Long-Term Outcomes of Children Undergoing Dialysis-treated AKI: Some Opinions and Prospects.

Shiyuan Wei

DOI: https://doi.org/10.1681/ASN.2021060862

IF: 14.978

2021-09-16

Journal of the American Society of Nephrology

Abstract:In a recently published article in JASN, Robinson et al. performed a retrospective study reporting the long-term kidney outcomes after childhood dialysis-treated AKI (DT-AKI). This topic is very important, and there is currently very limited evidence on this issue, especially in studies with a large population and long follow-up. The study found that pediatric survivors of DT-AKI had significantly increased long-term risks of kidney failure, death, major adverse kidney events, CKD, and hypertension, as compared with comparators. I believe the methods in this article are clear. However, I would like to point out some suggestions that may make the conclusions more convincing. Firstly, one of the biggest limitations of this article is a lack of a definition of AKI on the basis of laboratory tests, which may lead to miscoding of the study exposure. Although the authors have excluded children with an inborn error of metabolism or poisoning, survivors of DT-AKI may still include some patients without AKI, such as those with high potassium or other emergencies. However, if the researchers defined DT-AKI as a combination of an AKI diagnosis and the requirement of acute dialysis, instead of only using acute dialysis codes, it would greatly reduce their misclassification errors. Secondly, the choice of comparator group for the study appears to be discussed. Only 28 of 6,752 matched children were patients with AKI. Thus, the effect of DT-AKI on outcomes could be overestimated by selecting patients without AKI as the comparator group. The ideal comparator group should be patients with AKI but who are not on acute dialysis. Thirdly, death and kidney failure are in a competitive relationship and there is a significant difference in the incidence of death and kidney failure (approximately 3:1). Therefore, the effect of the composite end point may be largely due to death. As we all know, it is not easy to observe the end point of kidney failure in children during follow-up. I suggest that using the change of serum creatinine or eGFR as a surrogate end point may be more meaningful for kidney progression in children. Finally, I am quite confused about why the final model only included ten variables instead of the 31 variables listed in the baseline table 1. Did the authors have a process to select these variables? I think some variables, such as age, income, rural status, and admission to the intensive care unit, have a great influence on outcomes, which should be included in the final model. In summary, a larger sample size and more rigorous design studies are needed to further investigate the prognosis of children with DT-AKI.

Medicine