ATTAIN-1 oral GLP-1: is 12.4% at 36 weeks apples-to-apples?

I read it as treatment-regimen estimand, not completers. The placebo arm also lost some, so the 11% placebo-adjusted is the number I care about.

Where are you seeing 11% adjusted? The topline I saw only gave 12.4% versus placebo, no arm-level table yet.

That A1c drop in 14 weeks is interesting, but oral vs injectable comparisons get messy because titration and adherence differ wildly.

Exactly. If the oral arm has a 6% GI discontinuation rate, then the completers analysis will flatter it. ITT is the honest one.

Not sure the GI rate is really that high. The release says 6.1% overall, but most of those were mild and only 1.8% discontinued? Want to see the supplement before I buy that number.

36 weeks is short for a weight med. Apples to oranges.

Short trials are common for phase 3 topline, though. They usually follow with a 52-week extension. I'd wait for the full paper before drawing conclusions.

The apples-to-apples issue is the estimand. Some sponsors report treatment-regimen for maintenance doses, while older trials report completers. If ATTAIN-1 used treatment-regimen and still hit 12.4%, that's strong. But the placebo-adjusted difference is what matters because placebo groups in these trials often lose 2-4% from lifestyle support. Also, baseline BMI and diabetes status matter. Without those, the number is just a headline. I'd wait for the full table before comparing it to your 8.6% at 14 weeks.

The part people forget: 8.6% at 14 weeks is already above average for that timepoint — that's not nothing. A 36-week oral headline is a whole different timepoint, and it shouldn't be the thing messing with people's heads.

I disagree that 36 weeks is too short. It's enough to see the curve separating, and the GI dropout is the real signal. If people can't stay on it, efficacy is theoretical.

Fair, but adherence in a trial with weekly check-ins isn't real world. The oral pill might be easier for needle-averse folks, or harder if nausea hits daily.