Tirzepatide and the GIP paradox: why adding an "obesogenic" hormone worked

The GIP paradox

Glucose-dependent insulinotropic polypeptide (GIP) is secreted in the duodenum mainly after fat absorption. In 2002 a Kyoto group reported in Nature Medicine that wild-type mice on a high-fat diet hypersecreted GIP and accumulated visceral and subcutaneous fat, while mice lacking the GIP receptor (Gipr−/−) were protected from both obesity and insulin resistance; they used fat as their preferred energy substrate. The authors concluded that GIP links overnutrition to obesity and was a target for anti-obesity drugs¹.

The logical reading was to block the receptor, not stimulate it. Which is why it was surprising that industry bet on the opposite: a peptide that activates both the GIP and GLP-1 receptors, tirzepatide.

Antagonist or agonist: a debate still open

Amgen investigators developed antagonist antibodies against the GIP receptor. In diet-induced obese mice the antibody protected against weight gain and improved several metabolic parameters; in obese non-human primates weight loss was more pronounced, and it was enhanced when combined with GLP-1 agonists². The same Amgen strategy gave rise to maridebart cafraglutide, an antibody conjugate that combines GIP receptor antagonism with GLP-1 receptor agonism¹⁵, which the blog already covered in After semaglutide.

In 2020 the same group proposed a reconciliation: in adipocytes, chronic GIP receptor agonism desensitizes the receptor so that it ends up functioning like an antagonist³. In other words, in the laboratory both approaches can converge on a similar effect on adipose tissue.

Lilly investigators argue for another reading: GIP would be the primary incretin; in the central nervous system its receptor attenuates nausea and suppresses appetite; and in adipose tissue it would protect against ectopic fat⁷. In mice, a GIP receptor agonist reduced nausea-like behavior induced by a PYY analog without affecting hypophagia, and GIP and PYY receptors are expressed by the same neurons in the area postrema⁶. Both groups are employees of the companies developing each approach³⁷, and the debate is unresolved.

An imbalanced dual agonist

Tirzepatide (LY3298176) is a fatty-acid-modified peptide for once-weekly dosing. In its phase 1, with 142 study subjects exposed to tirzepatide, dulaglutide or placebo, the most frequent adverse effects were gastrointestinal, dose-dependent and mild to moderate in severity⁵.

Its pharmacology is not symmetric. By calculating occupancy of each receptor at clinically effective doses, Willard and colleagues showed greater engagement of the GIP receptor than of the GLP-1 receptor. At the GIP receptor it mimics native GIP; at the GLP-1 receptor it is biased toward cAMP generation over β-arrestin recruitment, with less receptor internalization. In islets, that bias was associated with greater insulin secretion⁴.

SURPASS-2: versus semaglutide in type 2 diabetes

In SURPASS-2, a 40-week open-label trial in 1,879 patients with type 2 diabetes (mean baseline HbA1c 8.28%), tirzepatide 5, 10 and 15 mg lowered HbA1c by 2.01, 2.24 and 2.30 percentage points, versus 1.86 with semaglutide 1 mg. All three doses were noninferior and superior. The weight difference versus semaglutide was −1.9 kg, −3.6 kg and −5.5 kg⁸. The SURMOUNT-1 results versus placebo are summarized on the product page and are not repeated here.

Obesity: head-to-head and withdrawal

SURMOUNT-5 compared the two drugs directly in 751 adults with obesity without diabetes, using each one's maximum tolerated dose (tirzepatide 10 or 15 mg; semaglutide 1.7 or 2.4 mg) for 72 weeks. Weight change was −20.2% with tirzepatide and −13.7% with semaglutide (P < 0.001), and waist change −18.4 cm versus −13.0 cm. The trial was open-label and funded by Lilly⁹.

SURMOUNT-4 measured what happens on stopping. After a 36-week open-label tirzepatide lead-in (mean loss 20.9%), 670 participants were randomized to continue or switch to placebo. From week 36 to 88, weight changed −5.5% with tirzepatide and +14.0% with placebo; 89.5% of those who continued kept at least 80% of their loss, versus 16.6% on placebo¹⁰. The blog already reviewed body composition and rebound in GLP-1s: muscle, rebound, eyes and pancreas.

SURPASS-CVOT: the first outcomes trial

SURPASS-CVOT randomized 13,299 patients with type 2 diabetes and atherosclerotic cardiovascular disease to tirzepatide (up to 15 mg) or dulaglutide 1.5 mg, a comparator with proven cardiovascular benefit. The composite of cardiovascular death, myocardial infarction or stroke occurred in 12.2% versus 13.1% (HR 0.92; 95.3% CI 0.83 to 1.01): noninferior (P = 0.003), not superior (P = 0.09). There were more gastrointestinal events with tirzepatide¹¹.

A prespecified exploratory analysis estimated the effect versus an imputed placebo derived from REWIND: tirzepatide was associated with fewer major adverse cardiovascular events (HR 0.72) and lower all-cause mortality (HR 0.61)¹². It is an indirect comparison, with several authors employed by the manufacturer, and does not replace a placebo-controlled trial. The other hard-outcome areas (sleep apnea, heart failure, liver) are in GLP-1s beyond weight.

Honest limits

• Gastrointestinal tolerability. In SURPASS-2, nausea affected 17 to 22% with tirzepatide and 18% with semaglutide; diarrhea 13 to 16% versus 12%. Serious adverse events were 5 to 7% versus 3%⁸. The antiemetic GIP hypothesis comes from animal models⁶.
• Discontinuation. In a cohort of 125,474 US adults starting liraglutide, semaglutide or tirzepatide, 64.8% of those without diabetes stopped within the first year; moderate or severe gastrointestinal events were associated with more discontinuation¹³.
• Weight regain. Stopping was associated with substantial regain of lost weight¹⁰.
• Cost. A US lifetime model estimated $197,023 per quality-adjusted life-year for tirzepatide, above usual thresholds; reaching $100,000 would require an additional 30.5% discount on the net price¹⁴.
• Funding. SURPASS-2, SURMOUNT-5 and SURPASS-CVOT were funded by the manufacturer, and SURMOUNT-4 had authors employed by the manufacturer⁸⁹¹⁰¹¹.