In certain corners of the Internet, the discourse over Ozempic is terrifying. 鈥淏ut the FDA says that [sic] inject Gila Monster venom weekly into yourself is healthy,鈥� one commentator sarcastically declares. 鈥淎re they lying?鈥�
鈥淪o, of course, the FDA wants to give it to six year-olds,鈥� says another. 鈥淔rom their totalitarian point of view, it鈥檚 better to blind them now than to wait until they鈥檙e old enough to cause trouble for the cabal.鈥� These comments were posted on聽聽penned for聽The Defender, the online publication of Children鈥檚 Health Defense. The latter was chaired for close to a decade by Robert F. Kennedy Jr, its most famous spokesman who is now the head of healthcare for the United States.
The bad arguments left on that post, dripping with conspiracy ideation, come from reports that the wonder drug Ozempic might be causing eye problems that lead to blindness. This example illustrates how the science sausage gets made and eaten: how rare side effects get measured in imperfect ways, leading to conflicting studies that can then be cherry-picked and weaponized online to scream at the other side.
So, what do we really know about Ozempic鈥檚 darker side? Should we be worried?
鈥淧rescribed鈥� is not the same as 鈥渢aking鈥�
In case you have been living under a rock, a brief history of Ozempic might be in order. In the early 1980s, scientists discovered that the gene coding for a hormone made in the body, glucagon, also coded for two shorter molecules, which were named glucagon-like peptide 1 and 2, or GLP-1 and GLP-2 for short. GLP-1 could increase insulin levels in the presence of sugar, which put it in the crosshairs of researchers looking for new treatments for diabetes. But GLP-1 was quickly chewed up in the blood by enzymes, and giving diabetic participants large doses of it made them throw up. What was needed was a molecule聽濒颈办别听GLP-1, something that would bind to the same receptor鈥攚hat we call a receptor agonist, like getting a slightly different key that opens the same lock鈥攂ut that would stick around longer and be better tolerated.
That molecule was found in the venom of the Gila monster, a North American lizard, and the version of it made in the lab was named exenatide (trade name of Byetta) and was approved by the Food and Drug Administration (FDA) in 2005 as an injectable drug for diabetes. This opened the door for more of these GLP-1 receptor agonists. As prescriptions for these increased, people reported also losing weight, which is how we end up with semaglutide which, under the trade name Ozempic, is used to treat diabetes, but under the trade name Wegovy is prescribed for weight loss. (For a detailed history, I recommend聽听蹿辞谤听STAT.)
The benefits of these drugs are undeniable, but when it comes to evaluating their risks, we run into challenges that researchers are familiar with but that the public at large may not be aware of. After all, when a drug like Ozempic is tested in a clinical trial, don鈥檛 they look for side effects? Yes, they do, but clinical trials do not recruit millions of participants. This means that the most common and obvious side effects are indeed detected, but rarer ones? They are much harder to pin down, and whichever approach a team of scientists chooses comes with limitations.
Take this聽聽out of Denmark. It tried to answer the question of whether or not semaglutide (Ozempic/Wegovy) was associated with a particular form of blindness, what聽The Defender聽article was focused on. This specific disease is called nonarteritic anterior ischemic optic neuropathy or NAION for short. It鈥檚 an 鈥渙ptic neuropathy鈥� because the optic nerve at the back of the eye is damaged; it鈥檚 鈥渋schemic鈥� because the area is not receiving enough blood; it鈥檚 鈥渁nterior鈥� as opposed to 鈥減osterior鈥� to denote that the damage is at the front of the optic nerve; and it鈥檚 鈥渘onarteritic鈥� because it does not involve harm to the arteries in the area. The first symptom of NAION is usually that one eye suddenly and painlessly goes blind as the optic nerve stops receiving enough blood. NAION is rare: doctors diagnose聽聽of it every year in the United States, which has a population of 340 million people. That NAION is so infrequent is a good thing for us, but a bad thing for researchers trying to figure out if semaglutide increases the risk of developing it.
The Danish researchers took everyone in Denmark with type 2 diabetes and, using the country鈥檚 patient registry, divided them into those who had been prescribed semaglutide and those who hadn鈥檛. They saw that those prescribed semaglutide had more than double the risk of having a diagnosis of NAION than the others. It was still very rare, but less rare than in people who had not been prescribed semaglutide.
I keep writing 鈥減rescribed鈥� because that is one of the important limitations of the study. Just because a doctor prescribes you Ozempic does not mean you will inject yourself regularly. In fact, we know that many people stop taking it because they experience significant nausea. Hence, we don鈥檛 know that the Danish diabetic patients who were prescribed semaglutide and who had NAION were actually injecting themselves with the medication.
The researchers also did not have access to data on smoking, body-mass index, or blood pressure, so they couldn鈥檛 make sure that the two groups they were comparing鈥攄iabetics with or without semaglutide鈥攚ere the same according to those variables. That鈥檚 a problem because we do not understand what causes NAION. There are聽辫辞迟别苍迟颈补濒听risk factors, like high blood pressure. If the Danish people prescribed semaglutide happened to have a higher blood pressure than those who were not, what caused the blindness: semaglutide or high blood pressure?
