Glucose-dependent insulinotropic polypeptide, usually shortened to GIP, is a hormone the gut releases after a meal. It signals the pancreas to release insulin when blood sugar is rising. For years, researchers have been curious whether giving extra GIP to people with type 2 diabetes could meaningfully improve blood sugar control, especially since some newer drug combinations pair GIP-like activity with another gut hormone called GLP-1.
A recent randomized, placebo-controlled trial published in The Lancet Diabetes and Endocrinology set out to answer that question directly. The researchers ran a six-week continuous subcutaneous infusion of native GIP, meaning the naturally occurring form of the peptide rather than a synthetic analog. They tested GIP alone and in combination with a GLP-1 receptor agonist in people with type 2 diabetes. The primary question was simple: would sustained exposure to higher GIP levels improve average blood glucose readings?
The short answer from the trial was no, at least not at the level the researchers were aiming for. But the design and the findings raise interesting questions about how GIP actually functions in the human body over extended periods, and what that might mean for future research into incretin-based approaches.
Trial design and participant characteristics
The trial was conducted at a single center in Hellerup, Denmark. Researchers enrolled 61 adults between the ages of 18 and 74, all of whom had been diagnosed with type 2 diabetes for at least six months and had relatively stable treatment regimens at the time of enrollment. Participants had a median age of 64 years and a mean body mass index of about 31.6 kilograms per square meter. All participants self-reported as White.
The study used a four-arm parallel design. Participants were randomly assigned to one of four groups in roughly equal numbers: placebo plus placebo, placebo plus GIP, a GLP-1 receptor agonist plus placebo, and a GLP-1 receptor agonist plus GIP. Both participants and investigators were blinded to treatment assignment throughout the study.
The protocol began with an eight-week run-in period in which participants either received the GLP-1 receptor agonist or a matching placebo. After that run-in, participants continued their assigned weekly injection and simultaneously started a continuous subcutaneous infusion of either GIP or saline placebo, delivered at a rate of 16 picomoles per kilogram per minute for six weeks. The primary outcome the researchers tracked was the change in 14-day average glucose concentration measured by continuous glucose monitoring from baseline to the end of the full 14-week treatment period.
GIP blood levels during the infusion
To confirm that the infusion was actually raising GIP in the bloodstream, the researchers measured fasting concentrations of both intact, biologically active GIP and total GIP at the end of treatment.
In the placebo-only group, fasting intact GIP averaged 7 picomoles per liter. In the group receiving GIP infusion without the GLP-1 drug, it rose to an average of 45 picomoles per liter. In the group receiving both the GLP-1 receptor agonist and GIP, it reached an average of 85 picomoles per liter. Total GIP concentrations followed a similar pattern, climbing from 14 picomoles per liter in the double-placebo group to 375 picomoles per liter in the GIP-only group and 527 picomoles per liter in the combination group.
This confirmed that the infusion was pharmacologically active and was successfully elevating circulating GIP. The question then was whether those elevated levels translated into measurable glucose improvements.
Primary glucose outcome
The trial had pre-specified a target improvement of 1.50 millimoles per liter in 14-day mean sensor glucose. Neither GIP arm reached that threshold.
When GIP was added to placebo, the estimated effect on mean glucose change was 0.80 millimoles per liter compared to double placebo. The 97.5 percent confidence interval for that estimate ranged from negative 0.18 to positive 1.80, and the p-value was 0.13. That means the result did not reach statistical significance.
When GIP was added on top of the GLP-1 receptor agonist, the estimated effect was even smaller, just 0.05 millimoles per liter compared to the GLP-1 drug plus placebo. The confidence interval ranged from negative 0.85 to positive 0.95, and the p-value was 1.00. In practical terms, adding GIP on top of ongoing GLP-1 therapy made essentially no measurable difference in continuous glucose monitoring readings.
The researchers acknowledged that dropouts complicated interpretation in the GIP-alone arm. Ten participants out of 61 discontinued the study overall, which represented about 16 percent of the enrolled group. Because the dropout rate affected group sizes, the authors noted they could not draw firm conclusions about GIP as a standalone add-on in the absence of a GLP-1 receptor agonist.
Safety and tolerability observations
The most commonly reported adverse event across all groups was injection site reactions, which affected 22 participants, representing about 36 percent of the total enrolled population. This was likely related to the continuous subcutaneous infusion apparatus required to deliver GIP around the clock.
Gastrointestinal side effects were more frequent in the arms that included the GLP-1 receptor agonist. In the double-placebo group, 60 percent of participants reported some gastrointestinal event. That proportion climbed to 69 percent in the GIP-only group, 73 percent in the GLP-1 drug plus placebo group, and 80 percent in the combination group. The pattern suggested the GLP-1 component was the primary driver of gastrointestinal effects, which is consistent with what earlier literature has shown for that class of compounds.
No safety signals unique to GIP were flagged as notable in the abstract. The trial funding came from Novo Nordisk.
What the results suggest about GIP biology
The findings add a layer of nuance to the ongoing scientific conversation about GIP. In animal models and some short-duration human experiments, GIP has shown effects on insulin secretion and fat tissue metabolism. But the literature also documents a well-known phenomenon called GIP resistance in people with type 2 diabetes, where the insulinotropic response to GIP is blunted compared to people without the condition.
This trial was designed partly to test whether overcoming that resistance through continuous, sustained elevation of GIP levels might restore a glycemic benefit. Based on the primary outcome, six weeks of native GIP infusion at the dose tested did not produce that effect, at least not at the magnitude the researchers set as clinically meaningful.
It is worth noting that the trial studied the native form of GIP, which is rapidly broken down in the body by an enzyme called dipeptidyl peptidase-4. Some researchers argue that modified, longer-acting GIP analogs might behave differently from native GIP because they maintain receptor engagement for extended periods. The results here do not necessarily predict what a stabilized GIP analog would do, a distinction the broader field continues to examine.
Early data across the incretin research space also point at the possibility that GIP's most relevant effects may involve tissues beyond the pancreas, including bone, the brain, and fat depots. Whether continuous GIP exposure influences those outcomes was not the focus of this trial, but it represents an area where future research may look more closely.
Limitations and what comes next
Single-center trials inherently limit the generalizability of findings. All participants in this study self-reported as White, which means the results may not reflect how GIP infusion would behave in more diverse populations.
The six-week infusion window, while longer than many acute GIP studies, may still be insufficient to capture delayed or gradual physiological adaptations. Additionally, the dropout rate of about 16 percent introduced uncertainty into the GIP-alone comparison arm, meaning the data from that subgroup should be interpreted with caution.
The trial also worked within a dose range that researchers believed to be physiologically relevant, but it remains an open question whether different dosing protocols or more targeted delivery approaches would shift the outcome. The literature suggests that the interaction between GIP signaling and GLP-1 receptor pathways is complex, and that the combination may need to be studied over longer timeframes or with different patient populations to fully characterize any additive effects.
For researchers and curious readers following the incretin peptide space, this trial is a useful data point. It provides one of the most rigorous and sustained tests of native GIP to date, and its null result on the primary glucose endpoint is itself informative, narrowing the scope of claims that can be made about this hormone in a well-controlled setting.
