Almost every practical question about these drugs — why the injection is weekly, why the dose ladder climbs in four-week steps, why a late dose is usually survivable, why stopping takes more than a weekend — comes back to one number per molecule, and that number is printed in section 12.3 of each approved label. It is also the number readers most often misuse, because a half-life sounds like a duration and is actually a rate. This page works through what the labels say, what the arithmetic on top of them predicts, and the three places where the labels and the arithmetic do not agree. The dose ladder itself is in the titration article, and what the dosing interval buys in outcomes rather than in concentrations is in the weekly-against-daily article.
A half-life is a slope, not a finish line
A half-life is the time it takes for the concentration in the blood to fall by half. That is a proportional statement, so it never reaches zero and it never depends on where you started. After one half-life 50% of a dose is left; after two, 25%; after three, 12.5%; after four, 6.25%; after five, about 3.1%. That last figure is where the familiar rule of thumb comes from: after four or five half-lives, roughly ninety-five-plus percent of a dose is gone, and whatever is left is usually too small to do anything.
The rule works in both directions, which is the part that surprises people. Start dosing a drug at a fixed interval and each dose adds to whatever remains of the last one, until the amount cleared between doses equals the amount put in. That plateau is steady state, and it arrives on the same clock: four to five half-lives. Time to clear and time to plateau are the same calculation, read forward and backward. Everything else on this page is a consequence of those two sentences, or an exception to them.
The six numbers
The Ozempic injection label gives semaglutide an elimination half-life of approximately 1 week, and attributes it to albumin binding, which slows renal clearance and protects the peptide from enzymatic breakdown.[1] The oral semaglutide label repeats the figure for the tablet.[2] Tirzepatide is approximately 5 days on the Mounjaro label, which adds “enabling once-weekly dosing”, and approximately 5–6 days on the Zepbound label in patients with overweight or obesity.[3][4] The same molecule carries two stated half-lives, because they were measured in two populations.
Dulaglutide is approximately 5 days.[5] Liraglutide is approximately 13 hours, which the label says makes it “suitable for once daily administration”.[6] Exenatide twice daily has a mean terminal half-life of 2.4 hours, and its label adds that concentrations are measurable for about ten hours after a dose — a little over four half-lives, which is the rule of thumb behaving itself.[7]
The benchmark that makes those numbers legible sits on the liraglutide label itself: native GLP-1 has a half-life of 1.5 to 2 minutes, cut down by dipeptidyl peptidase 4 and neutral endopeptidases almost as fast as the gut releases it.[6] Going from under two minutes to a week is roughly a five-thousand-fold extension, and it is the entire engineering achievement of the class. The receptor being hit is the same one the body already has.
Why weekly works, in one ratio
Whether an interval is sane depends on the ratio of the interval to the half-life, not on either number alone. Dose a drug once per half-life and the trough sits at half the peak, and the plateau is twice what a single dose would give. Dose it far more often than the half-life and the concentration piles up into a nearly flat line. Dose it far less often and each dose clears before the next one lands.
Semaglutide injected weekly is dosed at exactly one half-life, so at steady state its trough is about 50% of its peak and it accumulates about twofold over a single dose. Tirzepatide at five days dosed every seven is 1.4 half-lives per interval, giving a trough near 38% of peak and roughly 1.6-fold accumulation. Liraglutide daily at thirteen hours is 1.85 half-lives per interval — a trough near 28% of peak and about 1.4-fold accumulation. Those three figures are derivations from the labeled half-lives, not label statements, and all of them describe a drug whose concentration never falls far between doses. That is what “once weekly” is actually buying: not convenience, but a trough high enough that the receptor is never unoccupied.
There is exactly one place in this class where the calculation can be checked against a manufacturer’s own measurement. The Trulicity label is the only one that prints an observed accumulation ratio: approximately 1.56 at steady state.[5] The textbook value for a five-day half-life dosed every seven days is 1.61. The formula and the measurement agree to within three percent, on the one label that makes the comparison possible.
The drug in the class that never plateaus
Exenatide twice daily is the exception that shows what the ratio is doing. Twelve hours between doses against a 2.4-hour half-life is five half-lives per interval, so the trough sits near 3% of the peak and the accumulation factor is about 1.03. Functionally the drug clears completely between every injection, which is why its label does not describe a steady state at all — it describes concentrations being measurable for about ten hours and then not.[7] Each injection is an independent event timed to a meal, not a contribution to a plateau.
So “steady state” is not a property of a drug. It is what happens when the dosing interval is short relative to the half-life, and it stops happening when it is not.
When the plateau arrives, and what the dose ladder is waiting for
The labels put steady state where the arithmetic puts it. Semaglutide injected weekly: 4 to 5 weeks, which is four to five half-lives.[1] The tablet, taken daily: the same 4 to 5 weeks.[2] Tirzepatide: 4 weeks, or about 5.6 half-lives at the Mounjaro figure.[3][4] Dulaglutide: 2 to 4 weeks.[5] Liraglutide, at thirteen hours, plateaus in under three days.
