24.2% at 48 weeks vs 28.7% in Phase 3 - what gives?

We don't have lean mass numbers in the trial facts we have, so I wouldn't guess. But the glucagon piece is why retatrutide gets attention beyond GLP-1/GIP. That doesn't automatically mean better body composition. The CTgov entries for NCT05931367 and NCT06039826 don't give headline results, so we're stuck waiting on peer review. Also, NCT05548231 only enrolled 32 people, which is tiny for any body-comp claim.

Dose-response matters, but the glucagon piece is what I'm curious about. Triple agonist, so GLP-1/GIP/glucagon. I care less about scale weight if lean mass is tanking.

Not sure duration alone does it. If 28.7% is a top dose arm and 24.2% is a trial mean, that's apples to oranges.

Duration is my guess. Phase 2 looked at 48 weeks. Phase 3 may run longer. Knee OA and postmenopausal groups aren't the same as general obesity either.

Maybe, but completed trials are way bigger - 1946 and 2335 in two. Bigger n doesn't magically push mean weight loss up unless duration or dose differs.

The 28.7% is company-reported and not peer-reviewed, so I'd treat it as a headline, not like-for-like. The 24.2% was efficacy estimand. Different estimands alone explain a chunk.

Up to is doing a lot of work there; 28.7 sounds like a best-case arm, not the average.

Timeline mismatch is the first thing I’d check. Phase 2 says 48 weeks; the Phase 3 figures may be from different readout weeks, so 28.7 could be a later snapshot. I’ve tracked my own weekly averages for 9 months and the 6-9 month slope looked nothing like months 1-3. What week were those Phase 3 percentages measured at?

Baseline size changes what a percentage means. I started at 246 lbs; 10% was 24.6 lbs and my knees felt it fast. A 180-lb person losing 10% is 18 lbs. Those Phase 3 trials had very specific groups—knee OA, postmenopausal, small n—so the 28.7 could reflect who enrolled, not a stronger effect. Do you know baseline BMI for each?

Headline % is almost never the lived weekly experience.

Daily weigh-ins, but only Sunday averages get compared. Waist at the navel. Sleep and step count in the same note. Everything else is noise.

First three weeks: 6.8 lbs, mostly water. Weeks 12-20: 0.7 lbs/week, and the whole reason was weekend takeout and bad sleep, nothing mysterious. Weeks

I weigh every morning but only count Sunday after bathroom, naked, same scale. Funny thing: weeks my scale flatlined, my waist still dropped half an inch, so a 24% vs 28% headline would feel totally different depending on whether they measured scale or tape. Did the postmenopausal or knee OA groups track waist, or just pounds?

My sleep tracker averaged 5h42m for week 20 and my scale sat at the same tenth for 11 days. Then I got two 8-hour nights on a camping trip and dropped 2.3 lbs without changing my food. That kind of swing makes me wonder if Phase 3 sites even log sleep or stress, because 4.5 percentage points could be a lot of bad sleep.

Water weight can fake a 3-lb whoosh overnight; percentages don't care, but my pants definitely do.

Small n is my sticking point: 46 postmenopausal and 32 participants are tiny. If 6 people drop out, the mean can swing several points from one big responder. I’d rather see the completers’ curve and dropout reasons than a topline 28.7%. In my own log, one 5-lb water dump after a salty weekend made a 12-week average look way better than it was.

Wait, are the Phase 3 numbers at 68 weeks? If so you're comparing a 48-week slice to a longer run.

My husband lost 9 lb just from cutting beer; trials with diet support aren't pure drug effect.

I ran a 12-week office weight pol and the folks who dropped out were exactly the ones who'd have blown up the average. If Phase 3 reports completers or a different estimand, 28.7 vs 24.2 is noise until the missing-data rules match. Also check if baseline BMI was higher; heavier start often means bigger drop.

My waist tape moved 2 inches before the scale caught up, so I trust trends over exact numbers.

46 postmenopausal participants and 32 in that third one, those arms are way too small to average in with 445.

