This is the last post in the set I set out to write when I started this site, and I wanted to close by modeling something I've tried to do throughout: read the actual research carefully, rather than repeating headline numbers without context. STEP and SURMOUNT — the trial families behind semaglutide and tirzepatide — are as good a place as any to demonstrate what that looks like.
The headline numbers, and they're genuinely real
STEP 1, the pivotal trial for semaglutide, showed an average weight loss of 14.9% over 68 weeks, compared to roughly 2.4% with placebo. SURMOUNT-1, the equivalent trial for tirzepatide, showed average weight loss ranging from 16.0% at the lowest dose up to 22.5% at the highest dose over 72 weeks, against 2.4% with placebo. Nearly all tirzepatide patients — 89% at the lowest dose, 96% at the higher doses — achieved at least 5% weight loss, compared to 28% on placebo.
I want to be unambiguous about something before I get into the caveats: these numbers are real, they're substantial, and they represent a genuine advance in treating a condition that's historically been very difficult to treat pharmacologically. Nothing that follows is meant to undercut that.
The number you need to hold alongside the average
A trial average is the midpoint of a wide spread, not what every individual patient should expect. A post-hoc analysis of SURMOUNT-1 sorted patients by how quickly they responded: roughly 82% were "early responders," losing at least 5% of body weight within the first 12 weeks. The remaining 18% were slower to respond — but here's the detail I think matters most for patients who feel discouraged early: 90% of those slower responders still reached at least 5% weight loss by week 72, typically taking about 25 weeks to get there. A quiet first few months on these medications is common and doesn't predict a poor outcome.
Trial conditions are not the same as your actual life
This is the gap I think deserves the most attention, because it's the one most likely to create a mismatch between expectation and experience. Trial participants received highly structured support — regular visits, consistent dosing supervision, protocol-driven titration. Real-world data paints a somewhat different picture: one analysis found average real-world weight loss around 11.9% for patients who stayed on treatment, meaningfully below the 14.9% trial average, and discontinuation in the first three months running around 21% in real-world settings compared to roughly 7-11% within trials.
I don't read this as a reason for pessimism. I read it as the actual argument for structured, ongoing clinical support — the titration management, side-effect troubleshooting, and consistent follow-up I've described throughout this site as central to how I practice. The gap between trial results and real-world results isn't really about the drug performing differently; it's substantially about the difference between supervised, structured care and a prescription handed over with minimal follow-up.
Side effects are common — and usually not why people stop
Trial data reports a striking number here that's easy to misread in isolation: 89.7% of STEP participants and 80.5% of SURMOUNT participants reported experiencing some side effect. Read without context, that sounds alarming. Read alongside the discontinuation data, it looks different: only 4.3% to 7.1% of SURMOUNT-1 participants stopped tirzepatide specifically because of adverse events. The overwhelming majority of side effects — largely the GI symptoms I covered in an earlier post on this site — were tolerable enough that patients continued treatment through them.
One genuinely interesting detail from SURMOUNT-1: the placebo group had a higher overall discontinuation rate (26.4%) than any of the active treatment doses (14.3% to 16.4%). Patients experiencing side effects but real weight loss stayed in the trial more often than patients experiencing no side effects but also no meaningful results. That's worth sitting with — it suggests the actual driver of discontinuation, across the board, is more about efficacy than tolerability.
What direct comparison data now shows
More recently, SURMOUNT-5 provided something the field lacked for a while: a real head-to-head trial directly comparing tirzepatide and semaglutide in the same patient population, rather than comparing across separate trials with different populations. The result: tirzepatide produced greater average weight loss (20.2%) than semaglutide (13.7%) at maximum tolerated doses. This is genuinely useful data if you and your provider are choosing between the two medications specifically — though it's also worth knowing that some real-world analyses have found this advantage narrows or disappears when patients on tirzepatide stay at lower doses for cost reasons rather than escalating to where its advantage is most pronounced — another example of the trial-versus-real-world gap showing up in a different form.
What "reading honestly" means, put simply
None of what I've described here is about undermining trust in this data — quite the opposite. Understanding response variability, the trial-versus-real-world gap, and the actual relationship between side effects and discontinuation makes the data more useful, not less, because it lets you interpret your own experience against an accurate picture rather than an oversimplified headline number. This is the same standard I've tried to apply to every claim across this entire site — from the well-established GLP-1 efficacy data covered here, to the far more preliminary peptide research covered elsewhere, evidence deserves to be represented at the strength it actually has, not rounded up or down to fit a narrative.
Closing this series
This closes out the arc I set out to write — the Foundations pillar explaining the core biology, the Healthspan posts connecting inflammation, muscle, and emerging research, the clinical-reality posts drawing on real pattern recognition, and now, closing with the actual trial data behind the medication that's been a throughline across nearly all of it. I hope having all of it laid out this openly, caveats included, has been useful — that was the goal from the first post I wrote here.
Have questions about how your own results compare to what the trials actually showed?
No spam. Occasional new posts and updates from The Center for Metabolic Health.
Medical disclaimer: This content is provided for general educational and informational purposes only and does not constitute medical advice. It is not intended to diagnose, treat, cure, or prevent any condition, and it does not create a physician-patient relationship. Every patient's medical history, health status, and treatment needs are different. Always consult your own physician or qualified healthcare provider before starting, stopping, or changing any medication or treatment, and before making any decisions based on information found here. If you are experiencing a medical emergency, call 911 or go to your nearest emergency room.