One injection a week or one a day sounds like a question about convenience, and every seller treats it as one. The evidence says the interval matters less than the two molecules on either side of it, and in the place where the comparison has been run properly a good deal of the apparent difference was written into the protocol before anyone was dosed. The route question — needle against tablet — is a separate one, handled in the oral article.
The one randomized head-to-head
STEP 8 randomized 338 adults with overweight or obesity and without diabetes, in a 3:1:3:1 ratio, to once-weekly semaglutide 2.4 mg, to once-daily liraglutide 3.0 mg, or to matching placebos, for 68 weeks at 19 United States sites. Mean weight change was −15.8% with semaglutide against −6.4% with liraglutide — a difference of 9.4 percentage points (95% CI, −12.0 to −6.8; P < .001), or 8.5 kg in absolute terms (95% CI, −11.2 to −5.7). Pooled placebo was −1.9%.[1]
The responder thresholds separate further than the means do. A loss of 10% or more was reached by 70.9% against 25.6% (odds ratio 6.3; 95% CI, 3.5 to 11.2), 15% or more by 55.6% against 12.0% (odds ratio 7.9; 95% CI, 4.1 to 15.4), and 20% or more by 38.5% against 6.0% (odds ratio 8.2; 95% CI, 3.5 to 19.1), all at P < .001.[1] Those are large, real and correctly reported. They are also a comparison of two molecules at two different potencies, not a test of how often a syringe comes out of the refrigerator.
Part of the adherence gap was in the protocol
STEP 8 is quoted constantly for its discontinuation figures: 13.5% left semaglutide for any reason against 27.6% on liraglutide, and treatment stopped for adverse events in 3.2% against 12.6%.[1] Read the design before drawing the obvious conclusion from that.
Gastrointestinal adverse events were reported by 84.1% on semaglutide and 82.7% on liraglutide — within a point and a half of each other. Nausea ran 61.1% against 59.1%, and serious adverse events were actually more common on the weekly drug in one direction and the daily drug in the other, at 7.9% and 11.0%.[1] Side effects were not the difference.
What differed was the exit. A participant who could not tolerate semaglutide 2.4 mg was permitted to continue at 1.7 mg. A participant who could not tolerate liraglutide 3.0 mg discontinued treatment, and could then restart the titration from the beginning. The escalation itself ran 16 weeks on semaglutide and 4 weeks on liraglutide, and the trial report attributes liraglutide’s shorter time to discontinuation to that shorter ladder.[1] One arm had a dose to fall back to and four extra months to get there; the other had a door. The 9.4-point weight gap survives that observation comfortably. The tolerability gap does not survive it intact, and the mechanics of climbing either ladder are set out in the titration article.
The same two molecules, the opposite tolerability result
SUSTAIN 10 put once-weekly semaglutide 1.0 mg against once-daily liraglutide 1.2 mg in 577 adults with type 2 diabetes over 30 weeks. Glycated hemoglobin fell 1.7 points against 1.0 (estimated treatment difference −0.69 points; 95% CI, −0.82 to −0.56; P < 0.0001), and weight fell 5.8 kg against 1.9 kg (difference −3.83 kg; 95% CI, −4.57 to −3.09).[2]
The weekly drug won on both endpoints and lost on tolerability. Gastrointestinal disorders were reported by 43.9% against 38.3%, and adverse events leading to premature discontinuation ran 11.4% on the weekly drug against 6.6% on the daily one.[2] That is the reverse of STEP 8’s direction, in the same two molecules. The doses differ, the population differs and the duration differs — which is the point. There is no stable finding here that weekly dosing is better tolerated; there are two trials whose tolerability results disagree.
The cleanest test of the interval alone
Only one molecule has been randomized against itself on two schedules. DURATION-1 compared exenatide 2 mg once weekly with exenatide 10 µg twice daily in 295 adults with type 2 diabetes over 30 weeks. Glycated hemoglobin fell 1.9 points against 1.5 (95% CI, −0.54 to −0.12; p = 0.0023), and 77% against 61% reached a target of 7.0% or less (p = 0.0039). Weight reduction was similar in both arms.[3]
DURATION-5 repeated it in 252 adults over 24 weeks: glycated hemoglobin −1.6 against −0.9 points (P < 0.0001), weight −2.3 kg against −1.4 kg, and nausea in 14% on the weekly schedule against 35% on the twice-daily one.[4] That nausea contrast is the strongest single piece of evidence that less frequent dosing is easier to take.
Two caveats keep it from settling the question. The once-weekly arm is an extended-release microsphere formulation, so the comparison changes how the peptide is released as well as how often it is injected. And on the endpoint this site’s readers care about, the two schedules were close: the interval bought glycemic control and tolerability, not weight.