Honestly a 4.5 point gap across separate trials barely registers as a real discrepancy to me. In the twelve-week office weight pool I ran, people on basically the same plan landed anywhere from 1% to 11% down, and the group averages hid that completely. Within one trial the spread across individuals is huge, so comparing an efficacy estimand from Phase 2 against a company press release from Phase 3 is comparing two different kinds of numbers, not just two different populations. Add in that the OA knee cohort is a different enrollment filter entirely, and I mostly want to see the full numbers before treating 28.7 as the number that matters.

Something I keep coming back to: if the Phase 3 knee OA trial ran longer than 48 weeks, that alone could explain most of the jump, since the last eight weeks of a 48 week curve are usually the flattest part. That plus better retention than Phase 2 would do it, without needing the drug to behave differently at all.

The detail I got hung up on is that they're not the same kind of number. Efficacy estimand and a company-reported topline aren't directly comparable, and I say that as someone who's made that exact mistake arguing about this online. My scale did the same thing to me: flat for five months, then six pounds off over seven weeks with zero changes to my routine, which taught me that timing alone can swing a percentage point or two either way. So I'd want to know when each Phase 3 trial actually measured, because 52 or 60 weeks versus 48 could account for more of that gap than the populations do.

I stopped chasing trial percentages and started tracking waist at the same spot every Sunday; scale noise ate my progress.

My scale lied to me for six straight weeks: 194, 196, 193, 197, 194, 198, and I nearly blamed the med. What actually changed was my weekends. I was tight Monday to Friday, then Friday takeout, Saturday wine, Sunday brunch, and by Monday I'd see a 2-3 lb bump that made me feel like the whole week was wasted. I started weighing every morning after the bathroom and logging it, plus measuring my waist at the belly button every Sunday. The trend line showed I was still losing about 0.7 lb a week, but the weekend sodium and alcohol were hiding it on the scale. That was my unexpected turn: the stall wasn't a plateau, it was a tracking blind spot. I didn't go full monk either. I kept Friday dinner but split the entree and got a side salad, swapped the second glass of wine for sparkling water, and added a 20-minute walk after Saturday breakfast. Over the next 10 weeks I dropped 9 lb and 3 inches off my waist. My point is trial averages like 24% and 28% are group math over a year, not what my body does on a Tuesday after pho. I'd rather track my 10-day average and how my jeans fit than argue over a few points I can't control. If you're stuck, look at your weekends before you assume the number is wrong.

My break between patients is basically five minutes, so I’m catching up on this thread next to a stack of recall cards. I’m a dental hygienist. The 68-week endpoint is the thing I keep staring at. If 28.7% is at 68 weeks and Phase 2 was 48 weeks, that’s five extra months of grinding, so not apples to apples. I keep my own log in the notes app while the chair timer runs, and my 6-week numbers always look better than my 3-week numbers. Is that 28.7% from a later endpoint, or do Phase 3 folks have a different baseline BMI? That would explain a lot.

Tbh Are the Phase 3 topline numbers at 48 weeks, or a later landmark? If 28.7% is week 68+, you’re comparing exposure durations, not a Phase 2 vs Phase 3 discrepancy. I’d also check baseline BMI and whether the knee-OA cohort had pain-related activity limits skewing the result. I track morning weight and a 4-week rolling average; headline peaks don’t help me. What’s the Phase 3 reduction at the same 48-week mark? That’s the fair cross-trial read.

That “up to 28.7%” wording is doing a lot of work too—was that at 48 weeks or later? If Phase 3 ran longer, comparing a 48-week Phase 2 number to a later Phase 3 timepoint is like comparing my week 12 progress photos to week 60. I track a 10-day average instead of daily, and my trend often looked flat for two weeks before dropping. Does anyone have the actual week mark behind the 28.7%? Also curious if baseline BMI or joint-pain mobility differed in the knee OA group, since moving more comfortably could shift the average.

Tbh Not a stats person, but I’d want to know if those Phase 3 cohorts had the same baseline BMI range as Phase 2. Knee OA and postmenopausal groups can skew older/heavier, and mean % loss often looks bigger from a higher starting weight. Also, “up to” might be a completer slice rather than the full enrolled group. In maintenance I quit comparing my % to trial headlines and just track a 4-week rolling average plus waist-to-height—much less noisy. Do we know the baseline weights for those Phase 3 arms?

one thing i haven’t seen ppl mention: estimand/population. ph2 efficacy numbers can look different from a ph3 topline if one is all-randomized and the other is more on-treatment/completer-ish. do we know if the 28.7 came from the same estimand and same population mix? also the knee OA cohort is a confound—less pain = more movement/NEAT, so some extra loss could be step count, not just the drug. i track soda swaps in a sheet and my best week always looks way better than my 6mo trend lol.

