NCT04184622: Why do so many stall at 5 mg?

The trial had forced dose arms, not real-world self-selection. That alone makes 5 mg look different in forums.

Agreed. Also 2539 people is huge, but if raw data has no outcome numbers, we’re guessing about who plateaued where.

I’ve been on 5 mg for months and it still works for appetite. No plans to move unless my clinician says otherwise.

But that’s the thing—appetite suppression isn’t the only endpoint. Weight, waist, labs all matter. Forums flatten that.

I get why people blame plateaus on the dose, but Week 0–72 is a long run. Bodies change, activity changes, sleep changes, stress changes. I’ve watched my own weekly averages go flat for a month, then drift down again when work stress eased. That’s not a dose story. The trial can’t tell us much about that texture. What I’d want is better real-world data on maintenance after initial loss, not just arm labels. I'm not saying dose doesn't matter, just that arm labels alone are a blunt tool. The forum pattern might be real, but it might also be selection bias.

What real-world data would even be clean enough?

We need n and time on dose, not vibes.

I thought the 10 and 15 mg arms were about seeing dose-response, not necessarily saying everyone needs them.

Right, and placebo arm matters too. Without it, we forget some stalls are just normal variation.

My take: fourms overrepresent stalls because people doing fine don’t post as much. I’m guilty of that.

I’ve been tracking waist and a favorite pair of jeans since week 8. From week 12 to 20 the scale flatlined, but I lost 1.5 inches off my waist and went down one notch. If the trial only measured pounds, that’s a “stall” even when body composition is still shifting. So maybe some 5 campers are actually still changing shape.

My stall was a weekend problem, not a medication problem. Friday pizza and Sunday brunch undid Mon–Thu every single time.

The trial wasn’t built to show who needed higher arms—it randomized people to fixed arms and didn’t rescue nonresponders. So the real-world 5 crowd is probably a mix of side effects, cost, and “good enough.” Which tracks with what shows up in these threads: when nausea flattens a gym habit, a 20-minute after-dinner walk is often what gets the scale moving again, without touching anything else.

Water weight gaslit me for two weeks, then I dropped 3 lbs overnight. The scale is a liar, but the trend line isn’t.

Week 19 here and the scale sat on the same 1.5 lbs for six straight weeks — I nearly quit tracking. What actually moved was my belt: two notches in, and a pair of work pants that used to leave marks went from uncomfortable-by-2pm to just normal. So I switched to measuring my waist at the navel every Sunday instead of weighing daily, and it dropped 3.25 inches across that same flat stretch. The weird part is I'd been calling it failure, but my food log showed the same volume, just different food. Swapped an afternoon granola-and-yogurt habit for cottage cheese and a fistful of berries and the scale started moving about 10 days later. Not saying that's why, but I think 'stall' covers two very different situations — nothing changing, and things changing where the scale can't see it. Which one you're in changes what's worth doing next.

I'll be the contrarian here: my plateau months weren't about food at all. Frm roughly month 4 to month 7 I was averaging 5.5 hours of sleep because of a work crunch, and my tracking was clean — around 1,650 calories, 110g protein, 8k steps a day. Flat as a board. Then the crunch ended, I went back to 7.5 hours, changed literally nothing else, and lost 6 lbs in five weeks. My resting heart rate on my watch dropped 7 bpm in that same window, which is the only objective proof I have that anything was different. Stress and sleep are the two inputs nobody logs because they don't feel like diet inputs. If your log looks good and you're stuck, look at the eight hours before the log starts.

My tape measure has been more honest with me than my scale all year, and it's not even a smart one.

Nobody warned me the food noise comes back gradually, not all at once. I didn't notice until I caught myself planning dinner at 10am again. What pulled me out of a flat stretch was boring stuff — 30g protein within an hour of waking and a big salad before anything else at dinner. Four weeks, 5 lbs. Boring beats clever.

The boring variable for me was sleep. I logged it for 14 weeks and my flat stretches clustered on nights I averaged under 6 hours, four of my five stalls landed there. Weeks with 7.5+ I'd drop 1.2 to 1.8 lb changing nothing else. Not saying sleep is magic, but the log made it hard to argue with.

Honestly I don't think that trial can answer your question, and that's the part people miss. The arms weren't escalate-if-you-stall, they were fixed schedules, so the 5 mg group stayed put whether people were losing or not and the higher arms moved up on a calendar, not on response. The design removed the choice. What it shows is an average dose-response across thousands of people, which tells you almost nothing about any single person's stall. Anecdotally my own flat month turned out to be weekends, not dose, three weeks of Friday pizza plus a couple beers had me up 4 lb every Monday and back to baseline by Wednesday, net zero for a month. So I'd ask what changed in the six weeks before you plateaued, not which arm you'd have landed in.

