Chapter 9: Tirzepatide and Beyond: Dual and Triple Agonists

"Phase 2 is where drugs look their best. It is also where roughly half of them stop looking good." — a rule of thumb that has never let anyone down

Overview

Semaglutide was not the end of the story. It was the point at which the field realized how much room was left.

Tirzepatide arrived as a molecule that does something semaglutide does not: it activates two receptors — GIP and GLP-1 — with a single peptide. It produces larger effects on both glucose and weight. And it did so despite GIP having been written off as a therapeutic target for two decades, for reasons that were entirely sensible and that Chapter 7 §7.5 laid out.

Behind it is a pipeline of compounds hitting two, three, and sometimes more receptors, plus a completely different approach — small molecules that activate the GLP-1 receptor without being peptides at all, and which could therefore be swallowed.

This chapter has two jobs.

The first is to tell you what these compounds do, with the same discipline Chapter 8 applied to semaglutide: populations attached, both risk framings, and ratings that stop where the evidence stops.

The second is to teach the single most useful statistical skill in Part II, using the trial that makes it unavoidable. SURMOUNT-1 honestly reports two different weight-loss figures for the same dose over the same duration — about 21% and about 22.5% — and popular coverage quotes whichever is larger, almost always without saying which. That is not a scandal and nobody lied. It is a distinction called an estimand, Chapter 5 §5.9 introduced it, and this is where you learn to use it.

There is also a warning running through the chapter. Several compounds here have produced spectacular Phase 2 results. Phase 2 results are systematically optimistic, for reasons Chapter 5 §5.4 explained. The book flags the phase every single time, and by the end you should find yourself doing it automatically.

In this chapter, you will learn to:

• Explain unimolecular dual agonism and why it differs from taking two drugs
• State the tirzepatide results with their populations and their estimands
• Work the two-estimand problem completely
• Say what SURPASS-2's head-to-head comparison does and does not establish
• Distinguish a Phase 2 headline from a Phase 3 finding
• Read a pipeline announcement critically

Learning Paths

💊 GLP-1 — the whole chapter. §9.4 is the section that will most improve how you read coverage of this drug class. 🏋️ Performance — §9.4 and §9.10 transfer directly; the pipeline discipline applies to every compound in Part III. 🔬 Science — full read; §9.1 and §9.2 are the pharmacology. 🏥 Clinical — §9.4 and §9.5 are what your patients are quoting at you, and being able to name the estimand in one sentence is disproportionately useful.

9.1 One molecule, two receptors

Tirzepatide is a 39-amino-acid peptide, molecular weight about 4,800 Da, that agonizes both the GIP receptor and the GLP-1 receptor. It is described as a unimolecular dual agonist — one molecule, two targets.

That phrase deserves unpacking, because "hits two receptors" sounds like it should be equivalent to taking two drugs, and it is not.

WHY ONE MOLECULE IS NOT TWO DRUGS

  TWO SEPARATE DRUGS                    ONE DUAL AGONIST
  ──────────────────                    ─────────────────
  Drug A → receptor 1                   Molecule → receptor 1
  Drug B → receptor 2                            → receptor 2

  · two pharmacokinetic profiles        · ONE pharmacokinetic profile
    (different half-lives, different      — both receptors are engaged in the
     peaks, different clearance)          same temporal pattern, always

  · the RATIO of activity at the two    · the ratio is FIXED by the molecule's
    receptors varies over time and         structure and cannot drift
    with dosing

  · two products, two costs, two        · one product, one injection,
    injections, two adherence problems     one adherence problem

  · the ratio can be TUNED by dosing    · the ratio can only be changed by
    each drug separately                   making a different molecule

  The engineering consequence: a dual agonist is a bet on a specific RATIO of
  activity at two receptors. Getting that ratio right is the design problem.

How the molecule achieves it. Tirzepatide's sequence is based on GIP rather than on GLP-1 — a detail that surprises people — and is engineered to bind both receptors, with a fatty acid modification for albumin binding on the same principle as semaglutide (Chapter 4 §4.5). Its receptor activity is not balanced: it is considerably more potent at the GIP receptor than at the GLP-1 receptor, and whether that particular imbalance is optimal is not established.

