Every few months, a new head-to-head comparison of tirzepatide and semaglutide makes the rounds, and the headline is almost always some version of “Drug A beats Drug B.” Company-funded head-to-heads tend to frame that comparison in whatever light favors the sponsor’s own product. This one is different in a useful way: it’s an independent meta-analysis with no pharmaceutical funding, it pools ten separate studies instead of one, and it reports a real tradeoff instead of a clean winner.
Published September 9 in Clinical Obesity, it found tirzepatide produces meaningfully more weight loss than semaglutide. It also found tirzepatide comes with close to double the rate of serious adverse events. Both of those things are true at the same time, and neither one is the whole story.
What the analysis did
Researchers pooled ten head-to-head studies — three randomized controlled trials and seven retrospective cohort studies — covering 41,381 adults with overweight or obesity. Follow-up ranged from 24 to 72 weeks depending on the study. Everyone included had a BMI of 25 or higher; people with type 1 diabetes, a history of bariatric surgery, or active cancer were excluded.
This is a meta-analysis, not a single trial, which means it’s only as good as the studies feeding it — more on that below.
What it found
On weight loss, the gap was substantial:
- Tirzepatide produced 4.28 percentage points more weight loss than semaglutide, and 4.43 kg more in absolute terms
- Patients on tirzepatide were 43% more likely to lose at least 10% of body weight, 85% more likely to lose at least 15%, and roughly twice as likely to lose at least 20% — though at the 5% threshold, the two drugs performed about the same
- Tirzepatide also produced a slightly larger drop in A1c (a marker of average blood sugar), by 0.29 percentage points
On safety, the picture flipped: serious adverse events were reported in 5.7% of tirzepatide patients versus 2.9% of semaglutide patients — a rate ratio of 1.83. The paper doesn’t break down what counted as “serious” beyond the standard clinical definition (an event requiring hospitalization, causing lasting harm, or otherwise medically significant), which limits how much we can say about what the excess risk actually was. Notably, the two drugs had similar rates of people stopping treatment because of side effects — so whatever was driving the serious-event gap wasn’t simply more people quitting in discomfort.
What it doesn’t establish
Most of the data is observational, and the authors say so. Three of the ten studies were randomized controlled trials; the other seven were retrospective cohorts, and those seven contributed the bulk of the 41,381 patients. In a retrospective cohort, the people prescribed tirzepatide and the people prescribed semaglutide weren’t randomly assigned — a clinician chose between them based on factors like cost, insurance coverage, prior treatment history, and how sick someone already was. Some of that “excess” adverse event rate could reflect who tends to get prescribed which drug, not a property of the drug itself. The authors flag this directly: residual confounding “could not be ruled out.”
The studies didn’t agree with each other. The paper reports substantial heterogeneity across several of the efficacy outcomes, meaning the ten underlying studies didn’t produce consistent effect sizes. A pooled number can look precise while papering over real disagreement in what it’s pooling.
72 weeks is not long-term. The longest follow-up in any included study was 72 weeks — a year and a half. That’s enough time to compare early weight trajectories, but it’s too short to say anything about durability, rare late-onset side effects, or what either drug’s risk profile looks like at three or five years.
This is a comparison of branded, manufactured drugs. Every patient in every study was taking tirzepatide or semaglutide manufactured by the branded pharmaceutical companies that hold FDA approval for those products, at their standard concentrations and formulations. It says nothing about compounded semaglutide or tirzepatide, which NoBsRx provides provider-guided access to. Compounded preparations are made by state-licensed pharmacies for an individual patient and have not been studied in a trial like this one. A safety comparison between two branded, FDA-approved products doesn’t transfer to a compounded version of either. We’re not going to imply otherwise.
The part that’s easy to miss
A rate ratio of 1.83 sounds dramatic until you sit with the actual numbers: 5.7% versus 2.9%. Framed as absolute risk, that’s an increase of under three percentage points — real, and worth weighing, but a different picture than “almost double the risk” implies on its own. Relative and absolute risk are both true; they just answer different questions, and headlines tend to reach for whichever one sounds bigger.
None of this tells you what will happen if you take either medication. Average effects in a pooled analysis — even a well-run one — describe a population, not a person. Whether the added weight-loss benefit of one drug is worth its added risk profile, for your specific health history, is a clinical judgment that belongs to you and a licensed provider who can weigh your labs, your other conditions, and what you’re actually trying to accomplish. Our process starts with that kind of individualized clinical review before any medication is chosen, and program pricing is published up front rather than decided by a pooled average.
This post is general information, not medical advice. It does not describe a benefit or outcome of any NoBsRx program, and it is not a claim that either medication is safe, effective, or appropriate for any individual. Compounded medications are not FDA-approved. Treatment is never guaranteed — eligibility, diagnosis, and every treatment decision are made by independent licensed providers based on an individual clinical assessment, and availability varies by state. Talk to a licensed clinician before starting, stopping, or changing any medication.