A trial reported two positive outcomes and then told its readers not to lean on them [1]. That sentence is the reason to read it. The authors write that, given the underpowered nature of the study, the findings should be interpreted with caution — a caveat this site usually has to supply itself, as it did in the trial designed to tie.
The design is genuinely interesting. Patients with type 2 diabetes who had a minor ischemic stroke or a high-risk transient ischemic attack were randomized within 24 hours of symptom onset to receive liraglutide for 90 days on top of guideline-directed standard care, or standard care alone. Dosing escalated from 0.6 mg to 1.2 mg to 1.8 mg over the first two weeks.
Stroke recurred in 25 of the liraglutide patients (7.9%) and 44 of the control patients (13.8%), a hazard ratio of 0.56 with a confidence interval from 0.34 to 0.91. Functional outcomes followed: 87.3% of the liraglutide group reached a modified Rankin Scale score of 1 or less, against 77.8% of controls, an odds ratio of 1.95. Stroke prevention is a different claim from stroke treatment, a distinction this site drew in why preventing something is not treating it.
On safety, the abstract reports that symptomatic intracranial hemorrhage and all-cause mortality were low and similar between groups, without giving the counts. That is what the published summary says and this site will not supply numbers it did not print. Participants were Chinese and the authors frame their conclusion that way, so the absolute recurrence rates belong to that population.
What to do with a trial like this is the harder question. An underpowered positive result is the classic setup for an effect that shrinks on replication, and 636 patients is small for a stroke recurrence endpoint. It is also the first dedicated randomized test of this drug in acute stroke, and the direction agrees with the cardiovascular cohort evidence this site covered in the cohort that tested its own bias. Worth watching, not worth acting on alone.