A Model to Predict Risk of Incident CKD

Prediction tools incorporating self-reported health information may be useful in increasing awareness of chronic kidney disease (CKD), helping to identify modifiable lifestyle risk factors, and slowing disease progression. Ariana J. Noel, MD, MSc, and colleagues developed and validated a survey -based prediction equation to identify adults at risk for incident CKD. The model was described in the Clinical Journal of the American Society of Nephrology [2023;18(1):28-35].

The study cohort included adults with an estimated glomerular filtration rate (eGFR) ≥70 mL/min/1.73 m2 in Ontario, Canada. Eligible participants (n=22,200) completed a comprehensive general population health survey between 2000 and 2015. Prediction equations included age, sex, comorbidities, lifestyle factors, diet, and mood.

Models with and without baseline eGFR were created. Data on 15,522 patients in the UK Biobank were used to externally validate the models. The primary outcome was new onset CKD, defined as eGFR <60 mL/min/1.783 m2, within ≤8 years of follow-up.

Mean age of the cohort was 55 years, 58% were women, and baseline eGFR was 95 mL/min/1.73 m2. Nine percent of the cohort (n=1981) developed new-onset CKD during a median follow-up of 4.2 years. The final models included lifestyle factors (smoking, alcohol use, physical activity), and comorbidities (diabetes, hypertension, cancer).

The model was discriminating in both those with and those without a measure of eGFR at baseline (5-year c-statistic with baseline eGFR, 83.5; 95% CI, 82.2-84.9; and 5-year c-statistic without baseline eGFR measurement 81.0; 95% CI, 79.8-82.4). The model was well calibrated.

The corresponding values in the external validation cohorts were 78.1 (95% CI, 74.2-82.0) and 66.0 (95% CI, 61.6-70.4) in those with and without baseline eGFR measurement, respectively. Calibration was maintained in the validation cohorts.

In conclusion, the authors said, “Self-reported lifestyle and health behavior information from health surveys may aid in predicting incident CKD.”