Strategies to Reduce Mortality Among HIV-infected and HIV-exposed Children Admitted With Severe Acute Malnutrition

Background

HIV-infected and HIV-exposed-uninfected children (HEU) are at increased risk of developing malnutrition. Severely malnourished children have high mortality rates, but mortality is higher in those that are HIV-infected. Preliminary audits at the Mwanamugimu Nutrition Unit, Mulago Hospital, in 2014 showed that 43% of the severely malnourished children that died were HIV-infected/HEU, despite only 30% of admissions being HIV-infected/HEU, with deaths due to infections in 90% of cases.

Objectives

This study aims to investigate whether empiric use of an antibiotic with greater antimicrobial sensitivity (ceftriaxone) than standard-of-care (ampicillin plus gentamicin) will reduce mortality among HIV-infected/HEU children admitted to Mwanamugimu Nutrition Unit. Secondary objectives include: comparing length of hospitalization, weight-for-height, weight-for-age and height-for-age z-scores between ceftriaxone versus standard of care (ampicillin and gentamicin) treatment arms; ascertaining the pattern/antimicrobial sensitivity of pathogens among participants; determining the prevalence and factors associated with HIV-infection among severely malnourished children; and evaluating the pharmacokinetics (PK) of lopinavir/ritonavir (LPV/r) among severely malnourished HIV-infected children.

Methods

This will be an open label randomized controlled trial involving 300 children; 76 HIV-infected (current mortality - 33%) and 224 HEU (current mortality - 26%). The participants will be randomized to receive 1week of ceftriaxone (n= 150) or standard-of-care (ampicillin/gentamicin) (n=150), in addition to other routine care; the ratio of HIV-infected to HEU (1:3) in this sample is reflective of the current proportions of the HIV-infected and HEU children admitted at Mwanamugimu Nutrition Unit. The trial's primary outcome will be in hospital mortality. 300 randomised children provides >80% power to detect reductions in mortality from the expected 28% to 14%, allowing for 10% noncompliance/lost-to-follow-up in each group. Secondary outcomes will be: length of hospitalization; weight-for-height, weight-for-age and height-for-age z-scores; and pattern/antimicrobial sensitivity of pathogens. In addition, 280 severely malnourished children of unknown serostatus will be tested for HIV at admission to determine the prevalence and factors associated with HIV-infection. 280 children provide 80% power to determine the prevalence of HIV-infection. Furthermore, all the HIV-infected children on LPV/r will each provide sparse pharmacokinetic (PK) samples to evaluate the PK of LPV/r among malnourished children. In this PK sub-study, geometric means of steady-state LPV PK parameters [Area Under the Curve (AUC) 0-12h, maximum concentration (Cmax) and concentration at 12 hours after dose (C12h)] will be determined. The PK parameters (AUC 0-12h, Cmax, C12) will then be put in pharmacokinetic-pharmacodynamic (PK-PD) models to determine optimal doses for the study population.