Scholarlink

Background <p>Diabetic nephropathy (DN) is a prevalent and progressive complication of type 2 diabetes mellitus (T2DM), with a significant impact on morbidity and long-term renal outcomes. While conventional risk models often assess predictors at static time points, time-to-event (TTE) analysis offers a dynamic framework to evaluate the hazard of DN development over time.</p> Objective <p>This study aimed to develop and externally validate a parametric TTE model for DN in patients with T2DM.</p> Methods <p>A retrospective cohort of 251 T2DM patients from two tertiary hospitals in Malaysia was analyzed over a 7.2-year follow-up period. The outcome of interest was time to confirmed DN diagnosis, defined by persistent albuminuria. Parametric survival modelling using NONMEM 7.3.0 evaluated different baseline hazard structures, with covariate effects assessed through forward inclusion and backward elimination. Model selection was based on the likelihood ratio test, objective function value (OFV), Kaplan–Meier visual predictive checks (KM-VPC), relative standard error, and scientific plausibility. External validation was performed using data from 109 patients.</p> Results <p>The Gompertz hazard function best described the data. Every 1&#xa0;mmol/L increase in FBS was associated with a 25.8% increase in DN hazard (HR = 1.258; 95% CI 1.139–1.389), and each 1-mmHg increase in SBP was associated with an increase in hazard by 7.3% (HR = 1.073; 95% CI 1.043–1.103). Higher eGFR was protective, with each unit increase above 85.20&#xa0;mL/min/1.73 m<sup>2</sup> decreasing hazard by 4.1% (HR = 0.959; 95% CI 0.942–0.975). Both internal and external KM-VPCs demonstrated good model calibration and predictive accuracy.</p> Conclusion <p>This validated TTE model identifies SBP, FBS, and eGFR as key predictors of DN onset in T2DM. These findings support routine monitoring and control of modifiable risk factors and provide a foundation for future models incorporating therapeutic interventions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Time-to-event modelling of diabetic nephropathy in patients with type 2 diabetes mellitus: A parametric survival analysis with internal and external validation

  • 期刊论文

  • 发表日期：

    2026年5月25日

查看全文
  • Sohail Aziz,
  • Siti Maisharah Sheikh Ghadzi,
  • Sabariah Noor Harun

摘要

Background

Diabetic nephropathy (DN) is a prevalent and progressive complication of type 2 diabetes mellitus (T2DM), with a significant impact on morbidity and long-term renal outcomes. While conventional risk models often assess predictors at static time points, time-to-event (TTE) analysis offers a dynamic framework to evaluate the hazard of DN development over time.

Objective

This study aimed to develop and externally validate a parametric TTE model for DN in patients with T2DM.

Methods

A retrospective cohort of 251 T2DM patients from two tertiary hospitals in Malaysia was analyzed over a 7.2-year follow-up period. The outcome of interest was time to confirmed DN diagnosis, defined by persistent albuminuria. Parametric survival modelling using NONMEM 7.3.0 evaluated different baseline hazard structures, with covariate effects assessed through forward inclusion and backward elimination. Model selection was based on the likelihood ratio test, objective function value (OFV), Kaplan–Meier visual predictive checks (KM-VPC), relative standard error, and scientific plausibility. External validation was performed using data from 109 patients.

Results

The Gompertz hazard function best described the data. Every 1 mmol/L increase in FBS was associated with a 25.8% increase in DN hazard (HR = 1.258; 95% CI 1.139–1.389), and each 1-mmHg increase in SBP was associated with an increase in hazard by 7.3% (HR = 1.073; 95% CI 1.043–1.103). Higher eGFR was protective, with each unit increase above 85.20 mL/min/1.73 m2 decreasing hazard by 4.1% (HR = 0.959; 95% CI 0.942–0.975). Both internal and external KM-VPCs demonstrated good model calibration and predictive accuracy.

Conclusion

This validated TTE model identifies SBP, FBS, and eGFR as key predictors of DN onset in T2DM. These findings support routine monitoring and control of modifiable risk factors and provide a foundation for future models incorporating therapeutic interventions.