Two numbers circulate constantly as the result of the head-to-head trial between tirzepatide and semaglutide. The pair usually quoted is 20.9% and 14.9%. Neither number comes from that trial. Both come from separate studies, run years apart, for different lengths. No randomized comparison exists between them.
The actual head-to-head figures are 20.2% and 13.7%.
That gap between what the evidence says and what gets repeated is the reason for this page.
Tirzepatide is a synthetic 39-amino-acid peptide that activates two gut hormone receptors at once, the GIP receptor and the GLP-1 receptor. A fatty acid chain lets it bind to albumin in the blood. That binding extends how long the compound lasts and allows weekly dosing in clinical trials. It is the first approved compound in its class built against two incretin receptors rather than one.
The short answer on efficacy is that the human evidence is unusually strong and graded high throughout this page. Multiple large randomized trials, several of them double-blind, show substantial and sustained weight reduction. The same programs show large improvements in blood sugar control.
The limitation to carry through everything below is narrower than it sounds. All of this evidence describes one specific pharmaceutical formulation, manufactured and tested to a defined standard. It does not automatically describe anything else that shares the name.
This page grades each major trial, compares the two compounds only where comparable evidence exists, and says plainly where the evidence runs out.
What tirzepatide actually is
Tirzepatide is a laboratory-made peptide, 39 amino acids long. It was built on the sequence of one natural gut hormone, with activity against a second engineered into it. A C20 fatty diacid chain is attached to the molecule. That chain binds albumin, a protein in blood, which slows clearance. The result is a half-life long enough for weekly administration in trials.
The two receptors in its name have long titles. GIP stands for glucose-dependent insulinotropic polypeptide. GLP-1 is short for glucagon-like peptide-1. Both are incretin hormones, meaning the gut releases them after eating. Once released, they signal the pancreas and the brain.
Compounds that activate one receptor are called agonists. Tirzepatide activates two, so it is a dual agonist. That single design decision explains most of what follows.
Its regulatory position has widened four times. Current labeling covers type 2 diabetes in adults and in children aged 10 and older. It covers chronic weight management in adults meeting defined criteria. It covers moderate-to-severe obstructive sleep apnea in adults with obesity. And since August 28, 2026, it covers reduction of major cardiovascular event risk in adults with type 2 diabetes at high risk.
One important distinction needs drawing early on. Everything in this article describes the pharmaceutical product used in registered clinical trials. Research-use materials sold under the same compound name are a separate category. A later section explains exactly what the trial evidence can and cannot tell you about them.
How activating two receptors changes what happens
The honest version of this mechanism is more interesting than the simplified one. One half of it is well characterized. The other half remains poorly understood.
What the GLP-1 side does
Activating the GLP-1 receptor produces four effects that are established pharmacology rather than inference.
The pancreas releases more insulin, but only when blood glucose is already elevated. That dependence on glucose level matters clinically. It is why the compound carries a lower risk of driving blood sugar too low. Older agents push insulin release regardless of glucose level.
Less glucagon is also released, which is the hormone that tells the liver to produce sugar. The stomach empties more slowly, so food stays put longer. Receptors in brain regions governing appetite are activated, which reduces hunger signaling.
Those four actions are shared with every compound in the GLP-1 class, semaglutide included. On their own they do not distinguish tirzepatide from anything else.
What the GIP side adds
Here the evidence gets more specific and more frequently misreported.
A detailed pharmacology study characterized the receptor binding directly. It used albumin-free assays to isolate the molecule's intrinsic behavior. According to PubMed, that work found tirzepatide binds the GIP receptor with affinity comparable to native GIP. At the GLP-1 receptor, it binds roughly fivefold more weakly than native GLP-1 does (Willard et al., 2020).
This is worth stating carefully, because the opposite claim is common online. Tirzepatide does not bind the GIP receptor more tightly than the body's own GIP does. The two affinities are about equal. The imbalance in "imbalanced dual agonist" describes the relationship between the two receptors, not a contest with the natural hormone.
