Sohail Aziz (1), Siti Maisharah Sheikh Ghadzi (1), Shalini C Sree Dharan (2), Sabariah Noor Harun (1)
(1) School of Pharmaceutical Sciences, Universiti Sains Malaysia, Malaysia. (2) Hospital Sultan Abdul Halim, Sungai Petani, Kedah, Ministry of Health, Malaysia
Objectives:
Time-to-event (TTE) pharmacometrics analysis has an important role in the parametric quantification of the occurrence of diabetic nephropathy. Incorporating censoring event and time components in the TTE analysis allow unbiased and better estimates of survival function than the traditional regression methods. The present study focused on the development of TTE model of diabetic nephropathy among patients with type 2 diabetes mellitus (T2DM). Furthermore, the study evaluated the influence of covariates (discrete and continuous) on the progression of diabetes to diabetic nephropathy.
Methods:
The data was collected retrospectively from a tertiary care hospital comprising of 251 T2DM patients. The maximum follow-up period was about 7.2 years. Parametric TTE analysis was performed using NONMEM software. The event was defined by the development of diabetic nephropathy (persistent proteinuria for three visits). Three TTE models; exponential, Gompertz and Weibull hazard models were tested for base model. The effect of covariates on the overall hazard of developing diabetic nephropathy was investigated using stepwise approach. Univariate analysis was performed, followed by multivariate analysis. Multivariate analysis included the significant covariates from the univariate analysis. The likelihood ratio test (LRT) with a significance level (α) of 5% was used to discriminate between nested models. A drop of 3.84 with one degree freedom by the addition of covariate with the base model is considered statistically significant (p>0.05). The selection of final model was based on the lowest OFV, Kaplan Meier visual predictive check (KM-VPC) with 100 simulations, relative standard error (RSE) as well as the scientific plausibility. The final model was externally validated with a datset comprising of 109 T2Dm patients.
Results:
Out of 251 patients, 96 (38%) patients developed proteinuria. The best base model was Gompertz hazard model (h(t) = λ0 . eβ x t) with the baseline hazard of 1.006 /year (RSE=53%) and shape (probability density function) of 2.6671 (RSE=9%). The hazard of developing proteinuria decreases with increase in estimated glomerular filtration rate above 85.20 ml/min/1.73m2 (HR=0.959; 95% CI [0.975-0.942] while increase in fasting blood sugar over time above 7.4 mmol/L increase the hazard of developing proteinuria (HR=1.258; 95% CI [1.139-1.389]. Similarly, the increase in systolic blood pressure over time above 132 mmHg increase the risk of developing proteinuria in diabetic patients (HR=1.073; 95% CI [1.043-1.103].
H(t) = λ0 . exp [(SHP*(TIME) (θSBP * (SBP-132) + θFBS * (FBS-7.4) + θEGFR * (EGFR-85.20))
Conclusions:
The use of pharmacometrics approach in this study will help to generate new knowledge on the underlying mechanism of the progression of diabetic nephropathy in patients with T2DM by exploring patients and disease data over time and developing thereon so called pharmacometrics models. These approaches may not only improve the medical practitioners understanding about disease progression but will also be helpful in improving treatment strategies. Present study found that eGFR trend can be useful indicator for diabetic nephropathy while proper management of systolic blood pressure and fasting blood sugar may further help in minimizing the risks of developing diabetes associated proteinuria.
References:
[1] Beal, S. L., Sheiner, L. B. & Boeckman, A. J. NONMEM Users Guides. (1989–2006) (Icon Development Solutions, Ellicott City, MD, 2008).
[2].Barrett, J. S., Fossler, M. J., Cadieu, K. D. & Gastonguay, M. R. (2008). Pharmacometrics: A multidisciplinary field to facilitate critical thinking in drug development and translational research settings. J. Clin. Pharmacol. 48, pp. 632–649.
[3].Holford N. (2013). A time to event tutorial for pharmacometricians. CPT: Pharmacometrics and System Pharmacology, 2 (e43), pp.1-8.
[4].Møller, J. B.,Overgaard, R. V., Kjellsson, M. C., Kristensen, N. R., Klim, S., Ingwersen, S. H., & Karlsson, M. O. (2013). Longitudinal modeling of the relationship between mean plasma glucose and HbA1c Following antidiabetic treatments. CPT: Pharmacometrics and Systems Pharmacology, 2(10).
Reference: PAGE 30 (2022) Abstr 9948 [www.page-meeting.org/?abstract=9948]
Poster: Drug/Disease Modelling - Endocrine