fwiw
most of the sema-vs-tirz threads I read frame it as which one drops the scale faster, and that’s the wrong variable if you’re our age and trying to hold onto muscle. The actual difference on paper: tirzepatide adds GIP agonism on top of the GLP-1 action that sema runs alone. In animal work GIP receptor activity shows some independent effect on how fat vs lean tissue gets partitioned, which sema doesn’t replicate. Whether that translates to a meaningful lean-mass difference in real people is exactly what nobody posting these before-and-afters is measuring, because almost none of them have a DEXA baseline. Scale weight can’t tell you what came off. The part I’d gently push on with anyone choosing between them: the dose ramp is when GI side effects peak, and that’s also when slow-wave sleep takes the most damage, which is where most of your overnight growth hormone gets released. So some of the lean loss people blame on the compound may be a sleep confound, not a clean GIP-vs-no-GIP read. I logged training load, sleep and dose changes side by side for months. Exporting the lot as a PDF for my GP appt was what finally got her to take the pattern seriously. Get a baseline DEXA before you start. ymmv.
Slow-wave disruption was already my baseline after RYGB before tirzepatide entered the picture, which means the lean loss signal was impossible to read without logging dose timing and sleep in the same place. Adding a free-text note to each CareClinic entry next to dose date is what made the 11-week overlap actually visible rather than impressionistic.
logging dose timing and sleep in the same field is the right move, and “impressionistic rather than visible” is exactly the gap most people never close, so no argument on the build. where I’d push back is treating RYGB as just a sleep confound you solved by co-locating the data. bypass reroutes the foregut and changes incretin secretion directly, including endogenous GIP. so you’re not running tirz against a clean baseline the way someone with an intact gut is, you’re layering an exogenous GIP agonist on top of an already-remodeled GIP axis. that’s not a logging problem, it’s a different starting physiology, and it’s the part that makes your n=1 hard to read against anyone else’s. the one place it might actually cut in your favor: “slow-wave disruption was already my baseline.” if SWS was already compressed pre-tirz, the dose-ramp GH hit I described in the OP has less room to land on you specifically, because that variable was already pinned near the floor. which means the sleep axis is arguably more stable in your data than in someone going from intact sleep into the ramp, even while the GIP axis is murkier. two variables moving in opposite directions on cleanliness. what would tighten it: did you have any pre-tirz SWS marker, even a wearable estimate, to anchor “already my baseline” to a number? otherwise that’s the one piece still sitting on impression. not a dig, the 11-week overlap is more than most people log. ymmv.
leucine absorption post-RYGB is blunted and delayed, and the threshold for triggering muscle protein synthesis isn’t a slope, it’s a step function. some of the lean loss signal in these n=1s may be amino acid kinetics, upstream of either the GIP axis or the sleep axis this thread is working through.
the sleep confound is the right instinct but I’d watch the mechanism it leans on. “most of your overnight growth hormone gets released” during slow-wave is true, and steel-manning it, the ramp is genuinely when both GI and sleep disruption peak, so the timing lines up. where it gets shaky: the chain from a few weeks of fragmented ramp sleep to DEXA-measurable lean loss is long, and overnight GH pulses driving lean mass in adults over a titration window is a much bigger claim than “sleep got worse.” the thing I’d add is a third reader you skipped past, the intake floor. peak suppression stacked on a dose increase tanks protein 15-20g/day without you clocking it, and a few weeks below replacement reads as lean loss on a scan with no GIP effect and no GH story needed. that one’s actually testable without a sleep lab, just log protein against dose day. fwiw the dose-day reminder is the only reason I catch that suppression window before it eats my protein floor instead of after. baseline DEXA point stands though.