The same study found a second feature. At the GLP-1 receptor, tirzepatide favors one signaling route, cyclic AMP generation, over another, beta-arrestin recruitment. It also drives less receptor internalization than native GLP-1 does. That pattern is called biased signaling.
The authors proposed that the imbalance toward the GIP receptor, combined with that biased signaling, may account for the compound's efficacy. The word "may" is theirs, and it belongs in any accurate summary.
What the two together produce
Clinically, the combination outperforms GLP-1 activation alone. SURMOUNT-5 established that directly, and SURPASS-2 established it for blood sugar control. Both of those outcomes rest on high-strength evidence.
Attributing that advantage specifically to the GIP component is a different claim, and it is not established. The section below explains why that matters.
Why nobody can yet say how much of the effect comes from GIP
This is the clearest unresolved question in tirzepatide pharmacology, and almost no article covering the compound says so.
No human trial isolates GIP receptor activity at therapeutic exposure. There is no arm in which people receive only the GIP half of what tirzepatide does. Without that, the contribution of GIP activation to any individual outcome has to be inferred rather than measured directly.
The inference usually runs through the head-to-head comparison. Tirzepatide produced more weight reduction than semaglutide, semaglutide targets only GLP-1, so the extra receptor must explain the difference.
That reasoning is reasonable and insufficient. Three things differ between these two molecules at once. The number of receptors targeted differs. Binding affinity and signaling behavior at the shared GLP-1 receptor differ, since tirzepatide binds it more weakly and signals through it differently. The dose ranges are entirely separate scales, so no comparison runs at matched exposure.
Any one of those could contribute to the observed gap. A single randomized comparison cannot separate them.
Preclinical work offers several competing hypotheses. Animal and cell studies point toward effects on fat tissue handling of insulin, and toward a second route into brain appetite circuits. Those findings all sit at the preclinical level. They describe what happens in mice and in cultured cells, not what happens in people.
Settling the question would take trials designed to isolate receptor contributions in humans. No trials of that design currently exist.
Everything in this section is mechanistic interpretation, which is the weakest grade applied anywhere on this page. It is a scientifically reasonable explanation, not proof of a clinical outcome.
Every major tirzepatide trial, graded
The grading below separates what a study design can support from what people often claim it supports. High-strength evidence means well-conducted randomized human trials, read in light of their limitations. Moderate means credible human evidence with meaningful constraints. Limited means small, indirect, or inconsistently measured. Mechanistic interpretation means a plausible explanation that is not itself proof.
The final column is the one most often missing elsewhere.
- Trial: SURMOUNT-1 — Design: Phase 3 randomized, double-blind, placebo-controlled — Participants: 2,539 adults with obesity, no diabetes — Duration: 72 weeks — Primary endpoint: Percent weight change; 5% or more reduction — Main result: −20.9% at highest dose vs −3.1% placebo — Grade: High — Does not establish: Any comparison against an active drug
- Trial: SURMOUNT-2 — Design: Phase 3 randomized, double-blind, placebo-controlled — Participants: 938 adults with obesity and type 2 diabetes — Duration: 72 weeks — Primary endpoint: Percent weight change; 5% or more reduction — Main result: −14.7% at highest dose vs −3.2% placebo — Grade: High — Does not establish: That results in people without diabetes transfer to those with it
- Trial: SURMOUNT-4 — Design: Phase 3 randomized withdrawal — Participants: 670 randomized after lead-in — Duration: 36-week lead-in plus 52 weeks — Primary endpoint: Percent weight change, week 36 to 88 — Main result: −5.5% continuing vs +14.0% on placebo — Grade: High — Does not establish: What happens after planned, supported discontinuation