And it works, which was not obvious. Chapter 7 §7.5 explained why GIP had been set aside: its insulinotropic effect is blunted in type 2 diabetes and it appears to promote fat storage. Adding it should not have helped. That it does is the most interesting unresolved question in this chapter, and this book will not pretend to resolve it.

9.2 SURPASS and SURMOUNT

Two development programs, following the semaglutide pattern.

SURPASS — type 2 diabetes. A series of randomized trials against placebo and against active comparators, establishing glycemic control. Tirzepatide was approved for type 2 diabetes in 2022 (Mounjaro).

SURMOUNT — obesity. Establishing weight effects at higher doses. Tirzepatide was approved for weight management in 2023 (Zepbound).

Same molecule, two brand names, two indications — exactly the Ozempic/Wegovy pattern, and it confuses people for exactly the same reason.

SURMOUNT-1 is the trial this chapter turns on.

9.3 SURMOUNT-1

🔬 Read the Study — SURMOUNT-1

text FIGURE 9.3 — "Twenty-one, or twenty-two and a half" [real published trial] THE STUDY Randomized, double-blind, placebo-controlled. Adults with obesity, or overweight with at least one weight-related comorbidity, WITHOUT diabetes. Tirzepatide 5, 10, or 15 mg weekly vs placebo, 72 weeks. All arms received lifestyle intervention. Manufacturer-sponsored. THE QUESTION How much weight does tirzepatide produce over 72 weeks in this population, and does it differ by dose? WHAT IT SHOWS Substantial dose-dependent weight reduction. At the 15 mg dose: about −21% from baseline on the trial's treatment-regimen analysis, and about −22.5% on the analysis restricted to participants who remained on treatment. Placebo lost a small amount. THE LARGEST WEIGHT EFFECT DEMONSTRATED FOR ANY APPROVED PHARMACOTHERAPY AS OF THIS WRITING. THE DOESN'T It does not include people with diabetes (a separate trial, with smaller effects — the STEP 1/STEP 2 pattern repeats). It does not establish cardiovascular outcomes. It does not extend beyond 72 weeks. It does not tell you what happens on stopping. THE VERDICT ✅ for the weight claim in the studied population. THE LESSON TWO NUMBERS, ONE TRIAL, BOTH CORRECT. Which one you quote is determined by which question you are answering — and quoting one without naming it is systematically optimistic, because the larger figure is always the restricted analysis.

9.4 The two-estimand problem, worked

This is the section. Chapter 5 §5.9 introduced estimands; here they are unavoidable.

The situation. A 72-week trial. Some participants stop taking the drug — side effects, life circumstances, moving away, deciding they have had enough. Some stop and remain in the study for measurement. Some withdraw entirely.

What is "the result"?

There are two defensible answers, and they answer two genuinely different questions.

TWO QUESTIONS, ONE TRIAL

  ┌────────────────────────────────────────────────────────────────────────────┐
  │  THE TREATMENT-REGIMEN ESTIMAND                                            │
  │  "What happens if a doctor PRESCRIBES this?"                               │
  │                                                                            │
  │  Counts everyone as randomized, including those who discontinued.          │
  │  Reflects the real world, where some people will not tolerate a drug.      │
  │  Generally the primary analysis for regulatory purposes.                   │
  │                                                                            │
  │  SURMOUNT-1, 15 mg, 72 weeks →  about −21%                                │
  ├────────────────────────────────────────────────────────────────────────────┤
  │  THE EFFICACY ESTIMAND                                                     │
  │  "What happens if you TAKE it, as prescribed, for the full period?"        │
  │                                                                            │
  │  Restricted to participants who remained on treatment.                     │
  │  Reflects the drug's biological effect more directly.                      │
  │  SYSTEMATICALLY LARGER.                                                    │
  │                                                                            │
  │  SURMOUNT-1, 15 mg, 72 weeks →  about −22.5%                              │
  └────────────────────────────────────────────────────────────────────────────┘

  SAME TRIAL. SAME DOSE. SAME 72 WEEKS. BOTH PRE-SPECIFIED. BOTH REPORTED.
  Neither is "the" answer. Both are answers to different questions.

