A PREDICTION MODEL FOR CHRONIC KIDNEY DISEASE PROGRESSION...
Author(s)
Ruth Sim, PhD1, Yen Jun Wong, PhD2, Shaun W. Lee, PhD3.
1Monash, Jalan Lagoon Selatan, Bandar Sunway, Malaysia, 2Monash University Malaysia, Subang Jaya, Malaysia, 3Lecturer, Monash University, Subang, Malaysia.
1Monash, Jalan Lagoon Selatan, Bandar Sunway, Malaysia, 2Monash University Malaysia, Subang Jaya, Malaysia, 3Lecturer, Monash University, Subang, Malaysia.
OBJECTIVES: To develop a prediction model for future eGFR trajectories using routinely collected data from adults with CKD attending public hospitals in Malaysia.
METHODS: This prognostic study used longitudinal data collected between 2020 and 2024 from six Malaysian public hospitals. Adults aged ≥ 18 years with CKD and at least three eGFR measurements over ≥ 24 months were eligible. The primary outcome was the eGFR decline. Ten pre-specified baseline predictors were used: age, sex, history of diabetes mellitus, body mass index, HbA1c, systolic and diastolic blood pressure, serum cholesterol, serum creatinine and haemoglobin. A linear mixed-effects model with random intercepts and slopes was used to quantify the effects on eGFR trajectory.
RESULTS: A total of 435 eligible adults (median age 65 years [IQR 56-71]; 257 [57.6%] male) contributing 3,093 eGFR measurements were included in the primary model (mean 7.1 measurements per participant, maximum follow-up 7.9 years). Median baseline eGFR was 51.6 mL/min/1.73 m2 (IQR 36.7-75.3) and 63.8% were in stages G3a-G5. The most influential variables were: age (β = -0.74; 95% CI, -0.85 to -0.62 per year), male gender (+5.70; 2.62 to 8.77), systolic blood pressure (-0.14; -0.23 to -0.06 per mmHg), serum creatinine (-0.21; -0.23 to -0.19 per µmol/L) and haemoglobin levels (-0.11; -0.15 to -0.08 per g/dL), which were independently associated with baseline eGFR. The median estimated annual eGFR slope was -0.74 mL/min/1.73 m2/year. 22.1% of participants had predicted slopes meeting the commonly used threshold for rapid progression of <-3 mL/min/1.73 m²/year
CONCLUSIONS: A linear mixed-effects model using ten routinely collected baseline variables can be useful to predict eGFR decline among Malaysian adults with CKD. However, further validation is needed for the model to be used in routine clinical practice
METHODS: This prognostic study used longitudinal data collected between 2020 and 2024 from six Malaysian public hospitals. Adults aged ≥ 18 years with CKD and at least three eGFR measurements over ≥ 24 months were eligible. The primary outcome was the eGFR decline. Ten pre-specified baseline predictors were used: age, sex, history of diabetes mellitus, body mass index, HbA1c, systolic and diastolic blood pressure, serum cholesterol, serum creatinine and haemoglobin. A linear mixed-effects model with random intercepts and slopes was used to quantify the effects on eGFR trajectory.
RESULTS: A total of 435 eligible adults (median age 65 years [IQR 56-71]; 257 [57.6%] male) contributing 3,093 eGFR measurements were included in the primary model (mean 7.1 measurements per participant, maximum follow-up 7.9 years). Median baseline eGFR was 51.6 mL/min/1.73 m2 (IQR 36.7-75.3) and 63.8% were in stages G3a-G5. The most influential variables were: age (β = -0.74; 95% CI, -0.85 to -0.62 per year), male gender (+5.70; 2.62 to 8.77), systolic blood pressure (-0.14; -0.23 to -0.06 per mmHg), serum creatinine (-0.21; -0.23 to -0.19 per µmol/L) and haemoglobin levels (-0.11; -0.15 to -0.08 per g/dL), which were independently associated with baseline eGFR. The median estimated annual eGFR slope was -0.74 mL/min/1.73 m2/year. 22.1% of participants had predicted slopes meeting the commonly used threshold for rapid progression of <-3 mL/min/1.73 m²/year
CONCLUSIONS: A linear mixed-effects model using ten routinely collected baseline variables can be useful to predict eGFR decline among Malaysian adults with CKD. However, further validation is needed for the model to be used in routine clinical practice
Conference/Value in Health Info
2026-09, ISPOR Asia Pacific 2026, Bangkok, Thailand
Value in Health, Volume 55, Issue S1
Code
CO17
Topic
Clinical Outcomes
Topic Subcategory
Relating Intermediate to Long-term Outcomes
Disease
SDC: Geriatrics, SDC: Urinary/Kidney Disorders