Epidemiology of childhood acute kidney injury in England using e-alerts

Background Few studies describe the epidemiology of childhood acute kidney injury (AKI) nationally. Laboratories in England are required to issue electronic (e-)alerts for AKI based on serum creatinine changes. This study describes a national cohort of children who received an AKI alert and their clinical course. Methods A cross-section of AKI episodes from 2017 are described. Hospital record linkage enabled description of AKI-associated hospitalizations including length of stay (LOS) and critical care requirement. Risk associations with critical care (hospitalized cohort) and 30-day mortality (total cohort) were examined using multivariable logistic regression. Results In 2017, 7788 children (52% male, median age 4.4 years, interquartile range 0.9-11.5 years) experienced 8927 AKI episodes; 8% occurred during birth admissions. Of 5582 children with hospitalized AKI, 25% required critical care. In children experiencing an AKI episode unrelated to their birth admission, Asian ethnicity, young (<1 year) or old (16-<18 years) age (reference 1-<5 years), and high peak AKI stage had higher odds of critical care. LOS was higher with peak AKI stage, irrespective of critical care admission. Overall, 30-day mortality rate was 3% (n = 251); youngest and oldest age groups, hospital-acquired AKI, higher peak stage and critical care requirement had higher odds of death. For children experiencing AKI alerts during their birth admission, no association was seen between higher peak AKI stage and critical care admission. Conclusions Risk associations for adverse AKI outcomes differed among children according to AKI type and whether hospitalization was related to birth. Understanding the factors driving AKI development and progression may help inform interventions to minimize morbidity. Lay Summary Acute kidney injury (AKI) is a sudden drop in kidney function. This may have long-term health effects. In England, alerts triggered by rises in a person's blood creatinine levels are sent to the UK Renal Registry. We looked at a 1-year snapshot of alerts for children aged under 18 years across England. We found that children who were very young (under 1 year) or old (16-<18 years) were more likely to need critical care or die during an AKI episode compared with others. Longer length of stay in hospital was seen with increasing AKI alert severity. Children with an AKI alert during their birth admission had the highest risk of needing critical care and death, but this did not relate to AKI severity. This national work has detected factors linked to serious outcomes in AKI; by identifying them, we can now start to address them.