Stretching Doses to Save Money

At over a thousand dollars a month, people stretch these drugs — injecting every two weeks instead of every one. Two published papers argue this may keep most of the weight loss.^[1][2] The pharmacological reasoning is sound and worth understanding. The evidence behind it is a mathematical model and a case series of two people, and you should know that before deciding anything.

Why the idea is not silly

The intuition that halving your doses halves the effect is wrong, and understanding why is useful whether or not you ever change anything.

Semaglutide has an elimination half-life of roughly a week. That means a dose is still half present seven days later, and with weekly dosing the drug accumulates to a plateau over several weeks. Stretch to fortnightly and you do not drop to zero between doses — you settle at a lower plateau with wider swings. Since the relationship between drug level and appetite suppression is not a straight line, a lower plateau does not translate into a proportionally smaller effect.

You are not skipping the drug. You are running at a lower steady level.

That is the entire argument, and it is a reasonable one. It is also, so far, an argument rather than a measurement.

What the two papers actually contain

Both papers come from the same research group, which is normal for a new line of work and means these are not two independent confirmations of anything.

A 2026 case series from an unrelated group reports the same practice, and is covered separately in every-other-week dosing.^[3] Independent authors observing it is worth something. It does not change the shape of the evidence, which remains a handful of patients observed rather than randomized.

What the models leave out

• “An appropriate increase in dose size” assumes a dose you can deliver. Pens and vials come in fixed strengths. The regimen that preserves 100% of the effect in a simulation may not exist as a physical object.
• The tirzepatide modeling is borrowed. The authors built it by taking the semaglutide pharmacodynamic model and reparameterizing it to fit tirzepatide data — a fitted adaptation, not an independent model of that drug.
• Tolerability is not modeled. Titration schedules exist because larger jumps in exposure cause more nausea and vomiting. A bigger, less frequent dose is exactly the pattern those schedules are designed to avoid, and nothing here tested it.
• Nothing addresses diabetes. These are weight models. Someone whose glucose control depends on steady drug levels is in a different situation entirely.

A drug built for the interval, tested against a weekly one

The strongest evidence that fortnightly dosing can work is not about stretching an existing drug at all. Bofanglutide is engineered for a two-week interval, and a phase 2b trial put it head to head against weekly semaglutide in 272 people with type 2 diabetes.^[4]

A fortnightly injection beat a weekly one. That is a genuine demonstration that the interval is achievable — when the molecule is designed for it, with a half-life and a dose chosen to hold levels across fourteen days.

One more oddity worth noting: the doses do not order cleanly. Eighteen milligrams outperformed twenty-four on the same schedule — −2.28% against −1.94% — and the 24 mg fortnightly interval was the only arm whose interval touched zero. More is not reliably more, here or anywhere else in this class.

The public health argument, separately

The modeling paper makes a second claim worth separating from the first. Given a fixed national budget, it simulates treating a certain number of people weekly against treating twice as many fortnightly, and finds the second reduces population obesity and mortality considerably more.

That is an argument about resource allocation rather than about any individual’s treatment, and it does not depend on fortnightly dosing being as good — only on it being better than nothing, which is a far weaker requirement. It is a genuinely interesting policy point and it rests on the same unvalidated model.

If cost is the reason you are reading this

Cost ending treatment is the most common way these drugs fail, and it is a real problem that deserves better answers than a spreadsheet. The routes with actual evidence behind them are different from this one.

• A lower labeled dose has randomized evidence: reducing rather than stopping held much of the loss, covered in the maintenance trial.
• Handing over to a different agent was tested directly — a pill maintained 79.3% of the reduction against 37.6% on placebo, in the oral handover.
• What people actually do after stopping, and how little the average person regained, is in its own piece.
• Current prices by seller are on the price tracker, dated.

All of which is to say the fortnightly idea may well turn out to be right. Someone deciding today is deciding on a model and two people, and that is a materially different thing from deciding on a trial.