Kidney Week

Abstract: FR-PO0394

Risk Prediction Model for Development of CKD After AKI in Pediatric Patients

Session Information

• AKI: Biomarkers, Diagnostics, and Risk Prediction
  October 23, 2026 | Location: Exhibit Hall A, Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Acute Kidney Injury

• 102 AKI: Clinical, Outcomes, and Trials

Authors

• Frisby-Zedan, Jeanne, Ann & Robert H Lurie Children's Hospital of Chicago, Chicago, Illinois, United States

Jeanne Frisby-Zedan, MD, MS

• Robinson, Cal, The Hospital for Sick Children, Toronto, Ontario, Canada

Cal Robinson, MBChB, PhD, FASN

• Smith, Graham Douglas/D, Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada

Graham Douglas/D Smith, MSc, PhD

• Bota, Sarah E., Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada

Sarah E. Bota

• Ulrich, Emma Heydari, University of Alberta, Edmonton, Alberta, Canada

Emma Heydari Ulrich, MD

• Dart, Allison, University of Manitoba, Winnipeg, Manitoba, Canada

Allison Dart, MD

• Basu, Rajit K., Ann & Robert H Lurie Children's Hospital of Chicago, Chicago, Illinois, United States

Rajit K. Basu, MD, MS, FASN

• Chanchlani, Rahul, McMaster University, Hamilton, Ontario, Canada

Rahul Chanchlani, MD, MBBS, MS, FASN

Background

Pediatric acute kidney injury (AKI) is associated with long-term adverse sequelae including chronic kidney disease (CKD), end stage kidney disease (ESKD), and death. We aim to derive a risk prediction model for development of CKD among pediatric survivors of hospital-acquired AKI.

Methods

This study is a population-based retrospective cohort study of children 0-18 years in Ontario, Canada with hospital-acquired AKI from 1996-2024. The primary outcome is development of CKD or ESKD at 1, 3 and 5 years post-hospitalization. Patients with prior AKI, CKD, dialysis, and kidney transplantation, as well as death within 90 days of hospital discharge were excluded. Demographic, clinical, and laboratory risk factors were examined for associations with CKD+ESKD using multivariable logistic regression, which was subsequently used to develop timepoint-specific risk prediction models.

Results

Of the 7,289 patients included, followed at a median of 8.8 years, 18.6% developed CKD+ESRD, with 10.7%, 13.9%, and 15.6% developing the outcome at 1, 3 and 5 years, respectively. 5.8% died following hospitalization, 3.9% developed ESRD, 18.4% developed de novo hypertension, and 59% had re-hospitalization. In multivariable logistic regression modeling, glomerulonephritis, kidney replacement therapy (KRT), duration of hospitalization, age, and pre-existing hypertension increased odds of CKD+ESKD, whereas sepsis, rhabdomyolysis, and ECMO decreased odds for CKD+ESKD in the 1, 3 and 5-year models; AUC 0.72, 0.70, 0.70, respectively. In a subgroup of patients with laboratory data (n=2,003), KRT, lower baseline eGFR, and proteinuria increased odds of CKD+ESKD in the 1, 3 and 5-year models; AUC 0.75, 0.68, 0.70, respectively.

Conclusion

Readily available clinical and laboratory variables can moderately predict CKD/ESKD following pediatric AKI and may support use of risk-stratified follow-up models.

Funding

• Private Foundation Support