Mortality after emergency unit fluid bolus in febrile Ugandan children

• Published in: PloS one 2023-08, Vol.18 (8), p.e0290790-e0290790
• Main Authors: Rice, Brian, Hawkins, Jessica, Nakato, Serena, Kamara, Nicholas
• Format: Article
• Language: English
• Subjects: Analysis; Best practice; Biology and Life Sciences; Cardiac arrhythmia; Care and treatment; Children; Criteria; Diarrhea; Emergency medical care; Emergency medical services; Emergency medicine; Fatalities; Fluids; Health aspects; High income; Hospital patients; Hypothermia; Hypoxia; Malaria; Medical research; Medicine and Health Sciences; Medicine, Experimental; Missing data; Mortality; Patient outcomes; Patients; Pediatrics; Regression analysis; Resuscitation; Review boards; Sensitivity analysis; Sepsis; Statistical analysis; Sub-Saharan Africa; Training; Uganda; Variables; Vomiting
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0290790

Pediatric fluid resuscitation in sub-Saharan Africa has traditionally occurred in inpatients. The landmark Fluid Expansion as Supportive Therapy (FEAST) trial showed fluid boluses for febrile children in this inpatient setting increased mortality. As emergency care expands in sub-Saharan Africa, fluid resuscitation increasingly occurs in the emergency unit. The objective of this study was to determine the mortality impact of emergency unit fluid resuscitation on febrile pediatric patients in Uganda. This retrospective cohort study used data from 2012-2019 from a single emergency unit in rural Western Uganda to compare three-day mortality for febrile patients that did and did not receive fluids in the emergency unit. Propensity score matching was used to create matched cohorts. Crude and multivariable logistic regression analysis (using both complete case analysis and multiple imputation) were performed on matched and unmatched cohorts. Sensitivity analysis was done separately for patients meeting FEAST inclusion and exclusion criteria. The analysis included 3087 febrile patients aged 2 months to 12 years with 1,526 patients receiving fluids and 1,561 not receiving fluids. The matched cohorts each had 1,180 patients. Overall mortality was 4.0%. No significant mortality benefit or harm was shown in the crude unmatched (Odds Ratio [95% Confidence Interval] = 0.88 [0.61-1.26] or crude matched (1.00 [0.66-1.50]) cohorts. Adjusted cohort analysis (including both complete case analysis and multiple imputation) and sensitivity analysis of patients meeting FEAST inclusion and exclusion criteria all also failed to show benefit or harm. Post-hoc power calculations showed the study was powered to detect the absolute harm seen in FEAST but not the relative risk increase. This study's primary finding is that fluid resuscitation in the emergency unit did not significantly increase or decrease three-day mortality for febrile children in Uganda. Universally aggressive or fluid-sparing emergency unit protocols are unlikely to be best practices, and choices about fluid resuscitation should be individualized.