Tirzepatide cuts liver fat despite PNPLA3

A post hoc analysis of the SURPASS-3 MRI substudy, published in Liver International in September 2026, asks a question patients with fatty liver increasingly raise. If you carry the gene variant most strongly tied to fat buildup in the liver, does tirzepatide still clear that fat the way it does for everyone else? The authors grouped trial participants by whether they carried the PNPLA3 I148M allele, a common genetic variant linked to steatotic liver disease (fat accumulation in liver cells), and compared how liver fat changed over 52 weeks of treatment^[1]. Post hoc means the question was posed after the trial data were collected, which shapes how much weight the answer can carry.

The variant behind the question

PNPLA3 is a gene that encodes a protein involved in fat handling inside liver cells. An allele is one version of a gene, and your genotype is the pair of versions you carry, one from each parent. The I148M allele swaps a single building block in the PNPLA3 protein, and carriers are labelled CG or GG while non-carriers are CC. The variant was first tied to fatty liver in a 2008 genome-wide association scan in Nature Genetics, which found the G allele strongly associated with higher liver fat and with liver inflammation, and that people with two copies had more than twice the liver fat of non-carriers^[2]. It is the most replicated genetic risk factor in fatty liver disease, so it is the natural candidate to test against the liver fat effect seen with tirzepatide.

What SURPASS-3 MRI had already shown

The parent substudy, published in The Lancet Diabetes & Endocrinology in 2022, randomised 296 adults with type 2 diabetes and a fatty liver index of at least 60 to tirzepatide 5 mg, 10 mg or 15 mg once weekly, or to once-daily insulin degludec. The trial administered tirzepatide subcutaneously for 52 weeks. Liver fat was measured by MRI-PDFF, an MRI-based estimate of the percentage of liver tissue made up of fat. From a mean baseline liver fat content of 15.71%, the pooled tirzepatide 10 mg and 15 mg groups lost 8.09 percentage points of liver fat versus 3.38 points on insulin degludec, an estimated treatment difference of 4.71 points in tirzepatide's favour^[3].

What the new analysis adds

The new analysis split those same participants by genotype and re-ran the comparisons. The authors report that participants on tirzepatide had significant reductions in liver fat content, body weight and HbA1c (a blood marker reflecting average glucose over roughly three months), and generally greater improvement in blood lipids and liver enzymes, whether or not they carried the I148M allele. The findings were consistent across genotype subgroups, and the authors conclude that carrying the variant did not seem to affect tirzepatide-induced improvement in liver fat or in several cardiometabolic measures^[1]. In plain terms: the drug's liver fat effect looked similar in carriers and non-carriers.

That is a reassuring signal for the roughly one in four people of European ancestry, and an even higher share of people of Hispanic ancestry, who carry at least one copy of the allele. The 2008 scan found the variant was most common in Hispanic participants, the group with the highest fatty liver susceptibility in that study^[2].

How much weight this can carry

Several limits apply. The analysis is post hoc and explicitly exploratory, so it generates a hypothesis rather than confirming one. The parent substudy enrolled 296 people across four treatment arms; split again by genotype, each cell gets small, and small cells make it easy to miss a real difference or to see one that is not there. Randomisation was not stratified by genotype, so the subgroups are not protected by the trial's original design. And the population was insulin-naive adults with type 2 diabetes already on metformin with or without an SGLT2 inhibitor, so the result may not travel to people without diabetes or with more advanced liver disease^[3]. The honest reading is that nothing here suggests PNPLA3 carriers respond worse, not that equivalence is proven.

Where tirzepatide stands on liver disease

No regulator has approved tirzepatide for a liver indication; its approvals rest on type 2 diabetes and obesity. The class does have a foothold in liver disease through semaglutide: the US label for Wegovy includes treatment of metabolic dysfunction-associated steatohepatitis (MASH, the inflammatory stage of fatty liver disease) with moderate to advanced fibrosis in adults, an accelerated approval based on improvement of MASH and fibrosis that still requires confirmation of clinical benefit^[4]. The SURPASS-3 data speak to liver fat, an earlier and softer marker than inflammation or scarring, so they sit upstream of any label question. Our tirzepatide evidence notes track the indications as they stand.

Practical context

For readers, the takeaway is narrow but useful. In the largest MRI substudy of tirzepatide in type 2 diabetes, the most common fatty liver gene variant did not appear to stand between a patient and the drug's liver fat reduction^[1]. Genotype testing is not part of any prescribing decision today, and this analysis does not change that. What it does is remove one plausible reason to expect a weaker response, while leaving the harder questions, fibrosis, inflammation and long-term outcomes, to trials designed around them.

This article is educational and is not medical advice. Decisions about diabetes or liver treatment, including whether any medicine is appropriate, belong with a qualified healthcare provider who knows your history.