Why the efficacy figure is always larger. The participants excluded from it are disproportionately those who stopped — and people stop for reasons correlated with doing badly: side effects they could not tolerate, or an inadequate response. Removing them removes the lower tail of the distribution. This is systematic, not random, which is what makes an unnamed figure biased rather than merely imprecise.

And the size of the gap is itself informative. A large gap means many people did not stay on the drug, which is a clinically important fact about tolerability. A small gap means most people completed. Reporting only the efficacy figure discards that information entirely.

Which should you use?

If you are asking "should I start this drug?" — the treatment-regimen figure. It includes the possibility that you will be one of the people who stops, which is a real possibility with a real probability that the trial measured.

If you are asking "what does this molecule do to fat mass?" — the efficacy figure comes closer, though even it is imperfect because the people who remained are a selected group.

If you are reading a headline — assume the larger figure unless told otherwise, and adjust.

⚠️ Hype Check — "up to 22.5% weight loss"

This exact phrasing, or something close to it, appears constantly.

What's true: the figure is real, it is from a published trial, and it describes the highest dose.

What "up to" is doing: three separate things at once, and each one selects the most favorable option.

It selects the dose. 22.5% is the 15 mg result. Lower doses produced less. "Up to" quietly means "at the highest dose studied," which is not the dose everyone receives.

It selects the estimand. It is the efficacy analysis — restricted to people who stayed on treatment — not the treatment-regimen analysis a prescriber would want.

And it selects the population. Adults with obesity without diabetes. In type 2 diabetes the figures are meaningfully lower, exactly as with semaglutide.

The honest version: "In adults with obesity without diabetes, at the highest dose, over 72 weeks, with lifestyle support, mean weight loss was about 21% counting everyone assigned to the drug and about 22.5% counting only those who stayed on it."

Count the qualifiers. Five. That is why the short version wins, and it is Chapter 6 §6.5 in operation on a drug with excellent evidence.

9.5 Head-to-head: what SURPASS-2 does and does not establish

SURPASS-2 compared tirzepatide against semaglutide 1 mg in adults with type 2 diabetes over 40 weeks. Tirzepatide produced greater A1C reduction and greater weight loss at all three doses.

Head-to-head trials are genuinely valuable and comparatively rare — most trials compare against placebo, which answers a much easier question. So this deserves real weight.

And it deserves three specific qualifications.

The comparator dose. Semaglutide 1 mg was, at the time, the approved dose for type 2 diabetes. Higher semaglutide doses exist and are used. A comparison against the approved diabetes dose is legitimate; it is not a comparison against semaglutide at its maximum.

The indication. This was a diabetes trial. Extrapolating it to conclude tirzepatide is superior for weight management in people without diabetes is a population swap — and a separate head-to-head trial in obesity has since addressed that question, which is the correct way to answer it rather than by extension.

And the funder. Manufacturer-sponsored, by the manufacturer of the winning drug. Chapter 5's rule applies: funding changes the weight of evidence, not its truth value. The specific thing to check in a head-to-head is whether the comparator was used at an appropriate dose and schedule, and here the answer is defensible-but-arguable.

What it establishes: at the doses compared, in this population, over this duration, tirzepatide produced greater effects. That is a real finding and it is narrower than "tirzepatide is better."

💊 In the Clinic — "better" is not one question

A patient asking "which is better?" is asking a question that decomposes into at least six.

Better for what? Glycemic control, weight, cardiovascular outcomes, and quality of life are different endpoints with different evidence.

Better in whom? Diabetes and no diabetes produce different answers, consistently.

Better tolerated? Both classes produce gastrointestinal effects; comparative tolerability at equipotent doses is a genuine question.

