Semaglutide discontinuation rose 25 percent when US patients started on a high dose

A US claims study on semaglutide discontinuation finds that how the weekly injection is started affects whether people stay on it. Among 48,658 adults with diabetes or obesity who began injectable semaglutide, the 14.5 percent started on a high dose were 25 percent more likely to stop early. The paper, from the University of Texas Medical Branch in Galveston, appeared online in Diabetes, Obesity and Metabolism on 19 October. Semaglutide is approved, so nothing moves on our board, but the finding bears on every incretin that depends on slow dose escalation.

How the MarketScan cohort was built

The authors, led by L M Belalcazar of the endocrinology division at the University of Texas Medical Branch, used MarketScan, a large database of US commercial insurance claims, covering 2018 to 2021. They selected adults with diabetes, obesity or both who started injectable semaglutide and had no earlier incretin treatment. Each person was followed for 12 months. The starting regimen was sorted into two groups, low and high, based on the weekly amount on the first prescription. The authors describe the high start as an off label regimen, because it skips the gradual escalation that the product information sets out for new users. Claims show what was dispensed and paid for, not what people injected, so the study measures prescriptions filled.

The statistics follow a standard plan for this kind of question. Persistence over the year was estimated with Kaplan Meier curves, which track the share of people still filling prescriptions over time. A piecewise Cox model tested the risk of stopping early, and logistic regression looked for the factors that made a high start more likely. The models adjusted for age, obesity status, other illnesses and cost, which matters because money shapes whether people keep filling prescriptions. The study window ends in 2021, early in the use of semaglutide for weight management, and the cohort mixes people with diabetes and people with obesity. The abstract does not split results by brand or indication.

What a high starting dose did to semaglutide discontinuation

Starting high was not rare. Of the 48,658 adults in the cohort, 14.5 percent began on one of the higher weekly amounts. Those patients had lower persistence at one year, and their risk of stopping in the early period was 25 percent higher, with a hazard ratio of 1.25 and a 95 percent confidence interval of 1.17 to 1.34. The link held after adjustment for age, obesity status, other illnesses and cost. The authors call a high start a modifiable risk factor for early discontinuation. In plain terms, a step that the prescriber controls was tied to a measurable rise in people giving up on the drug in the first months.

In plain terms, a step that the prescriber controls was tied to a measurable rise in people giving up on the drug in the first months.

The study also asked who was most likely to get a high start. The odds were higher in the West and the South of the country, when the prescriber was an endocrinologist, and for patients who were male, younger, or had obesity rather than overweight. High starts were also more common in 2018, the first full year of the drug, than in 2019 to 2021, which suggests prescribers moved toward the labelled escalation as experience grew. The abstract does not give persistence percentages for each group, only the statistical tests. The likely mechanism is stomach side effects, which are worst when the dose rises quickly, but the claims data cannot show why any single person stopped.

How this fits the wider persistence problem

Semaglutide discontinuation is part of a broader pattern, because dropping off GLP 1 treatment is common in routine care. A 2025 cohort study in JAMA Network Open followed 125,474 US adults with overweight or obesity who started liraglutide, semaglutide or tirzepatide between 2018 and 2023. One year discontinuation was 46.5 percent among people with type 2 diabetes and 64.8 percent among those without it. More weight loss and higher income were linked to staying on treatment, while moderate or severe stomach side effects raised the risk of stopping. The new paper adds a specific lever to that picture. If side effects drive people away, the speed of the first dose steps is one of the few factors that a clinic can change directly.

The persistence question also shapes how new drugs are designed and tested. Corxel, for example, is pitching its oral GLP 1 candidate on tolerability, as our story on the Corxel SPAC merger explains, and tolerability has become a common way for newer candidates to set themselves apart. Uptake data tell a related story. A research letter covered in our report on oral semaglutide uptake found that the Wegovy pill quickly reached injection levels in new starts, which adds another route to the same molecule. Neither trend changes the core message of the Galveston claims study: staying on an incretin is not only a question of the molecule, but of how treatment begins.

Limits of a claims based look

Several limits apply. Claims data show dispensing, not use, and a gap in fills may reflect free samples, switching pharmacies or supply shortages rather than a decision to stop. The cohort is commercially insured, so it underrepresents older adults on Medicare and people without insurance. The data end in 2021, tirzepatide is not part of them, and prescribing habits have likely changed since as weight management use has grown. Observational studies can show association, not cause: doctors who start high may differ from other prescribers in ways the models cannot capture. The abstract also gives no data on weight or HbA1c, so it cannot say whether faster starts brought any benefit to offset the higher dropout.

For the board, this is a readout on an approved molecule with a flat signal. It does not alter any label or approval, and no regulator has acted on it. What it offers is practical evidence that following the labelled escalation is linked to better persistence. The next useful steps would be a study in more recent data that includes weight management users and tirzepatide, and ideally a randomised comparison of starting schedules. Readers can compare the persistence data with the JAMA Network Open cohort. They can also follow every incretin molecule by stage on our tracker, from early candidates to approved drugs like semaglutide.

Conclusion

This semaglutide discontinuation study ties a high starting dose, used for about one in seven new patients between 2018 and 2021, to a 25 percent higher risk of stopping early and lower persistence at one year. It is observational and based on claims, so it shows association rather than cause. Semaglutide stays approved, and the next checkpoint is newer data that include weight management users and tirzepatide. This story is research news, not medical advice.