A placebo arm does not necessarily mean participants received no active treatment. In SURPASS-5, background insulin is essential to interpreting the comparison. Primary report
Study snapshot
The 40-week randomized trial enrolled 475 adults with inadequately controlled type 2 diabetes using insulin glargine, with or without metformin. Tirzepatide or placebo was added to titrated glargine. HbA1c and weight outcomes favored the tirzepatide groups. Although 451 participants completed the trial, treatment discontinuation was more frequent in the tirzepatide groups than in the placebo group. Nausea and diarrhea were among the reported adverse events. Primary report
Our appraisal: two questions hidden in one result
First ask what differed between groups. Here, the randomized add-on differed; the existence of background insulin treatment did not. Calling the comparator “no treatment” would misrepresent the experiment and invite unsafe extrapolation.
Then ask what happened after participants stopped their assigned study drug. Remaining in follow-up and continuing treatment are separate events. A trial can have good follow-up completion while still having meaningful differences in drug discontinuation. A complete interpretation should explain whether outcome collection continued and how the analysis handled those observations.
For illustration, imagine a participant who stops an add-on because of nausea but returns for the final assessment. They can be a study completer and a treatment discontinuer simultaneously. This is an explanatory example, not an account of a particular participant.
Before publication, review insulin-adjustment rules, rescue treatment, the estimand, and the reasons for stopping treatment. These details help determine what the average result represents in practice.
Frequently asked questions
Was placebo the only treatment in the comparator arm?
No. Participants also received background glargine.
Was this a monotherapy study?
No. It tested an add-on strategy.
Is study completion the same as continued drug use?
No. Follow-up can continue after drug discontinuation.
Were tolerability events reported?
Yes, including nausea and diarrhea.
Can this report justify stopping basal insulin?
No. That is not the randomized comparison.
Why inspect the estimand?
It clarifies how events such as treatment discontinuation relate to the reported effect.
What source supports this summary?
The primary indexed abstract; complete methods and tables remain to be reviewed.