Accuracy of weight-based artesunate and artemether–lumefantrine dosing among children hospitalised with malaria in Kisumu County, Kenya: a multicentre clinical-record study

Abstract

Injectable artesunate reduces mortality from severe paediatric malaria, but only when the amount given matches the weight-specific recommendation. Current World Health Organization and Kenyan guidance specify 3.0 mg/kg per dose for children weighing under 20 kg and 2.4 mg/kg for those weighing 20 kg or more, whereas the earlier 2010 Kenyan guideline specified 2.4 mg/kg at all weights; follow-on artemether–lumefantrine is dosed by weight band. How closely routine inpatient practice matches these rules in western Kenya is not well described. Cross-sectional clinical-record study of 333 children aged 6–59 months hospitalised with malaria in eight public hospitals in Kisumu County, Kenya. Recommended doses were recalculated from each child’s recorded weight using the World Health Organization guidelines for malaria (2021) and the Kenyan national malaria guidelines (sixth edition, 2020), applicable when data were collected from September 2023. Dose error was the recorded minus the recommended dose, a negative value denoting a dose below the recommendation; signed, absolute and percentage errors were derived and doses classified as accurate (within a prespecified ±10% tolerance), below or above the recommendation. Artemether–lumefantrine tablet counts were compared with the recommended per-dose weight-band count. Proportions are reported with 95% confidence intervals. Confidence intervals for medians were obtained by percentile bootstrap. Results. Injectable artesunate was documented as first-line treatment, with a recorded dose, for 110 of the 333 children, and 97 of these were documented as continuing to artemether–lumefantrine; these formed the evaluable cohorts, and no record was excluded for a missing weight or dose. All 110 artesunate recipients weighed under 20 kg and required 3.0 mg/kg. No dose recorded on the treatment sheet fell within the tolerance; all 110 (100%, 95% CI 96.6–100) were below the recommendation. Median recorded dose was 2.37 mg/kg (IQR 2.22–2.52) and median absolute percentage dose error 20.9% (95% CI 18.1–22.8). Under a uniform 2.4 mg/kg rule, 98/110 (89.1%) would have been within ±10%. For artemether–lumefantrine, 86 children fell in the 5 to <15 kg band (one tablet per dose) and 11 in the 15 to <25 kg band (two tablets); 47/97 (48.5%, 95% CI 38.8–58.3) matched exactly, 47/97 exceeded and 3/97 (3.1%) fell below the recommendation; every discordance was a single tablet. Conclusion. Recorded artesunate doses were uniformly below the applicable recommendation, in a pattern consistent with the adult 2.4 mg/kg rule being applied to children under 20 kg. These findings describe an unimplemented dosing rule rather than random calculation error, and point to weight-band dosing charts and independent dose verification as low-cost remedies.