Observational studies have limitations
础苍听聽published in the journal聽贰测别听earlier this summer summarizes the muddy picture that emerges from the handful of imperfect studies looking at this question. One聽聽found a small risk increase, but some of the databases it used did not include patients over the age of 65. NAION risk increases with age, so they probably missed patients there. Meanwhile,听聽reported there was no increased risk of NAION with semaglutide鈥� but the researchers here lacked access to certain key variables and they only looked at white patients. Is this generalizable?
The big problem here is that this data is observational in nature. We can鈥檛 willingly put everyone on semaglutide or a placebo to see who goes blind in one eye and who doesn鈥檛. Instead, we have to look back at patient registries and try to find as much information as we can to interrogate this possible link, but these registries are often incomplete.
We see the same thing with the association between GLP-1 receptor agonists and macular degeneration, a much more common form of blindness in which people (usually older adults) lose sight at the centre of their vision but can still see around it. A聽聽showed a two-fold higher risk of macular degeneration for patients prescribed a GLP-1 receptor agonist鈥攁聽迟丑谤别别-蹿辞濒诲听higher risk when they used this medication long enough. But we don鈥檛 know how often they injected themselves, whether or not they were smokers, and what their sun exposure was like. (Smoking is a risk factor for macular degeneration and聽blue-light exposure is a contentious one.) This study also did not differentiate between GLP-1 receptor agonists, like semaglutide vs. tirzepatide, assuming they would all equally influence the risk of macular degeneration, whereas that may not be the case.
The story repeats itself for alleged increases to the risk of developing聽,听, and聽. An everything-but-the-kitchen-sink聽聽of GLP1- receptor agonists compared to other diabetes medications, where 175 health outcomes were assessed between the two groups, was published this year and it highlighted a number of slight risk increases with GLP-1 receptor agonists. A major limitation was that it was done exclusively among U.S. war veterans, who are older and mostly white men.
This is not me dismissing concerns out of hand to protect pharmaceutical companies; rather, this critical look at the limitations of what we know is exactly聽what scientists themselves do.聽No study is ever definitive. Researchers are encouraged to admit to the limitations of what they did, which they list in the discussion section of their papers. The goal is to give the reader an idea of the uncertainty that lingers at the end of this study, since all science can do is聽reduce鈥攏ot eliminate鈥攗ncertainty.
Has the media leaned heavily on the merits of semaglutide and its brothers while ignoring the risks? In some cases, yes. But the risks, either known or alleged at this point, need to be contextualized. Ozempic is prescribed to people with type 2 diabetes, typically after the drug metformin has been tried, and it may well increase the risk of a rare form of blindness (to be confirmed). Diabetes itself, however, comes with its own risk of blindness, namely diabetic retinopathy. This eye disease used to affect聽, but modern interventions have reduced the risk considerably.
Diabetes can lead to yet more聽: heart disease, nerve damage, stroke, and kidney disease. It鈥檚 easy to become fixated on the very small and still-debated risk of a medication while ignoring the much larger and clear risks of foregoing treatment. The former feels bigger because we choose this risk by taking a drug, while the latter is easier to downplay as it feels out of our hands, but it鈥檚 just an illusion.
The bottom line
The European Medicines Agency, having looked at all of this imperfect data,听聽last month that the information leaflet that comes with semaglutide (Ozempic, Wegovy and, in its pill form, Rybelsus) now include NAION as a side effect, qualifying its frequency as 鈥渧ery rare.鈥� The World Health Organization聽聽with this assessment; Health Canada and the U.S. FDA have yet to make statements of their own, as far as I can tell, though I suspect they are coming.
With the data we have so far, GLP-1 receptor agonists really are blockbuster drugs. They鈥檙e not only good for weight loss and managing diabetes, but evidence is accumulating that they protect from cardiovascular disease and聽聽shows that maybe they could help with addiction, Alzheimer鈥檚 disease, and a slew of other health issues鈥攃onfirmation in better studies is sorely needed. No drug, however, is side effect free. The gastrointestinal issues that can arise, especially at the beginning of treatment, are well known. Newer, rarer ones may crop up as more and more people use them and more and more studies are done looking for anything unusual.
What I don鈥檛 see is evidence of a global conspiracy in which drug regulators are pushing semaglutide on people to wipe out their sight. If anything, the EMA and WHO are being cautious by flagging a risk that is very low and is still not entirely certain in order to better inform the public.
But if you鈥檙e a conspiracy theorist, every study can be awkwardly fitted into place to prove that pharmaceutical companies and government agencies are in the business of causing harm.
Take-home message:
-Semaglutide (Ozempic/Wegovy/Rybelsus) and similar medications for diabetes and weight loss have been linked to an increased risk of a very rare form of blindness, meaning that the chance of developing this form of blindness would still be very rare but not quite as much
-The handful of studies done on this link come to different conclusions, with disagreements on how much the risk increases and some studies showing no increase in risk
-The problem is that all of these studies rely on incomplete databases of patients, and it is possible that any increase in the risk of developing this form of blindness is actually due to other variables that were not measured and compiled in these databases
