Boundary first: nothing here is medical advice. Dosing, side effects, whether a stall matters clinically, whether a medication is right for you — those are questions for the clinician who knows your case. This page is about one thing only: how to read a bathroom scale during a period of fast, intentional change. Trendline is an iPhone app that reads Apple Health; it measures nothing about medications and offers no guidance about them. Our health limitations page states these boundaries in full.
Fast change doesn't cancel daily noise
A morning reading is still one observation of total mass — water, food in transit, timing, and conditions included. Losing quickly doesn't switch that off; it just means the real signal is large enough that you'll sometimes see it through the noise anyway. The trap runs the other direction: after weeks of visible drops, a normal two-pound overnight bounce feels like failure, when it's the same fluctuation described in why daily weight changes when the plan is working — just landing on more sensitized eyes.
Trajectories flatten. Trials say so.
In the largest published trials of these medications, average weight curves fall steeply early and flatten over time — in the 68-week semaglutide STEP 1 trial, mean loss accumulated rapidly in the early months and plateaued toward the end of the trial period (Wilding et al., NEJM 2021). That's an average trajectory, not your promise — but it means a slowdown after months of fast movement is the documented shape of the curve, not necessarily a sign something broke. What a statistical trend view adds: it distinguishes "the trend has genuinely flattened" from "three noisy mornings in a row," which morning-to-morning comparison cannot do.
The same trial followed 327 of those participants for a year after treatment stopped. That extension is the part worth planning around, and it is rarely the part quoted: about two thirds of the loss came back.
| Timepoint | Mean change from baseline |
|---|---|
| Week 68 — treatment ends | −17.3% |
| Week 120 — one year after stopping | −5.6% |
Those are the two measured endpoints, not a curve. The trial reports where people were at each timepoint, so drawing a line between them would show a path nobody measured. What the numbers do say is that stopping is a transition with its own shape, and it is worth having a plan for it before you are in it — the TDEE calculator has an off-ramp planner that ramps intake back to maintenance a stated amount each week. None of this is medical advice, and whether or when to stop is a conversation with your clinician (Wilding et al., 2022, STEP 1 extension).
Composition estimates deserve extra caution right now
If your smart scale estimates body fat or muscle, remember what it's actually doing: passing a small current and inferring composition from resistance — an estimate that hydration alone can move (consistency, not accuracy). Periods of fast weight change and appetite change are exactly when hydration and glycogen are least stable, so treat composition percentages as rough context, not as a verdict on lean-mass retention. Questions about preserving muscle during rapid loss are clinician-and-training questions; a bathroom scale's estimate isn't the instrument that settles them.
A reading rhythm that survives the emotional weeks
- Weigh under consistent conditions — same scale, similar time — so readings stay comparable with each other.
- Judge direction on the trend, not the morning. A smoothed trend absorbs the bounce and shows whether change is holding; when your data is sparse, an honest tool says so instead of projecting confidence. The math Trendline uses is public on the methodology page.
- Let plateaus be data, not indictments. A flat trend across weeks is information worth bringing to your clinician — calmly, with the record in hand — rather than a nightly verdict on you.
Where your readings should live
However you weigh — smart scale sync or hand-entered readings — Apple Health is the neutral store that keeps your record independent of any one app. If a scale app frustrates you, the record can outlive it.
When Trendline launches, it will read your weight, body-fat, and lean-mass history from Apple Health and show the trend with its uncertainty stated — including the honest "not enough to say" when readings are sparse. Until then, the free weight trend calculator draws the same kind of slower signal from pasted readings, in your browser.
Try the idea
Paste your last several weeks of readings into the weight trend calculator. During fast change the gap between the jumpy raw line and the steady slower one is at its most vivid — and so is the difference in how the two feel to look at.
Sources
- Wilding et al. (2021), Once-weekly semaglutide in adults with overweight or obesity (STEP 1), NEJM
- NIDDK repository: PRIDE Lifestyle Behavior Change Manual (fluctuation literacy)
- Jeong et al. (2023), acute hydration effects on bioelectrical-impedance body composition
This page describes measurement interpretation only and was reviewed on July 20, 2026. It is not medical advice, and Trendline is not a medical device.