48-week 24.2% vs Phase 3 28.7% — what gives?

I keep a sheet too. The 1,946 and 2,335 trials are where I’d look for 28.7% context.

Population matters. The nkee OA n=445 cohort may have lower activity, so the curve could look different.

Glucagon receptor agonism is the wild card. More energy expenditure sounds good until glucose drifts.

The 24.2 vs 28.7 gap could just be estimand. Efficacy estimand isn’t company Phase 3 topline.

I’d be cautious comparing 24.2% to 28.7% directly. The Phase 2 number was at 48 weeks, while company Phase 3 reports may use different analysis sets, baseline BMIs, and discontinuation handling. Add the CTgov enrollment spread—32, 445, 46, 1,946, 2,335—and these cohorts are not interchangeable. Without peer review, I’d compare within trial only, not headlines. I’m not saying the higher number is hype; just that estimand and population matter. What would convince me is a full 48-week curve by baseline BMI and glycemic status, plus discontinuation rates. That would separate a design artifact from a genuine efficacy difference.

The n=32 Chinese trial tells us almost nothing.

Agreed, don’t mix cohorts. Want glucagon markers, not just weight.

I've been nerding out on this too, and the estimand thing flipped it for me. That 24.2% was an efficacy estimand at 48 weeks, which basically answers "what happens if everyone stays on and takes it as assigned" — phase 3 press releases tend to lead with the on-treatment or treatment-policy number, and those can diverge 3-5 points purely from how dropouts get counted. The enrollments don't look weird to me either: n=32 reads like a titration sub-cohort, 46 and 445 look like PK or dose-finding studies, and the 1,946/2,335 are outcomes trials where 48-week weight isn't even the primary endpoint. Baseline BMI and country mix matter a lot. I've logged weekly since week 4 and my 4-week rolling average sat flat for six straight weeks around week 30, then dropped 3.1 lbs in ten days. Short windows lie constantly.

Adding a boring variable nobody talks about: sleep. I ran 5-6 hours a night for two weeks during a work crunch and my weekly loss went from 1.4 lbs to basically 0.2, appetite came roaring back, scale didn't move for nine days. Got back to 7.5 hours and it unstuck within a week. If phase 3 sites have tighter follow-up and support than a small phase 2 cohort, that alone could be a pint or two of the gap.

Contrarian take: I'd trust the 24.2% over the 28.7% until there's body comp data. I'm at week 40, down 21% by scale, but my waist is down 7.5 inches and a scan around week 32 showed a real lean-tissue chunk in there that I'm not thrilled about. When my weekly loss rate got aggressive, the lean share got worse, not better. A faster-losing phase 3 population could absolutely post a bigger headline number that's partly water and muscle. Also "up to 28.7%" is the best arm of the best trial, not the average, and my own best 12-week stretch looked way better than my actual 40-week total. Has anyone seen waist-to-height data from these arms?

I'll push back on the enrollment thing — that spread isn't odd, it's just how these programs get built. A 46-person trial is almost certainly a PK or regional bridging study, 445 is dose-finding, and the two ~2,000 arms are outcomes trials where weight is a secondary readout. Different baseline BMIs, different countries, different food environments. Where I do think something's off is timeline. I'm at week 52 and my last 10 weeks have been pure maintenance, 0.4 lbs total, so a 48-week number and a 68-week number aren't comparable at all. If phase 3 is quoting a later timepoint than 48 weeks, that alone could be most of the 4.5-point gap.

I weigh every morning but I only look at the 7-day average. Daily noise for me has swung as much as 3.1 lbs. Week 34 was flat to the ounce versus week 28, then I dropped 2.4 lbs over nine days doing nothing differently. If the Phase 2 number came off single-visit weights and the Phase 3 readouts smoothed that out, a percent or two of your gap is just arithmetic, not biology.

The tape measure at my navel beat the scale for two straight months. Weeks 20 through 27 my weight sat inside a 1.2 lb band, but my waist went 41.5 down to 39.75 and I dropped a belt notch. If one trial protocol tracked waist and another only tracked pounds, the two numbers aren't describing the same thing. Are you logging anything besides the scale?

Restaurant weekends are my entire swing. Two dinners out with soy sauce and I'm up 4.2 lbs Monday, right back down by Thursday. That's roughly 2% of my body weight moving on salt and nothing else. If trial visits land on a fixed schedule, one cohort getting weighed after a holiday stretch versus a normal week could eat a real chunk of that gap you're chasing.

