Why is the 72-week tirzepatide trial still the benchmark?

72 weeks is long enough to show separation, but 2539 participants is what makes it hard to ignore. Smaller trials answer narrower questions.

Different endpoints, different sample sizes. Simple as that.

The AFib study isn't trying to prove weight loss. It's recurrence burden after ablation, so 200 might be enough for that signal.

I think people quote the 2539-participant trial because it had placebo and three dose arms over 72 weeks. Clean design. The 200-person study is a different question. Comparing them is like comparing a marathon to a sprint. Still, I wish more trials reported longer follow-up after the primary period.

Yeah — the part people forget when comparing options: 5/10/15 mg arms make dose-response way easier to talk about than a single-arm setup.

But n=200 isn't tiny for a focused endpoint. Depends on event rate.

Right, but event rate is doing a lot of work there. If recurrence is common enough, 200 can show a difference.

LY3457263 is also 200 and recruiting, but it's an add-on study in T2D on stable sema or tirz. Narrow question.

The rodent thyroid C-cell tumor signal keeps getting dragged into these threads. The weight-of-evidence analysis lists tirzepatide among seven approved GLP-1RAs across six domains and reports rodent thyroid C-cell tumors. That's not the same as human risk, and it doesn't tell us anything about trial size. I just think it's another reason people over-index on the big 72-week dataset: they want one study to answer efficacy, safety, and durability all at once.

I don't thikn one study can answer all three. That's the whole problem with forum debates.

What labs moved for you? I'm more interested in real-world trends than another trial-size argument.

The 72-week thing sticks because most of us live in 12-week insurance windows. I logged waist at 0, 12, 24, 52, 72. At week 12 I was down 9 lbs, week 24 down 22, week 52 down 31, then week 72 down 34. The last 20 weeks were mostly sleep, protein, and walking 7k steps. The surprising part: my best week wasn't early, it was week 58 after I fixed weekends. So long trials match real life plateaus.

72 weeks also catches wedding season, holidays, and vacations. Twelve-week streaks have a habit of dying at Thanksgiving — every time.

Meanwhile every coverage breakdown I've looked at goes 24 weeks and then just... stops. So that benchmark is basically fantasy.

I track waist not scale. Lost 3 inches by week 40, scale barely moved for a month.

I finally slept 7 hours for two weks and my appetite dropped before any scale change.

My scale lied every Friday for three months, so I switched to daily weigh-ins and a 14-day average. That average dropped 1.8 lbs when individual Fridays looked flat or up. The 72-week design lasts that long because short windows catch water swings, takeout nights, and stress and call it a plateau.

I hit 8k steps by walking during phne meetings; my resting heart rate dropped 11 bpm before the scale moved.

I lost only 4 lbs across 12 weeks and still went down two shirt sizes, which broke my brain. That stretch taught me the 72-week thing isn't about patience, it's about letting food habits get boring. I added 30g protein at breakfast, stopped grazing after 8pm, and walked 20 minutes after dinner. Weeks 16-28 the scale barely moved but my belt went from notch 4 to notch 2. Then weeks 29-40 dropped 13 lbs without changing anything. I still weigh daily, but I only judge the 14-day average. The benchmark feels long because it's the first timeline where consistency stops being a personality trait and just becomes Tuesday.

I quit reading trial lstings and started logging Sunday sodium instead. Weirdly more useful.

The 72 weeks sticks because almost everything we compare it to is a first-trimester snapshot. My own tracking over a stretch like that: 14 weeks, 11 lbs down, then flat for a month while squat went up 20 lbs and sleep went from 5.5 to 7 hours. The scale wasn't wrong, it just wasn't the whole timeline.

I meal-prepped twelve identical chicken lunches and was done with them by Wednesday. Twelve.

My own benchmark ended up being two identical 6-month blocks. Same meals, same gym schedule, but round one I still drank three nights a week and lost 8 lbs; round two, basically no alcohol, lost 19. That's why the long studies keep getting quoted, duration exposes what a 12-week window hides.

I stopped weighing daily and started measuring my waist every second Sunday — same spot, same tape, right after coffee. First 10 weeks: scale down 9 lbs, waist down 0.5 inches. Weeks 20 through 44: scale basically flat, waist down 3.25 inches. That's the reversal nobody warns you about, clothes changing while the number doesn't. In a 72-week window you actually capture that lag; in a 12-week study you'd call it a plateau and quit at week 14. I don't think 72 weeks is sacred, it's just long enough for the tape measure to catch up to the scale. If the question is whether something works at all, 72 weeks is overkill. If the question is whether it keeps working once life gets boring again, then yeah, you need a year-plus.

Totally get why 72 weeks feels like the gold standard — it covers a whole year of holidays, conferences, and airport cinnamon rolls. My personal benchmark is less clinical: can I get through a 3-flight day and still find a protein-ish breakfast? Current winning move is a collapsible bowl + spoon in my carry-on, then Greek yogurt and berries from whatever airport market. Not fancy, but it beats a pretzel the size of a steering wheel. Sharp question: for the road warriors here, what’s your one non-scale metric that survives hotel gyms and time zones? I’m considering “number of days I hit 8k steps purely from terminal walking.”