Better evidenced? Semaglutide has SELECT — a cardiovascular outcomes trial with a hard endpoint in a population without diabetes. Tirzepatide's cardiovascular outcomes program is ongoing as of this writing. On cardiovascular outcomes specifically, the evidence is not equivalent, and "produces more weight loss" does not substitute for it.

Better available? Both have experienced supply constraints. Availability is not a pharmacological property and it determines what actually gets taken.

Better affordable? Chapter 12.

Which of these matters most depends on the person, and that is a conversation with a clinician rather than a fact about molecules.

9.6 Retatrutide: a Phase 2 result that reads like a Phase 3

Retatrutide is a triple agonist — GLP-1, GIP, and glucagon receptors.

The glucagon addition is counterintuitive, since glucagon raises blood sugar (Chapter 3 §3.6). The rationale is that glucagon receptor agonism increases energy expenditure and has effects on hepatic fat, and that in the context of simultaneous GLP-1 and GIP agonism the glucose-raising effect is offset. Whether that reasoning is correct is exactly what trials are for.

The reported result: roughly −24% body weight at 48 weeks at the highest dose — in a Phase 2 trial.

That figure is larger than tirzepatide's. It is also, and this is the entire point of this section, a Phase 2 result.

WHY THE PHASE MATTERS — and why this book says "Phase 2" every time

  PHASE 2                              PHASE 3
  ───────                              ───────
  dozens to a few hundred              hundreds to thousands
  often shorter                        long enough for the real question
  frequently a favorable population    the population the drug is for
  multiple doses; the best is quoted   the dose that will be used
  designed to detect a SIGNAL          designed to MEASURE an effect

  CONSEQUENCE: Phase 2 effect sizes are SYSTEMATICALLY LARGER than the Phase 3
  results that follow. This is not fraud. It is regression to the mean plus
  selection plus small-sample variance, and it is one of the most reliable
  regularities in drug development.

  A −24% Phase 2 figure and a −21% Phase 3 figure are NOT comparable numbers,
  and treating them as a ranking is the single most common error in coverage
  of this pipeline.

📊 Evidence Rating — retatrutide for weight loss

Claim: Retatrutide produces substantial weight loss in adults with obesity.

Rating: ⚠️ Promising but preliminary.

Why: A Phase 2 result — roughly −24% at 48 weeks at the highest dose. Real human data, randomized, and genuinely striking. It is also a Phase 2 result, which is systematically optimistic, and Phase 3 programs are ongoing as of this writing.

What would change it: completed, adequately powered Phase 3 trials in the target population reporting both estimands, plus a cardiovascular outcomes trial. Either could move this to ✅ or could reduce the effect substantially.

Note this is the book's standing example of a ⚠️ that reads like a ✅. The number is bigger than tirzepatide's approved-drug number, which makes it feel stronger, and it sits on a much weaker evidentiary footing. If you can hold both of those at once, you have the chapter.

9.7 CagriSema and the amylin combination

Cagrilintide is a long-acting amylin analog. Amylin is a peptide co-secreted with insulin from pancreatic beta cells that slows gastric emptying and promotes satiety — a genuinely separate mechanism from the incretins, and Chapter 13 covers it properly.

CagriSema is cagrilintide combined with semaglutide: two peptides, two mechanisms, one product.

Note the architectural difference from tirzepatide. Tirzepatide is one molecule hitting two receptors. CagriSema is two molecules in one injection. The distinction matters for exactly the reasons §9.1's diagram gives — two molecules have two pharmacokinetic profiles and a ratio that can be tuned by adjusting each component, which is an advantage and a complication.

As of this writing, the combination has produced substantial weight effects in trials and its late- stage program has generated results that were, in some analyses, less dramatic than early figures had suggested — which is itself an instructive instance of the Phase 2 pattern. The program is ongoing and specific figures should be checked against current sources, since this book's information is current only to 2026.

Rating: ⚠️ — real human data, an active program, and not yet settled.