Cafe owner here, so I’m biased toward portion math: a 28.7% drop from a higher baseline BMI is a very different headline than 24.2% in a Phase 2 cohort with different starting weights. I’d want to see baseline BMI, sex mix, and completion/discontinuation rates before comparing. Knee OA and postmenopausal groups may also differ in muscle mass, activity, and appetite patterns. My own tracking got messier when I started weighing sauces and tasting spoons at

Check the clock instead. Phase 2 was 48 weeks. The Phase 3 topline I saw was 68 weeks. That’s 20 extra weeks on a curve that hadn’t flattened. I work nights and track weekly; my own loss was still dropping around month 6 before it stalled. You can’t compare a 48-week snapshot to a 68-week endpoint and call it inconsistent. What were the actual week counts in all four Phase 3 trials? If some are 48 and some are 68, the 28.7% is just the longest one.

Tbh I’m not a stats person either, but the first thing I’d check is time point. My own tracking taught me 48 weeks vs 68 weeks is a different story—my loss curve flattened hard after month 8, even with daily walks and cooking at home. If those Phase 3 numbers are from a later endpoint, that gap might be mostly calendar, not estimand. Do we know the exact week the 28.7% was measured at?

Not a stats person either, but the first thing I’d check is timepoint: is that 28.7% at 48 weeks, or at a later visit? If the Phase 3 trials ran longer, comparing that to the Phase 2 48-week number is like comparing my airport-day weight to my home-scale weight after a salt-heavy hotel week. I track in a notes app across time zones, and my own 2–3% swings are mostly timing, travel bloat, and which gym scale I’m standing on. So: same week, same visit window? Or are we mixing 48-week and end-of-trial numbers?

Also check the week. Phase 2 is 48 weeks. Phase 3 press releases often tout the best arm at final visit. “Up to 28.7%” is one group in one trial, not the average. That’s best-case marketing math. I work nights and track weekly. My 6-month drop beat my 3-month drop just because time passed. So unless the Phase 3 releases give a 48-week cut, you’re comparing a 48-week snapshot to a later best-case arm. Do they?

Wait, are the time points actually matched? I thought Ph2 24.2% was at 48 weeks, but the Ph3 28.7% figure was out to 68 weeks or so. If so, that's not "what gives"—it's just more runway. My own tracking (two kids, chaotic dinners) showed loss slowed a ton after month 6, but still crept down another few % over the next months. So a 20-week longer trial can easily add several points without any magic. Sharp follow-up: does the 28.7% have a week label? If it's 68 weeks vs 48, that's the whole answer.

What timepoint are those Phase 3 28.7% numbers actually at? I thought those readouts were 68 weeks, not 48. If so, that’s roughly five extra months of cumulative loss, and the Phase 2 48-week number shouldn’t match it. I track trial durations in a spreadsheet because mismatched endpoints always muddy these comparisons. Do you have the week-48 slice from the Phase 3 arms? That’s the apples-to-apples figure I’d want before calling it a real efficacy gap.

One thing I have not seen raised: the timepoints may not be comparable. In the Phase 2 obesity trial, the 24.2% figure was read out at 48 weeks, and the weight curve still looked like it had not fully flattened. The Phase 3 programs ran longer before their primary endpoint, so some of that extra 4 to 5 points could simply be additional weeks on treatment rather than a different drug effect or a different population.

Has anyone seen a published or company slide showing where the Phase 3 curves actually plateaued? If the two datasets were cut at the same week, I suspect the gap would shrink considerably. That would also explain why the knee osteoarthritis cohort, which ran longer, sits at the higher end.

one thing I haven’t seen mentioned: estimand and baseline BMI. On-treatment/efficacy vs treatment-policy/completers can move these by several points, especially if discontinuation rates differ. Also, if Phase 3 enrolled a heavier cohort, 28.7% could be similar absolute lbs to 24.2% in leaner Phase 2 folks. Do the releases give baseline weight/BMI and dropout rates? I track 7-day average weight and waist at home because lifting/creatine/sleep noise can fake a 2–3 lb swing; I’d want that context before comparing trial percentages.

