Both groups continued basal insulin
SURPASS-6 randomized 1,428 adults with type 2 diabetes receiving basal insulin. It compared weekly tirzepatide with mealtime insulin lispro, each added to insulin glargine, over 52 weeks. Treatment assignment was open-label. The primary comparison concerned HbA1c change, with tirzepatide doses pooled for the main estimate. [1]
This was an intensification question within diabetes care. It was not a study of people without diabetes choosing a weight-management treatment, nor a comparison of tirzepatide with all insulin therapy. The shared basal insulin is a defining part of what was tested.
Three outcomes worth keeping separate
Average HbA1c change was −2.1 percentage points with pooled tirzepatide and −1.1 with lispro. Weight changed by −9.0 kg and +3.2 kg. Hypoglycemia below the specified glucose threshold or severe hypoglycemia occurred at rates of 0.4 and 4.4 events per patient-year. Gastrointestinal symptoms were common with tirzepatide and usually mild or moderate. [1]
An event rate is not a percentage of participants. A person can have more than one event, and patient-years incorporate observation time. Relabeling 0.4 events per patient-year as 0.4% of people would be an error. The units deserve the same attention as the apparent size of the difference.
Noninferiority and superiority are different tests
The study first asked whether tirzepatide was not worse than the comparator by more than a specified margin for HbA1c change, and the reported result also met superiority. [1] That sequence concerns a defined glucose endpoint. It is not a blanket statement that one approach is preferable for every person or every outcome.
The open-label design and protocol-driven treatment management remain part of the evidence. A consumer should not use the comparison to infer how to substitute medicines, reduce insulin or reproduce the regimen. Those decisions depend on individual care rather than a research summary.
Pooling the tirzepatide groups means the central estimate summarizes several assigned regimens. It should not be quoted as an exact expectation for one specific dose. If a claim concerns an individual regimen, check whether the paper provides that separate analysis rather than inferring it from the pooled number.
Ask about the treatment strategy, not only the ingredient
Bring the trial to the clinician responsible for diabetes management and ask how its population and background care relate to your situation. Clarify which outcomes matter most and how they will be monitored. This encourages a discussion of the complete regimen instead of a simple winner-versus-loser framing.
For the placebo-controlled add-on study, see SURPASS-5. Our trial-reading guide explains why comparator and units are essential. The practical lesson is to preserve what the treatment was added to whenever an add-on trial is quoted.