Chapter 8 — Key Takeaways
Semaglutide
The core claims
Three modifications turn a two-minute hormone into a weekly drug. Aib at position 8 (blocks DPP-4); Lys→Arg at 34 (single attachment site, manufacturing only); C18 fatty acid at 26 (albumin binding, defeats renal clearance). Receptor activity unchanged — it is the same message, delivered for a week.
Four trial programs:
| Program | Established |
|---|---|
| SUSTAIN | Glycemic control in type 2 diabetes; and SUSTAIN 6 found cardiovascular benefit in a safety trial |
| STEP | Weight: ~−15% at 68 weeks without diabetes; ~−10% with diabetes |
| PIONEER | Oral semaglutide works, at ~1% bioavailability |
| SELECT | 20% relative / ~1.5 percentage points absolute MACE reduction over ~3 years, in established CVD + overweight, without diabetes |
The population lesson. STEP 1 and STEP 2 are the same drug, dose, and duration in different populations, producing different results. Quoting "15%" to someone with type 2 diabetes is the most common error in coverage of this drug.
SELECT changed what the drug is. It replaced a surrogate (weight) with a hard endpoint (events), which converts "reduces a risk factor" into "reduces events" — a categorically stronger claim, and the reason coverage and guideline positions shifted.
The adverse effects are the mechanism overshooting. Gastrointestinal effects come from slowed gastric emptying plus hindbrain receptor activation; they occur in a substantial minority and typically attenuate. Gallbladder disease and gastroparesis are real. The pancreatitis signal was appropriately raised and not confirmed at the feared magnitude. The thyroid boxed warning describes a rodent finding with a species-difference explanation and no demonstrated human signal — and the contraindication in MEN2 and medullary thyroid carcinoma is not theatre.
Lean mass is lost, as in any weight loss. Whether it produces functional decline is not well established; trials measuring physical function generally found it improved. Sarcopenic obesity is the population where it matters most and is least studied.
And weight returns when the drug stops.
The finding that reframes the drug
This is chronic therapy, not a course of treatment.
Semaglutide overrides an intact system rather than replacing an absent one (Ch 2 §2.8). The set point was never removed; it was outweighed. Stop, and it resumes.
That is not a failure of the drug. Nobody says statins don't work because cholesterol rises when you stop. The expectation that a weight drug should produce permanent change after stopping is applied to no other chronic therapy, and it comes from the moral framing of weight rather than from pharmacology.
It does, however, transform the cost question entirely. Chapter 12.
Evidence ratings issued in this chapter
| Claim | Rating |
|---|---|
| Semaglutide 2.4 mg for weight loss in adults with obesity or overweight-plus-comorbidity, alongside lifestyle support, for as long as treatment continues | ✅ |
| Semaglutide for glycemic control in adults with type 2 diabetes | ✅ |
| Semaglutide 2.4 mg for cardiovascular risk reduction in adults with established cardiovascular disease and overweight/obesity without diabetes | ✅ |
| Cardiovascular benefit in primary prevention (no established CVD) | NOT RATED — not studied |
Declining to rate is substantive. Absolute benefit scales with baseline risk; in a lower-risk population the same relative reduction produces a much smaller absolute benefit against identical harms. Extending a rating to a population a trial did not study is exactly the error the system exists to prevent.
And note what the weight ✅ does not cover: people who are not overweight, use without lifestyle support, durations beyond those studied, and every other indication (Chapter 10).
Numbers worth remembering
| Quantity | Value |
|---|---|
| Semaglutide half-life | ~1 week |
| STEP 1 (no diabetes, 68 wk) | ~−15% vs ~−2.4% placebo |
| STEP 2 (type 2 diabetes, 68 wk) | ~−10% |
| SELECT | 20% relative, ~8% → ~6.5% over ~3 years (~1.5 percentage points), ~17,000 participants |
| SELECT approximate NNT | ~65–70 over ~3 years |
| Oral bioavailability | ~1% |
| Approvals | 2017 injectable T2D · 2019 oral · 2021 weight management |
Key terms
SUSTAIN · STEP · PIONEER · SELECT · A1C · titration · gastroparesis · medullary thyroid carcinoma · MEN2 · lean mass · sarcopenic obesity · chronic therapy · indication
What you can now evaluate
- ✅ any quoted weight-loss figure — by asking for its population and its estimand
- ✅ the SELECT result, in both framings, with an NNT
- ✅ the thyroid warning, accurately, in neither direction
- ✅ why the nausea happens and why it fades
- ✅ what the muscle-loss concern does and does not establish
- ✅ why the weight returns, and why that is not a failure
- ❌ not yet: whether tirzepatide is better, and the trap in its trial. Chapter 9.
The one-sentence version
Semaglutide is the same message GLP-1 has always sent, delivered for a week instead of ninety seconds — and that single change in duration produced the best-evidenced metabolic drug of the last fifty years, whose benefits last exactly as long as you keep taking it.
Next: Chapter 9 — tirzepatide, the multi-agonists, and the trial that honestly reports two different answers.