I added lifting twice a week around week 30 and the scale went sideways for five weeks, average change 0.2 lb, while my pants got looser through the thighs and my resting heart rate dropped 6 bpm. If I'd gone by weight alone I'd have sworn I was stalled. It was actually the most on-track stretch of my year.

My own fail was ditching the food scale for three weeks because I was sure I knew my portions by then. My peanut butter servings had quietly doubled and I sat flat for 19 days. Pulled the scale back out, logged a week, and things started moving again. That stall wasn't my body, it was my eyeballs.

The thing that broke my last long flat stretch wasn't food, it was sleep — I averaged 5h40m a night for a month during a work crunch, and my weekly average weight didn't budge and my waist tape didn't move either. What's odd is the week I finally got 7-8 hours six nights in a row, I dropped 2.2 lbs changing nothing I ate. I've been logging my bedtime next to my morning weight, and the pattern is almost embarrassing: any week with three or more nights under 6 hours, my Friday weight is up 1-2 lbs. Some of that is obviously water, but it changed how I read a stall. Now when the scale freezes for ten days I check my sleep and my stress before I go tearing apart my food log. Also, eating late because I was up late was doing more damage than the actual calories, which took me way too long to notice.

Different angle: I keep belt notches and one pair of jeans as my real tracker. There was a five-week stretch where my weight sat inside a 1.5 lb window and I moved two notches anyway, shirts got looser, and that was the only reason I didn't quit that month. Scale is daily data for me now; belt is the monthly verdict.

trial can’t answer that. Arms were assigned, not chosen. Nobody “camped” at the middle dose by preference. Real world is different: people stop escalating for cost, supply, or side effects, then call it a plateau. The useful detail would be discontinuations vs dose reductions. Did the paper separate those? If not, it’s just group averages. My own flat stretches line up with 12-hour night rotations and gas-station food, not the shot. Tracked protein and steps, not vibes.

The trial probably can’t capture the “cleanup tax.” My stall wasn’t meals, it was kid-adjacent grazing: crusts, half bananas, abandoned sting cheese while clearing plates. Didn’t feel like eating because I was standing. I made a snack drawer with only pre-portioned stuff I actually like—edamame packs, turkey sticks, cuties—and a sit-down rule if I want it. The drawer gets restocked into lunchboxes every Sunday so it doesn’t become a free-for-all. Anyone else realize their flat stretch was really mindless bites off their kids’ plates?

The trial design mostly can’t answer cost vs side effects vs “enough,” because that arm is fixed—no rescue escalation. So real-world camping there is a self-selected mix: folks who feel fine and stop pushing, and folks who can’t tolerate more. Opposite phenotypes. My own spreadsheet: weekly weight, waist, protein g/kg, steps, 1–10 appetite score. My last scale flat stretch came with appetite scores already flat 4–6 weeks earlier while waist kept drifting down. So “stall” depends which column you’re staring at. The number I’d want from NCT04184622 is discontinuation reason by arm—AE-related vs not. Did anyone ever find that split?

Hi, brand new here (week one, sorry if this is a silly question). I keep wondering if the trial reported individual weight trajectories or just group averages? Like, how many people in that middle active group actually lost under 5% versus over 15% by week 72? A flat average could hide some folks still losing and others truly stuck. If there’s a supplementary figure with those categories, that might speak to the “stall” question better than the headline numbers. Has anyone seen that breakdown?

broke dorm take: what if the trial’s low-dose arm looks decent because everyone got regular visits and a food plan, while real-world camping happens in a microwave desert? I track protein per dollar (tuna, eggs, lentils, whatever yogurt is on sale) and my plateau almost always lines up with a protein slump, not appetite. The week I ran out of cheap protein and lived on ramen, the scale froze like it was also on a budget. Would love a subgroup by food access/dorm-kitchen situation. Anyone else see their “stall” track their grocery cart more than their hunger?

One thing the trial can’t really show is “camping,” because it reports mean weight change, not how long individuals sat at a plateau. I’m a dorm rat, so my version of a protocol is a whiteboard: cost per gram protein, veggie check, and whether I ate after 9 p.m. My longest flatline wasn’t mysterious—it was two weeks of dining-hall