ABSTRACT
-
Background/Aims
The first metabolic dysfunction-associated steatotic liver disease (MASLD) drug was approved with an unsatisfactorily small effect size. This study aimed to determine key factors impacting the cost-effectiveness of a new hypothetical MASLD drug as well as its treatment efficacy.
-
Methods
A Markov model reflecting the natural history of MASLD was developed, incorporating fibrosis progression, cardiovascular disease risk, and mortality. Treatment effect of drug X (with $20,000 of annual cost) was assumed to achieve a ≥1 stage fibrosis regression, with a 25% gap of effect size in regression rate over non-treatment in the first year. The incremental cost-effectiveness ratio (ICER) over a 20-year horizon was estimated. And sensitivity analyses were conducted to explore uncertainty and identify influential factors.
-
Results
In the base case analysis, drug X provided an incremental gain of 1.32 quality-adjusted life years (QALYs) and 1.20 life years compared to the non-treatment, with an ICER of $68,010/QALY–below the $100,000/QALY willingness-to-pay threshold, indicating that drug X treatment is cost-effective. Two-way sensitivity analysis further highlighted that the drug should achieve at least a 15% initial regression gap and maintain a minimum 3% sustained durability gap to remain cost-effective. In addition baseline fibrosis stage distribution also acted as an influencing factor.
-
Conclusions
Long-term sustained durability of the hypothetical drug, patient distribution based on baseline fibrosis stage, as well as initial treatment response rate are key factors that influence the cost-effectiveness of new MASLD drugs.
-
Keywords: MASLD; Metabolic dysfunction-associated steatotic liver disease; Drug Therapy; Cost-effectiveness analysis; Liver cirrhosis
Study Highlights
• We simulated a Markov model linking short-term fibrosis regression to long-term health and economic outcomes to quantify the key determinants of cost-effectiveness for new MASLD drugs.
• At $20,000/year, hypothetical drug X gained 1.32 QALYs over 20 years and was cost-effective versus no treatment (ICER, $68,010/QALY; WTP, $100,000/QALY).
• Key drivers included the initial fibrosis regression gap, the baseline fibrosis-stage distribution, and attenuation of the treatment effect over time.
• Two-way sensitivity analyses indicated cost-effectiveness when the initial regression gap exceeded 15% and a minimum 3% treatment-effect gap was sustained.
Graphical Abstract
INTRODUCTION
Metabolic dysfunction-associated steatotic liver disease (MASLD), a new term for nonalcoholic fatty liver disease (NAFLD), affects over 30% of the global population, with a regional prevalence ranging from 30.8% in Asia to 42.6% in the Middle East-North Africa [
1,
2]. The disease burden related to MASLD may be greater than current estimates considering not only the morbidity and mortality from major adverse liver-related outcomes but also cerebrovascular, cardiovascular diseases, and either hepatic or extrahepatic cancers. Lifestyle modification, including dietary changes, significant weight loss (≥7% of body weight), and enhanced physical activity, is strongly recommended as nonpharmacological treatment options for the management of MASLD [
3,
4]. Several studies have shown that significant weight loss due to lifestyle modification is associated with a decrease in steatosis and/or fibrosis [
3-
5]. However, achieving the recommended weight loss is challenging, with only a 15% success rate reported in a clinical trial [
5]; furthermore, this value is likely higher than achievable in the real world [
6].
The regression rate of hepatic fibrosis is considered the most important surrogate marker for predicting long-term outcomes in MASLD [
7]. For this reason, most clinical studies of the effectiveness of treatments for MASLD set the histologic primary endpoint as regression of hepatic fibrosis without worsening of metabolic-associated steatohepatitis (MASH). Resmetirom, the first FDA-approved pharmacological therapy for MASH, resulted in statistically significant improvement in fibrosis reversal defined as at least one stage of fibrosis reversal with no worsening of NAFLD activity score in a recent phase 3 trial (25.9% in the Resmetirom groups vs. 14.2% in the control group,
P<0.001) [
8]. Despite being the first drug to result in histological improvement in a large-scale phase 3 clinical trial, the efficacy of Resmetirom was limited by the gap in fibrosis regression between Resmetirom and the placebo of only 11.7% [
8]. The magnitude of statistically significant improvement observed in clinical trials may differ from what is considered clinically effective or meaningful in real-world practice. Defining the significant effect size of histological improvement is an important issue in MASLD drug development and remains a challenging task. In this context, cost-effectiveness analysis can provide valuable insights that could aid in interpretation of the significance of outcomes in phase 3 clinical trials. Moreover, estimating the long-term benefit of a drug based solely on a 52-week phase 3 trial is a complex endeavor. Both the initial therapeutic effectiveness and a durable response (i.e., a response sustained beyond the first year) are critical factors in determining the clinical utility of a drug. Furthermore, MASLD drugs may vary in their effectiveness in reducing cardiometabolic risk and the risk of other comorbidities. Due to the lack of evidence concerning the long-term benefits of a drug at the time of drug approval, decision-makers opt to rely on data obtained from well-controlled phase 3 clinical trials, which occur within a limited time frame. Bridging the gap between statistical and clinical significance remains a critical challenge.
This study aimed to identify factors that influence the cost-effectiveness of novel MASLD treatments using a Markov model to aid in the interpretation of MASLD/MASH clinical trial outcomes and estimate the applicability of novel MASLD drugs in real-world practice.
