The Cardiovascular Cost of Stopping

We recently covered what actually happens to people’s weight after they stop these drugs, and the answer was reassuring: an average of half a percent regained in a year. A review in Nature Reviews Endocrinology raises a different concern — that the cardiovascular cost of stopping may not show up in a weight number at all.^[1] It is a hypothesis, and the review is clear that it is one.

The argument

It has three parts, and each is worth separating because they carry different amounts of evidence.

1. Fluctuation itself may be harmful. Weight and HbA1c variability are established risk factors for cardiovascular and microvascular events, independent of where the values settle. A cycle of losing and regaining could be worse than either state held steady.
2. The benefit may not outlast the drug. The review notes these medicines lack anti-atherosclerotic and plaque-stabilizing effects — unlike statins, which change arterial structure. If the cardiovascular benefit depends on continued metabolic effect, it stops when the drug does.
3. Risk may be elevated acutely on withdrawal, rather than gradually returning to baseline.

Underneath all three sits the plainer point: re-developing obesity and losing glycemic control returns someone to the risk profile they had before, and time spent at higher risk accumulates.

What is missing, in the review's own words

Few data are available regarding the incidence of hard outcomes in people discontinuing such drugs.

That sentence governs everything above.^[1] Nobody has measured heart attacks, strokes or deaths in a population that stopped these drugs, against a comparable population that continued. The argument is built from mechanism and from what is known about variability in other contexts — which is a legitimate way to raise a question and not a way to answer one.

How it fits with the empirical picture

The two findings are not in conflict, and holding both is the right position. Weight regain after stopping is, on average and in ordinary care, modest — largely because most people do something else afterward. That is a real observation about a real population.

The cardiovascular question is different and unanswered. It is possible for average weight regain to be small while a subset who lose and regain repeatedly accumulate risk that no average captures. The individual variability noted in the discontinuation data leaves plenty of room for that.

There is also a middle option that this argument makes more attractive: not stopping, but reducing. The randomized evidence on lowering the dose rather than discontinuing is covered in the maintenance trial, and rescue therapy there was needed by 8% who stayed on full dose, 25% who stepped down, and 67% who stopped.

What to do with an unanswered question

• Treat cost as a clinical problem, early. If affordability is likely to end treatment, that belongs in the conversation before starting, not after stopping.
• Ask about reducing before stopping. A lower dose is a real option with randomized support behind it.
• Avoid the cycle if you can. Repeated stopping and restarting is the pattern this hypothesis is most concerned about.
• Do not read this as a reason to panic about a single interruption. Nothing here establishes that, and the review says as much.