This gets cited here weekly, usually second-hand, so it is worth setting out what it does and does not establish.
Steady state is the thing most people miss. The terminal half-life is about a week, so every dose step takes four to five weeks to fully express itself. Judging a step at day ten is judging the ascent, not the plateau, and it is the single commonest reason people escalate before they needed to.
Where I think it is weakest: the completion rate deserves as much attention as the headline, because a large effect among those who finished is a different claim from a large effect among those enrolled.
The question I want answered is what the dose-response curve actually looks like above 1.7mg, because the trial means hide how few people account for the extra loss. I would rather have one careful answer than five confident ones.
Note on sourcing:
Figures above are from the primary publication rather than the press summary. If a number here disagrees with one you have, post yours and we will work out which of us is reading a secondary source.