Titration reference
Pinned in #dosing-titration. The left-hand column is what the trials did. The right-hand column is what this room does, and it is not the same thing.
The trial ladders
| Step | Semaglutide (weekly) | Tirzepatide (weekly) |
|---|---|---|
| Weeks 1–4 | 0.25 mg | 2.5 mg |
| Weeks 5–8 | 0.5 mg | 5 mg |
| Weeks 9–12 | 1.0 mg | 7.5 mg |
| Weeks 13–16 | 1.7 mg | 10 mg |
| Weeks 17–20 | 2.4 mg | 12.5 mg |
| Weeks 21+ | 2.4 mg (ceiling) | 15 mg (ceiling) |
Trials escalate on a fixed calendar because a trial has to: a protocol cannot say “step when it feels right”. That is a constraint of clinical research, not a finding about the best way to titrate.
What this room actually does
| Trial schedule | Typical here | |
|---|---|---|
| Step interval | 4 weeks, fixed | 4–12 weeks, driven by symptoms |
| Holding a dose | A protocol deviation | Normal, expected, frequently indefinite |
| Top dose | Reached by design | Often never reached |
| Coming back down | Not studied | Common, especially at maintenance |
| Splitting a dose | Never studied | A minority do it and say there is no trial |
Half-life and steady state
Semaglutide's half-life is roughly seven days. After about five weeks at a given dose you are at approximately 94% of steady state for that dose. Two consequences the channel repeats constantly:
- Judging a new dose after one week is judging an incomplete exposure.
- “It stopped working after four days” is almost never pharmacokinetics. Appetite returning late in the week is extremely common and is not the dose failing.
Escalate on the calendar, or on symptoms?
The room has argued this annually since 2024 and has not resolved it, which is the honest answer.
Calendar is what the trials did, it gets you to an effective dose faster, and it is what most prescribers follow. Symptoms is what most of this room does: hold while the current dose is still doing something, step when it is not, and treat side effects as information rather than as an obstacle. Both are defensible. What matters is knowing which one you are doing and why.
The most common rule of thumb posted here: if a side effect is still present at day five after a dose, hold; if it clears by day three, you have room to step.
Holding is not failing
Most members who are still here after two years held at least one dose for two months or more, and several never went above the middle of the ladder. Going up spends headroom you may want later. If you are asking the channel whether to escalate, the answer that comes back most often is “not yet”.
Coming down, and maintenance
Reducing on purpose — for side effects, for maintenance, or because the dose above bought nothing — is a completely legitimate move and this channel says so more often than most places do. Members who have gone from a treatment dose to a maintenance dose over two or three months describe it as uneventful.
Lower-frequency maintenance — every ten days, fortnightly — is done by a minority here and has no trial behind it at all. Members report on it and that reporting is data, but it is not evidence. On stopping altogether, the trial to read is STEP 4 (JAMA 2021) for semaglutide and SURMOUNT-4 (JAMA 2024) for tirzepatide; both randomised withdrawal designs, and the regain in the withdrawal arms was not subtle.
Switching between compounds
There is no clean dose conversion between semaglutide and tirzepatide. Anyone who gives you one has made it up — the archive contains an unsourced conversion graphic that circulated for months before somebody traced it to nowhere. The room's position is to restart the ladder low on the new compound and titrate from there.
Gaps and missed doses
A missed week does not reset your titration. Going straight back to the top of the ladder after a supply gap of more than about three weeks is, however, how people get floored — several members restart one rung lower after a long gap. That is a personal rule, not a guideline.
Dose-day drift
With a weekly compound, a day or two of drift is mostly harmless. It still tends to creep: one member drifted from Sunday to Wednesday across a year and only noticed on checking their own log. Pick a day, write it down, and defend it.
Trials referenced on this page: STEP 1 (NEJM 2021), STEP 4 (JAMA 2021), STEP 8 (JAMA 2022), SURMOUNT-1 (NEJM 2022), SURMOUNT-4 (JAMA 2024), SURPASS-2 (NEJM 2021), SELECT (NEJM 2023).