Editorial: Beyond Clinical Trials—Real-World Data Suggest Usefulness of GLP-1 RAs in MASLD Treatment

Abstract

Metabolic dysfunction–associated steatotic liver disease (MASLD) is the most common liver disease globally and represents the hepatic manifestation of metabolic syndrome [1]. Since MASLD pathogenesis is strongly linked to insulin resistance, it is particularly prevalent in subjects with type 2 diabetes [2]. In the latter, glucagon-like peptide-1 receptor agonists (GLP-1 RAs) and sodium-glucose cotransporter 2 inhibitors (SGLT2i) constitute widely used treatment options that do not only improve glycaemic control, but also reduce cardiovascular events and overall mortality [3-5]. Despite their proven benefits in subjects with type 2 diabetes, their impact on MASLD remains to be convincingly demonstrated, which might be due to the difficulty to perform randomised controlled trials (RCTs) in hepatology. RCTs typically have a treatment duration of no more than a few years, which may not be sufficient given the slow pace of liver disease progression/regression [6-8]. Moreover, they rely on liver biopsy that captures only a very small portion of the liver and is therefore prone to sampling bias [7]. For example, Newsome et al. randomised 320 patients to the GLP-1 RA semaglutide versus placebo and demonstrated an improvement in the histological non-alcoholic steatohepatitis, but no impact on liver fibrosis that is often required by the regulators [9]. Despite the lack of liver-related approval, SGLT2i and GLP-1 RAs are routinely prescribed to MASLD subjects and hepatologists therefore need to advise the patients about their hepatic effects.

In this situation, real-world data (RWD) constitute the best available option. Despite their inherent limitations such as reliance on retrospective electronic health record (EHR) data and residual confounding that is due to their observational nature, they also have multiple advantages such as a much larger amount of follow-up years that enables them to use harder endpoints (Figure 1) [10]. In particular, the current study from Kuo et al. analysed EHR data collected outside of clinical trials to evaluate the association between a new prescription of SGLT2is or GLP-1 RAs and development of future major adverse liver outcomes, the latter being a composite of bleeding oesophageal varices, hepatic encephalopathy, clinically significant ascites, hepatocellular carcinoma and need for liver transplantation [10]. While studies such as the one by Newsome often rely on ~500 follow-up years, Kuo et al. focused on patients with MASLD and diabetes captured nearly 40,000 follow-up years per group, and the follow-up was twice longer than the one from Newsome et al. (Figure 1) [9, 10]. To minimise the risk of bias, they performed a propensity score matching and several supplementary analyses that do not only demonstrate the robustness of their findings but also provide useful insights into the metabolic efficacies of both drug classes such as their impact on serum lipid levels or HbA1c reduction (Figure 1). While the study provides useful evidence for everyday patient counselling, the relatively short follow-up of 2.6 years suggests that it primarily captures the disease progression in subjects with advanced liver fibrosis. Further analyses are therefore needed to compare the usefulness of SGLT2is and GLP-1 RAs in patients with lower fibrosis stages and to compare their efficacy in subjects without type 2 diabetes.

FIGURE 1

Open in figure viewerPowerPoint

Key findings and limitations of the study. Kuo et al. assessed real-world data (RWD) to compare the effects of glucagon-like peptide-1 receptor agonists (GLP-1 RAs) and sodium-glucose cotransporter-2 inhibitors in patients with metabolic dysfunction-associated steatotic liver disease and type 2 diabetes mellitus. The advantages of the RWD approach over randomised controlled trials are depicted. Novel findings are highlighted in gold, confirmatory results and study limitations are shown in silver and red, respectively. MALO, major adverse liver outcomes. Figure was created with BioRender.com.