seeing another week 9 sema post with strong scale numbers and genuinely curious whether there’s a DEXA behind them, because there rarely is. sema and tirz both hit GLP-1. tirz adds GIP agonism. GIP receptor activity in animal models shows some independent effects on fat vs lean tissue partitioning that sema doesn’t replicate. whether that translates to clinically meaningful differences in humans at clinical doses is not settled, but it’s also not what most people are tracking. ran sema 1mg for about 8 months, then switched to tirz (5mg, now 10mg) after getting a baseline DEXA. sema run: roughly 18lb total loss. DEXA split came out about 72% fat / 28% lean. not catastrophic but not what I’d call muscle-sparing for a woman lifting 3x per week. tirz is 5 months in, similar pace of loss compared to sema at the same timeframe. DEXA in 6 weeks. the variable people aren’t accounting for on either compound: nocturnal GH pulse. poor sleep suppresses it, and GI side effects plus appetite timing both disrupt sleep on GLP-1 class drugs. that’s not sema-vs-tirz, that’s a class effect that hits lean mass regardless of which you’re on. tracking weight without tracking sleep means you’re only reading one direction of that relationship. been logging weight, sleep quality, and estradiol dose together in one place. handed my endo a PDF export from the CareClinic health tracker at my last appt and she actually worked through it instead of just glancing at my chart. first time that’s happened. the comparison worth making: body composition at 6+ months, controlled for protein and calories. scale weight at week 9 is noise.
the nocturnal GH pulse piece is the thing I keep wanting to tattoo somewhere because nobody’s controlling for it. even before the sema/tirz question, GI disruption patterns alone shift sleep architecture in ways that compound over weeks. “scale weight at week 9 is noise” should be the flair on every progress post in this sub.
the “GI disruption patterns compound over weeks” framing is the part that makes the sema-vs-tirz comparison nearly impossible to run cleanly from observational data. dose ramp is when GI side effects peak, which is also when sleep architecture takes the most damage. slow-wave specifically is what gets compressed, and that’s where the majority of nocturnal GH secretion happens. so the lean mass split that shows up at DEXA month 5 or 6 has a meaningful contribution from month 1-2 sleep quality that nobody was tracking, and attributing that result to the compound when it’s partly a compromised GH axis during the window that matters most is exactly how people end up comparing protocols that weren’t actually running under the same conditions.
the 72/28 split is the part i’d push on. the case for it is fair, you got a real DEXA at the switch and 8 months of sema is a long enough window that the directional read isn’t crazy. but without a pre-sema DEXA the 28% lean number is a calculation against an assumed starting composition, not a measured one, and that assumption is doing a lot of work. for a woman lifting 3x/wk who probably had above-average lean going in, you can shift the inferred split 5-8 points either direction depending on what baseline you plug in. ymmv but i’d hold that specific number loosely until you’ve got two real measurements bracketing the same compound, which the 6-week tirz DEXA will start to give you. on the GH/sleep piece, steel-man is fair, sleep does suppress nocturnal GH pulse and GLP-1 GI side effects disrupt sleep for plenty of people. but calling it the variable people aren’t accounting for is a stretch when protein floor in the 48hr post-injection window is probably the bigger lean-mass lever at typical loss rates, and that one’s actually fixable without a sleep study. my food noise drops hardest days 5-6, protein intake tanks 15-20g without me noticing if i’m not pre-planning, and i’d close that gap before chasing GH. fwiw the GH contribution to adult lean preservation is also more modest than the bodybuilding lit assumes.
The slow-wave compression framing is sound, but month 1-2 isn’t a fixed window, and my GI disruption on tirz ran to week 11, which shifts that GH deficit period considerably for anyone trying to model when the axis actually recovered.
the sequence itself confounds the comparison. tirz is being measured against a body that already shed 18lb of mostly fat on sema, so partitioning at the second drop starts from a different baseline regardless of which compound is doing the work.
the 72/28 split on a year of sema with three weekly lifts is doing more work as a “not muscle-sparing” verdict than it actually licenses, i think. caloric restriction studies in non-resistance-trained populations land around 25-30% lean loss as the baseline, and lifting load tends to pull people toward the low end of that range rather than below it. that’s not a great outcome, but it isn’t evidence of an unusual lean mass insult from sema specifically, which is how the framing reads. the nocturnal GH pulse piece is the part i’d push harder on. the chain of inference there is: GLP-1 disrupts sleep, disrupted sleep blunts the GH pulse, blunted pulse drives clinically meaningful lean loss. each of those steps has some supporting data but the last one is the weakest. adult GH replacement studies at physiological doses show pretty modest body comp effects in healthy populations, and translating “blunted pulse” to “additional lean loss on top of the deficit” is more inferential than the post frames it as. protein intake and absolute resistance load almost certainly account for more of the variance at the magnitudes we’re discussing here. the comparison-at-6-months point i’m with you on completely, and the estradiol-as-third-axis logging is something i should be doing and am not. curious what the next DEXA shows on tirz, especially if your protein floor is comparable to the sema run.
