A photo of Susan Dell at the White House went viral, and the internet did what the internet does, splitting into two camps. One camp went straight to eating disorder speculation. The other camp went straight to Ozempic jokes.
Neither camp knows anything about her. That's the part that should bother people more than it does.
Daniel's asking us to use that moment as a launchpad, and he's explicit about not wanting to speculate about Susan Dell as a person. Fair enough. What he wants is the real thing underneath the internet's two favorite explanations. He says he knows relatively little about Ozempic, but he takes a stimulant for ADHD five times a day, and he knows that very strange feeling of your appetite being inexplicably missing. His words: not a particularly satisfying one.
And then it wears off.
And then it wears off, and you're ravenously hungry, and Daniel says there's something delightful about rummaging through the fridge for whatever calories you can find.
That's the tell, isn't it. For most people that rebound is an annoyance. For someone with anorexia, the absence itself is the point.
Which is the whole question. What does the data actually show about misuse in that population, how is the regulatory regime reacting as these cases get more prominent, and what does it feel like to have hunger chemically switched off. So let's take the internet's two explanations and look at what's actually underneath them.
Start with the vocabulary, because the two things get collapsed constantly and they're not the same. Anorexia nervosa is a fear of gaining weight plus a disrupted relationship with food. Body dysmorphia is different. It's a distorted body image, a mismatch between what's there and what the person sees. They overlap in a lot of patients, but one is about eating and the other is about perception.
And there's the atypical version.
Atypical anorexia is the one that trips people up, because the BMI sits within normal limits. So a person looks fine, and the clinician's instinct says fine. The literature is blunt about this. The risk of serious complications tracks the amount of weight loss, not the absolute weight. That's from a case report in BJPsych Open last year, and it means someone at a normal weight can be in serious danger.
Which kills the standard reassurance. If she's not underweight, she can't be that sick.
Maybe the single most costly misconception in the field. On scale, lifetime prevalence of anorexia runs as high as six point three percent in women and zero point three percent in men. Mortality from any cause is more than five times the general population. And that's from a New England Journal of Medicine perspective published this April.
Five times. That puts it near the top of the psychiatric mortality table, and I still don't think most people would guess it.
They wouldn't. Now put the other number next to it. One in eight American adults is currently on a GLP-1, and one in five has taken one at some point. That's not a niche drug anymore. That's a mass-market product with a mechanism that does, physiologically, a lot of what anorexia does. And I want to be careful with that framing because it sounds like a slogan, so let's actually walk the mechanism.
Walk it.
GLP-1 is an incretin hormone your gut releases after you eat. Semaglutide mimics it. Three things happen. It stimulates insulin release. It slows gastric emptying, so food sits in the stomach longer, which produces both the nausea and the sense of being full. And it crosses the blood-brain barrier and binds receptors in the nucleus of the solitary tract, which is where appetite and energy intake get dialed down.
So it's not a stimulant suppressing hunger as a side effect. The appetite suppression is the designed action, at the level of the brainstem.
Wendy Oliver-Pyatt, who runs Within Health and Galen Hope, put it flatly. These medications do the same things that actual anorexia does. And I think the reason clinicians keep reaching for that sentence is that eating disorder treatment is, at its core, a relearning. Patients have to reacquaint themselves with hunger cues, learn to trust them, eat in response to them. A drug that mutes those cues doesn't just sit alongside that work. It undermines it.
You're teaching someone to read a signal while a medication is turning the signal down.
And the patient is often relieved that it's quiet. That's the trap.
Okay. So what happens when people with eating disorders get hold of these drugs. There's actual survey data now.
There is, and it's the first study of its kind. Published in JAMA Psychiatry in June. They surveyed four hundred and thirty-six people with eating disorders. Thirty-two percent reported lifetime GLP-1 use, against roughly fifteen percent in the general adult population. Twenty-two percent were current users. Ten point one percent reported lifetime misuse, and misuse in that study meant taking more or less than prescribed, tampering with the injectors, or sharing medication. Nearly ten percent had used compounded products.
So the misuse isn't just dose drift. Tampering with an injector is a fairly deliberate act.
It is. And then the adverse events among the hundred and forty lifetime users are the part I'd linger on. Loss of appetite, eighty-one percent. Nausea, sixty-six. Diarrhea or constipation, fifty-six. Stomach pain, thirty-two. Headaches, twenty-two.
Eighty-one percent loss of appetite. In a general population that's the headline side effect everyone jokes about. In this population, that's the thing they came for.
Which is the whole problem in one number. That's not a tolerable side effect they're enduring. It's a feature. And the individual case reports show what that looks like at the sharp end. There's a study from Tulsa in Obesity Reviews where patients were restricting to three or four hundred calories a day. One patient went on a liquid diet of water and Diet Coke and did not eat for thirteen days. Lost twenty-one pounds in under two weeks.
Thirteen days. It stops being a diet at some point and becomes a different activity entirely.
Another participant described fasting six days and eating once a week. And then there's the adolescent case report from BJPsych Open. A girl lost roughly twenty kilograms in six months on zero point two five milligrams of semaglutide. That is the lowest starter dose. The floor of the dosing ladder. And her atypical anorexia worsened on it.