One gap I haven’t seen: timepoint. The Phase 2 24.2% is pinned to week 48. If the Phase 3 28.7% is a later week—68 or 80—that’s not really a mismatch, just a longer runway. I keep weekly averages in a spreadsheet, and my own 48-week trendline was still going down but flattening fast; week 48 to 68 was nowhere near linear. Does anyone have the exact week each Phase 3 top-line number came from? If it’s 68+, the comparison is apples to slightly riper apples.

Tbh Maybe I’m missing it, but are the Phase 3 28.7% figures from a longer time point than 48 weeks? If Phase 2 is week 48 and Phase 3 is week 68/80, that’s not really a discrepancy—just more runway. I’d want week-48 landmarks from Phase 3 before comparing. I’m not a stats person; I just know from my own empty-nester restart that my first six months looked different from month 12 once daily walks and cooking got consistent. A fair chart would line up the same weeks.

One thing I’d want to know: were the Phase 3 trials using a run-in diet/exercise program or more frequent check-ins? In the OA trial especially, once knees hurt less, daily steps can jump, and that can add a couple percent. Phase 2 often feels more “take it and get weighed.” If Phase 3 had structured lifestyle support, some of that 28.7% isn’t the molecule alone. I’m not saying it’s fake—just that “up to” plus support plus a different population makes the comparison muddy. Anyone seen whether the protocols mandated activity or diet counseling? That’s the variable I keep wondering about.

hi, brand new here and stll learning, so sorry if this is obvious. Could the starting weights be different? Even if two groups lose similar average pounds, a higher baseline BMI can make the percentage look bigger. I wonder if the knee OA group skews older, heavier, or less mobile, which might affect the numbers. I’ve only been tracking a week, but I already see how much my starting number changes what 5 lbs looks like as a percentage. Did the Phase 3 disclosures mention baseline BMI or whether they adjusted for it? Just trying to understand the apples-to-apples part. Thanks!

One thing I have not seen clarified: the Phase 2 24.2% was under the efficacy estimand, but the Phase 3 press releases may not be using that same estimand. Efficacy, treatment-regimen, and observed-only can diverge quite a bit once people discontinue or miss visits. Do the releases state which estimand they used and whether the 28.7% is observed or imputed? I ask because in my own weekly tracking, a four-week rolling average changes the picture a lot; endpoint definitions matter as much as the week. Baseline BMI and whether the trial enrolled people with osteoarthritis or postmenopausal status could also shift the average. I would love to see the full tables.

One thing I haven’t seen: whether the Phase 3 28.7% is a single trial’s primary endpoint or the best result across all four. “Up to” in a press release often cherry-picks the highest cohort, while the Phase 2 24.2% was one prespecified analysis. That makes the headline comparison apples-to-oranges, not necessarily shady. Does anyone have the per-trial breakdown? I’ve lost and regained enough to care about the boring middle numbers, not just the flashy top-line.

One thing I’d check: Phase 2 was multi-arm, so 24.2% may be a pooled/average figure across arms, while 28.7% is an “up to” from one of four Phase 3 trials. Those aren’t the same denominator. In maintenance I track weekly averages, not single timepoints, because one weird week skews the whole picture. Do you know if the 28.7% is a single trial’s top-line, or the best-performing arm across all four? If it’s the latter, the comparison is basically apples to oranges.

One piece I haven’t seen yet is the background lifestyle support and visit cadence. I track weekly averages, not daily weights, because my own numbers can swing a pound or two with sodium and cycle timing. In a trial, more frequent check-ins, standardized food logging, or a run-in phase can make the behavioral component stronger in Phase 3 than Phase 2. If the Phase 3 protocols had more intensive counseling or a different run-in, that could widen the gap beyond the drug itself. Do we know whether the Phase 3 trials specified a minimum lifestyle intervention or just encouraged it? That might be a missing piece.

Something I’d want to see: baseline body composition/waist and whether the Phase 3 pool skewed toward more knee OA or postmenopausal participants who may be less active. Scale weight can shift with inflammation/fluid, and OA trials often have NSAID use, which can muddy the picture. I track my own waist and scale separately for that reason. Also, were the 28.7% figures from a prespecified primary endpoint or an interim company readout? If the trial mix, regions, or baseline activity levels differ, that could explain a few points without anyone fudging. Gentle ask: do we know baseline waist or body-fat distribution in each?