MATERIALS AND METHODS
Ethics approval statement
The study protocol was conducted in accordance with both the Declarations of Helsinki and Istanbul and was approved by the Institutional Review Board of Hanyang University (IRB No. HY-2023-10-007). The requirement for informed consent was waived by the IRB due to the retrospective design of the study.
Study design and target population
A cost-utility analysis was conducted to evaluate the impact of treatment using hypothetical drug X on long-term clinical outcomes and the socioeconomic burden of MASLD. We constructed a Markov model, which was implemented using Microsoft Excel 2019, with nine health states simulating the natural course of MASLD (
Supplementary methods and
Fig. 1A). In our model, patients were able to move from one fibrosis state to another (depicted as a straight arrow) or stay in their current state (depicted as a recursive arrow) according to transition probabilities. We started with a hypothetical cohort of 40-year-old MASLD patients with fibrosis stages of F2 to F4 (F2 50%, F3 30%, and F4 20%) in the United States, considering the ongoing clinical trials including patients with F4 (
Table 1). Patients in this cohort were assigned to either the group treated with hypothetical drug X (treatment group) or the group who are untreated either with drug X or an intensive lifestyle intervention (ILI, non-treatment group). The length of the cycle was set to 1 year, and the time horizon was assumed to be 20 years. An annual discount rate of 3% was applied to all costs and outcomes. This study adhered to the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guidelines for economic evaluations (
Supplementary Table 1).
Treatment efficacy of drug X in terms of fibrosis improvement and the occurrence of cardiovascular disease
Hypothetical drug X was assumed to target patients with fibrosis stages of F2–F4 associated with MASLD. It was assumed to be an oral medication to improve liver fibrosis stages considering that recent novel pharmacological agent has targeted fibrosis improvement without worsening of MASH. The effectiveness of drug X on fibrosis improvement was defined as the gap in the annual rate of fibrosis regression between drug X and non-treatment (i.e., “Fibrosis regression rate per year by drug X in the treatment group”–“Fibrosis regression rate per year by natural course in the non-treatment group”) (
Table 1). In the base case analysis, the initial gap in the regression rate for the first year of medication was set at 25% (e.g., 32.9% and 7.9% of the treatment and non-treatment groups achieve one stage of fibrosis regression in patients with F2).
As there is obviously no data for our new hypothetical drug, we applied a waning treatment effect for our base-case analysis. Particularly, we assumed that it would be prudent to gradually decrease the effectiveness of drug X by 50% during the subsequent year as a conservative approach. This was done to reflect the inevitable waning of treatment effectiveness in terms of the tendency of patients to revert to their previous lifestyle habits and decreased compliance with drugs in real-world practice. Nevertheless, we assumed that the effectiveness of drug X would not decrease beyond the predetermined value of 2% (i.e., the gap of minimum sustained durability in regression rate) to be able to differentiate between the prognosis of patients in the treatment and no-treatment groups. In contrast to the conservative approach, we used to analyze treatment effectiveness, we also assumed that individuals who remained compliant with drug X would not progress beyond their initial fibrosis stage. Rates of fibrosis regression and progression up to the 5th year of treatment applied to this simulation are detailed in the
Supplementary methods and
Supplementary Table 2. An illustrative description of fibrosis stage movement over time by treatment is presented in
Fig. 1B and
Supplementary Fig. 1.
Given that drug X was assumed to target patients with F2–F4 fibrosis, we assumed that treatment discontinuation would be recommended if the patient’s fibrosis stage improved to F1 or F0. Worsening of liver disease with progression to DC or HCC were also considered indications for drug discontinuation. We assumed that individuals who ceased treatment with drug X would follow the natural history course of MASLD. We assumed that individuals who showed re-progression of fibrosis to F2 or higher would resume treatment in the following year despite stopping treatment the previous year.
The preventive effect of this hypothetical drug on cardiovascular disease (CVD) (ischemic heart disease, congestive heart failure, and stroke) was calculated as a relative risk ratio compared to the non-treatment group (0.8 times lower than non-treatment). As the representative effectiveness of newer MASLD drugs on the occurrence of CVD is limited, we assumed a relative risk ratio of 0.8 based on previous studies [
9,
10], assuming that it would be at least equal to or greater than that of ILI.
Transition probabilities of the natural history of MASLD
The non-treatment group followed the transition probabilities of natural regression and progression in fibrosis based on the fibrosis stage and the natural history of untreated MASLD (either pharmaceutical drugs or ILI) (
Table 2). Reference figures were obtained from a comprehensive review of published studies, government statistics, and our previous work (
Supplementary Tables 3,
4) [
10-
22]. The treatment group had the same transition probabilities as described above except that these probabilities were affected by the effectiveness of treatment at improving fibrosis and/or CVD risk.
Costs and utility weights
Based on previous studies [
23,
24], the cost of hypothetical drug X was assumed to be $54.7 per day, resulting in an approximate cost of $20,000 per year. Medical costs and utility weight for each health state were described in detail in
Supplementary methods and
Supplementary Table 5.