That one reframes the whole risk. You don't need a high dose or a long prescription history to tip someone who's already vulnerable, because the vulnerability is doing most of the work.
Right. And I should put the counterweight on the table too, because the story is double-edged. GLP-1s show real promise for binge eating disorder. There's a whole separate literature there with encouraging signals. But they're contraindicated for restrictive disorders. Same drug, opposite direction. And it's worth saying that the most rigorous controlled trial did not find greater binge reduction than placebo, despite greater weight loss. So even the promise side isn't settled.
Same molecule helps one eating disorder and harms another. That's not a side effect you can screen your way around with a warning label. That's a treatment decision that requires someone actually knowing what's in front of them.
And that gets us back to Daniel's experience, because I think it's the honest bridge here. He describes the appetite being inexplicably missing. Not hunger that's been satisfied, just gone. Then the rebound when it wears off.
I've watched him at the fridge. The man eats like someone who's making up for lost time.
What's interesting is that he called it unsatisfying, and I think that's exactly right for a normal person. You don't get the pleasure of the meal, and then you get the debt. For someone with anorexia, that absence is not a bug. It's the thing they've been chasing, sometimes for years. The Tulsa patient felt such a strong absence of hunger that she planned to fast six days a week. That's a person who found the off switch and decided to leave it off.
And the clinical guidance has started naming that. Extreme appetite loss as a monitoring trigger.
The Delphi panel flagged exactly that. Extreme appetite loss isn't a success metric in this population. It's a warning.
So the individual picture is coherent, and it's bad. What happens when you scale that to millions of prescriptions and a regulatory system that was never built for this?
The regulatory picture is the strangest part of the whole episode, because the alarm from clinicians and the signal from the regulators are out of sync. Start with the labels. FDA labels on these drugs do not list eating disorders as a side effect.
At all.
And when reporters asked the FDA directly whether it has concerns or is examining the issue, it didn't answer the question. It said it monitors the safety of drug products after they enter the market. Which is a sentence that says nothing.
That's the kind of statement you write when you'd rather not be quoted.
The EMA's pharmacovigilance committee went further and concluded no product label update was warranted on psychiatric risk. And the FDA similarly found no clear link between GLP-1s and suicidal ideation. So you have two major regulators looking at the psychiatric data and saying the signal isn't there.
And on the other side you have clinicians saying they're seeing it daily. Both of those can be true at once, which is the uncomfortable part. A population-level adverse event signal is a blunt instrument. It picks up things that are common and distinctive. It will absolutely miss a harm that concentrates in a subgroup.
And the screening gap is where that becomes concrete. Medical societies recommend routine eating disorder screening when you prescribe a GLP-1. In practice, such screenings are not routinely conducted. The American Academy of Family Physicians hasn't adopted screening recommendations at all. Their position is that the drugs are one tool and they reject generalized recommendations driven by a single clinical concern.
That's a defensible position in the abstract, and it collides with a clinic that's drowning in it.
Now the counterweight, because there's finally a consensus document. A modified Delphi consensus published in World Psychiatry this month. Forty-five panelists, average consensus ninety-one point seven percent. On the key question, whether GLP-1 receptor agonists should generally be avoided in active anorexia nervosa, atypical anorexia, and bulimia with marked restraint or weight suppression, the consensus was one hundred percent.
Total.
Total. Ninety-four point seven percent supported screening before prescribing. Eighty-four point two percent supported monitoring during and after treatment. So the expert opinion is about as unified as expert opinion gets. The problem is a consensus statement is a fax to clinicians who are already busy.
And a fax isn't a gate. Which leads us to telehealth.
The telehealth part is where the theory dies. There's a secret-shopper study in JAMA where researchers posed as patients on online vendor sites. The finding was that vendors often required no interaction with a health care professional at all. They relied on self-reported questionnaires. You check boxes, you pay, the medication ships.
There was a Washington Post investigation in June that put flesh on that.
It documented patients lying on telehealth forms to get GLP-1s. One woman, Stevee Williams, twenty-seven, was quoted saying, in reality, I just wanted to be smaller. That's not a fraud story. That's a description of the actual demand curve.
People are checking the boxes that unlock the prescription. And the boxes are the entire clinical evaluation.
Which is why Jaclyn Siegel at Villanova said her primary concern is that people with thinness-oriented disordered eating might be using these drugs to achieve rapid weight loss, and that it's concerning people could circumvent the standard safeguards. The safeguards, in a lot of the online channel, are a questionnaire.
So what does the scale of the harm actually look like. Are we talking dozens of cases, or something larger.
The NEJM perspective projected more than four hundred and twenty thousand people could develop an eating disorder with long-term use. That's just over one percent of the thirty-three million Americans who've reported taking them. There's a separate analysis of medical records that's more concrete. Out of more than sixty thousand people taking GLP-1s, one point two eight percent were diagnosed with an eating disorder within two years.
One point two eight percent sounds small until you put it on thirty-three million.
