Large-scale AI analysis reveals missed opportunities in albuminuria testing and disease-modifying therapy implementation

Abstract

Aims Albuminuria is a key diagnostic and prognostic biomarker of chronic kidney disease (CKD), associated with adverse cardiovascular and renal outcomes. Despite guideline recommendations, urine albumin-to-creatinine ratio (UACR) testing is infrequently performed in cardiology. This study assessed the uptake of UACR testing, the estimated prevalence of undiagnosed albuminuria, and the use of disease-modifying therapies in patients with cardio-kidney-metabolic (CKM) disease. Methods and results We conducted a retrospective cohort study of all adults seen at the cardiology department of a tertiary referral centre between 2019 and 2024. Data were extracted using CTcue, an AI-driven platform. Albuminuria was defined as UACR >= 30 mg/g. A weighted logistic regression model estimated albuminuria prevalence in untested patients. Among 77 351 patients (44.8% female, mean age 64.4 years), only 8.9% had a recorded UACR, of whom 46.4% had albuminuria. Testing rates were low across high-risk groups: 29.9% in diabetes, 21.7% in heart failure, and 13.7% in hypertension. In untested patients, the predicted prevalence of albuminuria was 36.6%, and highest in those with eGFR <30 mL/min/1.73m2 (70.0%), heart failure (47.8%), or diabetes (46.8%). Use of disease-modifying therapies was low, even among patients with confirmed albuminuria. In patients with documented vs. predicted albuminuria, 43.0% vs. 39.1% received renin-angiotensin system inhibitors, 12.8% vs. 5.7% received SGLT2 inhibitors, and <1% in both groups received finerenone. Conclusion Albuminuria is substantially underdetected in cardiology practice, possibly contributing to underuse of effective CKM therapies. Systematic UACR screening with structured treatment protocols may help close this gap and improve outcomes for patients with CKM disease.