The GLP-1 Timeline: When People Stop, and What Each Phase Actually Risks

• written by Perjan Duro

Most people picture a GLP-1 journey ending the way a marathon ends: you either finish or you gradually fade.

The cohort data does not look like that.

Discontinuation clusters. It arrives in waves, and each wave has a characteristic reason behind it. Knowing roughly which wave you are approaching is genuinely useful — not because a statistic predicts your outcome, but because the thing most likely to threaten continuity in month 2 is not the thing most likely to threaten it in month 10.

Quick answer

In a 2,306-patient academic obesity clinic cohort, cumulative discontinuation ran 14% by 3 months, 24% by 6 months, 35% by 9 months, and 50% by 12 months. Median persistence was 10.7 months.

That is not a slow, even trickle. It is a steep early slope, then a long grind — and the reasons shift as you move through it.

One finding is genuinely counterintuitive: in a large digital-service cohort, people who lost more than 10% of their body weight in the first month had 1.67× higher odds of dropping out. Fast early results were a risk marker, not a safety net.

Everything below is observational. These are patterns seen in groups of real patients, not rules about what will happen to you.

Before the map: what "phase" means here

These phases are drawn from cohort averages. Your treatment plan, your medication, your clinician's schedule, and your circumstances all shift the picture. Treat this as a map of where the terrain tends to get rough, not a forecast.

Also worth naming up front: the different cohorts disagree on exact figures, and that is normal. Samuels and colleagues reported 24% discontinuation by 6 months in an academic clinic; Van Laren and colleagues reported 31.7% within 6 months in a 1,374-patient real-world cohort. Different populations, different settings, different access conditions. The shape is consistent even where the numbers are not.

Months 0-1: the setup phase

What the data says. Very little discontinuation shows up in month 1 itself. In the Samuels cohort, the 14% figure is cumulative across the first three months — the first few weeks are mostly people getting started.

What actually threatens continuity here. Two things, and neither is about motivation.

The first is early tolerability. Adverse drug reactions were the single leading stated reason for discontinuation in the Van Laren cohort at 26.8% — ahead of every other category. That is a clinical matter. Side effects are something to describe to your clinician or pharmacist, with specifics and timing, as early as they become disruptive. They are not a test of grit, and nothing here should be read as advice to push through symptoms.

The second is the early-success paradox, and it deserves its own section.

The early-success paradox

Talay and colleagues analysed 19,693 people in a UK tirzepatide-supported digital obesity service. Losing more than 10% of body weight in month 1 predicted 1.67× higher odds of dropping out over the following year.

Read that carefully, because it is easy to over-interpret.

The study observed a prediction, not a mechanism. It found that a fast first month statistically flagged higher dropout odds. It did not establish why, and it did not show that losing weight quickly causes anyone to stop.

Several explanations are plausible and none of them are demonstrated by this study — they are hypotheses:

  • Rapid early loss may travel alongside more intense side effects for some people.
  • It may reset expectations to a pace no month can sustain, making month 3 feel like failure.
  • Some people may reach a personal target early and consider the job done.
  • It may partly reflect who signs up to a digital service and how quickly, rather than anything about the medication.

The honest position is that we do not know. What is defensible is the practical takeaway: an unusually strong first month is worth mentioning at your next appointment, not celebrating and then going quiet. It is information your clinician can use — about tolerability, about pace, about what the plan should look like next.

Months 1-3: expectation collision

What the data says. Cumulative discontinuation reaches 14% by 3 months.

What actually threatens continuity here. This is where the first month's pace stops repeating. Early change often includes shifts that do not continue at the same rate, and the second and third months frequently look flatter than the first, even when nothing has gone wrong.

The risk in this window is interpretive: reading a normal deceleration as evidence the plan has stopped working, and quietly disengaging before anyone in your care team knows there is a problem.

What is worth raising early:

  • side effects that are still disruptive rather than settling;
  • any symptom pattern you can tie to specific days or timings;
  • a mismatch between what you expected and what is happening, said plainly.

Months 3-6: the cost and logistics window

What the data says. Cumulative discontinuation roughly doubles from 14% to 24% between months 3 and 6 in the Samuels cohort. Van Laren's cohort reported 31.7% gone within the same window.

What actually threatens continuity here. The Van Laren reasons list is the most useful thing in this phase. Leading reasons for discontinuation were:

  • adverse drug reaction — 26.8%
  • cost — 14.4%
  • non-adherence — 11.2%

Cost is the one people are least likely to raise, and it is the one most likely to end a course silently. It is a pharmacist and clinician conversation — about coverage, about supply, about what the options are if the current arrangement stops being sustainable. Raising it three months before it becomes acute is a completely different conversation from raising it the week you cannot fill a prescription.