Analysis
Our economic model estimated costs, life-years (LYs), and quality-adjusted life-years (QALYs) gained from treatment vs. non-treatment over 20 years. The primary outcome was the incremental cost-effectiveness ratio (ICER) of treatment versus no treatment. Treatment with drug X was deemed cost-effective if the estimated ICER was <$100,000/QALY, which is the recommended willingness-to-pay (WTP) threshold in the US [
25]. ICER was calculated as follows:
Several sensitivity analyses were performed to assess the robustness of the results by altering applied values and assumptions to the base case analysis. One-way sensitivity analyses involved modifying various variables associated with the effectiveness of drug X (initial and durable response in regression rate of fibrosis, waning of the treatment effect, and ability of the drug to prevent CVD), fibrosis distribution among the population at baseline, and other analytic parameters such as adherence or age-specific risk [
26]. We also conducted sensitivity analysis using recent Phase 3 clinical data for Resmetirom [
8].
We performed a two-way sensitivity analysis to investigate the impact of variations in the estimated clinical effect parameters and their combinations on the cost-effectiveness of the hypothetical drug. We analyzed the interaction between the two key clinical outcomes associated with the hypothetical drug. This analysis allowed us to assess how changes in these clinical outcomes influenced the overall cost-effectiveness results. To generalize our study findings, we also conducted scenario analyses applying healthcare costs specific to South Korea (
Supplementary Table 6) and present these results in two-way sensitivity analyses. In addition, probabilistic sensitivity analysis was also performed (
Supplementary methods).
RESULTS
Base case analysis
For the base case analysis, we set the effect of the drug X treatment to be an initial gap of 25% in the annual rate of fibrosis regression between drug X and non-treatment, with a minimum sustained durability of 2%. Treatment with drug X was anticipated to cost $210,730 and to gain QALY by 11.38 (
Supplementary Table 7). Meanwhile, the non-treatment group was anticipated to cost $121,164 with a gain of 10.07 QALYs. Treatment with drug X led to an additional gain of 1.32 QALY and 1.20 LY per patient compared with the non-treatment strategy following the 20-year time horizon. Incremental costs between the two groups were $89,566 and the calculated ICER was $68,010/QALY. Considering a WTP threshold of $100,000/QALY, treatment with drug X was cost-effective in the setting of 25% treatment effectiveness in year 1 with the treatment effect waning by 50% per subsequent year in the base case analysis.
The results of the one-way sensitivity analysis revealed that the three most influential parameters were the initial gap in regression rate between the two groups, the distribution of fibrosis stage at baseline among the population, and the annual treatment effect waning percentage (
Fig. 2,
Supplementary Table 8). For example, an initial gap of 50% in the regression rate would lead to a decrease in the ICER to $28,885/QALY. Using drug X to treat only patients with stage F4 fibrosis was the most cost-effective strategy compared to the other scenarios targeting various groups of patients with different stages of fibrosis. In contrast, when only patients with F2 stage were targeted for treatment, the ICER was the highest among all analyses at $117,005/QALY. Additionally, the assumptions on the annual treatment effect waning also had a considerable impact on the ICER. In the scenario of maintaining initial treatment effect until 3rd year (i.e., treatment effect waning initiating in 4th year), the ICER decreased to $29,440/QALY. Under the assumption of 0% treatment effect waning, which indicates that the treatment effect gained with drug X in year 1 would remain the same afterwards, the ICER reached its lowest value of $16,482/QALY. Alternative scenario using transition probabilities by age for liver related events and extrahepatic events (i.e., CVD in our analysis) showed slightly higher ICER from $76,070 to $92,080 per QALY for cohort starting ages at 40 to 60, respectively. Expansion of the analytic time horizon, range of discount rates, treatment continuation in all fibrosis stages (including F0/F1 stage with lower adherence), medication adherence, and the minimum sustained durability rate of the two strategies considerably impacted the ICER. The effectiveness of drug X on preventing CVD had less impact on the ICER.
Applying the data from the clinical trial of Resmetirom in terms of the treatment effects and fibrosis distribution of the population, an ICER value of $72,638/QALY was calculated, which is not significantly different from that of the base case analysis value.
Two-way sensitivity analyses
In the two-way sensitivity analyses, we allowed wide variations of the treatment effects to reflect real-world practice where numerous factors affect treatment effectiveness compared to clinical trials. The initial treatment effect of drug X at year 1 (i.e., the initial gap in regression rate between the two groups) was allowed to vary from 10% to 50%. At the same time, the minimum sustained durability rates were allowed to vary from 2% to 10% or could remain at the initial gap of rates (i.e., no waning of treatment effect). Achieving the initial treatment (drug X) response of over 10% gap in fibrosis regression at year 1 and maintaining the minimum sustained durability of 3% over the years at an annual drug cost of $20,000 was cost-effective compared to non-treatment (
Fig. 3). The larger the initial gap in the effect between the two groups, the more cost-effective use of the drug became, even when the minimum sustained durability was very small. Conversely, a hypothetical drug with a higher minimum sustained durability was cost-effective even when the initial gap was relatively small. Additional two-way sensitivity analysis was performed by varying initial gap and annual treatment effect waning, and the hypothetical drug was likely to be cost-effective unless there was a 10% initial gap and more than 40% of effect waning at the same time (
Supplementary Fig. 2).