- Trial: SURMOUNT-5 — Design: Phase 3b randomized, open-label, active comparator — Participants: 751 adults with obesity, no diabetes — Duration: 72 weeks — Primary endpoint: Percent weight change — Main result: −20.2% vs −13.7% semaglutide — Grade: High, lowered for open-label design — Does not establish: Equivalence at matched maximum dose; any cardiovascular claim
- Trial: SURPASS-1 — Design: Phase 3 randomized, double-blind, placebo-controlled — Participants: 478 adults with type 2 diabetes — Duration: 40 weeks — Primary endpoint: HbA1c change — Main result: −1.87 to −2.07 points vs +0.04 placebo — Grade: High — Does not establish: Comparative efficacy against any active agent
- Trial: SURPASS-2 — Design: Phase 3 randomized, open-label, vs semaglutide 1 mg — Participants: 1,879 adults with type 2 diabetes — Duration: 40 weeks — Primary endpoint: HbA1c change — Main result: −2.01, −2.24, −2.30 vs −1.86 — Grade: High, lowered for open-label design — Does not establish: Comparison against semaglutide at obesity doses
- Trial: SURMOUNT-OSA — Design: Two Phase 3 randomized, double-blind, placebo-controlled trials — Participants: Adults with moderate-to-severe sleep apnea and obesity — Duration: 52 weeks — Primary endpoint: Apnea-hypopnea index change — Main result: Treatment differences of −20.0 and −23.8 events per hour — Grade: High — Does not establish: Any mortality or cardiovascular benefit in sleep apnea
- Trial: SURPASS-CVOT — Design: Phase 3 randomized, double-blind, active comparator — Participants: 13,299 with type 2 diabetes and established cardiovascular disease — Duration: About 4 years median — Primary endpoint: Cardiovascular death, heart attack, or stroke — Main result: 12.2% vs 13.1%, hazard ratio 0.92 — Grade: High — Does not establish: Superiority over the comparator; any placebo-relative magnitude
- Trial: SUMMIT — Design: Phase 3 randomized, placebo-controlled — Participants: 731 with obesity-related heart failure with preserved ejection fraction — Duration: 52 weeks — Primary endpoint: Cardiovascular death or worsening heart failure; symptom score — Main result: Reduced events; symptom improvement — Grade: High for that specific group — Does not establish: Benefit in heart failure without obesity
- Trial: SYNERGY-NASH — Design: Phase 2 — Participants: Adults with liver disease — Duration: Not stated — Primary endpoint: Disease resolution — Main result: 44% to 62% vs 10% placebo — Grade: Moderate, Phase 2 only — Does not establish: Any approved indication for liver disease
- Trial: Lean mass evidence — Design: Ten heterogeneous studies, mostly impedance-based — Participants: Varied — Duration: Varied — Primary endpoint: Body composition — Main result: 10% to 30% of lost weight was lean mass — Grade: Limited — Does not establish: Skeletal muscle loss; functional consequences; any head-to-head split
Three rows carry more interpretive weight than the table can hold.
SURMOUNT-5 was run as an open-label trial. Participants and investigators knew which compound they received. For outcomes shaped partly by behavior and expectation, including eating, that is a real limitation. Participants also received their maximum tolerated dose rather than a fixed one. In practice that meant tirzepatide at 10 mg or 15 mg, and semaglutide at 1.7 mg or 2.4 mg. Not everyone in either group reached the top of their range.
SURPASS-2 compared against semaglutide 1 mg. That is a diabetes dose, well below the 2.4 mg used for weight management. The trial cannot be read as a comparison at matched maximal dosing. Published HbA1c figures for it also differ depending on which statistical estimand is reported. The treatment-regimen estimand gives −2.01, −2.24 and −2.30. The efficacy estimand gives slightly larger reductions. Sources quoting one set rarely say which they used.
SURPASS-CVOT compared against an active drug, not placebo. The comparator already had established cardiovascular benefit. Tirzepatide met its non-inferiority target at P=0.003 and missed superiority at P=0.09. Those are different findings and should never be merged.
One fact applies to every trial above. All were funded by the manufacturer.
How tirzepatide and semaglutide compare on evidence that is actually comparable
One randomized trial has compared these two compounds directly in people with obesity and without diabetes. In SURMOUNT-5, mean weight change at 72 weeks was −20.2% with tirzepatide and −13.7% with semaglutide. The difference is 6.5 percentage points, and it was statistically significant at P<0.001.