Note also that "non-adherence" at 11.2% is the third reason, well behind the two structural ones. The dominant threats in this window are clinical and financial, not personal discipline.

Months 6-9: the quiet stretch

What the data says. Cumulative discontinuation moves from 24% to 35%.

What actually threatens continuity here. Nothing dramatic, which is the point. By this stage the novelty is gone, side effects have usually either settled or been addressed, and the routine is doing most of the work.

The characteristic risk is drift: appointments that slip, logging that stops, a refill that gets delayed by a trip and then by a busy week. Continuity in this phase erodes through logistics rather than decisions.

This is also where a long record starts to be worth more than any single measurement, because the questions worth asking have become longer-range ones — how the last three months compare to the three before, whether a symptom pattern is stable or shifting, how consistent the actual schedule has been versus the intended one.

Months 9-12: the one-year threshold

What the data says. Cumulative discontinuation reaches 50% by 12 months, with median persistence of 10.7 months. Roughly half of the cohort was still on treatment at a year, and the median person had stopped just before it.

In the Talay cohort, the 27% who stayed adherent for 12 months reported 22.6% weight loss, against 13.6% for the full cohort.

That gap is real and it is large. It is also, again, an association rather than proof of cause: people who tolerate a medication well, can afford it, and are seeing results are more likely to keep taking it, so continuity and outcome push on each other in both directions.

What actually threatens continuity here. Mostly the accumulated weight of everything above — plus, for some, the question of what comes next. That is a treatment-plan question, and it belongs with your clinician well before the twelve-month mark rather than at it.

The trend line is moving

One genuinely encouraging finding, at cohort level rather than individual level.

Marshall and colleagues analysed 33,607 initiators of high-potency, weight-loss-indicated GLP-1 medications in claims data. One-year persistence rose from 33% among 2021 initiators to roughly 61% for people starting in the first half of 2024. Persistence was higher for tirzepatide than for semaglutide.

Whatever the mix of causes — better preparation, better side-effect management, changing access conditions, changing populations — people starting more recently have been staying on treatment substantially longer than people who started in 2021. If you are reading older discontinuation statistics, they may describe a harsher environment than the current one.

The one positive predictor worth knowing

Across the Talay cohort's 19,693 people, the strongest positive predictor of 12-month adherence was consistent weekly self-weighing.

The necessary caution: this is an observational association in a single cohort. Weighing yourself weekly was not shown to cause anyone to stay on treatment. It may equally be a marker of people who were already engaged, already tolerating the medication well, or already in a stable routine.

What it does suggest is that a regular, low-drama review cadence tends to travel with continuity — and unlike side effects and cost, a review cadence is something you can set up yourself without a prescription or an appointment.

Weekly, not daily. Daily weight is mostly noise, and reacting to noise is its own source of discouragement.

What to raise, and when

Condensed to the practical version:

  • Month 1: tolerability, with specific timings. And an unusually fast start, if that is what is happening — it is worth a mention rather than silence.
  • Months 1-3: any gap between what you expected and what you are seeing, before it turns into disengagement.
  • Months 3-6: cost and supply, early, while there are still options. Persistent side effects, again, if they have not settled.
  • Months 6-9: the schedule itself — whether the actual cadence still matches the intended one.
  • Months 9-12: what the plan looks like beyond the first year.

Every one of these is a conversation with a qualified professional. None of them is a decision to make alone, and nothing here is guidance about starting, stopping, changing, or timing a medication.

The honest summary

Half of one large clinic cohort had stopped by twelve months, and the reasons were mostly not the ones people blame themselves for. Side effects led. Cost followed. Discipline came third.

Knowing which phase you are in does not change your odds. It changes what you bring to your next appointment — and raising a cost problem or a symptom pattern three months early is a very different conversation from raising it after a gap has already opened.

Velto implementation

Most of what is worth raising in these conversations is detail that is hard to reconstruct from memory: when a symptom started, how it tracked against treatment timing, how consistent the actual schedule has been, what the last three months look like next to the three before.

Velto keeps that in one timeline — treatments, symptoms, appetite, mood, weight, and habits — so a monthly review reads as a pattern, and an appointment starts from a record instead of a recollection.

Velto is a tracking companion. It does not treat, diagnose, or make dosing decisions, and it cannot change what any study reports.

References

Related reading

Medical Disclaimer: Educational content only. Not medical advice. For treatment decisions, consult a licensed clinician.
Back to Blog

Articles you might like

• written by Perjan Duro

What Studies Show About Weight Regain After Stopping a GLP-1

Weight regain after discontinuing a GLP-1 is one of the most consistent findings in the literature — and researchers describe it as physiology, not a failure of willpower. Here is what the trials and meta-analyses actually measured, what they could not answer, and how to bring your own record into the conversation.