Aerobic exercise interventions lead to MASH resolution in clinical trials: Pooled analysis using the MASH resolution index

INTRODUCTION Exercise training is a core management component of metabolic dysfunction–associated steatotic liver disease (MASLD) with broad hepatic and extrahepatic benefits.1,2 Despite the benefits of hepatic steatosis, even in the absence of clinically meaningful weight loss, the effects of exercise training on histological features of metabolic dysfunction–associated steatohepatitis (MASH) are uncertain due to the relative lack of randomized controlled trials.1,2 This may, in part, be due to the need for repeat liver biopsies to accurately assess these outcomes. Since liver fibrosis is a key prognostic measure for risk of all-cause and liver-related mortality,3 examining the impact of exercise training on validated noninvasive histological surrogates is imperative. There are several validated noninvasive tools to evaluate liver disease severity; however, most were developed and validated as screening and risk assessment tools, rather than to quantify intervention effects. We recently demonstrated that exercise training led to clinically meaningful improvements in imaging-based surrogates of liver histological change (≥30% relative reduction in MRI-measured liver fat).4 To address the ongoing need for a reliable, noninvasive assessment, with adequate sensitivity to detect the histological resolution of MASH, Loomba et al5 developed and validated the MASH resolution index (MASH-RI). This index outperformed changes in serum ALT and MRI-derived liver fat (MRI-PDFF), and provides an alternative endpoint to determine MASH resolution post-intervention, removing the need for paired liver biopsy in intervention studies. MASH-RI is comprised of baseline/change in MRI-PDFF, baseline/change in ALT, and baseline AST. We examined the efficacy of exercise training on the MASH-RI compared with standard care control. METHODS Our post hoc analyses pooled 3 primary randomized controlled trials,6–8 as detailed elsewhere.4 Briefly, moderate-to-high intensity aerobic exercise training was compared with standard care control for 12–20 weeks. Liver fat was measured by magnetic resonance spectroscopy6,7 or MRI-PDFF,8 and ALT and AST were assessed following an overnight fast at baseline and post-training. The MASH-RI was calculated as 0.520−0.003×baseline ALT (U/L)−0.024×(latest ALT (U/L)−baseline ALT (U/L))−0.048×baseline MRI-PDFF−2.571×([latest MRI-PDFF−baseline MRI-PDFF]/baseline MRI-PDFF)−0.039×baseline AST (U/L). A MASH-RI ≥−0.67 ruled in MASH resolution, and a MASH-RI of ≤−2.67 ruled out a resolution. Statistical analyses All analyses were conducted using Stata (v17, StataCorp). Between-group differences in MASH-RI were examined via independent t test (continuous data) and χ2 test (proportion achieving ≥−0.67). Odds of achieving MASH-RI ≥−0.67 were examined via separate penalized maximum likelihood logistic regression. Separate multiple linear regression examined the role of potential confounders (intervention duration, body mass index, waist circumference, and V̇O2peak) on MASH-RI (continuous). Associations between MASH-RI and change in weight and waist circumference were examined via the Pearson correlation coefficient. An α of 0.05 was adopted for all analyses. RESULTS Participant characteristics are reported in Supplemental Table S1, https://links.lww.com/HC9/B943. Participants were (mean [SD]) 52.3 (11.2) years, 33% female, body mass index 32.6 (4.2) kg/m2, and liver fat (absolute) 24.0 (15.7)%. All other pooled effects for exercise and control were reported previously.4 Pooled MASH-RI was −1.6 (1.4) among the total sample (Australia: −2.4 [1.1], UK: −1.4 [1.6], USA: −1.8 [1.1]). There was a between-group difference for MASH-RI (mean difference [95% CI]: −0.66 [−1.30, −0.01], p=0.046), with mean exercise scores closer to the MASH resolution threshold (−1.31 [−1.75, −0.86], n=45] than control (−1.97 [−2.42, −1.51]; Figure 1). Adjustment for intervention duration did not meaningfully change the strength, direction, or statistical significance of the effect estimate (adjusted β [95% CI]: −0.66 [−1.31, 0.01], p=0.047). Adjustment for baseline body mass index, waist circumference, or V̇O2peak did not meaningfully change the strength or direction of the effect estimate, yet these differences (adjusted β [95% CI]) were no longer statistically significant (body mass index: −0.59 [−1.24, 0.06], p=0.076; waist circumference: −0.55 [−1.19, 0.09], p=0.090; V̇O2peak: −0.60 [−1.25, 0.05], p=0.071).FIGURE 1: MASH resolution index for control (gray circles) versus