A group of physicians and Prader-Willi syndrome experts warned clinicians this week about a safety signal surrounding Vykat XR, the first drug approved to curb the relentless hunger of Prader-Willi syndrome and now sold by Neurocrine Biosciences. Since the drug's approval in March 2025, seven patients taking it have died, and more than 100 have suffered serious complications, mostly hospitalizations for swelling, respiratory, and cardiac problems, according to reports submitted to the Food and Drug Administration. The experts said their aim was to raise awareness of the risks for people with the syndrome who are starting the drug.
They also added a caveat that is easy to skim past and essential to understand: none of the deaths or severe events has been definitively linked to Vykat. That is not a legal hedge or an act of caution for its own sake. It points to the genuinely difficult problem sitting at the center of this story, and at the center of drug safety in general, which is the problem of telling harm caused by a drug apart from harm caused by the disease the drug is meant to treat.
The attribution problem
When a very sick person taking a new drug dies or ends up in the hospital, the hardest question is also the most basic one: was it the drug, or was it the disease? For Prader-Willi syndrome, that question is genuinely, technically hard to answer, because the syndrome itself carries a high baseline risk of exactly the kinds of events now being reported.
Prader-Willi drives an insatiable, unrelenting hunger that leads to severe and dangerous obesity, which in turn produces cardiac and respiratory complications, and the syndrome carries an elevated risk of early death even in patients who never take any drug at all. So the swelling, the breathing difficulties, and the heart complications appearing in these reports are also things that happen to people with Prader-Willi who never took Vykat, as a consequence of the disease itself. That overlap is precisely what makes attribution so difficult here. It is why the experts said the events have not been definitively linked to the drug, not to shield anyone, but because it is honestly hard to know which cause is at work in any given case.
What an adverse-event report is, and isn't
To read this news correctly, it helps to understand what the underlying numbers are. The seven deaths and the hundred-plus serious events come from the FDA's adverse event reporting system, a passive database that collects reports of things that happen to people while they are taking a drug. A report in that system means "this happened to someone on the drug." It does not mean "the drug caused this," and the distinction is not a technicality.
Such reports are shaped by several forces that pull against any simple reading. There is reporting bias: a newly approved, closely watched drug attracts far more adverse-event reports than a long-established one, and a death in a patient on a new drug is much more likely to be noticed and reported than the same death in a similar patient who is not. There is confounding: the disease itself produces events that look just like the ones being reported. And there is usually no denominator or control group, no easy way to compare the rate of these events among patients on the drug with the rate among comparable patients off it. For all these reasons, seven deaths and a hundred serious events constitute a signal, a reason to investigate carefully, rather than a finding that the drug is dangerous. The experts raising it are the safety system functioning as designed, not a verdict being handed down.
Why the signal still deserves to be taken seriously
None of that means the signal should be dismissed, and there are solid reasons to take it seriously. Diazoxide, the active ingredient in Vykat, has well-documented effects on the body's fluid balance and cardiovascular system; it is known to cause fluid retention and swelling and to affect the heart. That makes the reported complications biologically plausible as drug effects rather than far-fetched ones, and biological plausibility combined with a temporal association, events clustering after patients start the drug, is exactly the pattern that warrants serious scrutiny.
It still falls short of proof at the population level, because those same events are plausible as consequences of the disease. But "plausible, and worth rigorous investigation" is precisely the correct reading of the situation, sitting squarely between the two wrong ones. This is neither a confirmed danger nor a nothing to worry about. It is an open and serious question, and treating it as anything more certain in either direction would be a mistake.
Both mistakes are costly
What makes this genuinely hard, rather than merely uncertain, is that erring in either direction carries a real cost paid by real patients. Over-read the signal, and you risk frightening patients and families away from the first and only approved treatment for a merciless condition. Untreated Prader-Willi hunger is genuinely devastating on its own, dominating lives and endangering health, so scaring people off an effective therapy is not a safe default. Under-read the signal, and you risk missing a real harm and allowing patients to keep taking something that is injuring them.
The history of drug safety contains both kinds of error in abundance: real signals brushed aside for too long, and useful drugs abandoned over harms that turned out to belong to the disease rather than the treatment. There is no direction in which carelessness is safe. The responsible path is the one the experts actually took, which is to raise awareness, push for rigorous investigation, comparing the observed rate of events against what would be expected in Prader-Willi patients anyway, and looking for dose-response relationships and biological patterns, and to give clinicians and families the information they need to make individualized decisions. It is worth noting that the warning asks doctors to increase their awareness of the risks, not to stop prescribing the drug.
This is often how rare-disease safety gets learned
It helps to step back and see that this situation is not an aberration. It is characteristic of how safety works for rare-disease drugs, for reasons built into the structure of the problem. Small patient populations mean small clinical trials, which have limited statistical power to detect rare but serious harms before a drug is approved. The diseases in question are often serious and carry high baseline mortality, which muddies attribution in exactly the way it does here. And desperate unmet need creates real pressure to approve promising drugs on comparatively limited data.
The result is that rare-disease drugs tend to reach patients carrying more unresolved safety uncertainty than a typical drug for a common condition, and much of the safety picture necessarily fills in after approval, in real-world use. A post-market signal like this one is therefore not a sign that the system failed. It is the system doing what it must for a class of drugs whose safety cannot be fully established in advance, learned in the open, after approval, on the very patients who most needed the drug to exist. That is an uncomfortable truth, and often a defensible tradeoff for people with no other options, but it deserves to be stated plainly rather than discovered by surprise each time.
In the end, the news that seven people taking Vykat have died sounds like a verdict, and it is not one. It is a question, an urgent and important question, that the experts were right to raise and that now must be answered through the slow, unglamorous work of separating a drug's harm from the ravages of a merciless disease. For families living with Prader-Willi syndrome, this is an agonizing place to stand. The drug is the first real help for a hunger that dominates their children's lives, and now it carries a question mark. The honest thing to tell them is neither that it is killing people nor that it is fine, but that a signal has appeared, that it is being taken seriously and studied, that it has not been shown to be the drug's doing, and that the right response is a careful conversation with their own physician about their own situation, not panic, not a shrug, and above all not stopping a prescribed medicine on the strength of a news report. The larger truth is that for rare and deadly diseases, this is often how safety is learned: after approval, in the open, on the people who needed the drug most. Holding the hope and the caution at once, without collapsing into either, is the only honest way through.
Primary sources
- STAT, in reporting and analysis by Adam Feuerstein, for the statement issued this week by physicians and Prader-Willi syndrome experts warning clinicians of a safety signal, the report that seven patients taking Vykat XR have died and more than 100 have experienced serious adverse events, mostly hospitalizations for swelling, respiratory, and heart complications, per the FDA's adverse event reporting system since the drug's approval, the drug's role in curbing the intense hunger of Prader-Willi syndrome, its sale by Neurocrine Biosciences, the experts' stated aim of increasing awareness of the risks, and their explicit note that neither the deaths nor the severe side effects have been definitively linked to the drug.
- STAT's earlier reporting on the FDA's March 2025 approval of Vykat XR.
- General, well-established background on Prader-Willi syndrome, including its hyperphagia, associated severe obesity, cardiac and respiratory complications, and elevated baseline mortality, on the pharmacology of diazoxide, including its known effects on fluid retention and the cardiovascular system, and on the interpretation of passive adverse-event reporting systems, including reporting bias, confounding by the underlying disease, and the absence of a control group.