Furthermore, we adjusted for the drug cost to be $25,000/year, $30,000/year, or $40,000/year (
Supplementary Fig. 3). When the annual drug cost was set at $25,000, the initial gap in regression rate needed to be at least 25% for the drug to be cost-effective across all ranges of minimum sustained durable responses. When the cost of drug X was increased to $30,000/year, the initial gap in regression rate between the treatment and non-treatment groups needed to be 40% or higher or the minimum sustained durability had to be 3% or higher for the drug to be cost-effective for the treatment of MASLD patients with stage F2–F4 fibrosis.
Using healthcare costs specific to the Republic of Korea yielded similar results to the base case analysis (
Supplementary Fig. 4). When the same effectiveness estimates from the base case analysis were applied (25% of an initial gap in treatment response at Year 1 and 2% of the minimum sustained durable response over years), ICERs were calculated to be $11,517/QALY, $17,607/QALY, and $23,697/QALY in the settings of a daily treatment cost of $7.65, $9.56, and $11.47 (approximately ₩12,500, ₩10,000, and ₩15,000), respectively. Two of these ICERs were well below the acceptable threshold of ICER $23,322/QALY in the setting of the Republic of Korea.
Probabilistic sensitivity analysis results, which are depicted in a cost-effectiveness acceptability curve, indicated that treatment strategy was cost-effective across various ICER thresholds. For a threshold of $100,000/QALY, the probability of the treatment strategy being cost-effective compared with the non-treatment strategy was 74.7% (
Supplementary Figs. 5,
6). The curve demonstrated that the probability of treatment with drug X being more cost-effective than non-treatment occurred at a threshold of $69,000/QALY.
DISCUSSION
This study aimed to determine what significant treatment response data of a hypothetical drug would look like in real-world practice in terms of cost-effectiveness based on data from real randomized clinical trials. In this study, we found that waning of the treatment effect after the first year as well as the initial treatment response to a new drug (i.e., the initial gap in regression rate) both had a powerful impact on the cost-effectiveness of treatment, regardless of the regional setting (i.e., the United States and Republic of Korea). Furthermore, additional key factors (e.g., the distribution of fibrotic stages of patients) determining the cost-effectiveness of a hypothetical drug were identified in this study. We found that a hypothetical drug X, with an annual cost of $20,000, needs to have a minimum of 15% or higher initial response rates or have a minimum sustained durable response of 3% compared with non-treatment to be accepted as being cost-effective under a very conservative approach (i.e., a treatment waning effect of 50% per year in subsequent years).
Many new drugs for the treatment of MASH are in the pipeline, and it is common practice to use P-values to assess the effectiveness of treatment of a drug. However, the large number of patients enrolled in clinical trials could potentially amplify the gap between statistical significance and clinical significance (only 11.7% of the gap in treatment response for Resmetirom). Additionally, relying on short-term observation of histological outcomes may lead to overestimation of the effectiveness, where long-term durable response is crucial. Furthermore, the high cost of a new drug is a major barrier for the drug to be efficacious in real-world practice. This study emphasizes the importance of using a holistic framework to evaluate new pharmaceutical treatment options that not only consider P-values from clinical outcomes but also clinical significance prior to application of the drug in a real-world setting. Our results also highlight the critical need for minimum sustained durability to ensure treatment effectiveness, rather than focusing solely on short-term outcomes.
Several cost-effectiveness studies for new hypothetical treatments for MASLD have been conducted. Rustgi et al. [
27] found that the hypothetical treatment was not cost-effective compared to non-treatment (estimated ICER in the base case: >$200,000/QALY). Javanbakht et al. [
23] suggested that Resmetirom treatment would potentially be cost-effective ($53,929/QALY), and identified the maximum price for it to remain cost-effective. These studies primarily focused on estimating the maximum drug price that remained effective within defined ICER thresholds. In contrast, we aimed to reappraise the “clinically significant effect” of a hypothetical drug based on clinical trial results through the lens of cost-effectiveness by conducting comprehensive analyses. Furthermore, our conservative analytic framework including treatment effect waning and minimum sustained durability provides a balanced overview of the potential value of a new drug. Understanding the uncertainties associated with long-term treatment effectiveness could help policymakers to determine implementation of a new drug and decision risks [
28]. We obtained similar estimates to those reported previously when applying the closest conditions in input parameters, thresholds of cost-effectiveness, and analysis to our model. This further corroborates the robustness of our model.