Hi, brand-new here, sorry if this is obvious. One thing I’d want clarified: is the 28.7% from one specific Phase 3 trial, or a pooled “best of four” figure? “Up to” can hide a lot. Also, were all four in obesity-only participants, or did some include knee OA/OSA/T2D? Different baseline conditions and activity levels could shift average loss. I’m only week one and already learning my daily scale is noisy, so I’d feel better comparing like-for-like populations rather than the highest number.

New angle: comorbidity/functional status. The knee OA trial (NCT05931367) isn’t a clean obesity-only cohort—pain and mobility can tank NEAT, which blunts weight loss independent of appetite. If the 28.7% comes from a non-OA Phase 3 arm, comparing it to the Phase 2 obesity cohort is apples-to-oranges. I track steps, and my non-exercise activity drops hard on bad-joint days. Do we know which specific Phase 3 trial/arm generated 28.7%? That would settle a lot.

One thing I’d poke at: diabetes status. In this drug class, people with T2D often show smaller average weight loss than those without. If the Phase 2 obesity cohort was mostly non-diabetic, and some Phase 3 arms had a different T2D/prediabetes mix, that alone could widen the gap. Do we have baseline A1c or diabetes percentages for those Phase 3 trials? I track daily and only trust my 7-day trend, so I get wanting a clean comparison. But cross-trial averages are noisy; I’d look at the whole curve and subgroup tables, not just the headline 24.2 vs 28.7.

One thing I’d add to the spreadsheet: baseline diabetes/metabolic status. My own trendline shifts way more with A1c/fasting glucose than with calories alone. If the Phase 2 obesity cohort had a cleaner metabolic profile than the Phase 3 knee-OA/postmenopausal groups—or vice versa—mean % loss can move several points without any estimand/timepoint tricks. Do the Phase 3 baseline tables break out HbA1c, diabetes duration, or % with prediabetes? That’s the first column I’d compare before calling 28.7% vs 24.2% a real efficacy gap.

One thing that tripped me up in the old tirzepatide threads: the Phase 2 readout was a 48-week snapshot, while a lot of Phase 3 topline numbers land at 68 weeks. If that 28.7% is a 68-week endpoint, comparing it to 24.2% at 48 weeks is apples-to-oranges — the curve often hasn’t fully flattened by week 48. I’d want to see the week-48 intermediate from thoe Phase 3 trials before assuming a real efficacy gap. I track my own weekly weights and even a few extra months can change the average a lot. Does anyone know if the company’s 28.7% is specifically the 68-week timepoint, or is it a pooled/subgroup number?

Not an expert, just a 50-something empty nester tracking my own slow roll. One thing that jumps out: 24.2% was at 48 weeks, but the 28.7% figure may be at a later time point (68 weeks?), so it's not apples-to-apples. Also, trials in knee OA or postmenopausal groups may have different baseline muscle/fluid and mobility. I care less about scale now: my walks and cooking overhaul moved my waist more than pounds. Do the Phase 3 reports give a matching week-by-week curve, or just a peak number? That would settle a lot.

One thing I’m not seeing: diabetes status. Phase 2 obesity trials often exclude type 2 diabetes, and even a minority with T2D in Phase 3 can drag mean percent loss down—or make the comparison murky depending on completers. Do any of those four Phase 3 trials report baseline A1c or how many participants had T2D? I ask because I lost ~18% on a GLP-1, regained most of it, and my A1c/metabolic labs definitely changed how my weight tracked. If the Phase 3 pools include more T2D, the 28.7% vs 24.2% gap might be partly population, not just estimand or timepoint.

Tbh running a cafe, portion creep is my daily boss, so I track weight weekly and waist monthly. One thing I’d want to know: are those Phase 3 28.7% numbers at the same 48-week timepoint as the Phase 2 24.2%, or later? Even a few extra months can matter because loss often slows or plateaus. Also, did Phase 3 include more structured lifestyle or behavior support? That alone can nudge a couple percent. Not a scientist, just someone who sees how much time and support change results.