Total weeks actually on treatment is the column I'd want next to those percentages, not just 48 versus whatever the Phase 3 duration was. I stalled from week 29 to week 41, twelve weeks and 1.1 lbs total, then dropped 6.3 lbs over the next ten weeks with zero habit changes. Bodies stair-step, they don't slide. If the Phase 3 figure is a readout point where more people have already banked their own week 60, part of the higher number is just longer runway. The other part is likely who got enrolled, since a wider pool with higher starting BMIs makes percentage loss look better because there's simply more to lose. That's not extra effect. What I'd trust is pounds lost compared at the same week count per person.

Those 32- and 46-person enrollments are the tell for me. In a 32-person trial, one person dropping 40 lbs can swing the mean by over a point, so a flashy 28% can easily be a couple of super-responders plus small-sample noise. The 1,946 and 2,335 trials are the ones I’d actually weight, and they’re closer to the Phase 2 24.2%.

A 32-person arm is basically a group cat wearing lab coats.

My 4-week rolling average hides the daily chaos way better than any single morning number.

Sleep under 6 hours for three nights and my weekly loss basically flatlines.

Peer review or it’s just vibes with error bars, honestly.

The 1,946-person trial is the one I’d watch, not the 28.7% headline. I know bigger numbers sell, but my contrarian take is the 24.2% efficacy estimand is closer to what a real clinic sees, because it likely counts people who stopped or missed visits differently. Phase 3 can look hotter from run-in periods, more hand-holding, and completers-only summaries, and none of that is a new mechanism. Population matters too: if the Phase 3 cohort skews younger, heavier at baseline, or more adherent, the mean can move several points without any extra effect. The enrollment sizes you listed don’t tell us the estimand, so I’d want to know if those four trials pre-specified treatment policy or on-treatment. If they did treatment policy and still hit 28.7%, that’s genuinely interesting. If it’s completer data, it’s just a different denominator wearing a fancy suit.

I quit comparing my scale to trial percentages after week 36. Waist went from 42 to 36.5 inches, shirt size XXL to XL, scale only down 16%. Trial means are group averages; my body lost inches before pounds, then pounds caught up. If you’re tracking, add waist at navel and a pair of jeans that used to fit. It tells a cleaner story than daily weight when the water shifts.

My Monday scale is a liar. Friday fasted 214, Monday fasted 218, same calories-ish but restaurant sodium and carbs. I now only compare Friday-to-Friday and log waist Sunday. That explained a plateau that wasn’t a plateau. Trial averages smooth this out; individual weeks don’t.

The trial visits are basically a part-time job: monthly labs, food logs, dietitian calls, and someone weighing you in the same clothes. When I mimicked that for 12 weeks — 140g protein, 8k steps, no alcohol — I lost 11 lb. Without the check-ins I've lost 4 lb in the same stretch. That support alone can explain a few extra percent.

The sermorelin pricing gap keeps widening. Some vendors are dirt cheap, others are charging way more, and the spread is kind of nuts. Makes me wonder if it’s just markup or if there’s a real quality difference. If you’re shopping, don’t trust old posts or screenshots — check the live listings yourself. Curious whether anyone’s compared quality at the low end vs the high end.

I pulled the CTgov entries too—two of those tiny enrollments (32 and 46) are Phase 1/2, not Phase 3, so mixing them makes the set look weirder than it is. The 28.7% also reads like an on-treatment/completers number from one active arm, not the efficacy estimand. Do you know the week for that 28.7—72? If it’s 72 vs the 48-week 24.2, the gap is mostly time plus population, not magic. Also worth checking whether the Phase 3 trials excluded T2D or had different baseline BMI.

I might be misreading, but isn’t the 28.7% from the Phase 3 arms at 68 weeks, while the 24.2% Phase 2 number is the 48-week estimand? If so, a big chunk of the gap may just be time on treatment, not population. I’d still want the same estimand, same visit window, and transparency on early discontinuers. I track my own trend weekly, and my 6-month vs 12-month numbers honestly look like different studies. Do you know if any Phase 3 readout reports a 48-week landmark so we can line it up more fairly?

Hi—total newbie here, week one, so forgive the dumb question. Is the 28.7 possibly a completers/efficacy estimand while the 24.2 includes everyone randomized, even those who stopped early? I only ask because in my own food log, my “days logged” drops off hard when I feel icky, and that alone changes my weekly average. Also curious: did the Phase 3 trials prespecify a washout or off-treatment follow-up? That seems like it could inflate the top-line number if retention differed across the bigger trials. Thanks, and sorry for the ramble.

