Research methods
How to read metabolic peptide trials
Plain-English research-methods primer on reading trials of metabolic peptides and incretin medicines: phase, population, comparator, endpoints, hazard ratios, non-inferiority vs superiority, and the limits of cross-trial comparison.
Trials of metabolic peptides have produced eye-catching numbers — but the headline number is rarely the whole story. This primer walks through the moving parts every reader should hold in mind before quoting a result.
Phase tells you what the trial was designed to do
Phase 1 trials characterise pharmacokinetics and tolerability in small groups, typically of healthy volunteers. Phase 2 trials explore dose-response and proof-of-concept efficacy in the target population. Phase 3 trials are pivotal large-scale evidence for licensing — designed to confirm efficacy and characterise safety in the intended-use population. Cardiovascular outcome trials (CVOTs) and dedicated outcome trials (kidney, heart-failure, MASH) are usually a layer on top of phase 3, often required to make outcome claims rather than just biomarker claims. Phase 2 magnitude is not the same as phase 3 confirmation; phase 3 efficacy is not the same as outcome benefit.
Population: who was studied changes what the result means
Trial populations are defined by inclusion / exclusion criteria. SELECT enrolled adults with established cardiovascular disease and overweight or obesity but without diabetes; SURPASS-CVOT enrolled adults with type 2 diabetes and established atherosclerotic cardiovascular disease. Results from one population do not automatically generalise to another. The published abstract often makes this explicit; the headline often does not.
Comparator: placebo and active comparators tell different stories
A placebo-controlled trial measures effect over the baseline of natural history. An active-comparator trial (for example tirzepatide versus dulaglutide in SURPASS-CVOT) measures relative effect; non-inferiority in that setting does not mean superiority. SURMOUNT-5 compared tirzepatide with semaglutide 2.4 mg — a head-to-head between two licensed obesity treatments and very different from comparing either to placebo.
Primary vs secondary endpoints
The primary endpoint is the statistical focus of the trial — it is what the trial was powered to answer. Secondary endpoints add context but are not, on their own, evidence for licensing claims. Surrogate endpoints (HbA1c, body-weight change, MRI-PDFF, eGFR slope) inform decisions but are not the same as clinical outcomes (cardiovascular events, kidney failure, mortality). The research-methods page on surrogate vs clinical outcomes explores this further.
Hazard ratio and confidence interval
Cardiovascular and renal outcome trials report relative effects as hazard ratios with confidence intervals. A hazard ratio of 0.80 means a 20% lower hazard in the active arm relative to comparator. The 95% confidence interval is the range of plausible true values consistent with the data; if it crosses 1.00 the result is not statistically significant at conventional thresholds. The absolute effect (events per 1,000 patient-years) is often more clinically meaningful than the relative effect.
Non-inferiority vs superiority
Many CVOTs are designed for non-inferiority — the goal is to rule out a substantial worsening relative to comparator, not to demonstrate improvement. A non-inferiority result is not a superiority result. Headlines and editorial coverage sometimes blur this distinction; trial publications generally do not.
Absolute vs relative weight change
Trials sometimes report mean percent weight change from baseline, sometimes placebo-adjusted change, and sometimes the proportion of participants achieving categorical thresholds (≥5%, ≥10%, ≥15% body-weight reduction). These are not interchangeable. Cross-trial comparisons should use the same definition; otherwise the comparison can be misleading.
Duration matters more than you think
Most pivotal weight-management trials run 56–72 weeks. Two-year data (e.g., STEP 5) and randomised-withdrawal designs (e.g., SURMOUNT-4) tell you something about durability and post-discontinuation trajectory, but no pivotal obesity trial yet matches the multi-decade follow-up available for established cardiovascular drugs. Short-duration trials cannot fully establish long-term safety.
Cross-trial comparison: usually a bad idea
Indirect comparisons between SURMOUNT-1 and STEP 1 are tempting but methodologically weak. Different populations, comparators, endpoints, trial sites, eras and behaviours mean apparent differences may be artefact rather than signal. Head-to-head trials (SURPASS-2, SURMOUNT-5) and network meta-analyses with explicit methodology are the right places to read off comparative effect — and even then, the assumptions matter.
Who funded the trial and who wrote it up?
Pivotal metabolic peptide trials are typically manufacturer-sponsored. This is normal — but it is also a reason to read the study with the published statistical analysis plan and independent registry record alongside the manufacturer's publication. Sponsor bias is real and well-documented across pharmacology; the research-methods page on sponsor bias explores it.
Key takeaways
- Always know the phase, population, comparator and primary endpoint before quoting a result.
- Non-inferiority is not superiority.
- Surrogate endpoints are not clinical outcomes.
- Cross-trial comparisons need explicit methodology; otherwise they mislead.
- Phase 3 efficacy is not outcome benefit; outcome benefit is not lifetime safety.
- Formal statistical training.
- Individual treatment, prescribing or dosing advice.
- Permission to skip the original publication of any trial cited.
Sources are tagged by class. Regulatory documents, peer-reviewed publications and trial registry records are preferred. Manufacturer disclosures and conference material are accepted where peer-reviewed publication is not yet available; preprints are flagged.