9.8 Orforglipron and the small-molecule threat

Everything in this book so far has assumed that activating the GLP-1 receptor requires a peptide. That assumption is being tested.

Orforglipron is a small molecule — not a peptide — that activates the GLP-1 receptor. If it works, the consequences are substantial:

It can be swallowed as an ordinary tablet. No permeation enhancer, no 1% bioavailability, no fasting protocol. Chapter 1's size spectrum explains why: a small molecule is in the region where oral absorption is routine.

It can be manufactured by conventional chemical synthesis at conventional cost. Chapter 32 explains why peptide manufacturing is expensive; small molecules largely escape that.

And it would not require cold chain. Chapter 4 §4.8's cost stack loses several lines.

Why this is hard. Peptide receptors like GLP-1R are class B GPCRs with large extracellular domains that grip an extended peptide ligand (Chapter 7 §7.3). Getting a small molecule to activate such a receptor — rather than merely block it — is a genuine medicinal chemistry achievement, and it was long considered close to impossible.

Status as of this writing: in late-stage clinical development, with reported results that support meaningful glycemic and weight effects. Not approved as of 2026. Whether the effects match the injectable peptides, and what the tolerability profile looks like, are the open questions.

📊 Evidence Rating — orforglipron for weight loss and glycemic control

Claim: Orforglipron, an oral small-molecule GLP-1 receptor agonist, produces clinically meaningful weight loss and glycemic control.

Rating: ⚠️ Promising but preliminary, as of 2026.

Why: Real human trial data supporting meaningful effects, and a late-stage program. Not approved as of this writing, and comparative efficacy against the injectable peptides is not settled.

What would change it: completed Phase 3 with approval, plus head-to-head data against an injectable agonist and a cardiovascular outcomes program.

Why it matters disproportionately: if an oral small molecule matches the injectables, the cost structure, the access picture, and the compounding market described in Chapters 6 and 12 all change substantially. This is the most consequential pipeline entry in the chapter and it is not a peptide.

9.9 The pipeline as of 2026

A date-stamped snapshot, which is the only honest way to present a pipeline.

Note the two entries that should give you pause. GIP receptor agonists and GIP receptor antagonists are both being pursued for obesity. Chapter 7 §7.5 flagged this, and it remains the most honest available indication that the mechanism is not understood.

9.10 How to read a pipeline announcement

The practical skill this chapter exists to build.

READING A PIPELINE ANNOUNCEMENT — six questions

  1. WHAT PHASE?                     Phase 2 figures are systematically larger than
                                     the Phase 3 results that follow. If the phase is
                                     not stated, assume 2 or earlier.

  2. WHICH ESTIMAND?                 If not named, assume the larger (efficacy) one.

  3. WHICH DOSE?                     "Up to" means the highest dose studied, which may
                                     not be the dose that gets approved or tolerated.

  4. WHICH POPULATION?               Diabetes or no diabetes changes the answer by a
                                     third. Every time.

  5. WHAT ENDPOINT?                  Weight and A1C are surrogates. Cardiovascular
                                     outcomes are not. Which was measured?

  6. WHO IS ANNOUNCING, AND WHERE?   A press release, a conference abstract, and a
                                     peer-reviewed publication are three different
                                     things (Ch 5 §5.10). Many conference abstracts
                                     never publish in full.

  A compound that answers all six well may still fail. A compound that cannot
  answer them has told you where its evidence stops.

And one habit worth building: when a new compound's headline number exceeds an approved drug's, your first thought should be "what phase?" rather than "that's better." In this therapeutic area, in this decade, that reflex will be right far more often than it is wrong.

The three announcement genres

It also helps to know what kind of document you are reading, because each has different obligations.

A press release is written by a company for investors and journalists. It is not required to report a full result, it is not peer-reviewed, and it will lead with the most favorable defensible figure. It is not thereby false — securities regulation imposes real constraints on material misstatement — but it is selective by design, and the selection is what you are reading.

