W.WENG1, S.Y.WONG1, Y.ANG2, H.X.NG2, C.K.LIM3, S.C.YEO1
Tan Tock Seng Hospital1, NHGHQ2, NHG Polyclinics3
Accurate identification of individuals at risk of developing CKD may improve clinical care and delay progression. 2 risk prediction equations (Nelson, 2019) used to estimate the risk of incident estimated glomerular filtration rate (eGFR)<60 ml/min/1.73m2 in diabetic and non-diabetic patients have not been validated locally. We aim to validate the non-diabetic equation in our local population and develop further prediction models.
Demographics, clinical and laboratory data of hypertensive non–diabetic patients with baseline eGFR ≥60ml/min/1.73m2 who were on follow up with NHG polyclinics between 2010 to 2015 were collected. 5-year risk of incident eGFR<60ml/min/1.73m2 was predicted using the non-diabetic equation. Variables included age, sex, race, eGFR, history of cardiovascular disease(CVD), ever smoker, BMI and albuminuria. Models employed Cox proportional hazard regression and were evaluated by the Harrell C-statistic.
35,043 patients were included in the study. 49.1% were female and mean age was 58.0 years. Baseline mean eGFR was 90.8ml/min/1.73m2. 3461 (9.9%) patients developed eGFR<60ml/min/1.73m2. Older age, lower baseline eGFR, male gender, history of CVD, ever smoker, elevated BMI were risk factors for eGFR<60ml/min/1.73m2. Applicability of the existing equation was limited by missing albuminuria, absence of black race and exclusion of non-hypertensive patients in our cohort. Nonetheless, the model without the 3 variables demonstrated C-statistic of 0.85 (95%CI:0.85-0.86). Novel models including race, blood pressure and history of heart failure demonstrated similar results.
The adapted risk prediction equations demonstrate similar high discrimination in our population, despite utilising fewer variables. They have potential clinical applicability and should be validated in a future cohort.