Those absolute figures are the honest starting point.
About that 47% figure
The same result often appears as "47% greater weight loss." That number is arithmetically correct. Dividing the 6.5-point gap by semaglutide's 13.7% gives 47.4%.
It is a relative expression of the absolute difference, and relative framing makes any gap sound larger. A 6.5-point difference between two large effects is meaningful without inflation. The relative figure should never appear without the absolute numbers beside it.
Where the circulating numbers come from
The figures "20.9% versus 14.9%" are widely presented as the head-to-head result. Pages ranking at the top of search results for this exact comparison repeat them.
They are not the head-to-head result.
The 20.9% comes from SURMOUNT-1, a 72-week placebo-controlled trial of tirzepatide alone. The 14.9% comes from STEP-1, a 68-week placebo-controlled trial of semaglutide alone. Different trials, different populations, different durations, and no randomization between them.
Comparing them produces an indirect comparison, which is a weaker form of evidence than the direct trial that actually exists. Using indirect figures while calling them head-to-head results is the most common factual error on this topic.
The comparison across twelve dimensions
- Dimension: Molecular identity — Tirzepatide: 39-amino-acid peptide, fatty acid modified — Semaglutide: Modified GLP-1 analogue — Comparability: Direct
- Dimension: Receptor targets — Tirzepatide: GIP and GLP-1 — Semaglutide: GLP-1 only — Comparability: Direct
- Dimension: GIP receptor activity — Tirzepatide: Affinity comparable to native GIP — Semaglutide: None — Comparability: Tirzepatide only; contribution to outcomes not established
- Dimension: GLP-1 receptor activity — Tirzepatide: About fivefold weaker than native GLP-1, biased toward cAMP signaling — Semaglutide: Full GLP-1 receptor agonism — Comparability: Comparable in kind, different in degree
- Dimension: Trial populations and designs — Tirzepatide: SURMOUNT and SURPASS programs — Semaglutide: STEP and SUSTAIN programs — Comparability: Direct only within SURMOUNT-5 and SURPASS-2
- Dimension: Weight loss outcomes — Tirzepatide: −20.2% at 72 weeks — Semaglutide: −13.7% at 72 weeks — Comparability: Direct randomized evidence
- Dimension: Diabetes outcomes — Tirzepatide: −2.01 to −2.30 HbA1c points — Semaglutide: −1.86 points at 1 mg — Comparability: Direct, but only against a diabetes dose
- Dimension: Gastrointestinal adverse events — Tirzepatide: Predominant, mostly mild to moderate, mainly during escalation — Semaglutide: Same pattern — Comparability: Comparable in pattern; finer rate differences not established
- Dimension: Discontinuation and tolerability — Tirzepatide: Few discontinuations in trials — Semaglutide: Similar — Comparability: Partly comparable; widely quoted percentages are often unsourced
- Dimension: Lean mass findings — Tirzepatide: 10% to 30% of weight lost, mostly impedance-measured — Semaglutide: Not comparably measured — Comparability: Not directly comparable. No verified randomized head-to-head measurement
- Dimension: Additional indications — Tirzepatide: Sleep apnea; cardiovascular risk reduction in type 2 diabetes — Semaglutide: Cardiovascular risk reduction in obesity without diabetes — Comparability: Comparable by regulatory status only
- Dimension: Key uncertainties — Tirzepatide: GIP attribution; long-term safety; open-label head-to-head design — Semaglutide: Separate evidence base — Comparability: Both carry unresolved questions
What this comparison cannot settle
The two compounds have cardiovascular evidence of different kinds, and the difference is easy to blur.
Semaglutide has placebo-controlled evidence of reduced major cardiovascular events in people with obesity and established cardiovascular disease but without diabetes. Tirzepatide's cardiovascular trial ran in a different population, people with type 2 diabetes and established atherosclerotic disease. That trial also used an active comparator rather than placebo, and it demonstrated non-inferiority rather than superiority.
