Early detection of chronic kidney disease in low-income and middle-income countries: development and validation of a point-of-care screening strategy for India

Abstract

INTRODUCTION: Although deaths due to chronic kidney disease (CKD) have doubled over the past two decades, few data exist to inform screening strategies for early detection of CKD in low-income and middle-income countries. METHODS: Using data from three population-based surveys in India, we developed a prediction model to identify a target population that could benefit from further CKD testing, after an initial screening implemented during home health visits. Using data from one urban survey (n=8698), we applied stepwise logistic regression to test three models: one comprised of demographics, self-reported medical history, anthropometry and point-of-care (urine dipstick or capillary glucose) tests; one with demographics and self-reported medical history and one with anthropometry and point-of-care tests. The 'gold-standard' definition of CKD was an estimated glomerular filtration rate <60 mL/min/1.73 m2 or urine albumin-to-creatinine ratio ≥30 mg/g. Models were internally validated via bootstrap. The most parsimonious model with comparable performance was externally validated on distinct urban (n=5365) and rural (n=6173) Indian cohorts. RESULTS: A model with age, sex, waist circumference, body mass index and urine dipstick had a c-statistic of 0.76 (95% CI 0.75 to 0.78) for predicting need for further CKD testing, with external validation c-statistics of 0.74 and 0.70 in the urban and rural cohorts, respectively. At a probability cut-point of 0.09, sensitivity was 71% (95% CI 68% to 74%) and specificity was 70% (95% CI 69% to 71%). The model captured 71% of persons with CKD and 90% of persons at highest risk of complications from untreated CKD (ie, CKD stage 3A2 and above). CONCLUSION: A point-of-care CKD screening strategy using three simple measures can accurately identify high-risk persons who require confirmatory kidney function testing.