditions such as fever, hypoxemia, pain, or exposure [5, 12, 13]. Irrespective of the exact prevalence and mortality, shock is an important clinical syndrome in African children as it has a high mortality and given its severity has a relatively high prevalence. A more clinically relevant bedside definition of paediatric shock is however urgently needed.
Aetiology
Critically ill children such as children with shock, in our study and others, are often diagnosed with multiple underlying diseases [4, 6]. This study underlines that data is limited on the aetiology shock in African children, in part due to limited diagnostic capacity in low-resource settings which makes differentiation of underlying aetiologies challenging and based on clinical characteristics only [14].
Sepsis and septic shock are recognized as an important contributor to childhood mortality worldwide [15, 16]. In this study, children with shock diagnosed with sepsis as their main diagnosis indeed had the highest odds of death. Mortality in these children was higher than reported in Kenyan children with suspected septic shock (63% vs. 23%) [6]. Diagnosing sepsis is however challenging in most low-resource settings like Malawi due to limited microbiological testing. In our study site blood cultures are available and were performed in a third. Still sepsis was not commonly diagnosed (5.7%) and in an even smaller proportion a pathogen was isolated from blood cultures, although antibiotics were administered in almost 45% of children with shock. A possible explanation could be prior antibiotic administration which may have lowered prevalence of bacteraemia in our population. The organisms we found are in line with prior research from African settings [17, 18].
Gastroenteritis was the second diagnosis associated with high odds of death. Children with gastroenteritis were excluded from the FEAST trial, however Mbevi et al. reported that the majority of children with shock had diarrhoea [6]. Whether these children have gastroenteritis or gastrointestinal symptoms due to systemic illness is less clear, which makes choosing the appropriate treatment strategy challenging [19].
Respiratory disease was the most common diagnosis in this study, which included bronchiolitis, asthma and severe pneumonia, and these children had the lowest odd of dying. A similar proportion of children with shock were diagnosed with respiratory tract infection in the FEAST trial (42%) and by Mbevi et al (46%) [4, 6]. A potential explanation may be a pulmonary infection as the cause of circulatory failure. Alternatively, one could state that children with respiratory distress (and/or hypoxia) often have an increase in heart rate and respiratory rate and therefore may be “falsely” labelled as shocked applying the FEAST criteria. The low mortality in these children suggests that the latter may apply in our setting and possibly also in other studies [4].
Limitations
Limitations of this study include the use of the main discharge or death diagnosis in a setting that allowed us to do some diagnostic testing, however, we did not perform comprehensive testing to assess all aetiologies of shock in all children. Furthermore, the use of routinely collected data from paper hospital admission files, including variables needed for inclusion criteria, contributed to the missing data. This might have resulted in an underestimation of prevalence and mortality, as previous studies have shown that files of children who died had more missing data, possibly due to prioritization of care over recording admission information [20]. Furthermore, not all admitted children could be included for further analysis of underlying aetiology, as parents or guardians did not give their consent or were discharged before consent (174 of 679 eligible children, 25.6%). This might have led to an overestimate of mortality in this group and an underestimate of diseases that could have led to faster recovery.
Conclusion
Considering the high mortality, and relatively high prevalence, we conclude that shock is an important clinical presentation in Malawi. Our data underline that there is a need for a clinically relevant bedside definition for shock, not only to more precisely define the prevalence but more importantly to improve the clinical approach of shock in African children and facilitate research that can be easily compared. Our data suggest sepsis and gastroenteritis are common diseases associated with poor outcome. As aetiological data on shock in African children are limited and a comprehensive diagnostic study is lacking, these data contribute to our understanding of paediatric shock.
Supporting information
S1 Table. Definitions for shock used in the WHO ETAT guideline, FEAST trial and this study.
https://doi.org/10.1371/journal.pgph.0002282.s001
(DOCX)
S2 Table. Description of variables used in this study.
https://doi.org/10.1371/journal.pgph.0002282.s002
(DOCX)
S3 Table. Diagnostic categories of the included children (N = 505) and which main discharge diagnosis was considered for the diagnostic categories.
https://doi.org/10.1371/journal.pgph.0002282.s003
(DOCX)
S4 Table. Diagnostic categories in children who died (n = 75).
https://doi.org/10.1371/journal.pgph.0002282.s004
(DOCX)
S1 Checklist. Inclusivity in global research.
https://doi.org/10.1371/journal.pgph.0002282.s006
(DOCX)
S2 Checklist. STROBE statement—Checklist of items that should be included in reports of observational studies.
https://doi.org/10.1371/journal.pgph.0002282.s007
(DOCX)
Acknowledgments
We want to thank the doctors and nurses of Queen Elizabeth Central Hospital who cared for the children, and the children and their parents or guardians who gave consent to be included in the study.
References
- 1.
Levels & Trends in Child Mortality: Report 2021, Estimates developed by the United Nations Inter-agency Group for Child Mortality Estimation. New York: United Nations Inter-agency Group for Child Mortality Estimation (UN IGME); United Nations Children’s Fund, 2021.