What adherence data actually show
A meta-analysis of seven studies covering 75,159 adults with type 2 diabetes compared weekly and daily injectable GLP-1 receptor agonists on proportion of days covered. Once-weekly dosing carried an 11% lower risk of non-adherence (risk ratio 0.89; 95% CI, 0.83 to 0.95), with heterogeneity of I² = 89%.[5] A German prescription database of 5,449 patients starting weekly exenatide and 24,648 starting daily liraglutide found median proportion of days covered of 0.88 against 0.77, and adjusted odds of reaching 80% coverage of 1.78 (95% CI, 1.62 to 1.96).[6]
An 11% relative reduction is a real effect and a modest one. Set it against the spread inside the weekly class. A United States claims analysis of 56,715 treatment-naive adults found 360-day persistence of 67.0% on weekly semaglutide, 56.0% on weekly dulaglutide, 40.4% on daily liraglutide and 35.5% on weekly exenatide, with every pairwise comparison against semaglutide significant at p < 0.001.[7] The worst-persisting drug in that study is a weekly one, below the daily one, and the gap between the best and worst weekly agents is 31 percentage points — far larger than anything the schedule explains.
In weight loss, the floor is lower than either figure
Those persistence studies are in diabetes. In obesity the numbers fall. A commercial claims analysis of 4,066 adults with obesity and without diabetes who started a GLP-1 in 2021 found persistence of 46.3% at 180 days and 32.3% at one year, mean proportion of days covered of 51.0%, and only 27.2% adherent by the 80% threshold. By product, one-year persistence was 47.1% on weekly semaglutide and 19.2% on daily liraglutide.[8]
That is a wide gap and it still confounds molecule with schedule, since the two products differ in effect size, in price and in insurance treatment as well as in how often they are injected. What it establishes plainly is that two thirds of people starting one of these drugs for weight are off it inside a year, which changes what a trial percentage means for any individual reader — and what happens next is the subject of the withdrawal article.
What daily dosing still has going for it
The pharmacokinetics are the substantive argument, and the labels state them. Liraglutide has an elimination half-life of approximately 13 hours, which both its labels give as the reason it is suitable for once-daily administration. Semaglutide has a half-life of approximately one week, and the Wegovy label states it will be present in the circulation for about five to seven weeks after a last 2.4 mg dose. Tirzepatide’s is approximately five to six days.
That difference decides what happens when something goes wrong. A reaction to a daily drug is largely gone within a few days of the last injection. A reaction to a weekly one persists for more than a month after the decision to stop, which is the constraint behind the pre-procedure planning described in the surgery article and behind the handling rules in the missed dose article. Daily dosing also titrates more finely — the liraglutide ladder moves in 0.6 mg steps at weekly intervals — and its labels instruct re-titration from 0.6 mg if more than three days have elapsed, a shorter forgiveness window but a cheaper one to serve.
There is a population argument too, though it is narrower than it is usually stated. Liraglutide is labeled for glycemic control from age 10 in type 2 diabetes, which is younger than any weekly product reaches. For weight, the daily and weekly products both reach age 12. The adolescent liraglutide trial randomized 251 participants aged 12 to under 18 and found a difference in body-mass index standard-deviation score of −0.22 (95% CI, −0.37 to −0.08; P = 0.002), with 43.3% against 18.7% achieving a 5% or greater body-mass index reduction, treatment discontinuation for adverse events in 10.4% against none on placebo, and a rebound of +0.15 in that score after stopping (95% CI, 0.07 to 0.23).[9] The wider pediatric picture is in the adolescents article.
The comparison that has never been run
No randomized trial has compared tirzepatide with any once-daily GLP-1 receptor agonist. A 2026 Bayesian network meta-analysis states that head-to-head comparisons across the three agents are lacking and estimates indirectly that tirzepatide 10 mg and 15 mg produce 12.86 and 13.95 percentage points more weight reduction than liraglutide 3.0 mg; that analysis was funded by tirzepatide’s manufacturer.[10] The only randomized comparison at the top of the class, SURMOUNT-5, is weekly against weekly: tirzepatide −20.2% (95% CI, −21.4 to −19.1) against semaglutide −13.7% (95% CI, −14.9 to −12.6) over 72 weeks in 751 adults.[11]
So the honest summary is narrow. Weekly dosing is associated with modestly better adherence across diabetes populations, and the two molecules most people are choosing between happen to have the weekly one in front on weight by a wide margin. Neither of those facts is a finding about the calendar. A reader picking a schedule is really picking a molecule, and the molecule is what the evidence separates.
Every figure above came from a trial or label of an FDA-approved product. Compounded semaglutide, tirzepatide and liraglutide are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before dispensing, and a compounded preparation sold on a weekly schedule has no pharmacokinetic data of its own establishing that the schedule fits it. How the figures on this site are established before publication is set out in the methodology.