28% lean on that sema split while lifting 3x/week is the number I’d be fixating on, not the 18lb total. the nocturnal GH pulse point is the mechanism I see most consistently skipped in these threads, and you’re right that it’s a class effect, not a tirz-vs-sema distinction. fwiw I genuinely cannot separate my GI-disrupted sleep from my toddler-disrupted sleep in my own data right now, so the confounders don’t resolve cleanly, but the direction still tracks. protein at 120-130g on good days and I’d still want a DEXA before I called any of my current loss “mostly fat.”
one variable that doesn’t get pulled into these comparisons enough: protein distribution under appetite suppression. delayed gastric emptying plus reduced hunger drives a lot of people from 3-4 protein feedings down to 2, and the leucine threshold work suggests MPS caps per-meal somewhere around 30-40g for women in this age range. you can hit the same daily total and still under-stimulate synthesis if it’s two big hits instead of four spaced ones. that’s not the GH pulse mechanism you’re describing, it’s an MPS frequency mechanism, and on tirz specifically the appetite curve makes hitting four feedings genuinely hard in the first 36 hours post-shot. separate methodological flag for the 6-month DEXA: test-retest precision for lean mass on the same machine runs roughly 1-1.5 lb depending on hydration state and positioning. on 18 lb total loss, a 28% lean split is ~5 lb, so a meaningful chunk of that signal sits inside the noise floor for a single before/after. same machine, same time of day, fasted, similar hydration state matters more than people realize. ymmv but I scheduled mine for first thing in the morning and held off on creatine loading the week prior because the water shift alone can move the lean number by a pound. curious whether your endo is tracking IGF-1 alongside the estradiol piece. that’d be the closest proxy you have for whether the GH suppression is actually showing up in your case vs being a theoretical worry.
the nocturnal GH pulse framing is doing more work than it should here. the case for it is plausible: GH peaks in slow wave sleep, GLP-1 GI symptoms disrupt sleep architecture, less SWS means less pulsatile GH, less IGF-1 support for muscle. but the magnitude question is unaddressed. chronic sleep restriction in healthy adults moves IGF-1 modestly, and the lean mass effect of that delta over a 6 month weight loss arc is not the same order of magnitude as protein intake or training stimulus. the more parsimonious explanation for the 28% lean fraction is the size of the caloric deficit times protein distribution. 72/28 isn’t actually that bad for unassisted weight loss in a 40+ woman. caloric restriction reviews I’ve read put typical lean fraction loss at 20-30% without resistance training and roughly 10-15% with structured lifting plus adequate protein. 28% with 3x/wk lifting means either the protein wasn’t landing where it needed to (appetite suppression compressing meals matters a lot here) or the training stimulus didn’t progress under the deficit. neither of those is GH. also “tirz adds GIP agonism” doing the lean preservation work in animal models is where I’d want to see the dose comparison more carefully. comparable weight loss between sema 1mg and tirz 5/10mg isn’t matched GLP-1 receptor occupancy, so any GIP-attributable difference in your DEXA split is confounded by the GLP-1 occupancy curve too. that’s the part of most sema vs tirz comparison posts that frustrates me. you’d need matched GLP-1 occupancy to isolate the GIP contribution, and at clinical doses that’s not what either compound is doing. agreed the 6 month DEXA is the right read and scale weight at week 9 is noise. would also pull in MAGE or postprandial AUC if you have CGM access, bc T2D vs euglycemic baseline changes the lean mass conversation in ways protein and calories don’t fully cover. what’s your protein g/kg actually landing at on tirz with the appetite suppression? that’s the variable I’d want pinned down before attributing anything to GH pulse.
the GIP agonism mechanism is legit and I’ve read enough on it to not dismiss it, but the SURMOUNT-1 lean mass substudy data doesn’t really back the human translation the way the animal model framing implies. tirz users in that trial still lost meaningful lean mass - the fat/lean split wasn’t dramatically better than STEP 1 sema numbers once you account for starting body composition differences. “animal models show independent fat partitioning effects” has a long track record of not surviving dose translation in humans, and “not settled in humans at clinical doses” is doing a lot of work in your framing. the sleep/GH variable I’d keep, that part’s solid. what I’d actually push back on is weighting it above protein timing on injection day specifically - MPS signaling is competing with peak GI suppression on the lowest-appetite day of your week, and that compounds over a 6-month run in a way a weekly protein average won’t catch. the confounders don’t resolve just bc you’re tracking them, but day-by-day protein on injection days is the one most people aren’t separating out. curious whether your logs break it down that way or if you’re averaging across the week.