Line those up against the escalation schedules and the four-week step stops looking arbitrary. Tirzepatide starts at 2.5 mg and rises after at least four weeks on each dose; semaglutide climbs on the same four-week rhythm. In both cases the interval between dose increases is approximately the interval required to reach the plateau of the dose already in use. The ladder is not waiting for tolerance in some vague sense; it is waiting for the current dose to finish arriving. Anyone climbing faster than that is raising a dose whose full effect they have not yet felt.
Liraglutide is the outlier in the other direction. Its steps are weekly, which at a thirteen-hour half-life is about thirteen half-lives per step — three times more conservative, in pharmacokinetic terms, than semaglutide’s four-week step at four half-lives. The daily drug climbs on the proportionally slower clock, not the faster one.
One half-life, three persistence windows
Here the arithmetic and the labels come apart, and the labels are right. Semaglutide has a single stated half-life of about one week across every product. But the Ozempic injection label says semaglutide “will be present in the circulation for about 5 weeks after the last dose”,[1] the oral label says about five weeks after the last tablet,[2] and the Wegovy label says “about 5 to 7 weeks after the last injectable dose of 2.4 mg or 7.2 mg or oral dose of 25 mg”.[8]
Same molecule, same slope, two different answers, and the reason is the starting height. A larger dose begins its decay from a higher concentration, so it takes longer to fall below whatever threshold counts as “present”. Half-life fixes how fast a drug falls; the dose fixes how far it has to fall. This is why a reader who has memorized “five half-lives and it’s gone” will contradict the obesity label of the same drug they are taking. The slope is a property of the molecule; the finish line is a property of the prescription.
The washout that is not a half-life calculation
Every semaglutide label instructs discontinuation at least 2 months before a planned pregnancy. The Ozempic injection gives the reason as “the long washout period for semaglutide”; the Wegovy label gives it as “the long half-life of semaglutide”.[1][2][8] Two months is roughly 8.7 half-lives, at which point about a quarter of one percent of the last dose remains. The four-to-five-half-life rule would have given five weeks. The label is nearly twice as conservative, because the cost of being wrong is not symmetrical.
The part that decides the argument is an absence. Neither tirzepatide label states any pre-pregnancy interval, and neither contains the word “washout” or the phrase “long half-life” anywhere — at a labeled half-life of five to six days, longer than liraglutide’s by a factor of ten. Dulaglutide, at the same five days, carries no such interval either. Three long-acting agents; one washout instruction. Whatever produced the two-month figure, it was not a rule applied consistently across the class from the pharmacokinetics. What the instruction means for someone planning a pregnancy, and the separate contraceptive question that only tirzepatide raises, are in the pregnancy and contraception article.
The question the arithmetic cannot answer at all
The other long gap people are told to take is before general anesthesia or deep sedation, and half-life has almost nothing to do with it. The risk in question is residual gastric contents, which is a consequence of delayed gastric emptying rather than of plasma concentration, and the labels are unusually blunt about the state of the evidence: available data are insufficient to say whether modifying fasting instructions or temporarily discontinuing the drug would reduce retained gastric contents.[3][8] No approved label in the class derives a pre-procedure hold from a half-life, because the half-life does not predict the thing being avoided. The perioperative evidence, and the guidance that replaced the original blanket hold, are in the surgery and anesthesia article.
A missed dose, in units of half-lives
The reason a late weekly injection is usually uneventful is visible in the trough figures. A semaglutide user who is three days late has let the concentration fall to roughly 74% of its usual trough rather than to nothing, and a tirzepatide user three days late is at about 66%. Nobody in either case has washed out; they have slipped part of the way down one half-life. That is why the labeled windows are measured in a handful of days and not in hours. The windows themselves differ by product in ways the half-life does not explain, and are set out in the missed dose article; the percentages here are derivations from the labeled half-lives, offered to explain the shape of those rules rather than to replace them.
What a compounded vial inherits from the label
A compounded preparation of the same peptide, correctly made and correctly concentrated, would be expected to share the molecule’s pharmacokinetics, because half-life is a property of the compound and its albumin binding rather than of the manufacturer. That is the honest version of the claim. What does not carry over is the verification: the figures above come from human pharmacokinetic studies submitted for a specific product at a specific concentration, and the sellers listed on the compounded semaglutide board dispense preparations that are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before they reach a patient. A vial whose actual peptide content is unverified has an unverified starting height, and every calculation on this page begins with the starting height.
The one thing to carry away is the asymmetry. Four to five half-lives gets a drug to its plateau and gets most of it out again. But the labels round in whichever direction the harm lies: about five weeks to say the drug is still present, two months before a pregnancy, and no number at all where the evidence has not been gathered.