SURPASS-2: Tirzepatide Outperforms Semaglutide 1 mg in Head-to-Head Diabetes Trial

In the first large head-to-head trial of two weekly injectable incretin-based drugs, all three doses of tirzepatide lowered blood sugar and body weight more than semaglutide 1 mg — with a similar side-effect profile.

What the researchers did

SURPASS-2, led by Juan Pablo Frías and published in the New England Journal of Medicine in 2021, was the first large trial to compare two weekly injectable incretin-based therapies directly against each other. Tirzepatide activates both the GIP and GLP-1 receptors; semaglutide targets GLP-1 alone. The biological question — does dual agonism add anything over GLP-1 agonism — was also a commercial one, since both drugs compete for the same patients.

The trial enrolled 1,879 adults with type 2 diabetes whose blood sugar was inadequately controlled on metformin alone, at sites across North America, South America, Europe, and Asia. At baseline, participants had a mean HbA1c of 8.28 percent — clearly above target despite metformin. They were randomized to tirzepatide at 5, 10, or 15 mg, or to semaglutide at 1 mg, each injected once weekly for 40 weeks. Tirzepatide was started at a low dose and escalated in steps to the assigned maintenance dose, a strategy designed to improve gastrointestinal tolerability.

The trial was open-label: patients and investigators knew which drug and dose they were receiving. The primary endpoint was the change in HbA1c — the three-month average of blood glucose — from baseline to week 40. Eli Lilly, the maker of tirzepatide, funded the study.

What they found

All three tirzepatide doses reduced HbA1c more than semaglutide 1 mg. The estimated reductions were 2.01 percentage points with tirzepatide 5 mg, 2.24 points with 10 mg, and 2.30 points with 15 mg, compared with 1.86 points for semaglutide. The differences were statistically significant at every dose, meaning even the lowest tirzepatide dose beat the comparator. Consistent with the average reductions, a greater share of participants on tirzepatide reached the common treatment target of an HbA1c below 7 percent, and more reached near-normal glucose levels, than on semaglutide.

Body weight followed the same pattern. Participants on tirzepatide lost between 1.9 kg (at the 5 mg dose) and 5.5 kg (at the 15 mg dose) more than those on semaglutide, with a clear dose-response gradient across the three tirzepatide arms.

The most common adverse events were gastrointestinal — nausea, vomiting, and diarrhea — and they were broadly similar between the groups, occurring mainly during the dose-escalation period and easing over time. Discontinuation because of adverse events was uncommon overall, though somewhat more frequent at the higher tirzepatide doses than with semaglutide. Hypoglycemia was rare in all arms, as expected for drugs that act glucose-dependently in patients not on insulin or sulfonylureas.

Limitations

Three caveats frame how these results should be read.

First, the open-label design. Because patients and clinicians knew the treatment assignments, expectations and behavior could have influenced outcomes, particularly subjective ones like reported side effects and dietary habits. Objective endpoints such as HbA1c are less vulnerable to this bias, but not entirely immune to it.

Second, the comparator dose. Semaglutide was tested at 1 mg, the dose approved for type 2 diabetes at the time — but a 2 mg dose has since been approved, and higher doses are used for obesity treatment. SURPASS-2 therefore does not answer how tirzepatide compares against semaglutide at its maximum approved dose.

Third, duration. Forty weeks is long enough to establish glycemic and weight effects, but not long enough to assess cardiovascular outcomes, durability of weight loss, or rare adverse events. Dedicated cardiovascular outcome trials are the appropriate instrument for those questions.

What this means

Within its limits, SURPASS-2 established that dual GIP/GLP-1 agonism lowers HbA1c and body weight more than GLP-1 agonism alone at the doses studied, without a meaningfully worse gastrointestinal burden. The trial was a pivotal step toward tirzepatide’s regulatory approvals for type 2 diabetes and anchored the broader SURPASS program, which has since compared tirzepatide with other standard therapies and in earlier-stage disease.

It does not, on its own, tell any individual patient which drug is right for them. Tolerability, cardiovascular profile, cost, availability, and coexisting conditions all enter that decision — one to make with a physician, not from a single trial’s headline numbers. And for readers without diabetes: these results come from a population with established type 2 diabetes on metformin and should not be extrapolated to weight loss in metabolically healthy people.