Predictors associated with critical care need and in-hospital mortality among children with laboratory-confirmed COVID-19 infection in a high HIV infection burden region

INTRODUCTION: Despite the extra mortality associated with COVID-19 death globally, there is scant data on COVID-19-related paediatric mortality in Sub-Saharan Africa. We assessed predictors of critical care needs and hospital mortality in South African children with laboratory-confirmed SARS-CoV-2 infection in region with high HIV infection burden. METHODS: We conducted a secondary multicentre analysis of the AFREhealth cohort (a multinational, multicentre cohort of paediatric COVID-19 clinical outcomes across six African countries) of children admitted to the Inkosi Albert Luthuli, a quaternary hospital in KwaZulu-Natal, South Africa, with confirmed RT-PCR between March 2020 and December 2020. We constructed multivariable logistic regression to explore factors associated with the need for critical care (high care/ intensive care hospitalisation or oxygen requirement) and cox-proportional hazards models to further assess factors independently associated with in-hospital death. RESULTS: Of the 82 children with PCR-confirmed SARS-CoV-2 infection (mean ± SD age: 4.2 ± 4.4 years), 35(42.7%) were younger than one year, 52(63%) were female and 59(71%) had a pre-existing medical condition. Thirty-seven (45.2%) children required critical care (median (IQR) duration: 7.5 (0.5–13.5) days) and 14(17%) died. Independent factors associated with need for critical care were being younger than 1 year (aPR: 3.02, 95%CI: 1.05–8.66; p = 0.04), having more than one comorbidity (aPR: 2.47, 95%CI: 1.32–4.61; p = 0.004), seizure (aPR: 2.39, 95%CI: 1.56–3.68; p < 0.001) and impaired renal function. Additionally, independent predictors of in-hospital mortality were exposure to HIV infection (aHR: 6.8, 95%CI:1.54–31.71; p = 0.01), requiring invasive ventilation (aHR: 3.59, 95%CI: 1.01–12.16, p = 0.048) and increase blood urea nitrogen (aHR: 1.06, 95%CI: 1.01–1.11; p = 0.017). However, children were less likely to die from COVID-19 if they were primarily admitted to quaternary unit (aHR: 0.23, 95%CI: 0.1–0.86, p = 0.029). CONCLUSION: We found a relatively high hospital death rate among children with confirmed COVID-19. During COVID-19 waves, a timely referral system and rapid identification of children at risk for critical care needs and death, such as those less than one year and those with comorbidities, could minimize excess mortality, particularly in high HIV-infection burden countries.