Nevertheless, cautious interpretation of our findings is required, given the assumptions employed in the Markov model in this study. First, we assumed that the cost of a new drug X for MASLD would be comparable to that of antiviral medication for chronic hepatitis B for it to be widely used in clinical practice. Thresholds for clinical significance may vary depending on regional contexts and actual pricing of new MASLD drugs. However, this study was not aimed at evaluating the cost-effectiveness of a particular MASLD drug, so this assumption on the cost of drug X may not undermine the primary objective of this study. Second, assumptions regarding the treatment effects of the new hypothetical drug were set to be within the range of feasible values for initial treatment responses at year 1, minimum sustained durability of fibrosis regression, treatment effect waning, and the effects of the drug on the prevention of CVD. It is challenging to anticipate consistent annual treatment effectiveness, even with sustained pharmacological intervention. Our sensitivity analyses (
Fig. 2 and
Supplementary Table 8) clearly demonstrated that annual treatment effect waning percentage and minimum sustained durability in regression rate significantly affected the 1-year regression rate. We will be able to incorporate more accurate acceptable ranges in our model, when long-term data on the effectiveness of therapeutic agents become available in the future. Third, we also assumed that fibrosis regression rates during treatment would be the same regardless of the initial fibrosis stage of the patient (F2–F4). There are several new drugs that target MASH patients with cirrhosis in phase 2 clinical trials [
29]. However, determining different values for treatment effect based on initial stage of fibrosis is challenging due to the paucity of reliable data on this in the literature. Fourth, we assumed that pharmaceutical treatment with drug X could improve cardiovascular outcomes, underscoring the strong impact of extrahepatic clinical outcomes on determining the cost-effectiveness of MASLD treatment [
10]. However, the real world is more likely to be complex. Various CVDs may improve, worsen, or remain unaffected by MASLD pharmacotherapies, and the management of CVD itself may influence MASLD progression. To address this uncertainty, we incorporated a distinct CVD morbidity state into our Markov model. Additionally, we assumed a relative risk reduction of 0.8 for the effect of drug X on CVD based on the previous data. Nevertheless, one-way sensitivity analyses suggested that this assumption did not substantially impact the ICER. Consequently, the interaction between CVD and drug X treatment was unlikely to be a major determinant of cost-effectiveness when it is compared to the direct therapeutic effects of drug X and the baseline fibrosis distribution within our model. We acknowledge that our model represents a simplified framework, capturing only selected risks of MASLD and CVD. While it is a limitation that multiple metabolic disorders could not be simultaneously accounted for, incorporating numerous simultaneous effects would require extensive assumptions, which could compromise the model’s credibility and robustness. Clinical trials on various drugs with distinct mechanisms of action (e.g., selective thyroid hormone receptor-β [THR-β] agonists, glucagon-like peptide-1 [GLP-1] receptor agonists, and farnesoid X receptor [FXR] agonists) are currently underway [
8,
30,
31]. The impact of these drugs on extrahepatic outcomes (e.g., weight loss, control of glucose or lipids, prevention of major cardiovascular outcomes, and development of extrahepatic cancers) could differ depending on the mechanisms of action of the drug. Studies using sophisticated models that reflect the unique characteristics of new medications are needed. However, considering the higher risk of patients with MASLD who have coexisting cardiometabolic risk factors [
32,
33], treatment strategies that include these high-risk patients are likely to be more cost-effective. Fifth, we assumed that a ≥1-stage improvement in fibrosis necessarily leads to improved long-term outcomes. It is evident that fibrosis stage is the most important predictor of long-term outcomes. However, improvements in fibrosis do not fully offset risks conferred by either the present fibrotic burden or persistent cardiometabolic comorbidities. Further research with longer follow-up and validation of histologic surrogates would be needed to refine this assumption. Sixth, treatment discontinuation due to adverse events or reduced adherence in terms of long-term treatment may further diminish the effectiveness and cost-effectiveness of new MASLD drugs when they are compared to those of the trial-based estimates. Nevertheless, the cost-effectiveness remained positive under various assumptions on adherence as shown in our sensitivity analyses. Importantly, beyond these assumptions, our study also provides meaningful insight into treatment strategies: it may be more cost-effective to initiate treatment in patients with advanced fibrosis (particularly stages≥F3) compared with those with F2 fibrosis. Meanwhile, treatment initiation might be more appropriately considered in F2 patients who present with higher-risk metabolic features. This finding underscores the utility of our modeling approach, which enables stratification by fibrosis stage and comorbidity burden. It may provide practical insights for prioritizing MASLD treatment. We acknowledge that applying the same treatment effect to both F3 and F4, due to limited stage-specific data, may have led to an overestimation of cost-effectiveness in advanced fibrosis.
In conclusion, initial treatment response, long-term sustained durability of the drug, and patient distribution of baseline fibrosis stages are key factors that affect the cost-effectiveness of MASLD drugs. An initial fibrosis regression gap of at least 15% or higher and sustained durability with a gap of at least 3% compared to placebo over a 20-year time horizon are appropriate drug efficacy targets.
FOOTNOTES
-
Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
-
Authors’ contributions
Study concept and design: All authors. Acquisition of data: Huiyul Park, Ji-hyeon Park, Dae Won Jun and Hye-Lin Kim. Analysis and interpretation of data: All authors. Drafting of the manuscript: Ji-hyeon Park and Eileen L. Yoon. Critical revision and final approval of the manuscript: All authors. Statistical analysis: Ji-hyeon Park and Hye-Lin Kim. Study supervision: Dae Won Jun and Hye-Lin Kim
-
Acknowledgements
This work was supported by the research fund of Hanyang University (HY-202100000003681). This research was also partially supported by a grant from the KIST Institutional Program (Project No. 2E33111-24-042) and by a grant from the Patient-Centered Clinical Research Coordinating Center (PACEN) funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HC23C0058). This research was also supported by the “Regional Innovation System & Education (RISE)” through the Seoul RISE Center, funded by the Ministry of Education (MOE) and the Seoul Metropolitan Government (2025-RISE-01-027-01). The funding sources had no role in study design, implementation, data collection, analysis, and interpretation or the preparation, review, or approval of the manuscript.
-
Conflicts of Interest
The authors have no conflicts to disclose.
SUPPLEMENTARY MATERIAL
Supplementary material is available at Clinical and Molecular Hepatology website (
http://www.e-cmh.org).
Supplementary Figure 1.