- 2. Ayieko P, Ogero M, Makone B, Julius T, Mbevi G, Nyachiro W, et al. Characteristics of admissions and variations in the use of basic investigations, treatments and outcomes in Kenyan hospitals within a new Clinical Information Network. Arch Dis Child. 2016;101(3):223–9. Epub 2015/12/15. pmid:26662925.
- 3.
Pocket Book of Hospital Care for Children: Guidelines for the Management of Common Childhood Illnesses. Geneva: World Health Organization; 2013.
- 4. Maitland K, Kiguli S, Opoka RO, Engoru C, Olupot-Olupot P, Akech SO, et al. Mortality after fluid bolus in African children with severe infection. The New England journal of medicine. 2011;364(26):2483–95. Epub 2011/05/28. pmid:21615299.
- 5.
Updated guideline: paediatric emergency triage, assessment and treatment. Geneva: World Health Organization, 2016.
- 6. Mbevi G, Ayieko P, Irimu G, Akech S, English M. Prevalence, aetiology, treatment and outcomes of shock in children admitted to Kenyan hospitals. BMC medicine. 2016;14(1):184. Epub 2016/11/17. pmid:27846837.
- 7. Houston KA, George EC, Maitland K. Implications for paediatric shock management in resource-limited settings: a perspective from the FEAST trial. Critical care (London, England). 2018;22(1):119. Epub 2018/05/08. pmid:29728116.
- 8. Prin M, Itaye T, Clark S, Fernando RJ, Namboya F, Pollach G, et al. Critical Care in a Tertiary Hospital in Malawi. World J Surg. 2016;40(11):2635–42. Epub 2016/10/22. pmid:27230398.
- 9. Gondwe MJ, Henrion MYR, O’Byrne T, Masesa C, Lufesi N, Dube Q, et al. Clinical diagnosis in paediatric patients at urban primary health care facilities in southern Malawi: a longitudinal observational study. BMC Health Serv Res. 2021;21(1):150. Epub 2021/02/17. pmid:33588848.
- 10. Vandenbroucke JP, von Elm E, Altman DG, Gøtzsche PC, Mulrow CD, Pocock SJ, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med. 2007;4(10):e297. Epub 2007/10/19. pmid:17941715.
- 11. Ahmad S, Ellis JC, Kamwendo H, Molyneux EM. Impact of HIV infection and exposure on survival in critically ill children who attend a paediatric emergency department in a resource-constrained setting. Emerg Med J. 2010;27(10):746–9. Epub 2010/06/03. pmid:20515902.
- 12. Duke T. What the African fluid-bolus trial means. The Lancet. 2011;378(9804):1685–7. pmid:21683436
- 13. Southall DP, Samuels MP. Treating the wrong children with fluids will cause harm: response to ’mortality after fluid bolus in African children with severe infection’. Arch Dis Child. 2011;96(10):905–6. Epub 2011/07/01. pmid:21715393.
- 14. Yadav H, Shah D, Sayed S, Horton S, Schroeder LF. Availability of essential diagnostics in ten low-income and middle-income countries: results from national health facility surveys. Lancet Glob Health. 2021;9(11):e1553–e60. Epub 2021/10/10. pmid:34626546.
- 15. Weiss SL, Peters MJ, Alhazzani W, Agus MSD, Flori HR, Inwald DP, et al. Surviving Sepsis Campaign International Guidelines for the Management of Septic Shock and Sepsis-Associated Organ Dysfunction in Children. Pediatr Crit Care Med. 2020;21(2):e52–e106. Epub 2020/02/08. pmid:32032273.
- 16.
Global report on the epidemiology and burden of sepsis: current evidence, identifying gaps and future directions. Geneva: World Health Organization, 2020 9240010785.
- 17. Iroh Tam PY, Musicha P, Kawaza K, Cornick J, Denis B, Freyne B, et al. Emerging Resistance to Empiric Antimicrobial Regimens for Pediatric Bloodstream Infections in Malawi (1998–2017). Clin Infect Dis. 2019;69(1):61–8. Epub 2018/10/03. pmid:30277505.
- 18. Assies R, Snik I, Kumwenda M, Chimalizeni Y, Langton J, van Woensel JBM, et al. Etiology, Pathophysiology and Mortality of Shock in Children in Low (Middle) Income Countries: A Systematic Review. J Trop Pediatr. 2022;68(4). Epub 2022/07/08. pmid:35796755.
- 19. Akech S, Ayieko P, Gathara D, Agweyu A, Irimu G, Stepniewska K, et al. Risk factors for mortality and effect of correct fluid prescription in children with diarrhoea and dehydration without severe acute malnutrition admitted to Kenyan hospitals: an observational, association study. The Lancet Child & adolescent health. 2018;2(7):516–24. Epub 2018/07/05. pmid:29971245.
- 20. Olson D, Davis NL, Milazi R, Lufesi N, Miller WC, Preidis GA, et al. Development of a severity of illness scoring system (inpatient triage, assessment and treatment) for resource-constrained hospitals in developing countries. Trop Med Int Health. 2013;18(7):871–8. Epub 2013/06/14. pmid:23758198.