For me 72 weeks matters because it shows whether people stay on it, not just peak loss. 2539, placebo-controlled, multiple arms. That’s a stress test. My sharp question: what was the week-72 completion rate, and how did they handle dropouts? If it’s low, the benchmark is shakier than the headline. Night shift taught me the real endpoint is maintenance: same sleep window, same eating window, steps before shift. Track that shift cycle after shift cycle and you learn more than any short graph.

Yeah, I think the benchmark isn't the 72 weeks—it's the 2,539-person placebo arm plus multiple fixed arms. That's what lets you trust the separation in responder categories. A 200-person post-ablation study can't replicate that, even if it runs longer. I keep a "completers only" tab beside my all-in tab; my 18-month trendline gets way prettier if I quietly drop travel weeks, which is why retention is the stat I want. Sharp follow-up: does anyone have the week-72 completion/discontinuation split? If it's 85% vs 70%, the tail of that curve reads very differently.

The 72-week thing feels benchmark-y to me because it covers the boring middle after the initial drop—the part where the scale stops being a useful daily scoreboard. I switched to a rolling 30-day average plus a calendar streak of 20-minute after-dinner walks. If the average is flat but the streak is intact, I don’t spiral. Are you asking why 72 weeks became the standard duration, or why it still feels more convincing than newer, smaller trials? For me the durability part is what matters.

Small-town pharmacy angle: 72 weeks is basically two New Year’s resets, so it catches the January deductible/prior-auth scramble and the summer fair/potluck season. That always felt like the real benchmark to me — not just the big headcount, but that it runs long enough to include the boring middle where motivation dips and the pharmacy has to deal with a formulary change. I’m curious: do you all think it’s the 72-week length that matters most, or the fact it had enough people to make the results stick? A smaller 200-person study just can’t answer that same everyday-life question for me.

The 72-week design matters to me because it’s long enough for the scale to stall while other stuff keeps shifting. I’m 14 months in, same weight for 11 weeks, but my resting HR dropped 6 bpm and I added 15 lbs to my squat. A 12-week trial would’ve called that a fail. Does anyone else see nonscale progress lag the scale by months?

I nerd out on trial design too. What makes 72 weeks feel like the benchmark to me is it captures the boring middle—weeks 30 to 60—when novelty wears off. I track gym visits per month and whether I’m still prepping lunches by then. That middle stretch is where my past attempts died. The ablation study is 200 people wth a totally different endpoint, so it’s apples and oranges. Sharp follow-up: did the 72-week one include a maintenance phase or any post-treatment follow-up? That durability piece is what I’d want next to the on-treatment results.

For me the 72-week thing is less about the final percentage and more about curve shape: in the trials, the placebo-adjusted gap often keeps widening into the second half of the year, so shorter studies mostly capture the steep early phase. NCT07382024 is asking an arrhythmia question, so it can’t really compete as an obesity benchmark. What I’d want next is a long extension with body-composition tracking—how much of any regain after stopping is fat

New angle: 72 weeks isn’t just biology math, it’s 18 months of real-life seasons. For me that’s two back-to-school resets, a summer with kids home, holidays, birthdays, and at least one “snack drawer becomes dinner” week. That’s the benchmark because it tests whether habits survive chaos, not just a smooth lab calendar. My concrete tracking: I note each week whether the snack drawer has a protein-ish option within reach or if it’s just crackers. The trial may not capture that, but it’s what I’d want a long study to reflect. Did they collect any lifestyle

the thing that makes 72 weeks fel like a benchmark to me is it’s long enough to cover a full annual cycle plus holidays. My own scale graph has this ugly November–January ridge, and a 52-week study would just call it noise. 72 weeks forces the protocol to show whether the trend line holds when real life smacks it. Also 2,539 people means enough to look at slow responders still trending down around week 68, which is my whole deal. Do they publish per-participant trajectory data, or only group means? I’d love to see how many were still losing at the end.

Contrarian take: 72 weeks gets credit, but for me the benchmark is really n=2539 with 4 arms. My own spreadsheet has 18 months of daily weights; the first 12 weeks are pure water/glycogen noise. Once I switched to a 10-day moving average, my trendline didn’t stabilize until ~week 60. A 200-person AF ablation recurrence study can’t give you that kind of precision, even if it ran 72 weeks. Does anyone else weigh the sample size/design more than the duration? Or am I overfitting my own n=1?

72 weeks is the bar because it’s long enough that dropouts and missing data actually matter. Shorter “real world” posts are mostly survivors who stuck with it. I want to know if the week-72 number was completers-only or all randomized with proper handling. That changes how much I trust it. Night shift here: I track weight after sleep, waist monthly, and whether I ate at 3am. Concrete question: did they prespecify how they counted people who stopped the drug but stayed in the study? That’s the part I never see quoted.

for me the 72-week thing matters less as calendar math and more as a test of the boring middle. like weeks 30-55, when the scale slows and old habits sneak back in. i run a dumb spreadsheet for soda swaps and late-night snack triggers, and my longest streak is 41 weeks before a new game launch nuked my sleep. i’d rather see trials report how many people are still hitting their behavior goals at wk72 than another average. did anyone here actually track a full 72 weeks without the wheels coming off?

Honestly, the part that makes 72 weeks feel like the benchmark to me is retention. A big trial is only impressive if they didn’t lose half the room by month 10. I’ve tracked my weight weekly for years, and my first 72 weeks were so noisy that the real trend only showed up after. Did the publication break out dropout rates by arm? That’s the number I’d want before comparing it to a 200-person study with a totally different endpoint.