In a US claims analysis, patients who started tirzepatide were more likely to hit their blood-sugar and weight goals than patients who started semaglutide, and the accompanying math put tirzepatide's cost per success lower too. Every author on the paper works for Eli Lilly, which makes tirzepatide, or for the research firm Lilly hired.
Tirzepatide ↗ is the molecule behind Mounjaro and Zepbound. It hits two gut-hormone receptors at once, the GIP receptor ↗ and the GLP-1 receptor ↗, where semaglutide ↗, sold as Ozempic, hits only the second. Head-to-head trials have already shown tirzepatide drives larger drops in blood sugar and weight. The open question is what happens outside a trial, where people skip doses, switch drugs and fall off treatment. The new study, published in the Journal of Managed Care and Specialty Pharmacy ↗, used insurance claims and lab results to look.
The setup was a retrospective match, run on claims and lab data from the Healthcare Integrated Research Database by Carelon Research under contract to Lilly. Adults with type 2 diabetes who started tirzepatide (Mounjaro only) or semaglutide (Ozempic only) between May 2022 and May 2023 were paired by propensity score, then split by whether they had used a GLP-1 drug before. That left 10,702 matched pairs of first-timers and 5,577 pairs who had been on a GLP-1 drug already. Over 12 months, tirzepatide starters stuck with treatment more: adherence ran 60 percent versus 45 percent in the first-timers, and persistence, meaning no gap of 45 days or more, ran 62 percent versus 47 percent. More of them also hit every blood-sugar and weight target measured.
Then comes the number the paper is built to deliver: cost per responder. It is a simple division, the average pharmacy spend divided by the share of patients who reached a given target. On the strictest goals the gap looks enormous. To get one first-time patient down to an A1c below 5.7 percent, essentially a non-diabetic blood-sugar level, cost $45,300 with tirzepatide and $65,300 with semaglutide. Among patients who had already tried a GLP-1 drug, that figure was $107,700 versus $270,300. For getting one patient to lose at least 15 percent of body weight, it was $62,900 versus $124,400 in first-timers and $110,200 versus $326,800 in the experienced group.
Those three-to-one spreads are real arithmetic, and that is the problem with reading them as three-to-one value. Cost per responder divides by the fraction of people who hit the target, so when that fraction is tiny the number explodes. Only 6 percent of the experienced semaglutide patients reached an A1c under 5.7 percent; dividing a year of drug cost by 0.06 is what produces a $270,300 figure, not a special failure of the molecule at the margin. The authors say so themselves, in the fine print: for the looser, more common A1c goals, cost per responder was similar between the two drugs. The eye-popping gaps sit precisely where the fewest patients actually live.
The rest of the caveats are the standard ones for claims research, and they cut the same way. This was not a randomized trial, so propensity matching balances the traits the researchers could measure and leaves the ones they could not, including why a given patient landed on one drug rather than the other and what their copay looked like. Adherence differences can reflect formulary design and out-of-pocket cost as much as the drug itself. And the party choosing which targets to feature, and paying for the analysis, sells the drug that wins.
None of this makes the underlying signal wrong. Tirzepatide's advantage on weight and blood sugar is well established, and better adherence is a genuine finding worth its own study. The caution is narrower: cost per responder is becoming a favorite tool for payers deciding which drug to cover, and it bends hard on the choice of threshold and denominator. When a manufacturer picks both, the metric deserves to be read as an argument, not a measurement. Both drugs anchor peptidemodel's GLP-1 receptor ↗ map, with tirzepatide reaching across to GIP ↗ as well.