A conference abstract has been reviewed, usually lightly, and presented to specialists. It is more constrained than a press release and much less than a paper. Chapter 5 §5.10 noted that a meaningful fraction of conference abstracts never publish in full, and the ones that do not are disproportionately negative or flawed. An abstract from three years ago with no subsequent paper is a signal.

A peer-reviewed publication has been through review, carries a methods section you can interrogate, and is expected to report pre-specified endpoints and the analysis populations. It is the only one of the three that can be checked against a registry entry.

None of these is the trial. The trial is a protocol, a dataset, and a statistical analysis plan, and what you are reading is a summary of it written for a purpose. Knowing whose purpose is most of the work.

What "the pipeline is exciting" is doing

One final caution, because it applies to this chapter as much as to any marketing document.

A pipeline table like §9.9's is genuinely informative and it is also a selection. It lists the compounds that have advanced far enough to be named, which means it lists the survivors of every earlier round of attrition. The compounds that entered Phase 1 in this area and stopped do not appear, and there are many more of them than there are entries in that table.

This is Chapter 6 §6.3's survivorship problem, operating on drug development rather than on testimonials. The visible pipeline is filtered for compounds that have so far succeeded, and reading it as "look how much is coming" overstates what will arrive. A more accurate reading is: these are the candidates that have not yet failed, and roughly nine in ten of those entering human trials never reach approval.

That framing is less exciting and it is the one that has been right for the last fifty years.

A note on tolerability across the class

One thing the pipeline discussion tends to omit: as effect sizes have grown, so have the gastrointestinal effects.

The mechanism explains why. Chapter 7's action ③ — slowed gastric emptying — is one of the routes by which these drugs reduce intake, and it is also where the nausea comes from. A compound that produces a larger effect on intake is, generally, producing a larger effect on that action too. The benefit and the burden scale together, and there is no obvious reason to expect a compound to escape the trade-off by getting more potent.

Across the class, the pattern is consistent: nausea, vomiting, diarrhea, and constipation are the dominant adverse effects; they occur in a substantial minority; they are usually mild to moderate; they are managed by stepwise titration; and they typically attenuate over weeks while the metabolic effects persist (Chapter 2 §2.8 — tolerance is a property of drug-tissue pairs).

Two things this means for reading trial results.

Discontinuation rates are a real endpoint. A compound with a spectacular efficacy figure and a high dropout rate has told you something important, and the gap between its two estimands is where it shows up. §9.4's point that the size of the gap is informative is not a technicality — it is how tolerability appears in the numbers.

And blinding degrades. A drug producing noticeable gastrointestinal effects is difficult to blind in practice, whatever the protocol says. Chapter 5 §5.5 flagged this for semaglutide; it applies with at least equal force to compounds producing larger effects. For weight — measured on a scale — this matters little. For patient-reported outcomes, it matters more.

🩺 Safety and Risk — the multi-agonists specifically

The dual and triple agonists carry the same class considerations described for semaglutide in Chapter 8 §8.7 — the gastrointestinal profile, gallbladder effects, the rodent thyroid C-cell finding and its contraindication in medullary thyroid carcinoma and MEN2, and the practical aspiration issue around anesthesia.

What is genuinely less established for the newer compounds is exposure time. Semaglutide has more patient-years behind it than tirzepatide, which has more than retatrutide, which has essentially none outside trials. That ordering is not a statement about their relative safety — it is a statement about how much anyone could know.

Adding a third receptor adds a third set of tissues where the drug acts. Glucagon receptor agonism in particular has effects on hepatic glucose output and on energy expenditure that are the point of including it, and whose long-term consequences in humans are, as of this writing, unstudied at scale.

None of this predicts a problem. It states what is and is not currently knowable, which is the distinction Chapter 5 was written to preserve. Individual decisions belong with a clinician.

📊 Evidence Rating — tirzepatide for weight loss in adults with obesity

Claim: Tirzepatide produces substantial weight loss in adults with obesity, or overweight with a weight-related comorbidity, without diabetes, alongside lifestyle support, for as long as treatment continues.

Rating: ✅ Strong clinical evidence.

