Lilly scheduled a tirzepatide symposium at this year’s European diabetes meeting (EASD, September 30) built around a large cardiovascular trial called SURPASS-CVOT. The trial was published in the New England Journal of Medicine, and its headline has been summarized in a few different ways, some of them more generous than the numbers allow.
Here is what it found, and the statistical distinction that decides how much to read into it.
What the trial did
SURPASS-CVOT enrolled adults with type 2 diabetes and established atherosclerotic cardiovascular disease, meaning they already had heart or vascular disease. About 13,000 people were randomized to weekly tirzepatide (titrated up to 15 mg) or weekly dulaglutide (1.5 mg), another GLP-1 medication, and followed for a median of about four years. Average age was 64, average BMI was about 33, and 29% of participants were women.
The primary outcome was the first occurrence of cardiovascular death, heart attack, or stroke.
What it found
The primary outcome occurred in 12.2% of the tirzepatide group and 13.1% of the dulaglutide group, a hazard ratio of 0.92. The 95% confidence interval ran from 0.83 to 1.01.
That interval crosses 1.0, which is the line for “no difference.” So the trial met its goal of showing tirzepatide was noninferior to dulaglutide, but it did not show superiority (p = 0.09). Gastrointestinal adverse events were more common with tirzepatide, at 42.5% versus 35.9%, while overall adverse event rates were similar.
What “noninferior” actually means
A noninferiority trial asks a narrower question than most headlines assume. It asks: is the new treatment not meaningfully worse than the comparator, within a margin the researchers chose in advance? Here the margin was a hazard ratio of 1.05. Passing that test is a statement about cardiovascular safety. It is not a statement that one drug protects the heart better than the other.
Some coverage leaned on the 0.92 point estimate to say tirzepatide “bested” dulaglutide. The confidence interval is the honest summary: the data are compatible with a modest benefit and with essentially none.
Secondary results, including all-cause mortality, weight, and kidney measures, were reported as favoring tirzepatide, but they were not tested under the same strict statistical controls as the primary outcome. They are worth watching, not worth treating as settled.
What the trial does not establish
It did not include a placebo group. Both arms received an active GLP-1-class medication, so the trial cannot tell you how much either one lowers risk compared with no treatment.
The comparator was a submaximal dose. Some commentators noted that dulaglutide 1.5 mg is not the highest dose available, which can matter when a comparison is close.
The population is specific. Participants had type 2 diabetes and existing cardiovascular disease, with an average diabetes duration of about 15 years. The results do not describe people without diabetes, people without established heart disease, or people using a medication for weight management. The authors also point to limited global diversity among participants and some imbalances between groups.
Trial averages are not individual forecasts. A hazard ratio describes a group. Whether any given person’s cardiovascular risk changes on a medication depends on their history, other treatments, and clinical context, which is a conversation for a licensed provider.
The part that applies to compounded medications
SURPASS-CVOT studied branded, manufactured tirzepatide made under an approved application, in a patient population defined by diabetes and heart disease. That evidence covers that product and does not transfer to a compounded preparation.
Compounded semaglutide and tirzepatide are prepared by state-licensed pharmacies for an individual patient. They are not FDA-approved products, and none of them has been through a cardiovascular outcomes trial like this one. If a company cites SURPASS-CVOT as a reason to buy a compounded product, it is stretching the study well beyond what it tested.
Why a null-ish result is still useful
A trial that comes back “noninferior” can still be informative. For people with diabetes and heart disease, it is reassuring that a newer, more potent drug did not carry a cardiovascular penalty compared with an established one. That is a safety finding, and a reasonable one.
It just isn’t a reason to expect more. If you want to know how treatment decisions get made and what is and isn’t covered by evidence, our FAQ and how it works pages lay out the process, including the point that every prescribing decision belongs to an independent licensed provider.
This post is general information, not medical advice. It does not describe a benefit of any NoBsRx program and is not a claim that any medication will protect your heart or improve any health outcome. Compounded medications are not FDA-approved. Treatment is never guaranteed. Eligibility and every treatment decision are made by independent licensed providers based on an individual clinical assessment, and availability varies by state. Do not start, stop, or change any treatment based on an article. If you have heart disease or diabetes, talk to a licensed clinician.