Document Type



Paediatrics and Child Health (East Africa)


Introduction: Acute kidney injury (AKI) is a frequent life-threatening complication in hospitalized children. Emerging data suggest AKI is a heterogeneous condition that varies based on the underlying cause and is composed of distinct phenotypes. The objective of this study was to define AKI phenotypes using proposed classification systems in Ugandan children hospitalized with hypoxemia and to evaluate differences in phenotypes by malaria infection.

Methods: Between 2019 and 2021, 2402 Ugandan children <5 years of age hospitalized with hypoxemia were enrolled in a cluster randomized trial of solar powered oxygen delivery across 20 districts in Uganda. At enrollment, urine NGAL was measured using a point-of-care lateral flow test with a positive test defined as a level ≥150ng/mL. Malaria was assessed using a threeband rapid diagnostic test. In an extended sub-study, 491 children had creatinine measured to define AKI. AKI was defined using a single creatinine measure at enrolment and phenotypically characterized using two acute dialysis quality initiative (ADQI) proposed AKI phenotypes. The AKI biomarker definition incorporated urine NGAL into the KDIGO definition[group 1, no AKI; group 2, subclinical AKI (biomarker positive); group 3, AKI; group 4, biomarker positive AKI]. The ADQI sepsis AKI phenotype groups stage 1 AKI as sepsis phenotype (SP)-1 irrespective of biomarker status and differentiates severe AKI (stage 2/3) based on biomarker positivity where severe AKI that is biomarker negative is (SP2) and severe biomarker positive AKI is SP3.

Results: Overall, 491 children were included in the extended study with AKI defined and uNGAL measured. The median age was 1.3 years (interquartile range, 0.7 to 2.3) and 53.8% of children were male. There were 4 deaths (0.8%) and 24 children required transfer to a higher-level health facility (4.9%). Among children included, 91.2% met a clinical definition of pneumonia and 49.5% were positive for malaria. The frequency of creatinine defined AKI was 32.0% (157/491) and 36.5% (179/491) were biomarker positive. AKI was associated with a 3.24-fold increase in mortality (95% CI 0.34 to 31.4) but underpowered to show a difference. In children without malaria, 17.7% were biomarker positive and AKI negative (subclinical AKI, 44/248) while 37.5% of children had AKI (93/284) of whom 39.8% (37/93) were biomarker positive. In children with malaria, 14.0% had subclinical AKI, 34/243), 59.3% had AKI (144/243) with 44.4% of AKI cases biomarker positive (64/144). Children with malaria had a higher frequency of AKI compared to children without malaria (59.6% vs. 37.6%, p<0.001) but comparable frequency of a positive biomarker test (41.3% vs. 36.2%, p=0.10). Using the sepsis phenotype criteria, 16.3% of children had SP1, 17.9% were SP2 and 14.1% were SP3. When evaluating the sepsis phenotype by malaria status, children with malaria were more likely to have SP2 (23.1% vs. 12.9%) and SP3 (18.1% vs.10.1%) compared to children without malaria (p<0.001).

Conclusions: In this population of children hospitalized with hypoxemia across 20 health centers in Uganda, KDIGO-defined AKI was more common in children with malaria. While there was no difference in the AKI-biomarker classification based on malaria status, children with malaria were more likely to have severe phenotypes of AKI.

Publication (Name of Journal)

Kidney International Reports


Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Included in

Pediatrics Commons