exercise (black circles). Solid circles represent individual data points. Solid lines represent group means. Short dashes represent the MASH resolution “rule in” threshold (−0.67). Long dashes represent the MASH resolution “rule out” threshold (−2.67); *p<0.05. Abbreviations: AU, arbitrary unit; MASH, metabolic dysfunction–associated steatohepatitis.The proportion of patients achieving the MASH resolution threshold (≥−0.67) favored exercise (n=16 [36%] vs. n=6 [19%]) yet was not significant (p=0.130). The odds of achieving resolution were ~2-fold greater (OR [95% CI]: 2.2 [0.8, 6.3]) for exercise compared with control, yet it was not significant (p=0.143). There was a weak negative association between the MASH-RI and change in weight (r=−0.425, p<0.001) and change in waist circumference (r=−0.354, p=0.002). DISCUSSION Our data support a potential histological benefit of aerobic exercise training using the newly proposed MASH-RI. Short-term aerobic exercise training led to better MASH-RI than control, yet on average, did not surpass the MASH resolution threshold. While not statistically significant compared with control, over one-third of exercising participants met the threshold for MASH resolution. This represented a 2-fold greater likelihood of achieving MASH resolution than control. These data suggest that longer-term exercise training would likely achieve MASH resolution for most patients. Due to the limitations of a threshold-based score to represent longitudinal change, we are unable to infer the clinical significance and implications of a −0.66 unit difference in MASH-RI between exercise and control. Notably, this difference would be considered clinically meaningful based on the established distribution-based method threshold of 0.5 SD of baseline values. Given the time course for fibrosis regression and MASH resolution, approaching the threshold for resolution is likely meaningful, particularly in the context of the relatively short durations of our exercise interventions. Resolution may take longer in those with more severe liver disease at baseline, and differences in the MASH-RI below the cut-off threshold may be clinically meaningful yet undersell the utility of the intervention. For example, participants in the Australian cohort had, on average, higher body mass index, waist, and steatosis with lower cardiorespiratory fitness and MASH-RI score (−2.4) than the USA and UK cohorts, and considerable variability was observed across both exercise and control groups. Notably, exploratory analyses for those who achieved the resolution threshold showed no indication that intervention duration or these participant characteristics explained the odds of resolution. Further examination of predictors of resolution is warranted. The potential benefit of exercise training for histological improvement is supported by trials using surrogate markers of liver histology4,9. We showed a ~5-fold higher likelihood of achieving a clinically meaningful imaging threshold (≥30% relative reduction in liver fat), associated with a 2-point improvement in NAFLD activity score, with exercise compared with control. This is in line with a pilot study (n=16) of moderate-to-vigorous aerobic exercise training in which 58% and 67% of participants regressed one fibrosis or ballooning stage, respectively, after 12 weeks, despite no significant change in NAFLD activity score.10 Our data show that 36% of exercising participants met the threshold for MASH resolution (vs. 19% in control). Notably, improvements in histology or histology surrogates with exercise training have occurred in the absence of meaningful weight loss,4,9,10 which is important given management guidelines promote substantial weight loss of at least 7%–10% for histological improvement in MASH. Therefore, aerobic exercise training appears to be a potent weight-neutral management strategy for MASLD, and strategies targeting long-term adherence should be prioritized. The newly validated MASH-RI advances the previous screening biomarkers’ ability to inform intervention efficacy. However, further understanding of the use of the MASH-RI on a continuous scale would enable a better understanding and applicability of its use in this context. While short-term exercise interventions appear to provide a greater resolution of adverse histological change (MASH) than control, more research with larger sample sizes is needed to ascertain the dose–response relationship (including exercise training frequency, intensity, and duration) between exercise and MASH resolution and the influence of individual characteristics on achieving resolution.