And then the honest part, which is the research gap. A scoping review in the Journal of Eating Disorders in June screened eighty records and found only nine eligible studies. Most were small. Most focused only on binge eating. And none of them looked at people using the injections without medical supervision.
None.
None. So the population most likely to be misusing these, the ones buying online with no doctor in the loop, is the population with zero data. And there's no validated eating disorder screening tool specific to GLP-1 users as of this fall. Siegel's line is that we need a whole new set of tools to understand what an eating disorder might look like for a person on one of these drugs.
Because the standard screening instruments were written for people without a pharmacological appetite suppressant on board.
Right. If the drug is suppressing appetite, the question "are you eating less than you used to" has a completely different meaning. The instrument doesn't know that.
And the big causal question, does the drug create the disorder or unmask one that was already there, is still open.
Wide open. Evelyn Attia at Columbia put the honest version of it. We don't know enough to truly predict how patients with eating disorders may fare on these medications, but the worry is that someone with a history of the pursuit of thinness may experience worsening symptoms, or a return of previously quiescent ones. That's a careful statement and it leaves the question exactly where it is.
Which is the version of the science I actually trust more than a confident answer.
Now the cultural layer, because this is where it stops being a pharmacology story. The internet noticed something in that photo. The term is Ozempic face. Facial hollowing, volume loss. It can show up as early as eight to twelve weeks on the drug, which is faster than most people would guess.
And the aesthetic complaint and the clinical warning are two readings of the same phenomenon.
That's the thing I keep circling back to. The hollowing is a marker of rapid weight loss. Rapid weight loss is exactly the thing that signals danger in an eating disorder. So the internet's rude joke and the clinician's alarm are describing the same face. One is laughing and one is worried, and they're both pointing at the same rapidity.
That's unsettling when you lay it out like that. The evidence of harm is legible from the outside, we've just trained ourselves to read it as style.
And then there's Rebecca Boswell at the Princeton Center for Eating Disorders describing eighth-graders singing the Ozempic jingle in unison. Her word for it was unnerving. Rebecka Peebles at Monte Nido said they have at least three people presenting daily who are on a GLP-1, not for diabetes, but because they were put on it for weight loss.
Eighth-graders with the jingle memorized. That's the cultural normalization arriving before the clinical guidance does.
Elizabeth Wassenaar at the Eating Recovery Center said these medications are being sold as no big deal, when in fact they are quite a big deal. That's the whole tension in one sentence. Regulators see no clear signal, clinicians see a crisis, and the culture has already decided it's a lifestyle product.
No big deal is doing an enormous amount of work in that sentence.
It's not the drugs.
I said it's not the drugs. You've been talking like the problem is the molecule. The molecule's fine. It does what it says on the label. The problem is the whole thing runs on the assumption that people will tell the truth about their own bodies, and nobody does. The telehealth form is a joke. Anyone who's ever been desperate to be smaller knows exactly which boxes to tick. That's not a gap in the screening. That's the design.
You're saying the screening gap is a feature.
I'm saying it's not a gap, Corn. It's working exactly the way the business needs it to work.
That's the bit that doesn't show up in a consensus statement. The panel can recommend screening, but screening is a form a patient fills in about themselves.
I knew a guy ran a wellness clinic. Maybe two thousand nine. Fax machine, white coat, stack of pre-signed prescription pads. That was the operation. He'd prescribe to anyone who could pay the consult fee. Business was respectable enough until the weight loss drugs hit, and then it wasn't respectable at all, it was a queue out the door.
Pre-signed pads.
He'd sign a hundred at a time and have the girl at the front fill them in.
Is that even legal?
Legal? It was a fax machine. Nobody asked.
Nobody asked because nobody was in a position to ask. That's the same structure as the online vendors.
The clinic got shut down eventually. Not by the board. By his mother. She worked at the state licensing office and she recognized his handwriting on a complaint form.
His own mother.
She called him at home.
Who filed the complaint?
I did. He shorted me on a referral fee.
That's the ending. That's how it ends.
It's how it ended. The fee was two hundred dollars. I never saw it.
The pads never showed up in any of the investigations, did they.
No. Nobody asked.
Okay. So where does that leave us.
Where it leaves us is that the questions the episode opened with are open. Whether GLP-1s cause eating disorders or simply unmask them is undetermined. There's no validated screening tool for this population. And there are no studies at all on misuse among people who never went through a doctor in the first place.
The numbers on both sides are moving at the same time. The drugs keep scaling, one in eight adults and climbing, while the clinical alarm gets louder and the regulatory response stays quiet. Those two curves aren't going to reconcile on their own.
The line I'd leave people with is that the question was never whether the drugs work. They work. The question is what happens when they work exactly as intended, in a person whose relationship with hunger was already broken.
That's the version that doesn't fit in a jingle.
It doesn't. Thanks to Hilbert Flumingtop for producing, and for the referral fee story nobody asked for.
If this was your kind of episode, go back for episode thirty-five, The Privacy Gap. This has been My Weird Prompts. Send us your own prompt on Telegram at t dot me slash MWP listener bot.
We'll be back soon.