Neither result translates into the other's population. Saying one compound "has better cardiovascular data" oversimplifies two findings that answer different questions.
What the evidence shows about stopping
Weight returns after stopping, and the trial establishing this is more specific than most summaries of it.
SURMOUNT-4 began with a 36-week open-label period in which all participants received tirzepatide. Mean weight reduction over that period was 20.9%. Participants were then randomized either to continue or to switch to placebo for a further 52 weeks.
From week 36 to week 88, the continuing group lost a further 5.5%. The group switched to placebo gained 14.0%. The published between-group difference was 19.4 percentage points, derived from unrounded values.
A second measure makes the pattern clearer. Among those who continued, 89.5% maintained at least 80% of the weight they had lost during the lead-in. Among those switched to placebo, 16.6% did.
This is high-strength evidence for the question it answers, which is what happens when treatment is withdrawn abruptly in a trial setting.
It does not describe planned discontinuation with clinical support, dietary structure, or gradual reduction. No trial in this program tested that scenario. No conclusion about it can be drawn from these numbers.
What lean mass findings do and do not mean
This is where the available evidence is weakest and the confident claims are loudest.
A 2026 narrative review examined ten studies covering randomized trials, observational studies and modeling analyses. According to PubMed, it found that weight reduction was predominantly fat mass. Reductions in lean mass, fat-free mass or skeletal muscle mass appeared in all studies. Lean mass accounted for roughly 10% to 30% of total weight lost. The authors describe that range as broadly in line with the expected physiological response to weight loss of any cause.
Three limitations in that review matter more than the number itself.
Most studies used bioelectrical impedance rather than DXA scanning. Those methods measure different things and are not interchangeable.
Lean mass is not the same as muscle. A second 2026 review states this directly. DXA-derived lean mass and impedance-derived fat-free mass both include water, organ tissue and connective tissue. Losing lean mass on a scan is not the same finding as losing skeletal muscle.
Functional outcomes were not assessed in these studies. No strength testing and no physical performance measurement were reported. The clinical significance of the measured reductions therefore remains uncertain.
One widely repeated claim deserves direct correction. A specific head-to-head split, around 26% lean mass loss on tirzepatide against around 45% on semaglutide, appears on pages ranking highly for this comparison. It is presented as DXA substudy data and carries no citation.
No verified randomized head-to-head measurement supports a specific lean-mass percentage difference between these two compounds. The honest classification is insufficient evidence, not a number.
This section grades as limited evidence.
Side effects, serious risks and the boxed warning
Gastrointestinal effects dominate the safety picture across every trial in the program. Nausea, diarrhea, vomiting, constipation and reduced appetite were the most commonly reported events. Severity was mild to moderate in the majority of cases. These events clustered while doses were being increased rather than during steady treatment.
SURPASS-2 reported adverse event rates directly. Nausea occurred in 17% to 22% of tirzepatide participants against 18% on semaglutide. Diarrhea occurred in 13% to 16% against 12%. Vomiting occurred in 6% to 10% against 8%. The profiles were broadly similar between the two compounds.
Treatment discontinuation due to adverse events stayed low. In SURMOUNT-1 it ran from 4.3% to 7.1% across tirzepatide groups, against 2.6% on placebo.
The boxed warning
Tirzepatide carries a boxed warning concerning thyroid C-cell tumors, including medullary thyroid carcinoma.
The evidence behind it comes from rodent studies. Rats given tirzepatide and similar compounds developed thyroid tumors. Whether this applies to humans is not known, and current labeling states exactly that.
The phrase "not established in humans" causes genuine confusion, so it is worth unpacking. The risk has not been ruled out. Neither has the risk been demonstrated in people. What the phrase means is that the rodent finding has not been shown to translate, leaving the question open. Regulators treat an unresolved signal of this kind conservatively, which is why it appears as a boxed warning rather than a footnote.
Labeling contraindicates use in anyone with a personal or family history of medullary thyroid carcinoma, or with multiple endocrine neoplasia syndrome type 2.