Why: SURMOUNT program; at 15 mg over 72 weeks, about −21% on the treatment-regimen analysis and about −22.5% on the efficacy analysis, in adults without diabetes, with lifestyle support. Approved for weight management in 2023. The largest weight effect demonstrated for any approved pharmacotherapy as of this writing.

What would change it: loss of effect during continued treatment, or a serious harm at multi-year use.

What this rating does NOT cover: cardiovascular outcomes — the outcomes program is ongoing as of this writing, and tirzepatide does not currently have the hard-endpoint evidence semaglutide has from SELECT. Producing more weight loss does not substitute for that.

📊 Evidence Rating — tirzepatide for glycemic control in type 2 diabetes

Claim: Tirzepatide improves glycemic control in adults with type 2 diabetes.

Rating: ✅ Strong clinical evidence.

Why: The SURPASS program, including active-comparator trials; approved for this indication in 2022; greater A1C reduction than semaglutide 1 mg in a head-to-head at all three doses.

What would change it: essentially nothing for the A1C claim. Note that A1C is a surrogate, and the cardiovascular outcomes question is separate and open.

📋 Your Evidence Dossier

This chapter teaches comparison discipline: two entries, one indication, the same standard applied to both.

The temptation when comparing is to reach for a single ranking. Resist it. A structured comparison produces something more useful than a winner.

COMPARISON WORKSHEET — apply to any two peptides for one indication

                              Peptide A          Peptide B
  ─────────────────────────────────────────────────────────────
  The exact claim             ..................  ..................
  Population studied          ..................  ..................
  Duration                    ..................  ..................
  Effect size                 ..................  ..................
    — which estimand?         ..................  ..................
    — which dose?             ..................  ..................
  Endpoint: surrogate/hard    ..................  ..................
  Head-to-head data?          ..................  ..................
    — comparator dose fair?   ..................  ..................
  Cardiovascular outcomes     ..................  ..................
  Adverse effect profile      ..................  ..................
  Rating                      ..................  ..................
  WHAT DIFFERS THAT ISN'T
  THE MOLECULE                ..................  ..................

That last row is the one that does the work. Two compounds frequently differ in trial duration, population, comparator, endpoint, and funder — and attributing all of the difference to the molecules is the most common comparison error.

Worked demonstration — semaglutide vs. tirzepatide for weight loss

                        SEMAGLUTIDE 2.4 mg        TIRZEPATIDE 15 mg
  ───────────────────────────────────────────────────────────────────────
  Claim                 weight loss in obesity    weight loss in obesity
  Population            overweight/obesity,       obesity or overweight+
                        no diabetes (STEP 1)      comorbidity, no diabetes
  Duration              68 weeks                  72 weeks
  Effect                ~−15% from baseline       ~−21% (treatment-regimen)
                                                  ~−22.5% (efficacy)
  Estimand quoted       treatment-regimen         BOTH — and coverage
                                                  quotes the larger
  Endpoint              surrogate (weight)        surrogate (weight)
  Head-to-head          SURPASS-2 (in DIABETES,   same trial, tirzepatide
                        vs semaglutide 1 mg)      favored
  CV outcomes           ✅ SELECT — hard endpoint  ONGOING as of this writing
  Adverse effects       GI-dominant, class effect GI-dominant, class effect
  Rating                ✅ weight, ✅ CV            ✅ weight, CV NOT RATED
  WHAT DIFFERS THAT     4 weeks' duration;        the head-to-head was in
  ISN'T THE MOLECULE    slightly different        diabetes, at a comparator
                        inclusion criteria        dose below semaglutide's max
  ───────────────────────────────────────────────────────────────────────
  HONEST CONCLUSION: tirzepatide produces greater weight loss. Semaglutide has
  cardiovascular outcome evidence that tirzepatide does not yet have. These are
  different advantages and neither is "better" without specifying for what.

Your task

Run the comparison worksheet on two peptides in your dossier that are used for the same purpose.

If you cannot find two, that is informative — it means your dossier's compounds are not competing for the same indication, and the comparison discipline applies differently.

