What happens to GLP-1 results outside the trials

The headline numbers for GLP-1 drugs come from randomised trials: about 15 percent weight loss on semaglutide 2.4 mg in STEP 1 over 68 weeks, and about 21 percent on tirzepatide 15 mg in SURMOUNT-1 over 72 weeks. Those trials gave people free drug, regular visits, lifestyle counselling and a protocol that pushed everyone to the top dose. Real-world studies use pharmacy claims and electronic health records to ask what happens without those supports. All of the studies below are observational. They describe what happened, not what the drug can do.

How long people stay on treatment

Prime Therapeutics, a US pharmacy benefit manager, has published the most cited persistence numbers from its commercial claims data. These are company analyses, not peer-reviewed papers, though the methods of its first-year study were published in the Journal of Managed Care and Specialty Pharmacy.

The two-year analysis, released in July 2024, followed 3,364 commercially insured adults with obesity and no diabetes who started a GLP-1 drug in 2021. A person counted as persistent until a gap of 60 days without supply. Persistence was 47.1 percent at 180 days, 28.9 percent at one year and 14.8 percent at two years. Weekly semaglutide did best at two years (Wegovy 24.1 percent, Ozempic 22.2 percent) and daily liraglutide worst (Saxenda 7.4 percent, Victoza 7.0 percent). Only 16.6 percent were adherent, defined as having drug on hand for at least 80 percent of days, and 26 percent switched between GLP-1 products.

The three-year update, released in June 2025, followed 5,780 adults who started between January 2021 and March 2022. Eight percent, about one in 12, were still on treatment at three years, rising to 14 percent for high-potency obesity drugs such as Wegovy. Thirty-eight percent switched products along the way. There was one encouraging trend: one-year persistence rose from 33 percent for people starting in 2021 to 63 percent for those starting in the first quarter of 2024, which Prime attributed mainly to the end of the drug shortages.

How much weight people lose in practice

A Cleveland Clinic study in the journal Obesity (2025), led by Hamlet Gasoyan, used health records from Ohio and Florida for 7,881 adults without type 2 diabetes who started injectable semaglutide (6,109) or tirzepatide (1,772) between 2021 and 2023. More than 20 percent stopped within three months and 32 percent stopped between three and twelve months. Most, 80.8 percent, were on low maintenance doses. Average weight loss at one year was 8.7 percent overall: 3.6 percent for early stoppers, 6.8 percent for later stoppers and 11.9 percent for those who stayed on. People who stayed on a high maintenance dose lost 13.7 percent on semaglutide and 18.0 percent on tirzepatide, close to the trial results.

A follow-up from the same group asked why people stopped. In a random sample of 288 patients who discontinued within a year, 47.6 percent stopped because of cost or insurance, 14.6 percent because of side effects, 11.8 percent because shortages meant they could not fill the prescription, 2.4 percent because they switched to a compounded version and 1.7 percent because weight loss was not enough. For 11.1 percent the record gave no reason.

Tirzepatide against semaglutide in the clinic

Before the head-to-head SURMOUNT-5 trial, a JAMA Internal Medicine study by Rodriguez and colleagues (2024) compared the two drugs in health records from a collective of US health systems. After matching, it included 18,386 adults with overweight or obesity who started the diabetes-labelled versions, Mounjaro or Ozempic, between May 2022 and September 2023; 52 percent had type 2 diabetes. People on tirzepatide were more likely to reach 5 percent (HR 1.76), 10 percent (HR 2.54) and 15 percent (HR 3.24) weight loss. The on-treatment difference was 2.4 percentage points at three months, 4.3 at six months and 6.9 at twelve months, in favour of tirzepatide. Gastrointestinal adverse events were similar. More than half of each group stopped during follow-up: 55.9 percent on tirzepatide and 52.5 percent on semaglutide. The randomised comparison is on the semaglutide vs tirzepatide study page.

Why real-world numbers come in lower

• People stop. Trials keep most participants on drug for the full period. In practice most people stop within two years, and weight loss falls with time on treatment. The weight regain page covers what happens after stopping.
• Lower doses. Trials titrate to the maximum dose. In the Cleveland Clinic cohort four in five were on low maintenance doses, through choice, tolerance, supply or cost.
• Cost and supply. Nearly half of stops in the reasons study were about money or coverage, and shortages through 2022 to 2024 interrupted treatment. See GLP-1 costs.
• Different measures. Trials report a pre-set estimand. A treatment-policy estimand counts everyone randomised; an efficacy estimand imagines everyone stayed on drug. Real-world studies often report on-treatment results, which flatter, or all-starters results, which do not. Compare like with like.
• Different people and less support. Trials exclude many conditions and provide counselling and frequent visits. Real-world weights come from whatever visit happened to record one.

The consistent finding across all of these: people who stay on treatment at a full dose get results close to the trials. The gap is mostly about staying on, not about the drug working differently outside a trial. The full trial list is on GLP-1 clinical trials, and maintenance is covered on the maintenance dose page.

Logging it

Persistence is visible in a log. Doses taken, missed weeks, dose steps, refill dates, weight and side effects on one timeline show whether a stall is the drug, the dose or the gaps. Peptide AI records all of it on the free tier; see the free GLP-1 tracker.