Sustained starter doses moved the scale 2.2% on semaglutide and 5.5 on tirzepatide
Patients who stayed on the lowest approved dose of semaglutide for at least six months lost an average of 2.2 percent of body weight at 12 months, against 5.5 percent for patients who stayed on the lowest approved tirzepatide dose.
The gap was statistically significant at every measured timepoint, according to the matched cohort study published in Biology Methods & Protocols.
Of 490,072 patients with at least one semaglutide prescription, 814 met the sustained low-dose definition of at least three distinct 0.25 mg prescriptions spanning at least six months; of 322,442 tirzepatide patients, 1,016 met the analogous 2.5 mg criterion.
The head-to-head comparison matched 534 patients per arm on age, sex, baseline body mass index, and baseline type 2 diabetes.
Reuters reported that average weight loss at one year on recommended higher doses was 20.2 percent with tirzepatide and 13.7 percent with semaglutide in an earlier head-to-head trial.
The study itself compared sustained lowest-dose semaglutide against high-dose semaglutide and found mean percent weight change of -1.76 versus -12.03 at 12 months.
Adverse event rates diverged by drug.
Over 24 months, the study found higher event probabilities in the tirzepatide arm for constipation (32.58 versus 22.41 percent), acute kidney injury (2.68 versus 0.43 percent), lumbar disc disease (3.30 versus 0.29 percent), muscle cramps (8.34 versus 4.00 percent), and dyspnea on exertion (9.64 versus 6.80 percent).
The semaglutide arm showed higher rates of otitis (5.60 versus 2.05 percent), diaphoresis (5.54 versus 3.24 percent), and ankle swelling (6.92 versus 3.74 percent).
None of those signals survived correction for multiple comparisons.
After Benjamini-Hochberg adjustment across 461 evaluated conditions, all q values exceeded 0.5, and the authors wrote that the individual associations should be treated as hypothesis-generating rather than confirmatory.
Across 104 structured diagnosis outcomes, no comparison reached statistical significance.
Observational studies of this kind cannot prove cause and effect, according to study leader Venky Soundararajan of nference.
He told Reuters that the two medicines are not equivalent and that patients on low doses should know both the limited weight loss on offer and the adverse event profile.
Prescription records cannot confirm that medication was dispensed, filled, or administered, so adherence measures such as proportion of days covered could not be assessed.
Fills obtained outside participating health systems, including through telehealth or compounding pharmacies, were not captured.
Cohort sizes were small enough to preclude formal subgroup and sensitivity analyses, and the authors noted the study population is likely not fully representative of the US or global GLP-1 patient population.
The authors are employees of nference, which conducts research collaborations with biopharmaceutical companies whose products feature in the study.
The paper states none of those companies funded or had any role in the work, and that the research received no external funding.