I keep coming back to duration. If the 28.7% figure comes from 68 or 72 weeks, then comparing it with the 48-week Phase 2 number is not quite apples to apples. My own weekly averages looked very different at week 48 versus week 68, even when I was doing the same things. I would also want baseline BMI and diabetes status broken out, since a population with more obesity-only participants and fewer type 2 diabetes participants can post a prettier mean. Do you know whether the Phase 3 trials used the same estimand and discontinuation handling? That would explain more to me than the enrollment sizes.

Apartment turnover work means I see exactly how much takeout gets abandoned in kitchens, so portion weirdness is pretty much my daily scenery. Anyway, I’ve been digging through CTgov and can’t find whether the bigger Phase 3s used an active run-in or a placebo lead-in. If they did, early GI dropouts get screened out, and the completers’ average ends up rosier. The other wrinkle is site mix: more high-takeout or food-desert enrollment in the 1,946 and 2,335 cohorts than the earlier trial could shift background diet and adherence. That doesn’t automatically mean the design explains it. I’m only saying 24.2% vs 28.7% may be apples to oranges. Has anyone seen run-in details in the Phase 3 protocols?

I’m a broke dorm kid, so my only useful tracking skill is averaging out caeteria sodium swings. What I haven’t seen in the thread: those teeny enrollments (32, 46) smell like PK/food-effect sub-studies or single-site arms, not the main efficacy cohorts. If the 28.7% comes from the 1,946/2,335 groups, fine, but lumping all five CTgov numbers together is apples-to-dorm-microwave. Sharp follow-up: do the big Phase 3s pre-specify a per-protocol or efficacy estimand that drops early discontinuers? Because if yes, that alone could explain a chunk of the gap without any population difference.

the enrollment mix stuck out to me too—two tiny trials and two huge ones probably shouldn’t carry equal weight. I’m more curious whether the Phase 3 protocols standardized diet/exercise or left it site-dependent. In my own tracking, my 48-week loss was 21%, then I dropped another 4% just from restarting meal prep and a 10-minute after-dinner walk. If lifestyle support varied a lot by site, that alone could nudge the average. OP, do you know if they required a minimum step or protein target? And what habit has stayed easiest for you when the scale stalls?

One thing I’d pull from the protocols is how intercurrent events are handled. The Phase 2 24.2% may have been an efficacy estimand, censoring people after they stop treatment. The 28.7% in a press release could be an on-treatment or completer snapshot from a more selected group, which lifts the average. I track weekly averages in a spreadsheet, and the same thing happens: my all-weeks average is much lower than my consistent-weeks average. Do the Phase 3 registry entries specify the same estimand and similar baseline A1c/diabetes proportions? That would explain a lot.

Tbh I track my own trendline in absolute kg, not just %, because % hides baseline. At 110 kg, 24.2% is 26.6 kg; at 85 kg it’s 20.6 kg. If Phase 3 enrolled heavier folks, mean % can drift up even with similar absolute loss. Do the Phase 3 tables list baseline BMI and weight by arm? That’s the first column I’d check. My spreadsheet taught me a 1% difference can be one salty weekend and a bad sleep week.

Tbh I’m in maintenance after losing on one of these, and what moved my own tracking was waist and protein, not the headline mean. Phase 2 is often a smaller, more selected population; Phase 3 spreads across sites and regions, so baseline BMI, sex mix, and lifestyle support can shift the average by several points. My sharp question: are all four Phase 3 trials using the same estimand and missing-data handling as Phase 2? If yes, the jump is mostly population. If not, it’s not apples-to-apples until peer review.

One angle I haven’t seen yet: timepoint. Phase 2 is 48 weeks, while Phase 3 obesity readouts often use 68–72 weeks. If 28.7% is a later primary endpoint, it’s not apples-to-apples—weight-loss curves can still drift down even when weekly loss is tiny. I track my own weekly weights and cumulative % keeps creeping up just from time, so I now compare slope at matched weeks, not final totals. Has anyone found a 48-week interim in the Phase 3 supplements, or are they only reporting the later timepoint? That could explain a chunk without needing a totally different population.

Before comparing 24.2% and 28.7%, I’d want the exact week of the Phase 3 endpoint. Phase 2 readout was 48 weeks; if the company figure is 68 or 72 weeks, that’s not a discrepancy—that’s just more time on drug. Also check whether Phase 3 required a standardized lifestyle/behavioral run-in or had a different background program. I log weekly weight and waist, and my own trend flattens hard after month 9; 48-week data don’t extrapolate linearly. What’s the actual assessment week for the 28.7%?

Not a stats person, but one thing I didn’t see in the thread: trial lifestyle support. Phase 2 protocols often just say “counseling encouraged,” while Phase 3 might mandate dietitian visits or a protein target. As a dorm kid living on microwave rice and canned tuna, my own trendline swings way more when I actually hit decent protein and walk between classes than when I don’t. So if the Phase 3 protocols bake in structured nutrition/exercise, 28.7% isn’t pure drug effect. Does anyone know if those four protocols specify mandatory lifestyle intervention? That’d be my cheap first place to look.