Other serious risks named in labeling
Current labeling identifies the following serious risks:
- Pancreatitis
- Gallbladder disease
- Acute kidney injury, usually following dehydration from gastrointestinal effects
- Serious allergic reactions
- Worsening of diabetic retinopathy
- Low blood sugar when combined with insulin or sulfonylureas
- Aspiration risk under anesthesia, related to delayed stomach emptying
Everything in this section describes the labeled pharmaceutical product. It is not a monitoring protocol, and it is not guidance for use.
What the research shows beyond weight and blood sugar
Four research areas get discussed together while sitting at four completely different stages. Separating them by status is the point of this section.
Obstructive sleep apnea, approved
Two 52-week randomized, double-blind, placebo-controlled trials tested tirzepatide in adults with moderate-to-severe obstructive sleep apnea and obesity. One enrolled people not using positive airway pressure therapy. The other enrolled people who were.
Baseline severity was high, with a mean apnea-hypopnea index of 51.5 and 49.5 events per hour. In the first trial, the treatment difference against placebo was −20.0 events per hour. In the second it was −23.8. Both results reached a P value below 0.001. Hypoxic burden, inflammatory markers, systolic blood pressure and patient-reported sleep outcomes also improved.
This supported the first approval of any medication for this condition.
A common explanation is that the benefit is simply weight loss relieving pressure on the airway. A 2025 post-hoc analysis found that improvements did track achieved weight reduction, which supports that reading. The same analysis states explicitly that further research is needed to separate weight-driven from weight-independent effects. Confident single-cause versions repeated elsewhere go beyond what their own source claims.
Cardiovascular risk in type 2 diabetes, approved August 2026
SURPASS-CVOT randomized 13,299 adults with type 2 diabetes and established atherosclerotic cardiovascular disease. The design was double-blind, and the comparator was an active drug with proven cardiovascular benefit rather than placebo.
The primary composite endpoint of cardiovascular death, heart attack or stroke occurred in 12.2% on tirzepatide and 13.1% on the comparator. The hazard ratio was 0.92, with a 95.3% confidence interval of 0.83 to 1.01. Non-inferiority was met at a P value of 0.003. Superiority was not met, at P=0.09.
The FDA approved a cardiovascular risk reduction indication on August 28, 2026, on the strength of that result.
A difference in all-cause mortality also appeared. It was not a primary endpoint and was not adjusted for multiple comparisons, so it is exploratory and should be read that way.
Heart failure with preserved ejection fraction, trial results but not an approved indication
The SUMMIT trial enrolled 731 people with obesity-related heart failure with preserved ejection fraction. Tirzepatide reduced the risk of cardiovascular death or worsening heart failure, and it improved symptom scores.
A prespecified secondary analysis found the benefit consistent between women and men, with hazard ratios of 0.66 and 0.61 and no significant interaction.
This is high-strength evidence for that specific obesity-related phenotype. It is not an approved indication, and it does not extend to heart failure without obesity.
Liver disease, Phase 2 only
A Phase 2 trial reported disease resolution rates of 44% to 62% against 10% on placebo, with signals of fibrosis improvement.
Phase 2 means earlier-stage evidence, smaller numbers, and results that have not been confirmed in a Phase 3 program. This is not an approved indication.
To state it once plainly, sleep apnea and cardiovascular risk reduction in type 2 diabetes are approved indications. Heart failure with preserved ejection fraction and liver disease are not.
Why reported tirzepatide prices are not comparing the same thing
Price figures for tirzepatide vary enormously, and the variation is usually not about the compound at all. It is about which category is being priced.
At least four distinct categories get quoted interchangeably. The first is the list price of a prescribed pharmaceutical product. A second is what a person actually pays after insurance, which differs by plan and by indication. Manufacturer direct-pay pricing forms a third, and it differs again. Research-use materials form a fourth, and they are not prescribed products and are not priced as them.
A figure drawn from any one of those categories tells you almost nothing about the others. Comparisons that move between them without saying so are comparing incompatible things.
