Could a blood sample help explain why one person loses more weight than another after switching to tirzepatide? A new Japanese study offers a possible clue, but it has not produced a test that can guide that decision. The distinction is between finding an association in a group and reliably predicting what will happen to an individual.
The paper published online on 31 October 2026 examined 90 adults with type 2 diabetes at one hospital. Higher expression of a gene in certain blood cells was associated with greater weight reduction over six months after a treatment switch. The result is worth following because it asks about differences between people, rather than simply reporting another average weight change.
What was measured in the blood
The cells were monocytes, a type of immune cell. The researcher measured messenger RNA for the glucose-dependent insulinotropic polypeptide receptor, abbreviated GIPR. This is a measure of gene expression, not a count of functioning receptors throughout the body. Receptor protein levels and activity in organs such as the liver were not established by this measurement.
Everyone in the main analysis had already used dulaglutide or semaglutide for at least six months. They then received tirzepatide, starting at 2.5 mg weekly and increasing to 5 mg after a month. Those doses describe the study’s treatment course, not instructions for readers. Our tirzepatide profile provides the broader medicine context.
The group’s mean age was 57.5 years and mean BMI was 27.6. These were people with established type 2 diabetes, not a sample representing everyone seeking obesity treatment. In the main results table, average weight was 74.9 kg before the switch and 72.3 kg six months later. That before-and-after change cannot isolate the effect of switching from other influences.
An association is not a prediction tool
The relationship between higher GIPR expression and greater percentage weight reduction remained after adjustment for several baseline characteristics, including age, diabetes duration, BMI, HbA1c and the previous medicine. Adjustment makes the finding more informative, but it cannot eliminate every difference between people or establish causation.
Crucially, the paper’s discussion says that predictive performance, clinically useful cutoffs and independent validation were not assessed. There is no tested threshold here separating likely responders from nonresponders. Nor does the study show that choosing treatment using this measurement improves outcomes.
Selection also matters. The main analysis included 90 of 169 screened patients and excluded, among others, people whose other medicines changed or whose tirzepatide dose increased beyond the selected regimen. The paper reports additional sensitivity analyses, but those cannot make this selected group representative of every person who starts or stops treatment. The main text and tables were available for this review; the separate supplemental tables were not inspected.
Why a switching trial answers a different question
There was no parallel group continuing its previous medicine in the new study. Looking back at weight changes during earlier GLP-1 treatment does not replace that comparison: the time periods and treatment circumstances differ.
For context, the separate SURPASS-SWITCH randomized trial, published in 2025, assigned 282 adults with inadequately controlled type 2 diabetes either to switch to tirzepatide or to increase dulaglutide. Its abstract reports greater average glucose and weight reductions with switching. It was an open-label trial funded by Eli Lilly.
The trial registry specifies 40 weeks and escalation toward different maximum tolerated doses. Those results should not be ranked against this six-month, 5 mg observational cohort. More fundamentally, a randomized comparison can address average treatment effects without validating a blood marker for individual selection. Our guide to randomized and real-world evidence explains that difference.
The liver finding needs its own caution
Higher GIPR expression was also associated with a larger decline in FIB-4, an index calculated from age, liver enzymes and platelet count. A lower index does not by itself demonstrate that liver scarring has reversed. This study did not include direct fibrosis assessment by imaging or liver biopsy, and changes in the blood-test components could contribute to the score change.
The laboratory experiments offer another clue, with a different limit. They exposed monocytes to native GIP, not tirzepatide, and found reduced inflammatory responses under the tested conditions. They did not use receptor blockade or genetic inhibition to establish that a specific receptor mechanism caused the clinical weight changes.
What would make the marker useful?
A convincing next step would need more than another statistically significant association: reproducible measurements, prospective testing in independent groups, useful prediction beyond ordinary clinical information, and evidence that using the result improves decisions. The current study does not establish those requirements.
For now, this is a lead for research into treatment-response differences. It does not establish a clinical role for ordering this assay or changing medicines because of its result. Readers can use our trial-reading guide to separate the population, comparison and outcome, and our appointment questions to keep treatment discussions grounded in their own care needs.



