Model-based meta-analysis of relationship between pioglitazone and histological outcomes in non-alcoholic steatohepatitis patients
Quyen Thi Tran(1), Ngoc-Anh Thi Vu(1), Jung-woo Chae(1,2,*), Hwi-yeol Yun(1,2,*)
(1) College of Pharmacy, Chungnam National University, South Korea; (2)Bio-AI Convergence Research Center, Chungnam National University, South Korea; *: Those of authors contributed equally as correspondence
Objectives:
Non-alcoholic fatty liver disease (NAFLD) is a global pandemic health problem with a prevalence of approximately 30% worldwide population[1]. Around a quarter of NAFLD cases can progress to non-alcoholic steatohepatitis (NASH) and severe complications, such as fibrosis, cirrhosis, and hepatocellular carcinoma. However, currently, there has been no approved FDA-drug for NASH. The treatments are focusing on lifestyle modifications, such as diet, exercise, and weight loss. Pioglitazone, a commonly used type 2 diabetes mellitus drug, is currently used as an off-label for treatment NASH according to guidelines of the Japanese Society of Gastroenterology[2], EASL[3], AASLD[4]. Our study aims to perform a model-based meta-analysis to examine the quantitative efficacy of pioglitazone in improving histological parameters in patients with NASH.
Methods:
A comprehensive search was performed in Pubmed and clinicaltrials.gov and selected criteria-matching studies. Histological endpoints (including steatosis, inflammation, ballooning and fibrosis) were main outcomes. Data on histological endpoints were collected from literature and used for creating virtual data because of sparse data. Then model for virtual histological dataset were performed using logistic model. Model evaluations were carried out by visual predictive check and bootstrap method. The procedures were performed in NONMEM (version 7.5.0) and R (version 4.2.1).
The logit for the probability of a patient having histological score equal or greater to a specific score m (m = 0,…,4) was given by
Logit[P(Yij ≥ m)] = baseline(m) - f(Dose,t) + eta(i)
where denotes the histological score for the ith patients at the jth time point; m is the histological score (for steatosis and inflammation, m = 0,1,2,3; for ballooning, m = 0,1,2; for fibrosis, m = 0,1,2,3,4); baseline (m) is the baseline conditions for Y ≥ m; f(Dose, t) is the function describing the drug effect, and eta(i) is the inter-individual variability that follows normal distribution with mean zero and variance omega2.
The probability distribution of histological scores is calculated by inverse logit of the probability as below:
P(Yij≥m)] = (exp(baseline(m)+f(Dose,t)+eta(i))/(1+exp(baseline(m)+f(Dose,t)+eta(i))
The probability of each score was derived from upper level of cumulative probability as below:
P(Yij≥m)] = P(Yij≥m)] - P(Yij≥(m-1))]
Results:
A total of 425 studies were initially identified. After screening title, abstract and full-text, eight studies matched with our criteria and were included in the final analysis. Logit model was used to analyze four categories’ outcomes. The results showed that using pioglitazone improved all four histological parameters. For example, the improvement of steatosis is increased when dose of pioglitazone increases and duration of treatment increases, which was presented below equation:
f(Dose,t) =Dose * Emax*Time/(EC50 + Time)
Where, baseline(m≥1)=2.9, baseline(m≥2)=0.95, baseline(m=3)=-0.34, slope=0.054, EC50=0.778. For ballooning, scoring system of 0-2 was used, thus baseline(m≥1)=1.12, baseline(m=2)=0.58, slope=0.0835, EC50=17.4.
For inflammation and fibrosis, effect of pioglitazone was represented as linear relationship:
f(Dose,t) = Slope * Dose *Time
For inflammation, score from 0-3 was used, hence baseline(m≥1)=2, baseline(m≥2)=-0.3, baseline(m=3)=-3.32, slope=0.00159. For fibrosis, grade of 0-4 was used, then baseline(m≥1)=0.903, baseline(m≥2)=-0.457, baseline(m≥3)=-1.394, baseline(m=4)=-2.677, slope=0.00159.
Conclusions:
The administration of pioglitazone led to histologic improvement in subjects with nonalcoholic steatohepatitis. The model of pioglitazone was developed well to present its effect on NASH histological outcome. Further, this model can be used to compare efficacy of NASH-targeted treatment and confirmed effects of new NASH treatments.
References:
[1] K. Riazi et al., “The prevalence and incidence of NAFLD worldwide: a systematic review and meta-analysis,” Lancet Gastroenterol. Hepatol., vol. 7, no. 9, pp. 851–861, 2022.
[2] S. Watanabe et al., “Evidence-based clinical practice guidelines for nonalcoholic fatty liver disease/nonalcoholic steatohepatitis,” J. Gastroenterol., vol. 50, no. 4, pp. 364–377, 2015, doi: 10.1007/s00535-015-1050-7.
[3] G. Marchesini et al., “EASL-EASD-EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease,” J. Hepatol., vol. 64, no. 6, pp. 1388–1402, 2016, doi: 10.1016/j.jhep.2015.11.004.
[4] N. Chalasani et al., “The diagnosis and management of nonalcoholic fatty liver disease: Practice guidance from the American Association for the Study of Liver Diseases,” Hepatology, vol. 67, no. 1, pp. 328–357, 2018, doi: 10.1002/hep.29367.