Pumpkin-Seed Peptide Supplement: What a Small Human Trial Actually Found

A small human trial has finally tested a pumpkin-seed peptide supplement. The headline result is not the main result.

A pumpkin-seed peptide supplement produced a striking number in a small subgroup: people with higher starting body fat who took it lost about 2.3 kilograms more than placebo over 12 weeks. But that subgroup was carved out after the trial was underway, contained just 13 people, and was not the study’s overall result. Across the full group, the supplement did not significantly change weight, BMI, or measured fat mass compared with placebo. The distinction is the story.

The study, published September 30, 2026, is notable for a reason beyond its marketing potential: it puts a standardized food-derived peptide preparation into a randomized human trial. It also shows how quickly an exploratory signal can be turned into a promise it has not earned. Read the paper.

What researchers actually tested

The 12-week study randomized 50 adults with overweight to 600 milligrams a day of IRK-19, a pumpkin-seed-derived preparation, or an equal daily dose of wheat dextrin placebo. Participants and outcome assessors were blinded. Everyone also received dietary counseling aimed at modest weight reduction. Forty-nine participants completed the intervention, and 48 had usable week-12 DXA body-composition measurements.

That design is stronger than a testimonial or an uncontrolled before-and-after report. Randomization and blinding help reduce bias; DXA provides a more direct body-composition measure than a bathroom scale. But 50 participants and 12 weeks still make this an early, small study. It cannot tell us whether an effect persists, applies to a broader population, or is large enough to matter in everyday care.

The result for the whole group was essentially flat

At week 12, the adjusted difference between IRK-19 and placebo in body weight was minus 0.05 kilograms, with a 95% confidence interval from minus 1.03 to plus 0.92 kilograms (p=0.915). The BMI difference was minus 0.09 kg/m² (p=0.649). Total fat mass was 0.30 kilograms higher in the supplement group, on average, with a confidence interval that included no difference (p=0.390). Trunk fat and visceral fat measures also showed no statistically clear separation.

In plain language: this trial did not show that the preparation reduced weight or fat in the overall population studied. That is not proof that the product has no effect; a small trial can miss a real, modest effect. But it is the main comparison the reader should see before any subgroup result.

Where the promising number came from

After looking at the data, the researchers examined a subgroup whose baseline DXA fat mass was at least 28.4 kilograms. That cutoff was the 75th percentile within this particular sample, not a previously validated clinical threshold. It identified 13 people: six assigned to IRK-19 and seven to placebo.

Within that group, the adjusted week-12 difference favored IRK-19 by 2.30 kilograms for body weight (95% CI, 4.08 to 0.51 kilograms lower; p=0.018) and 0.75 kg/m² for BMI (95% CI, 1.48 to 0.02 lower; p=0.045). The estimated difference in total fat mass was 1.60 kilograms lower, but its confidence interval ran from 3.23 kilograms lower to 0.04 kilograms higher (p=0.055). That last interval includes no effect.

The apparent pattern is interesting. It is not confirmation that people with higher body fat respond better. The subgroup was exploratory and chosen after the fact; many outcomes were examined without adjustment for multiple comparisons, which increases the chance of a nominally positive result arising by chance. More importantly, the formal tests asking whether treatment response actually changed with baseline fat mass did not reach statistical significance for weight, BMI, or fat mass. The authors themselves call for a new trial with the groups specified in advance.

‘Pumpkin seed peptide’ does not mean one proven active molecule

IRK-19 was a complex preparation, not a purified peptide drug. The study reported 66.2% crude protein in the material; that measurement is not the same as total peptide content. Six selected peptide markers together accounted for 2.79% by weight. The researchers did not establish which, if any, component caused a clinical effect, and the trial tested the whole preparation rather than individual sequences.

Earlier laboratory work on pumpkin-seed proteins produced hydrolysates with enzyme-inhibition activity in test-tube assays. Those experiments provide a reason to investigate the material, not evidence that eating or taking it changes weight or treats a disease in people. Digestion, absorption, dose, processing, and the exact mixture all matter. Results from one standardized batch should not automatically be applied to pumpkin seeds, pumpkin seed oil, or supplements with different ingredients.

