Related Experiment Videos
Association of Variability of Monthly Continuous Glucose Monitoring-Derived Metrics With Diabetic Kidney Disease in
Byeongjae Kang1,2, Rosa Oh3, Taeyoung Kim1
1Medical AI Research Center, Research Institute for Future Medicine, Samsung Medical Center, Seoul, Republic of Korea.
Insights
The standard deviation of monthly time in range (TIR) and time in tight range (TITR) are key indicators for diabetic kidney disease (DKD) in type 1 diabetes (T1D). Machine learning models effectively identify DKD risk using these continuous glucose monitoring metrics.
Area of Science:
- Endocrinology
- Nephrology
- Data Science
Background:
- Hemoglobin A1c (HbA1c) is standard for glycaemic control but misses glycaemic variability.
- Continuous glucose monitoring (CGM) offers longitudinal data for improved diabetes management.
- Diabetic kidney disease (DKD) is a major complication in type 1 diabetes (T1D).
Purpose of the Study:
- Investigate the discriminatory capacity of longitudinal CGM metrics.
- Identify CGM metric patterns associated with DKD in T1D.
- Utilize machine learning (ML) for DKD risk prediction.
Main Methods:
- Analyzed 1-year CGM data from 282 T1D patients.
- Defined DKD by persistent albuminuria or reduced eGFR.
- Developed and compared ML models (LightGBM, XGBoost, Random Forest, Logistic Regression).
Main Results:
- LightGBM model achieved highest performance (AUROC=0.91).
- Standard deviation (SD) of monthly time in range (TIR) and time in tight range (TITR) were most influential features.
- SD of monthly TIR and TITR remained elevated even in early DKD.
Conclusions:
- SD of monthly TIR and TITR strongly associate with DKD in T1D.
- CGM metrics offer improved DKD discrimination beyond HbA1c.
- ML integration of longitudinal CGM data enhances DKD assessment.
Background:
While HbA1c is the standard for monitoring long-term glycaemic control, it fails to capture glycaemic variability. We investigated the discriminatory capacity of longitudinal continuous glucose monitoring (CGM) metrics and identified CGM metric patterns associated with diabetic kidney disease (DKD) in individuals with type 1 diabetes (T1D) using machine learning (ML).
Methods:
We analysed cross-sectional data from 282 T1D patients with 1-year consecutive CGM data. DKD was defined by persistent laboratory abnormalities (urine albumin-creatinine ratio ≥ 30 mg/g or estimated glomerular filtration rate < 60 mL/min/1.73 m2) confirmed by at least two measurements within the 1-year period. LightGBM, XGBoost, Random Forest, and Logistic Regression (LR) were developed. Feature importance was assessed using SHAP analysis.
Results:
The LightGBM model achieved the highest performance (AUROC = 0.91 [95% CI, 0.88-0.93], F1 score = 0.65). All tree-based ML models outperformed the LR model. SHAP analysis identified the standard deviation (SD) of monthly time in range (TIR) and time in tight range (TITR) as the most influential features. In contrast, the CV of sensor glucose did not differ significantly between groups (p = 0.416). Even in the early DKD subgroup, the SD of monthly TIR and TITR remained significantly elevated.
Conclusion:
The SD of monthly TIR and TITR is strongly associated with DKD in T1D, whereas the CV of sensor glucose is not. ML-based integration of these longitudinal metrics offers improved discrimination of concurrent DKD status beyond conventional glycaemic markers.
Related Concept Videos
Diabetic Nephropathy
Drug Dosing in Renal Diseases: Estimation of Glomerular Filtration Rate Based on Serum Creatinine Concentration
Chronic Kidney Disease II: Clinical Manifestations
Drug Dosing in Renal Diseases: Measurement of Glomerular Filtration Rate
Chronic Kidney Disease III: Interprofessional Care
Type I Diabetes III: Clinical Manifestations