• NONALCOHOLIC STEATOHEPATITIS
• LIVER BIOPSY
• META-ANALYSIS

We have read with interest the excellent review by Trauner and Fuchs on the novel therapeutic approaches that are currently developed for treating non-alcoholic steatohepatitis (NASH).1 Although there are no licensed pharmacotherapies for NASH, long-acting fibroblast growth factor-21 (FGF-21) analogues are being evaluated to treat NASH because FGF-21 is a pleotropic liver-derived hormone regulating lipid metabolism, insulin sensitivity and energy homeostasis, all mechanisms closely implicated in NASH development.2

To quantify the magnitude of the possible hepatoprotective effects of FGF-21 analogues, we systematically searched three electronic databases from the inception date to 1 September 2023 to identify phase-2 randomised controlled trials (RCTs) examining the efficacy of FGF-21 analogues on the US Food and Drug Administration defined histological endpoints for conditional approval of new drugs for NASH, that is, NASH resolution without worsening of fibrosis or ≥1 stage fibrosis improvement without worsening of NASH. We also calculated the fragility index (FI) to establish the robustness of the trial’s data.3 More details of the systematic review are reported in online supplemental material.

We included five phase-2 placebo-controlled RCTs4–8 involving 602 obese adults with biopsy-confirmed NASH and stages F1–F4 fibrosis, most of whom had diabetes (446 randomly assigned to FGF-21 analogues and 156 assigned to placebo). Online supplemental figure 1 shows the results of the literature research and study selection. The main characteristics of the eligible RCTs are summarised in table 1. Three RCTs6–8 were in phase 2b, whereas two4 5 were in phase 2a with histological endpoints available only for a subset of individuals. Figure 1 shows the treatment effects of FGF-21 analogues on resolution of NASH without worsening of fibrosis (figure 1A, n=3 RCTs6–8) or ≥1 stage fibrosis improvement without worsening of NASH (figure 1B, n=5 RCTs4–8) compared with placebo. Treatment with once-weekly subcutaneous FGF-21 analogues for 16–48 weeks resulted in a significantly higher percentage of patients with NASH resolution with no worsening of fibrosis, or ≥1 stage fibrosis improvement without worsening of NASH than placebo (especially using common-effects models). However, it should be noted that the FI for both histological endpoints was small, suggesting that the findings are weak. Specifically, the FI for the ≥1 stage fibrosis improvement was 3 (ie, only three participants from the active-comparator arm should be reassigned to the placebo arm to change the result from significant to non-significant), while the FI for the resolution of NASH was 8.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Forest plots and pooled estimates of the effects of long-acting FGF-21 analogues on the histological resolution of NASH with no worsening of fibrosis ((A), n=3 RCTs6–8) or the improvement in fibrosis of one stage or more without worsening of NASH ((B), n=5 RCTs4–8) compared with placebo. The pooled and individual effect sizes for all RCTs are expressed as OR and 95% CIs, as estimated by both common-effects and random-effects models. For the active-comparator group (ie, participants using FGF-21 analogues), we calculated the individual effect sizes of each RCT by combining the treatment effect data of variable dosages of FGF-21 analogues into a single group (to avoid including individuals in the placebo group several times in the analysis). With regard to this, it is important to note that before combining the data of the active-comparator group for each RCT, we have previously tested in a multivariate meta-regression analysis that neither the different type nor the different dosage of FGF-21 analogue used in each RCT were significant predictors of the observed effects of this drug class on the histological liver endpoints. FGF-21, fibroblast growth factor-21; NASH, non-alcoholic steatohepatitis; RCT, non-alcoholic steatohepatitis.

The FI measures the robustness (or fragility) of the results from a clinical trial using dichotomous outcomes. The FI represents the minimum number of participants whose status needs to change from an ‘event’ to a ‘non-event’ (or vice versa) so that the results switch from statistically significant to non-significant. Whether few participants are needed to hamper the significance of a result, the strength of evidence for affirming the superiority of a specific treatment over a placebo might be questionable. When we looked at the efficacy of pioglitazone or semaglutide in achieving resolution of NASH or ≥1 stage fibrosis improvement without worsening of NASH, we found that the FI of pioglitazone for NASH resolution with no worsening of fibrosis was 11, while the FI for ≥1 stage fibrosis improvement was 6.9 In a phase-2b placebo-controlled RCT testing semaglutide,10 we found that the FI of semaglutide for the resolution of NASH without worsening of fibrosis was 8.

In conclusion, the results of this meta-analysis suggest that FGF-21 analogues are a promising treatment option for adults with biopsy-confirmed NASH and fibrosis. However, based on the FI of the trials’ data, uncertainty remains about the robustness and clinical benefit of FGF-21 analogues. Future large phase-3 RCTs with long-term follow-up are needed to have more robust findings.