Baseline DEXA before you start is the right call, and “scale weight can’t tell you what came off” is exactly the gap. But a baseline only fixes the read if you’re actually starting fresh. Most people choosing between these are switching, sema first then tirz, and the second molecule gets measured against a body that already dropped weight on the first one. That’s a different partitioning baseline no matter which drug is doing the work, and the GIP-vs-no-GIP signal gets buried inside the sequence. The clean version is a DEXA at the switch itself, not just at the original start, and almost nobody has both draws. On the sleep confound, I’d run it the other direction too. The dose-ramp window where slow-wave sleep takes the hit isn’t the same calendar week on a switch as it is on a true start, so the timing you’d pin the lean loss to doesn’t line up cleanly either. I’ll flag my own window though, I’m sema only, compounded, so I’ve never run the switch and can’t speak to the tirz side from experience.
the sleep confound is the part i’d underline twice, because it’s the piece everyone skips. i actually can’t cleanly separate it in my own data either, GI-disrupted slow wave sleep on the dose ramp stacked on top of toddler-disrupted sleep, and there’s no untangling those two without a sleep lab i’m never going to get. so the overnight GH pulse argument cuts both ways: it’s real, but it’s also a class effect, not a tirz advantage. anyone on sema is taking the same hit to slow wave sleep during their ramp. it doesn’t help you pick between the two. where i’d push a little is the GIP-partitioning-to-DEXA leap. the animal work showing GIP shifting fat vs lean is interesting, but when you actually look at the human numbers, tirz users still lost meaningful lean mass, and the fat-to-lean split wasn’t dramatically cleaner than the sema trials once you account for people starting at different body comps. so the “GIP saves your muscle” story is doing more work in the marketing than the data supports. “scale weight can’t tell you what came off” is exactly right, and i’d extend it: the trial DEXA data we do have doesn’t show the gap the animal model predicts either. the PDF-for-the-GP move is underrated though. i log dose changes, weight and a rough symptom score and the visual trend lines were what got my OB to actually engage instead of nodding, you can see the ramp weeks line up with the bad stretches at a glance in a way a spreadsheet column never communicates. none of this is proof, the lean mass question genuinely isn’t settled for either compound at our doses. but if someone’s choosing between them expecting GIP to be muscle insurance, that’s not what the human data shows yet. baseline DEXA before you start is the one thing in your post i’d put in bold. ymmv.
the timing between GI peak and sleep decline during ramp’s tight, but correlation doesn’t tell you if GI symptoms caused the sleep hit or if the dose ramp itself was just stressing the system. ngl separating those needs more granular tracking than most people manage. DEXA baseline solves the lean mass variable though, and you’re right that’s the actual gap in the existing posts.
the sleep confound is the part I’d push on, not because the chain is wrong but because the magnitude isn’t there. each link is plausible (GI symptoms wreck SWS, less SWS means less pulsatile GH, less IGF-1) but when I went looking for the actual numbers, chronic sleep restriction in healthy adults only moves IGF-1 modestly, and the lean mass delta you’d get from that over a few months is hard to source at the size the GH framing implies. it reads bigger than it measures. the variable I’d put ahead of it is protein distribution. delayed gastric emptying plus the appetite drop quietly collapses a lot of people from 3-4 protein feedings down to 2, and the leucine threshold work caps MPS somewhere around 30-40g per meal for women our age. you can hit the same daily total and still under-stimulate synthesis. that’s a cleaner mechanical explanation for lean loss than the sleep pathway, and it’s actually loggable. same caution on the training side of your log. “logged training load” is the right instinct but load needs to be volume-load (sets x reps x load) to be comparable week to week, otherwise it’s a checkbox that says you trained, not that the stimulus was adequate. my DEXA showed ~2.1 lb lean loss at month 9 while I was lifting the whole time, just not at enough volume to matter, which is exactly the gap a coarse log hides. logging protein per-meal instead of per-day is annoying enough that I only stuck with it once the check-in flow was fast, but that’s the granularity where the answer actually lives. baseline DEXA, agreed, no argument there.
the sleep confound is a genuinely good one to stack here, and I don’t see it raised nearly often enough, so credit for that. The piece I’d add sits one step earlier though: even once someone gets the DEXA you’re asking for, a lean-loss number reads worse than it is until you put the calorie-restriction baseline next to it. In non-resistance-trained people, plain restriction lands somewhere around 25-30% of total loss as lean before you’ve added any compound at all. So a DEXA showing lean drop on tirz or sema isn’t automatically a GIP-vs-GLP-1 partitioning story, it might just be the restriction floor everyone hits. That’s the comparator most “not muscle-sparing” verdicts quietly skip. I log dose, sleep and training in one place in CareClinic, and being able to drop a short note on each entry is what let me see which weeks were actually the confounded ones. Get the DEXA, agreed, just hold the baseline beside it.