Cumulative proportion of patients with metabolic dysfunction-associated steatotic liver disease (MASLD) who achieve fibrosis regression (hypothetical scenario). Cumulative proportion of patients who achieved improvement of fibrosis over time was shown based on the scenario described above.
Supplementary Figure 2.
Two-way sensitivity analysis by treatment effect waning and initial gap. ICER threshold range classification by color; red indicates an ICER over $100,000/QALY; green, an ICER between $50,000 and $100,000/QALY; blue, an ICER under $50,000/ QALY. ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-year; Tx, treatment.
Supplementary Figure 3.
(A) Treatment cost of $68.5 per day ($25,000) per year; (B) Treatment cost of $82.2 per day ($30,000) per year; (C) Treatment cost of $109.6 per day ($40,000 per year). Two-way sensitivity analysis by regression rate and cost. ICER threshold range classification by color; red indicates an ICER over $100,000/QALY; green, an ICER between $50,000 and $100,000/QALY; blue, an ICER under $50,000/QALY. ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-year.
Supplementary Figure 4.
(A) Treatment cost of $7.65 (approximately 10,000 KRW) per day ($2,791 per year); (B) Treatment cost of $9.56 (approximately 12,500 KRW) per day ($3,489 per year); (C) Treatment cost of $11.47 (approximately 15,000 KRW) per day ($4,187 per year). Scenario analysis for the setting of South Korea. ICER threshold range classification by color: red indicates an ICER over $23,322/QALY (=₩30,500,000/QALY); green an ICER between $15,293 (=₩20,000,000) and $23,322/QALY; blue an ICER under $15,293/QALY. 1 US$=₩1307.76 (2023 exchange rate). ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-year; KRW, Korean Won (₩).
Figure 1.Markov model and schematic flow of hypothetical scenario. (A) Markov model. (B) Schematic flow of fibrosis stage over time. A simplified scenario of changes in fibrosis stage that does not consider worsening of the fibrosis stage, development of serious complications (HCC or DC), or death. Only people in the non-treatment group progress in terms of fibrosis stage. HCC, hepatocellular carcinoma; DC, decompensated cirrhosis; CVD, cardiovascular disease.
Figure 2.Tornado diagram for one-way sensitivity analysis. *Value in parentheses for each label indicates following: (base case; lower bound-upper bound). Red (blue) colored bar indicates the result from the upper (lower) bound sensitivity analysis. If there is a single value, it indicates that we tested single alternative scenario. CVD, cardiovascular disease; RR, relative risk; ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-years.
Figure 3. Two-way sensitivity analysis by initial gap and minimum sustained durability in regression rates. ICER threshold range classification by color: red indicates an ICER over $100,000/QALY (not cost-effective); green indicates an ICER between $50,000 and $100,000/QALY (cost-effective); blue indicates an ICER under $50,000/QALY (highly cost-effective). ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-year.
Table 1.Characteristics of the target population and treatment efficacy for base case analysis
Table 1.
|
Parameter |
Value |
Range |
|
Starting age (yr) |
40 |
- |
|
Fibrosis distribution (%) |
|
|
|
F2 |
50 |
0–100 |
|
F3 |
30 |
0–100 |
|
F4 (liver cirrhosis) |
20 |
0–100 |
|
Pharmaceutical efficacy |
|
|
|
Treatment efficacy for hepatic fibrosis |
|
|
|
Progression rate (%) |
0 |
- |
|
Regression rate in gap* (%) |
|
10–50 |
|
At Year 1 |
25 |
|
|
At Year 2 |
12.5 |
|
|
At Year 3 |
6.3 |
|
|
At Year 4 |
3.1 |
|
|
At Year 5+ (final) |
2 |
2–10 |
|
Treatment effect waning |
|
|
|
Reduction ratio compared to the previous year |
0.5 |
0.5–1.0 |
|
Treatment efficacy for CVD prevention |
|
|
|
Relative ratio of CVD risks (treatment vs. non treatment) [9,10] |
0.8 |
0.5–1.0 |
Table 2.Transition matrix based on the natural history of MASLD
Table 2.
