Quick number for context: I’ve been tracking my CGM response on injection day vs day 6 (trough) for about 7 months. The delta tells you more about whether your dose is actually working than the scale does. Saw a post today about someone hitting goal weight on starter dose the whole time, food noise completely gone. Which made me want to write out something I don’t see discussed enough: the trough concentration argument for not titrating. Tirzepatide’s half-life is roughly 5 days. Weekly dosing means by day 5-6 you’re at your lowest plasma concentration before the next injection. At 2.5mg, that trough is lower than at 5mg or 7.5mg – obviously. But the relevant question isn’t “how low is the trough” in absolute terms. It’s “is food noise returning at trough?” Most titration logic is calendar-based: 4 weeks at X, move to X+2.5. That protocol exists because the trials were designed that way. It isn’t a guarantee that you personally need the next increment. Here’s the trough test I actually use:
- Days 1-3 post injection: appetite suppression obvious, doesn’t tell you much
- Days 5-7: this is the signal window
- Is food noise returning? Are meals feeling harder to stop? CGM showing more post-meal excursions? If days 5-7 feel basically the same as days 1-3 from a behavioral standpoint, you’re probably at therapeutic threshold on that dose. Titrating further adds GI exposure and cost without adding proportional benefit. The person who stayed at 2.5mg and lost 30 lbs is a clinical example of trough still doing meaningful work. The mistake I see is people titrating to 7.5mg or 10mg bc “that’s what the schedule says” when their trough at 5mg was already suppressing food noise effectively. I can’t separate this from individual variation in GIP/GLP-1 receptor sensitivity – some people genuinely need higher doses to get the same trough effect. But that’s an argument for checking the signal, not for defaulting to escalation. Track your days 5-7. That’s your real dose data.
suppressed food noise at trough doesn’t necessarily mean it stabilized though - one’s the medication, one’s learning. they could feel identical in days 5-7 but mean different things downstream.
The distinction holds for self-reported food noise, but the CGM side of the trough signal is harder to attribute to behavioral learning. The case for your point is real. Months of appetite suppression do train behavior. You learn what satiety feels like, you internalize new portion habits, and by month 4 or 5 that conditioning could look pharmacologically indistinguishable on subjective days 5-7 reports. I don’t dismiss that. If someone is tracking trough solely through hunger ratings, the “is this the drug or is this learned” problem is legitimate. But gastric emptying rate and incretin-mediated insulin secretion do not adapt through conditioning. When my Libre 3 shows the same attenuated post-meal curve on day 6 as day 2, that’s not a learned response. The glucose delta doesn’t know what day of my injection cycle it is. A flat CGM line at trough is pharmacological activity, not habit formation. So “feel identical but mean different things downstream” is accurate if the trough signal you’re relying on is subjective. It’s less accurate if you’re tracking objective glucose response. The CGM is specifically what makes the trough signal usable as dose data rather than just vibes. Without it, your concern about attribution is well-placed.
‘A flat CGM line at trough is pharmacological activity, not habit formation’ - sure, but activity being there doesn’t mean it has to stay that tight to keep behavior stable. which is still the actual question in the trough test. glucose flatness at day 6 proves the drug’s working. doesn’t prove you need suppression that tight to stay stable. you still have to know - if the pharma loosened up, would your learned satiety actually hold? the CGM tells you what’s happening day 2 versus day 6. not whether your behavior would if it didn’t.
anyway.