Predicting response to GLP-1–based therapy in MASLD/MASH: an evidence-tiered, endpoint-specific framework built on host genotype and gut microbial features
- 1Department of Infectious Diseases, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University
Abstract
Incretin-based therapy is now central to treating metabolic dysfunction-associated steatohepatitis (MASH), but the benefit is uneven. In the ESSENCE trial, semaglutide resolved steatohepatitis without worsening fibrosis in 62.9% of patients; placebo did so in 34.3%. No one can yet tell in advance which patient will respond. Weight loss explains most of the difference between trials, but it cannot be prescribed. This review asks whether biology measured before the first dose can predict who will benefit. Hepatocytes carry almost no GLP-1 receptor, so the liver benefit arrives indirectly, through weight loss and through gut–liver signaling. Both routes can be influenced from outside the liver, which points to two measurable domains: host genotype, fixed and assayable at baseline, and gut microbial features. For a dietary prebiotic, baseline microbial features have already forecast response, with an externally validated area under the curve of 0.74 (95% CI 0.66–0.81). Nothing equivalent has been tested for a GLP-1 receptor agonist. The genotype evidence comes from dietary weight loss rather than from a drug. Both pillars therefore rest on interventions other than incretin therapy, and the framework grades them accordingly. We assemble this evidence into a framework with two rules. Findings are graded by how mature the evidence behind them is. Liver and cardiovascular benefits are predicted separately, because the two do not track together. The output is a probability rather than a label. One conclusion is already usable: a predicted poor liver response is not, on its own, a reason to withhold or stop a GLP-1 receptor agonist.
Related
- Predicting atropine treatment response heterogeneity in pediatric myopia using explainable machine learning: an on-treatment longitudinal monitoring framework
- Pharmacometabolomic signatures as individualised biochemical response evidence: towards a precision jurisprudence framework for forensic toxicology and criminal responsibility
- Dose- and endpoint-dependent effects of perioperative esketamine in children: evidence from a targeted review
- LLM-assisted evidence audit of late-stage cancer incidence as a screening trial endpoint
- Microbial traits have minimal power for predicting death from bloodstream infection
BibTeX
@article{zhu2026predicting,
title = {{Predicting response to GLP-1–based therapy in MASLD/MASH: an evidence-tiered, endpoint-specific framework built on host genotype and gut microbial features}},
author = {Guojian Zhu and Yucao Huang and Yun Huang and Wenlong Yang},
journal = {Frontiers in Endocrinology},
year = {2026},
month = oct,
doi = {10.3389/fendo.2026.1962068},
url = {https://www.frontiersin.org/articles/10.3389/fendo.2026.1962068/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Predicting response to GLP-1–based therapy in MASLD/MASH: an evidence-tiered, endpoint-specific framework built on host genotype and gut microbial features
AU - Guojian Zhu
AU - Yucao Huang
AU - Yun Huang
AU - Wenlong Yang
PY - 2026
DA - 2026/10/27
JO - Frontiers in Endocrinology
DO - 10.3389/fendo.2026.1962068
UR - https://www.frontiersin.org/articles/10.3389/fendo.2026.1962068/full
AB - Incretin-based therapy is now central to treating metabolic dysfunction-associated steatohepatitis (MASH), but the benefit is uneven. In the ESSENCE trial, semaglutide resolved steatohepatitis without worsening fibrosis in 62.9% of patients; placebo did so in 34.3%. No one can yet tell in advance which patient will respond. Weight loss explains most of the difference between trials, but it cannot be prescribed. This review asks whether biology measured before the first dose can predict who will benefit. Hepatocytes carry almost no GLP-1 receptor, so the liver benefit arrives indirectly, through weight loss and through gut–liver signaling. Both routes can be influenced from outside the liver, which points to two measurable domains: host genotype, fixed and assayable at baseline, and gut microbial features. For a dietary prebiotic, baseline microbial features have already forecast response, with an externally validated area under the curve of 0.74 (95% CI 0.66–0.81). Nothing equivalent has been tested for a GLP-1 receptor agonist. The genotype evidence comes from dietary weight loss rather than from a drug. Both pillars therefore rest on interventions other than incretin therapy, and the framework grades them accordingly. We assemble this evidence into a framework with two rules. Findings are graded by how mature the evidence behind them is. Liver and cardiovascular benefits are predicted separately, because the two do not track together. The output is a probability rather than a label. One conclusion is already usable: a predicted poor liver response is not, on its own, a reason to withhold or stop a GLP-1 receptor agonist.
SN - 1664-2392
ER -