|
To health state (tx+1) |
|
F0 |
F1 |
F2 |
F3 |
F4 |
DC |
HCC |
CVD (F0) |
CVD (F1) |
CVD (F2) |
CVD (F3) |
CVD (F4) |
LRM |
CVM |
|
From health state (tx) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
F0 |
97.35 |
1.55 |
- |
- |
- |
- |
0.0004 |
1.1 |
- |
- |
- |
- |
- |
- |
|
F1 |
7.1 |
81.09 |
10.7 |
- |
- |
- |
0.011 |
- |
1.1 |
- |
- |
- |
- |
- |
|
F2 |
- |
7.9 |
79.42 |
11.54 |
- |
- |
0.04 |
- |
- |
1.1 |
- |
- |
- |
- |
|
F3 |
- |
- |
7.9 |
81.27 |
9.1 |
- |
0.34 |
- |
- |
- |
1.39 |
- |
- |
- |
|
F4 |
- |
- |
- |
- |
85.5 |
6.59 |
3.78 |
- |
- |
- |
- |
1.8 |
2.33 |
- |
|
DC |
- |
- |
- |
- |
- |
76.22 |
3.78 |
- |
- |
- |
- |
- |
20 |
- |
|
HCC |
- |
- |
- |
- |
- |
- |
86.95 |
- |
- |
- |
- |
- |
13.05 |
- |
|
CVD |
- |
- |
- |
- |
- |
- |
0.0004 |
89.74 |
1.55 |
- |
- |
- |
- |
8.71 |
|
CVD |
- |
- |
- |
- |
- |
- |
0.011 |
7.1 |
73.48 |
10.7 |
- |
- |
- |
8.71 |
|
CVD (F2) |
- |
- |
- |
- |
- |
- |
0.04 |
- |
7.9 |
71.81 |
11.54 |
- |
- |
8.71 |
|
CVD (F3) |
- |
- |
- |
- |
- |
- |
0.34 |
- |
- |
7.9 |
73.95 |
9.1 |
- |
8.71 |
|
CVD (F4) |
- |
- |
- |
- |
- |
6.59 |
3.78 |
- |
- |
- |
0 |
78.59 |
2.33 |
8.71 |
|
LRM |
- |
- |
- |
- |
- |
- |
- |
- |
- |
- |
- |
- |
100 |
- |
|
CVM |
- |
- |
- |
- |
- |
- |
- |
- |
- |
- |
- |
- |
- |
100 |
Abbreviations
cardiovascular-related mortality
incremental cost-effectiveness ratio
intensive lifestyle intervention
metabolic-associated steatohepatitis
metabolic dysfunction-associated steatotic liver disease
nonalcoholic fatty liver disease
quality-adjusted life-year
REFERENCES
- 1. Henry L, Paik J, Younossi ZM. Review article: the epidemiologic burden of non-alcoholic fatty liver disease across the world. Aliment Pharmacol Ther 2022;56:942-956.
- 2. Kim HY, Yu JH, Chon YE, Kim SU, Kim MN, Han JW, et al. Prevalence of clinically significant liver fibrosis in the general population: a systematic review and meta-analysis. Clin Mol Hepatol 2024;30(Suppl):S199-S213.
- 3. Cusi K, Isaacs S, Barb D, Basu R, Caprio S, Garvey WT, et al. American Association of Clinical Endocrinology clinical practice guideline for the diagnosis and management of nonalcoholic fatty liver disease in primary care and endocrinology clinical settings: co-sponsored by the American Association for the Study of Liver Diseases (AASLD). Endocr Pract 2022;28:528-562.
- 4. European Association for the Study of the Liver; European Association for the Study of Diabetes; European Association for the Study of Obesity. EASL-EASD-EASO Clinical Practice Guidelines on the management of metabolic dysfunction-associated steatotic liver disease (MASLD): executive summary. Diabetologia 2024;67:2375-2392.
- 5. Vilar-Gomez E, Martinez-Perez Y, Calzadilla-Bertot L, Torres-Gonzalez A, Gra-Oramas B, Gonzalez-Fabian L, et al. Weight loss through lifestyle modification significantly reduces features of nonalcoholic steatohepatitis. Gastroenterology 2015;149:367-378e5.
- 6. Hernandez Roman J, Patel S. Why do lifestyle recommendations fail in most patients with nonalcoholic fatty liver disease? Gastroenterol Clin North Am 2020;49:95-104.
- 7. Angulo P, Kleiner DE, Dam-Larsen S, Adams LA, Bjornsson ES, Charatcharoenwitthaya P, et al. Liver fibrosis, but no other histologic features, is associated with long-term outcomes of patients with nonalcoholic fatty liver disease. Gastroenterology 2015;149:389-97.e10.
- 8. Harrison SA, Bedossa P, Guy CD, Schattenberg JM, Loomba R, Taub R, et al.; MAESTRO-NASH Investigators. A phase 3, randomized, controlled trial of resmetirom in NASH with liver fibrosis. N Engl J Med 2024;390:497-509.
- 9. Choi S, Kim K, Kim SM, Lee G, Jeong SM, Park SY, et al. Association of obesity or weight change with coronary heart disease among young adults in South Korea. JAMA Intern Med 2018;178:1060-1068.
- 10. Park H, Yoon EL, Kim M, Kwon SH, Kim D, Cheung R, et al. Cost-effectiveness study of FIB-4 followed by transient elastography screening strategy for advanced hepatic fibrosis in a NAFLD at-risk population. Liver Int 2024;44:944-954.
- 11. Period life table, 2020, as used in the 2023 trustees report. Social Security Administration web site, <https://www.ssa.gov/oact/STATS/table4c6.html#fn1>. Accessed 25 Sep 2023.
- 12. Cancer facts & figures 2021. American Cancer Society web site, <https://www.cancer.org/research/cancer-facts-statistics/all-cancer-facts-figures/cancer-facts-figures-2021.html>. Accessed 25 Sep 2023.
- 13. Age-adjusted death rates for selected causes of death, by sex, race, and Hispanic origin: United States, selected years 1950-2019. Centers for Disease Control and Prevention web site, <https://www.cdc.gov/nchs/hus/data-finder.htm?year=2020-2021&table=Table%20SlctMort>. Accessed 25 Sep 2023.
- 14. Cardiovascular disease death rate by age group, United States, 2020. Our World in Data web site, <https://ourworldindata.org/grapher/cardiovascular-disease-death-rate-age-group-who-mdb>. Accessed 25 Sep 2023.
