Tirzepatide's half-life is about five days, Lilly researchers report, explaining once-weekly dosing

A population pharmacokinetic analysis built from 19 pooled clinical studies puts tirzepatide's half-life at roughly five days, a property the authors say supports once-weekly subcutaneous dosing ^[1]. The paper was published February 14, 2024, in the journal CPT: Pharmacometrics & Systems Pharmacology ^[1].

Tirzepatide, the molecule in the U.S. brands Mounjaro and Zepbound, is described in the paper as a first-in-class glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1) receptor agonist ^[1]. The authors are Karen Schneck and Shweta Urva of Eli Lilly and Company's Global PK/PD & Pharmacometrics group in Indianapolis ^[1].

Their model is a two-compartment model with first-order absorption and elimination ^[1]. It reports a mean half-life of 5.4 days, an apparent clearance of 0.061 L/h, and 1.7-fold accumulation of the drug when doses are given weekly ^[1]. The published abstract describes the half-life as approximately 5 days ^[1]. Mean absolute bioavailability is about 80%, a value incorporated into the model as a fixed parameter ^[1].

The analysis pooled data from healthy participants, people with kidney impairment, people with liver impairment, and people with type 2 diabetes enrolled in phase 1, phase 2 and phase 3 trials ^[1]. Those trials included SURPASS-1 through SURPASS-5 and the Japanese trials SURPASS J-mono and SURPASS J-combo ^[1]. Doses across the studies ranged from a single 0.25 mg dose up to 15 mg once weekly, and samples were collected after up to 109 weeks of continuous treatment ^[1].

Blood levels were measured with a validated liquid chromatography-mass spectrometry assay with a quantification range of 2 to 500 ng/mL; about 16% of collected samples fell below that lower limit and were excluded from the analysis ^[1]. Clinical pharmacology studies used serial sampling, while phase 2 and phase 3 trials used sparse sampling, with participants in three phase 3 studies randomly assigned to sampling windows of 1 to 24 hours, 24 to 96 hours, or 120 to 168 hours after a dose ^[1]. Modeling was done with NONMEM 7.4.2 ^[1].

A covariate analysis found that no dose adjustment was necessary for kidney impairment, liver impairment, body weight, sex, age, race or ethnicity ^[1]. Body size was handled with what the authors call a semimechanistic allometry model ^[1]. The approved regimen described in the paper starts at 2.5 mg once weekly, can rise to 5 mg after 4 weeks, and can be increased further in 2.5 mg increments after at least 4 weeks on a dose, up to a maximum of 15 mg once weekly ^[1]. The paper says the same simulation work supports the labeling instruction that a missed weekly dose can be taken within 4 days.

Why it matters for patients

A drug's half-life is the main reason a weekly shot schedule works. At about five days, tirzepatide clears slowly enough that one injection keeps drug levels steady across a seven-day week, which is why the labeled schedule is once weekly rather than daily ^[1]. A half-life in that range also means a dose change does not reach its full effect immediately; the 1.7-fold accumulation reported in the model describes how levels build when weekly doses are repeated ^[1].

Because the analysis found no meaningful effect from body weight, age, sex, race, ethnicity, kidney function or liver function, it supports the idea that most adults taking the drug follow the same weekly schedule rather than individualized dosing ^[1]. That is a statement about drug levels in the body, not a promise about how well the drug works or how well it is tolerated for any one person. The paper does not report whether differences in blood levels explain differences in weight loss, blood sugar control or side effects between patients.

What happens next

The paper is open access and presents no new clinical trial results, and it does not announce any label change or new study ^[1]. The authors write that the model can be used to predict tirzepatide exposure for various scenarios or populations, which could matter for groups that were not directly studied ^[1]. Whether the model will be used to support dosing in children, during pregnancy, or in other groups has not been disclosed.