One point matters more than any number. The price of a research-use material carries no information about its formulation, purity, sterility, stability or biological equivalence to the product used in clinical trials. Cost is not a quality signal in either direction.
This section deliberately names no vendor and quotes no figure, because doing either would turn an explanation into sourcing guidance.
What trial evidence can and cannot tell you about a research vial
This section exists because the distinction gets blurred constantly, usually in whichever direction suits the page making the claim.
Every clinical result on this page describes one thing. It describes a pharmaceutical product manufactured under defined controls, tested to a defined standard, and reviewed by a regulator. The trial outcomes attach to that product.
No evidence establishes that a research-use material sold under the same compound name shares that product's formulation, purity, sterility, stability or biological activity. Not some evidence, weak evidence, or contested evidence. None. The studies that would establish it have not been done.
That is a statement about what has been tested, not an accusation about any supplier.
What a certificate of analysis establishes
A certificate of analysis is useful and frequently over-read. It can establish identity, meaning the material is the compound it claims to be. Purity by a stated analytical method, usually HPLC, is within its scope. So is content measured against a stated specification.
Safety is not within its scope, and neither is efficacy. Sterility falls outside it unless sterility was specifically tested and reported. Correct biological activity is also outside it, because chemical purity and biological function are different properties. A purity figure says nothing about how a material behaves in a living system.
Reading one properly is a skill worth having. How to read a peptide certificate of analysis covers what each section means. How to choose a research peptide supplier covers the questions worth asking before ordering anything.
The research-use boundary
Research-use materials are supplied for laboratory and in-vitro research by qualified personnel. They are not drugs, supplements or cosmetics. Human and animal use falls outside that boundary, as do diagnostic, therapeutic and preventive purposes.
Two things that label does not do are worth naming. It does not resolve every regulatory question about these compounds, which is a more complicated area covered in our guide to peptide legality. Nor is it a justification for providing human-use guidance elsewhere while disclaiming it here.
This article contains no dosing, no reconstitution instructions, no administration guidance and no sourcing recommendations. That is a deliberate position, not an oversight.
What the evidence still does not establish
Strong evidence in some areas does not mean strong evidence everywhere. Five significant questions remain genuinely open.
How much of the effect comes from GIP activation. No human trial isolates it, so the attribution rests on inference rather than measurement.
Long-term safety beyond the durations trials covered. The longest trials here ran 72 to 88 weeks, with the cardiovascular trial following participants about four years. People may use these compounds for far longer than any trial has observed.
Whether lean mass changes matter functionally. Measured reductions are consistent across studies. Strength and physical performance were not assessed.
How the two compounds compare on cardiovascular outcomes. They have been studied in different populations against different comparators. No trial answers the comparative question.
Durability past the longest published follow-up. What happens after five or ten years of continuous use has not been studied, because the data do not yet exist.
Two design issues apply across the comparison evidence. Both head-to-head trials, SURMOUNT-5 and SURPASS-2, were open-label. Blinding matters most for outcomes influenced by behavior and expectation, and weight is exactly such an outcome.
Every trial cited here was also funded by the manufacturer. That is normal for registration trials and does not invalidate them. It is a fact readers should have.
Absence of evidence is not evidence of absence. These are open questions, not demonstrated problems.
The short version
The human evidence for tirzepatide on weight and blood sugar is high-strength and unusually consistent. The direct comparison against semaglutide gave 20.2% against 13.7% over 72 weeks, a real 6.5-point difference, in an open-label trial. Weight regain after withdrawal is well documented and substantial. Sleep apnea and cardiovascular risk reduction in type 2 diabetes are approved indications supported by Phase 3 evidence.
Three questions about tirzepatide remain genuinely unsettled. Why the dual mechanism outperforms single-receptor activation is inferred rather than demonstrated. Lean mass changes are measured, but their functional significance is unknown. Long-term safety beyond roughly four years has not been studied.
Everything above describes a pharmaceutical product used in registered trials. Research-use materials are supplied for laboratory investigation only. No evidence establishes that they reproduce the properties of the product those trials tested.