- 15. SEER relative survival rates by time since diagnosis, SEER Cancer Statistics. 2000-2019. National Cancer Institute web site, <https://seer.cancer.gov/statistics-network/>. Accessed 25 Sep 2023.
- 16. Dulai PS, Singh S, Patel J, Soni M, Prokop LJ, Younossi Z, et al. Increased risk of mortality by fibrosis stage in nonalcoholic fatty liver disease: Systematic review and meta-analysis. Hepatology 2017;65:1557-1565.
- 17. Estes C, Anstee QM, Arias-Loste MT, Bantel H, Bellentani S, Caballeria J, et al. Modeling NAFLD disease burden in China, France, Germany, Italy, Japan, Spain, United Kingdom, and United States for the period 2016-2030. J Hepatol 2018;69:896-904.
- 18. Estes C, Razavi H, Loomba R, Younossi Z, Sanyal AJ. Modeling the epidemic of nonalcoholic fatty liver disease demonstrates an exponential increase in burden of disease. Hepatology 2018;67:123-133.
- 19. Im JH, Lee KS, Kim KY, Hong NS, Lee SW, Bae HJ. Follow-up study on mortality in Korean stroke patients. J Korean Med Assoc 2011;54:1199-1208.
- 20. Noureddin M, Jones C, Alkhouri N, Gomez EV, Dieterich DT, Rinella ME; NASHNET. Screening for nonalcoholic fatty liver disease in persons with type 2 diabetes in the United States is cost-effective: a comprehensive cost-utility analysis. Gastroenterology 2020;159:1985-1987.e4.
- 21. Orci LA, Sanduzzi-Zamparelli M, Caballol B, Sapena V, Colucci N, Torres F, et al. Incidence of hepatocellular carcinoma in patients with nonalcoholic fatty liver disease: a systematic review, meta-analysis, and meta-regression. Clin Gastroenterol Hepatol 2022;20:283-292.e10.
- 22. Younossi ZM, Tampi RP, Nader F, Younossi IM, Cable R, Srishord M, et al. Hypothetical treatment of patients with nonalcoholic steatohepatitis: potential impact on important clinical outcomes. Liver Int 2020;40:308-318.
- 23. Javanbakht M, Fishman J, Moloney E, Rydqvist P, Ansaripour A. Early cost-effectiveness and price threshold analyses of resmetirom: an investigational treatment for management of nonalcoholic steatohepatitis. Pharmacoecon Open 2023;7:93-110.
- 24. Tice JA, Suh K, Fahim SM, Carlson JJ, Richardson M, Herce-Hagiwara B, et al. Resmetirom and obeticholic acid for Non-Alcoholic Steatohepatitis (NASH); draft evidence report. Institute for Clinical and Economic Review web site, <https://icer.org/assessment/nonalcoholic-steatohepatitis-2023/>. Accessed 10 Jun 2024.
- 25. Neumann PJ, Cohen JT, Weinstein MC. Updating cost-effectiveness--the curious resilience of the $50,000-per-QALY threshold. N Engl J Med 2014;371:796-797.
- 26. Pennisi G, Enea M, Romero-Gomez M, Viganò M, Bugianesi E, Wong VW, et al. Liver-related and extrahepatic events in patients with non-alcoholic fatty liver disease: a retrospective competing risks analysis. Aliment Pharmacol Ther 2022;55:604-615.
- 27. Rustgi VK, Duff SB, Elsaid MI. Cost-effectiveness and potential value of pharmaceutical treatment of nonalcoholic fatty liver disease. J Med Econ 2022;25:347-355.
- 28. Trigg LA, Melendez-Torres GJ, Abdelsabour A, Lee D. Treatment effect waning assumptions: a review of national institute of health and care excellence technology appraisals. Value Health 2024;27:1003-1011.
- 29. Noureddin M, Rinella ME, Chalasani NP, Neff GW, Lucas KJ, Rodriguez ME, et al. Efruxifermin in compensated liver cirrhosis caused by MASH. N Engl J Med 2025;392:2413-2424.
- 30. Kumar J, Hasan M, Mohsin S, Alzaher MH, Nagar T, Jamil A, et al. Assessing the efficacy of farnesoid X receptor agonists in the management of metabolic dysfunction-associated steatotic liver disease: a systematic review and meta-analysis: efficacy of Farnesoid X receptor agonists in metabolic dysfunction-associated steatotic liver disease: systematic review and meta-analysis. Clin Res Hepatol Gastroenterol 2025;49:102530.
- 31. Sanyal AJ, Bedossa P, Fraessdorf M, Neff GW, Lawitz E, Bugianesi E, et al. 1404-0043 Trial Investigators. A phase 2 randomized trial of survodutide in MASH and fibrosis. N Engl J Med 2024;391:311-319.
- 32. Sorrentino P, Terracciano L, D’Angelo S, Ferbo U, Bracigliano A, Vecchione R. Predicting fibrosis worsening in obese patients with NASH through parenchymal fibronectin, HOMA-IR, and hypertension. Am J Gastroenterol 2010;105:336-344.
- 33. Younossi ZM, Tampi RP, Racila A, Qiu Y, Burns L, Younossi I, et al. Economic and clinical burden of nonalcoholic steatohepatitis in patients with type 2 diabetes in the U.S. diabetes care 2020;43:283-289.
, Dae Won Jun1,2
