Most papers report a relative risk and leave you to work out what it buys. This one does the arithmetic and prints the answer, which turns out to be the most honest thing in it — the step most results never take.
The setup
Trigger finger is a thickened tendon catching as it slides through a pulley in the palm. Diabetes makes it commoner. The first treatment is a steroid injection, and when that fails the next step is surgery to release the pulley.
Adults with type 2 diabetes who had that injection were pulled from a database. [1] Of 69,222, some 5,871 were on a GLP-1. Matching left 5,863 per arm.
What happened
At six months, nothing. 6.1% of GLP-1 users had gone on to surgery against 6.8% without.
At twelve months the gap opened: 10.6% against 11.9%, relative risk 0.89, 95% CI 0.81 to 0.99, p = 0.027. At twenty-four months, 14.2% against 15.7%, relative risk 0.90, 95% CI 0.83 to 0.98, p = 0.016.
How thin it is
The twelve-month interval runs to 0.99. One more surgery in the treated arm and it crosses.
The six-month result was flat, so the effect only appears later, which is either a real slow-acting benefit or the kind of pattern that shows up when you look at three timepoints. The authors call it small and statistically significant, and both halves of that are doing work.
Who this is about
Adults with type 2 diabetes who already had a steroid injection in a finger. That is a narrow group.
Nobody without diabetes is in it, and nobody who had not already reached the injection stage. The same limits apply here as to the foot outcomes in that population, and the exposure is a prescription record rather than a randomization, which is what every surgical database study on this site has in common.
Why it is here
Not as a reason to buy. Nobody should take a GLP-1 to keep a finger out of an operating room.
It is here because a number needed to treat is the unit a buyer can actually think in, and almost nothing published about these drugs gives you one.