Tirzepatide vs semaglutide for A1c: read SURPASS-2 separately from weight-loss trials

Start with the question the trial actually asked

If you search for tirzepatide versus semaglutide, the results can make separate trials sound like one contest. SURPASS-2 asked about blood-glucose control in adults with type 2 diabetes whose diabetes was inadequately controlled with metformin. Its main comparison was the change in glycated hemoglobin, usually called HbA1c or A1c. I would keep that purpose beside every result. It is useful evidence for a diabetes discussion, but it cannot independently settle which obesity treatment, payment arrangement or telehealth practice is right for someone without diabetes.^[1] For the separate obesity comparison, read our SURMOUNT-5 analysis rather than merging the study populations.

Who was studied, and for how long

The open-label phase 3 study randomized 1,879 participants and followed the treatment comparison for 40 weeks. Tirzepatide was studied at weekly doses of 5, 10 or 15 mg, against semaglutide 1 mg weekly. The paper reports a mean starting A1c of 8.28%. Those details describe the research setting; they are not instructions for starting or changing medication. Open-label means the treatment identity was known, which is relevant when assessing expectations and reported experiences. The specified comparator dose also places a clear limit on the comparison.^[1]

Read the A1c result in percentage points

In the original publication analysis, mean A1c reductions were 2.01, 2.24 and 2.30 percentage points for the three tirzepatide groups and 1.86 points for semaglutide. A reduction from 8% to 6% is two percentage points; it is not simply a 2% relative decrease. The trial found tirzepatide noninferior and superior on its primary A1c comparison. That describes a group result under the study conditions. A clinician still needs to interpret the starting value, treatment target, other medicines and tolerability for an individual.^[2]

Why an A1c result is not a daily glucose reading

NIDDK explains that A1c reflects average blood glucose over roughly the preceding three months. It differs from a glucose reading taken at one moment. Averages also do not describe every high or low during that period, and some medical conditions can affect interpretation. For an appointment, I would bring the dated laboratory report rather than a remembered number. Ask whether any other results or monitoring are needed to interpret it. A trial mean should not become a promised laboratory target for someone whose clinical history is different.^[3] Combination treatment adds a separate issue: GLP-1s and hypoglycemia explains why the accompanying medicines matter.

Keep weight outcomes in their own column

SURPASS-2 also measured body weight, but kilograms and A1c percentage points are different endpoints. A table that blends the two into a single effectiveness score hides which outcome is being judged. Likewise, the longest numerical reduction is not a complete account of side effects, treatment burden or sustained use. When I compare study summaries, I first mark the population, endpoint and time point. If those three items differ, I do not treat the figures as a direct ranking, even when the medicines have familiar names.

Do not use this study to select your own dose

The research doses identify what was tested. They do not supply a conversion chart between medicines, a restart schedule or a reason to pursue the highest dose. Ask the prescriber how the proposed regimen relates to the evidence and what should happen if it is difficult to tolerate. If a commercial program advertises a treatment change as an upgrade, ask what clinical question the change is supposed to answer. An offer with a larger number on its label does not establish a better fit. Before changing treatment, use our semaglutide-to-tirzepatide switching guide to organize the discussion, not to calculate a dose.

Separate clinical choice from insurance and service choice

After a clinician identifies an appropriate option, there is still a purchasing question: which practice manages follow-up, which product is prescribed, and what will the patient actually pay? A diabetes trial cannot establish a particular telehealth company’s ability to provide a product or obtain insurance approval. I would ask the practice to distinguish the visit charge, membership and pharmacy expense in writing. If a quote assumes insurance coverage, ask what happens if that assumption is wrong. Research evidence and an insurer’s benefit decision are separate documents. Our LifeMD review examines service and payment questions; it does not show that a particular prescription will be covered.

The comparison I would take to an appointment

Bring three short questions: does this study resemble my situation, which outcome are we trying to improve, and how will we judge whether the plan is working? Keep any relevant A1c report with the question list. If you are comparing service quotes afterward, retain the same medication and payment scope in each budget. The useful conclusion is narrower than a universal winner: SURPASS-2 provides a defined diabetes comparison. Its strengths come from knowing exactly what was compared, and its limits remain important when moving from the paper to a personal decision. If two summaries use different analysis definitions, keep those definitions attached to their numbers. Rounding is not the only reason figures can differ, and silently averaging them would create a result the trial never reported. Our complete-cost guide helps keep the service comparison separate from the clinical one.

Explore relevant provider reviews

Use these reviews to compare the published services and terms. Inclusion is a research starting point, not a recommendation for treatment.

Continue your research

• Semaglutide and smoking: craving changes are not proof of quitting
• Semaglutide and NAION: rare vision risk, absolute numbers and urgent symptoms
• Mounjaro and heart health: what SURPASS-CVOT proved against dulaglutide