Somewhere in a poison control center, a phone rings, and the person on the other end has been injecting their weight-loss medication every single day. They were supposed to take it once a week. Worse, they started at the full dose rather than easing in. The call is logged, coded in near real time by a specialist trained in clinical toxicology, and fed into a national database. Multiply that scene by thousands, year after year, and you get one of the stranger public health signals of the decade: a drug celebrated for shrinking waistlines was flooding the country’s poison hotlines.
For most of the 2010s, those hotlines fielded somewhere between 1,000 and 1,500 calls a year about this entire class of drugs. By 2023, the number had climbed past 8,000. Something had changed, and it had a date.
The class in question is the glucagon-like peptide-1 receptor agonists, GLP-1RAs for short, the family that includes semaglutide, the active ingredient sold as Ozempic and Wegovy. These drugs mimic a gut hormone, slowing how fast the stomach empties, dulling appetite, nudging the body toward satiety. They began life as diabetes treatments. Then, on 4 June 2021, the US Food and Drug Administration approved once-weekly semaglutide for chronic weight management, and a medical product became a cultural one.
Jordan Miller wanted to know whether the timing was coincidence or cause. As an undergraduate at the University of Texas at San Antonio, she went looking for a research question and found one hiding in the noise of poison center data.
“One of them was this quite odd category of semaglutide,” recalls David Han, the Romo Endowed Professor in the university’s Department of Statistics and Data Science, who supervised the work. “We suspected that the call volume was skyrocketing because of the misuse and mishandling of this drug and that it may be attributed to the FDA approval of this drug for weight management.” That was the hunch. The job was to prove it.
A Different Crowd, A Different Story
Working with poison-center colleagues from the Long School of Medicine, the team pulled twelve years of records from the National Poison Data System, every human GLP-1RA exposure logged between 2012 and 2023. They drew a line at 1 July 2021, just after the FDA’s decision, and asked what looked different on either side of it. The answer was: nearly everything.
Before the cut-off, the records held 3,113 exposures. After it, 6,920, roughly 2.2 times as many. Semaglutide, which had accounted for about a quarter of cases beforehand, now dominated at 64.2 per cent. And the people behind the calls had changed. The typical patient got younger (the mean age of new cases dropped by about a decade, to 47.5) and far more likely to be female, with women rising from 68.9 to 78.2 per cent of cases. This was no longer a diabetes story.
“In that figure that tracks the increase by specific drug, I wasn’t expecting semaglutide to be so incredibly dominant,” Miller said. “I figured that it would lead the pack, but it was staggering. On the other hand, it makes sense with all the media attention.”
Han frames the divide bluntly. “When the GLP-1[RA] drugs are being sold to diabetic patients, that’s a completely different story versus when the drug is used for weight management,” he said. A long-managed diabetic population is one thing. A vast new wave of first-time users, many self-directed and primed by relentless press, is quite another.
The Errors Hiding in the Numbers
Here is the part that ought to reassure and unsettle in equal measure. The vast majority of these exposures were not overdoses in any dramatic sense, and they were not deliberate. They were unintentional therapeutic errors, the mundane mistakes of people getting the mechanics wrong, and most produced nothing worse than nausea, vomiting or a churning stomach. To run a proper statistical autopsy, the researchers modelled the quarterly call counts with segmented Poisson regression, a technique built to spot whether a known event bends a trend or just rides existing growth. It bent. Semaglutide exposures accelerated by an extra 9.9 per cent per quarter after the approval, while the rest of the drug class barely budged, a divergence that points squarely at the new weight-loss indication rather than background creep. Two mistakes turned up again and again: people injecting daily when the drug is meant for once a week, and people starting at the full dose instead of titrating up slowly. By raw count, “wrong dose” led the tally at over 5,000 cases.
“Can you imagine something you’re supposed to trickle up to, and you’re going full blast and seven times more often than you’re supposed to?” Miller said. Put that way, the surge stops being mysterious.
The clinical fallout was mostly mild, yet the system felt it anyway. The share of cases that ended up in, or referred to, a healthcare facility rose from 23 per cent before approval to 33.5 per cent after, even though the symptoms themselves rarely turned serious. Drug shortages during this period muddied things further, pushing patients toward compounded products of uncertain strength and forcing awkward switches between formulations, any of which can breed error. There was one death in the dataset, involving tirzepatide started weeks after liposuction, though the researchers are careful to say causality cannot be pinned down from a single report. No cases of pancreatitis surfaced at all, a small note of reassurance among more than ten thousand records.
What the study cannot do is settle every why. Poison center records are voluntary, thin on dosage and weight, and shaped by the same media frenzy that may have lowered the threshold for picking up the phone in the first place. The signal is real; its precise composition is blurrier. Still, the throughline holds. A great many of these calls traced back to confusion that a clearer conversation at the prescriber’s office or the pharmacy counter might have prevented.
That is where Han lands, and where the work points next. “We need to better educate the public because how this drug behaves in our body and its long-term safety are not yet fully understood,” he said. The drugs are not going anywhere; the user base keeps swelling. Whether the next wave of new patients learns to trickle up rather than go full blast may decide if those hotlines keep ringing, or finally start to quiet.
Frequently Asked Questions
Why did poison control calls spike if the drug itself isn’t especially dangerous?
The surge came mostly from how people were using semaglutide, not from the drug’s inherent toxicity. The two most common mistakes were injecting it daily instead of weekly and starting at the full dose rather than easing in, errors that flooded in as millions of first-time users adopted it for weight loss. Most cases produced only mild stomach symptoms, but the sheer volume of confused new users drove the numbers up sharply.
Is it true that most of these cases were serious overdoses?
No. The overwhelming majority were unintentional therapeutic errors, not deliberate or severe overdoses, and most patients experienced nothing worse than nausea or vomiting. What changed was that a larger share, about a third, still ended up in or referred to a healthcare facility, suggesting the burden fell more on the system than on individual patients’ health.
How did researchers prove the FDA approval actually caused the jump?
They used a statistical method called segmented Poisson regression, which tests whether a specific event bends a trend beyond what ordinary growth would predict. The model showed semaglutide exposures accelerating by an extra 9.9 per cent each quarter after the 2021 approval, while the rest of the drug class barely moved. That divergence pointed to the new weight-loss indication rather than background growth in prescribing.
What’s the simplest way to avoid becoming one of these statistics?
Two habits prevent most of the documented errors: inject only once a week, and start at the lowest dose before gradually increasing it. The researchers argue these mistakes are largely preventable with clearer guidance from prescribers and pharmacists at the point of dispensing. Since the long-term safety profile of these drugs is still being mapped, cautious dosing matters more than ever.