Rates of Laboratory Adverse Events by Course in Pediatric Leukemia Ascertained Using Automated Electronic Health Record Extraction: A Retrospective Cohort Study Report from the Children’s Oncology Group

• Authors: Tamara Miller, Emory University; Kelly D. Getz, University of Pennsylvania; Yimei Li, University of Pennsylvania; Biniyam G. Demissei, University of Pennsylvania; Peter C. Adamson, University of Pennsylvania; Todd A. Alonzo, University of Southern California; Evanette Burrows, Children's Hospital of Philadelphia; Lusha Cao, Children's Hospital of Philadelphia; Sharon M. Castellino, Emory University; Marla H. Daves, Baylor College of Medicine; Brian T. Fisher, University of Pennsylvania; Robert Gerbing, Children's Oncology Group; Robert W. Grundmeier, University of Pennsylvania; Edward M. Krause, Children's Hospital of Philadelphia; Judy Lee, Children's Healthcare of Atlanta; Philip J. Lupo, Baylor College of Medicine; Karen r. Rabin, Baylor College of Medicine; Mark Ramos, Children's Hospital of Philadelphia; Michael E. Scheurer, Baylor College of Medicine; Jennifer J. Wilkes, University of Washington; Lean E. Winestone, University of California at San Francisco; Douglas S. Hawkins, University of Washington; M. Monica Gramatges, Baylor College of Medicine; Richard Aplenc, University of Pennsylvania
• Journal Title: The Lancet Haematology
• Volume: Volume 9, Number 9
• Publisher: Elsevier | 2023-09-01, Pages 678-688
• Type of Work: Article | Post-print: After Peer Review
• Publisher DOI: http://doi.org/10.1016/S2352-3026(22)00168-5
• Subjects: Health Sciences, Public Health; Health Sciences, Oncology
• Keywords: adverse events; automated laboratory AE ascertainment; leukemia

Abstract

Background: Adverse events (AEs) are often misreported on clinical trials, leading to an incomplete understanding of toxicities. We aimed to test automated laboratory AE ascertainment and grading (ExtractEHR) to demonstrate scalability and define AE rates for children with acute myeloid and acute lymphoblastic leukemia (AML, ALL). Methods: Patients aged 0–22 years at Children’s Healthcare of Atlanta (CHOA, 01–01-2010–01-11–2018), Children’s Hospital of Philadelphia (CHOP, 01–01-2011–31-12–2014), and Texas Children’s Hospital (TCH, 01–01-2011–31-12`2014) were included. ExtractEHR acquired, cleaned, and graded laboratory data per CTCAEv5 for 22 commonly evaluated AEs with numerically-based CTCAE definitions. Descriptive statistics tabulated AE frequencies. ExtractEHR-ascertained AEs were compared to manually reported Children’s Oncology Group (COG; AAML1031, NCT01371981; AALL0932, NCT02883049) AEs for trial-enrolled patients. Findings: Laboratory results on 1077 patients (AML: 166, ALL: 911; CHOA: 583; CHOP 200; TCH 294) with 4611 courses (AML: 549, ALL: 4062) were extracted, processed, and graded. For AML patients, 86/166 (52%) were female, 96/166 (58%) were white, and 132/166 (80%) were non-Hispanic. Among ALL patients, 406/911 (45%) were female, 596/911 (65%) were white, and 641/911 (70%) were non-Hispanic. AML patients experienced the most AEs during induction I and intensification II. Hypokalemia (1/6 (17%) to 75/156 (48%) courses) and alanine aminotransferase (ALT) increased (13/134 (10%) to 27/156 (17%) courses) were the most prevalent non-hematologic AML AEs. ALL patients experienced the greatest number of AEs during induction and maintenance (prevalence ≥10% for 8 AEs; Induction and Maintenance: Anemia, Platelet Count Decreased, White Blood Cell Count Decreased, Neutrophil Count Decreased, Lymphocyte Count Decreased, ALT Increased, Hypocalcemia; Induction: Hypokalemia; Maintenance: AST Increased, Blood Bilirubin Increased). ExtractEHR laboratory AE rates were substantially higher than COG-reported AE rates for AEs with at least 2% prevalence. Interpretation: ExtractEHR is scalable and accurately defines laboratory AE rates for pediatric acute leukemia that are higher than COG-reported AEs. These rates can be used for comparisons between therapies and to counsel patients regarding chemotherapy risks for patients treated on or off of trials. ExtractEHR AE ascertainment provides a novel mechanism for improving laboratory AE trial reporting.