An evaluation of the prevalence, characteristics, outcomes and predictors of shock in children admitted to paediatric wards at Queen Elizabeth Central Hospital in Blantyre, Malawi

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
Kamuzu University of Health Sciences
Shock is an important pathophysiological mechanism of death in children in low-resource settings. Despite this, limited data is available on the prevalence, causes and outcome of shock in children in these settings. We performed a prospective study to assess the prevalence, aetiology and risk factors of death in children with shock. This data will be essential to improve guidelines and interventions to reduce shock related mortality in children. Methods: A prospective descriptive study was performed from 1st February 2019 to 31st January 2020 at Queen Elizabeth Central Hospital (QECH) paediatric department. ALL paediatric admissions screened for shock as defined by FEAST were recruited. The criteria used were impaired consciousness and/or respiratory distress in combination with at least one sign of impaired perfusion: a capillary refill time (CRT) >3 seconds, cold peripheries, weak radial pulse volume and/or severe tachycardia. The WHO definition of shock was used as a comparative definition. Demographic, clinical, laboratory and outcome data were collected from the patient records. Predictors for death were assessed using univariate and multivariate models. Results: Out of 12840 admissions, 679 had shock resulting in a prevalence of 5.3%. Of these 505 were included in the study of 15/439 (3.4%) fulfilled the more stringent WHO criteria for shock. The median age was 17 months and ranged from 2 months to 16 years. Respiratory distress was reported in 397/488 (81.4%), fever in 383/495 (77.4%), vomiting or diarrhoea was reported in 183/484 (37.8%) and 127/478 (26.6%) respectively and severe malnutrition was documented in 39/471 (8.3%). Severe anaemia (Hb ≤5g/dL) was present in 19/334 (5.7%), and 67/395 (17.7%) tested positive for malaria. HIV prevalence was vi 27/358 (7.5%) and blood cultures were positive in 8/176 (4.5%). The main clinical diagnoses on discharge were viral/reactive lower respiratory tract diseases 211/470 (44.9%), pneumonia 89/470 (18.8%), gastroenteritis 64/470 (13.6%) and presumed sepsis 57/470 (12.0%). Overall mortality in shocked children was 79/679 (11.6%). We constructed two multivariate models aimed at a) predicting outcome, and b) assessing disease associated outcomes. Clinical factors predictive of death were low coma score (AOR = 4.9, 95% CI = 2.2 - 11.1), delayed CRT (AOR = 3.5, 95% CI = 1.4 – 8.5) and dehydration (AOR 5.9, 95% CI 3.2 – 11.1). The main clinical diagnoses of children that died were presumed sepsis 34/76 (44.7%), gastroenteritis 21/76 (23.7%), severe malaria 13/76 (17.1%), severe malnutrition 9/76 (11.8%) and meningitis 8/76 (10.5%). In the explanatory model for causative factors, having a diagnosis of presumed sepsis (AOR = 9.9, 95% CI = 4.1 – 23.8) or gastroenteritis (AOR = 3.7, 95% CI = 1.8 – 7.4) was associated with increased mortality, while having viral/reactive airways disease was not associated with death (AOR = 0.02, 95% CI = 0.005 – 0.079). Conclusion: Shock is a common diagnosis in children seen at QECH in Malawi using the modified FEAST criteria, affecting 1 in 18 admissions. The actual prevalence of children with shock may be lower as the FEAST definition we used may have overestimated shock prevalence. This was evidenced by the fact that approximately 40% of those that qualified had a purely respiratory condition and only 3.4% of those who qualified fit the very strict WHO definition. This discrepancy underlines the lack of a valid bedside definition for shock in children in Malawi. In our population, mortality was high at 11%. The children who present with a low BCS, delayed CRT and dehydration were more likely to die. The diagnoses of sepsis, gastroenteritis and malaria were contributors to death. This might underscore the importance of their prompt treatment.