A population-based prognostic study published in JAMA Oncology found cisplatin treatment to be associated with a predictable decrease in the estimated glomerular filtration rate. According to Grant et al, their findings place patients with lower baseline kidney function at the highest risk for developing chronic kidney disease.
“Although cisplatin has extended countless lives, some patients are left with irreversible kidney damage through tubular cell apoptosis, inflammation, and ischemic injury,” the investigators commented. “These results can be used in clinical practice to inform consent and guide prevention strategies.”
Study Details
The investigators developed prediction models based on a retrospective cohort study of patients who received cisplatin for nonhematologic cancer in an outpatient setting between July 2014 and June 2017. The models were tested on two cohorts: a temporal test cohort of patients from Ontario, Canada, who initiated treatment between July 2017 and June 2020, and an external test cohort from a single center in the United States.
A total of 9,521 patients were included, of whom 9,010 were not diagnosed with chronic kidney disease (defined as estimated glomerular filtration rate < 60 mL/min/1.73 m²) before treatment.
Key Findings
Among those without chronic kidney disease at baseline, 13.6% developed the condition after treatment with cisplatin, 0.9% developed grade 4 or worse disease, and 0.2% required dialysis. The estimated glomerular filtration rate decreased by a mean of 8.1 mL/min/1.73 m2 (95% confidence interval [CI] = 7.8–8.4 mL/min/1.73 m2). A simple spline-based regression model based solely on the baseline estimated glomerular filtration rate was found to predict chronic kidney disease in the temporal (AUC = 0.80, 95% CI = 0.78–0.82) and external (AUC = 0.73, 95% CI = 0.66–0.78) test cohorts after treatment. Similarly, according to the investigators, such a model predicted the estimated glomerular filtration rate after treatment (temporal: mean absolute error = 12.6 mL/min/1.73 m2, 95% CI = 12.3–13.0 mL/min/1.73 m2; external: mean absolute error = 14.3 mL/min/1.73 m2, 95% CI = 13.2–15.5 mL/min/1.73 m2). Compared with the univariable models, complex machine-learning systems incorporating all features seemed to fail to improve predictions.
Robert C. Grant, MD, PhD, of Princess Margaret Cancer Centre, University Health Network, Toronto, is the corresponding author of the JAMA Oncology article.
Disclosure: The study was funded by the Institute for Clinical Evaluative Sciences (ICES), which is funded by an annual grant from the Ontario Ministry of Health and the Ministry of Long-Term Care, and others. For full disclosures of the study authors, visit jamanetwork.com.