The safety data deserve the same attention as the weight data

The paper also reports laboratory differences that deserve follow-up. At week 12, the adjusted fasting-glucose difference was 9 mg/dL higher in the IRK-19 group than placebo (95% CI, 1.93 to 16.07; p=0.014). The adjusted triglyceride difference was 70.16 mg/dL higher (95% CI, 9.90 to 130.41; p=0.024), and two participants receiving IRK-19 had triglyceride readings at or above 500 mg/dL. These are small-sample findings across multiple laboratory comparisons, so they do not prove the supplement caused harm. They do make a blanket ‘no safety concerns’ reading inappropriate.

The investigators reported no recorded clinical adverse events or serious adverse events during the 12 weeks. That is useful, but it does not erase abnormal laboratory measurements. A study this small and short is not equipped to establish long-term safety or detect uncommon harms. The paper says further assessment of metabolic laboratory abnormalities is warranted.

Who funded the study matters

The preparation’s developer, Greenyn Biotechnology, supplied IRK-19 and funded the study. Two authors were company employees. The paper says the sponsor participated in study conception and design, interpretation, manuscript preparation support, and the decision to submit. The reported statistical analyses and cross-checking were performed by another author, but sponsor involvement remains part of the evidence readers should weigh. Funding does not automatically invalidate results; transparency helps readers judge where independent replication is needed.

What this means for peptide science—and for consumers

Food-derived bioactive peptides are a broad research area. A 2026 review surveys the path from laboratory studies to human trials, where translation depends on the preparation, dose, and endpoint being tested. This pumpkin-seed study is a useful early human data point, not a clinical endorsement of a product category. The careful takeaway is: a standardized preparation has been tested in a small randomized trial; the overall weight and fat results were not significant; an unplanned subgroup produced a hypothesis worth testing; and metabolic findings need attention.

For a stronger answer, the next study should enroll enough people to test a prespecified baseline-adiposity hypothesis, use a clear primary endpoint, correct for multiple comparisons, report adherence and dietary changes in detail, and monitor glucose, triglycerides, liver and kidney measures over longer follow-up. Independent investigators should replicate the work. Until then, the 2.3-kilogram subgroup estimate is a lead for research, not a result to bank on.

In the United States, FDA rules for dietary-supplement structure/function claims require truthful, substantiated statements and do not allow a supplement to claim to diagnose, treat, cure, or prevent disease; FDA does not pre-approve those claims. A trial result—especially an exploratory one—does not by itself establish that a marketed product is effective or FDA-approved.

Sources

Primary randomized trial: Lin YC et al. International Journal of Molecular Sciences (2026), 27(19):8813. https://doi.org/10.3390/ijms27198813

PubMed record for the primary trial (PMID 42884056): https://pubmed.ncbi.nlm.nih.gov/42884056/

Lin CT et al. Protein identification and potential bioactive peptides from pumpkin (Cucurbita maxima) seeds. Food Science & Nutrition (2024). https://pmc.ncbi.nlm.nih.gov/articles/PMC11317681/

Santos-Sánchez G et al. Bioactive peptides: from preclinical to clinical studies. Current Opinion in Clinical Nutrition and Metabolic Care (2026). https://pubmed.ncbi.nlm.nih.gov/41076731/

ClinicalTrials.gov study record NCT05542160: https://clinicaltrials.gov/study/NCT05542160

U.S. FDA: Structure/Function Claims: https://www.fda.gov/food/nutrition-food-labeling-and-critical-foods/structurefunction-claims

Evidence grade: C — One small, 12-week randomized trial with a null overall body-composition result; the positive finding was post hoc, based on 13 participants, uncorrected for multiple comparisons, and not supported by significant treatment-by-adiposity interactions. Industry funding and laboratory signals warrant independent replication and longer safety follow-up.

Medical disclaimer: This article is for educational purposes only and is not medical advice. It does not recommend taking IRK-19, pumpkin-seed peptide supplements, or any peptide product. Supplements are not substitutes for individualized care. Discuss weight-management products and abnormal glucose or triglyceride results with a qualified healthcare professional. Oria provides research education and access to licensed medical professionals; individual care decisions should be made with a clinician.