Pay particular attention to the last row. For most pairs you will find that at least one difference is not about the molecules at all — a longer trial, an easier population, a weaker comparator, a different endpoint. Those differences produce apparent superiority that has nothing to do with pharmacology, and noticing them is the entire skill this section teaches.

Conclusion

Tirzepatide is a unimolecular dual agonist: 39 amino acids, about 4,800 Da, activating both the GIP and GLP-1 receptors with a fixed ratio built into its structure. One molecule, one pharmacokinetic profile, one injection — which is genuinely different from taking two drugs. It produces larger effects on weight and glucose than GLP-1 agonism alone, despite GIP having been reasonably written off as a target, and nobody can yet say why.

SURMOUNT-1 is the trial that makes estimands unavoidable. At 15 mg over 72 weeks in adults with obesity without diabetes: about −21% counting everyone assigned to the drug, and about −22.5% counting only those who stayed on it. Same trial, same dose, both pre-specified, both correct. The efficacy figure is systematically larger because the people excluded from it stopped for reasons correlated with doing badly — so an unnamed figure is biased rather than merely imprecise.

SURPASS-2 found tirzepatide superior to semaglutide 1 mg in type 2 diabetes at all three doses. That is a real finding, and it is narrower than "tirzepatide is better" — different indication, comparator at the approved diabetes dose, manufacturer-funded.

Behind them: retatrutide, a triple agonist with a roughly −24% Phase 2 result that is not comparable to a Phase 3 number; CagriSema, two molecules with two mechanisms in one injection; and orforglipron, a small molecule that is not a peptide at all and that could change the cost and access picture entirely if it works.

And the standing warning. Phase 2 figures are systematically larger than the Phase 3 results that follow. When a pipeline compound's headline number exceeds an approved drug's, the first question is what phase, not is it better.

Chapter 10 takes the class somewhere unexpected: the heart, the kidney, the liver, and possibly the brain — where one molecule earns five different ratings, two of them 🔬.

Key Terms

Unimolecular dual agonist — a single molecule that activates two different receptors, with a fixed activity ratio determined by its structure.

Polyagonist — a molecule activating three or more receptors.

SURPASS — the tirzepatide development program in type 2 diabetes.

SURMOUNT — the tirzepatide development program in obesity.

Head-to-head trial — a trial comparing two active drugs directly, rather than either against placebo.

Non-inferiority — a trial design asking whether a treatment is not meaningfully worse than a comparator, rather than whether it is better.

Superiority — a trial design asking whether a treatment is better than its comparator.

Retatrutide — an investigational triple agonist at GLP-1, GIP, and glucagon receptors.

Cagrilintide — a long-acting amylin analog.

Amylin analog — a compound mimicking amylin, a peptide co-secreted with insulin that slows gastric emptying and promotes satiety.

Orforglipron — an investigational oral small-molecule GLP-1 receptor agonist; not a peptide.

Phase 2 optimism — the systematic tendency for Phase 2 effect sizes to exceed the Phase 3 results that follow, arising from small samples, favorable populations, best-dose selection, and regression to the mean.

Spaced Review

1. (Ch 5) SURMOUNT-1 reports about −21% and about −22.5% for the same dose. Explain why both are correct, why the second is systematically larger, and what the size of the gap tells you.

2. (Ch 7) GIP was written off as a target for two sensible reasons, and adding GIP agonism helps anyway. State both reasons and at least two candidate explanations for the discrepancy.

3. (Ch 8) Tirzepatide produces greater weight loss than semaglutide; semaglutide has a cardiovascular outcomes trial that tirzepatide does not yet have. Explain why "produces more weight loss" does not substitute for the second, using Chapter 5 §5.6.

4. A pipeline compound reports a larger weight effect than an approved drug. State the first question you would ask and explain why it is first.

5. (Ch 1, Ch 4) Orforglipron is a small molecule rather than a peptide. Using Chapters 1 and 4, list four consequences of that difference and say which matters most for access.