One thing I haven’t seen mentioned: baseline BMI and prediabetes/T2D mix. In my own logs, my first 10% came off fast when my fasting insulin was higher, then stalled hard; a friend with similar starting weight but no metabolic flags lost slower but steadier. If Phase 3 skewed toward higher baseline BMI or fewer diabetes exclusions, that alone could explain part of the 24.2 vs 28.7 gap. Do you know if the Phase 3 protocols reported baseline BMI by arm and whether they excluded folks with T2D? That feels like a cleaner comparison than headline enrollment numbers.

Budget dorm rat here, so I track every can of tuna like it’s a stock. One thing I’d want before comparing 24.2% and 28.7%: is that Phase 3 number an all-randomized ITT figure or a completer/efficacy-population figure? A 24% ITT can look like 28% if dropouts are excluded, especially in open-label trials. Also baseline BMI and geography can shift mean %. What’s the baseline BMI in the Phase 3 arm that hit 28.7, and is it using the same estimand as the Phase 2 24.2%? Enrollment size alone won’t answer that.

I’d want baseline weight/BMI, not just the % headline. Percent reduction scales with starting weight: 24.2% off 220 lb is ~53 lb; 28.7% off 260 lb is ~75 lb. If the Phase 3 cohort started heavier, the same absolute loss can look bigger as a percentage. Also raw vs placebo-adjusted matters. Do the press releases give baseline weight and placebo change? As a daily weigher, I compare my 10-day trend to my own baseline, not trial percentages—noise is loud, but the slope is the signal.

Not a stats guru, but I’d want baseline weight and SDs before comparing those means. 24% off 115 kg is ~27.6 kg; 28.7% off 105 kg is ~30.1 kg, so the gap moves around depending on starting point. I switched my own log from % to kg plus 7-day trend because early water weight made percentages look wild. Also, “up to 28.7%” usually means best arm or best subgroup, not the whole trial average. Do the Phase 3 releases give CIs and baseline BMI? That’d tell us if it’s a real jump or just a different denominator.

One thing I’d pull from older trial threads: baseline weight/BMI and diabetes mix. A 300-lb cohort losing 24% is 72 lb; a 240-lb cohort losing 28.7% is ~69 lb. Same-ish absolute loss, very different percentage. I track waist and protein too, and my own % swings a lot based on start weight. If Phase 3 skewed heavier and excluded T2D, that could explain a chunk before we blame trial quality. Do the CTgov entries list baseline BMI/diabetes exclusion? That’s the first table I’d want.

The 32 and 46 enrollments are almost certainly not efficacy cohorts—probably PK or titration substudies. So the real comparison is 445 vs 1,946/2,335. I’d want baseline A1c and metformin use before believing 28.7%. In my own spreadsheet, excluding prediabetics shifts the mean by ~2.5 points. If Phase 3 enrolled a leaner, less diabetic population, that alone could close the gap. Was the company’s 28.7% on-treatment or ITT?

Hi, total newbie here, sorry if this is obvious. One thing I haven’t seen asked: are both percentages using the same missing-data rule? Sometimes a Phase 2 “efficacy estimand” and a company Phase 3 topline can handle dropouts/discontinuations differently, so one may be on-treatment-ish and the other all-randomized. Also, were baseline starting weights similar? A higher baseline can make the same pound loss look like a bigger %. Just trying to learn what’s actually comparable.

One thing I always check on company slides: the estimand/intercurrent-event handling. A 24.2% Phase 2 “efficacy estimand” might count people who stopped, while a 28.7% Phase 3 press-release number could be on-treatment/completers or a hypothetical strategy. That alone can bridge a few points without any real population difference. Also, the tiny 32/46/445 enrollments are probably early-phase PK/tolerability, not obesity efficacy, so I wouldn’t let them muddy the comparison. Do the Phase 3 releases say whether they’re ITT or completers? Typing this from an airport gate, so I’ve got time to nerd out.

Tbh one thing I’d check is baseline body weight/BMI. A 28.7% mean can look bigger partly because Phase 3 obesity cohorts often start heavier; 28% of 115 kg is a much bigger absolute loss than 24% of 105 kg. More sites/countries also means lifestyle support and food environment aren’t identical. I’m in maintenance now, and I care more about pounds kept off at two years than the headline percentage. Do the Phase 3 reports show average starting weight and completion rates, or just the topline percent?