Who could have seen that coming… and what does it have in common with other drug side-effects?
15 July 2024
Just under two weeks ago, researchers from Massachusetts Eye and Ear at Harvard Medical School, Boston published a study which found an association between the use of semaglutide (marketed as Ozempic and Wegovy), and increased risk of an untreatable eye condition that can cause blindness.
The condition, nonarteritic anterior ischemic optic neuropathy (NAION), is “the second most common form of optic neuropathy and a significant cause of blindness among adults“. Its incidence rate is between two and ten cases per 100 000, so although it is the second most common cause of blindness due to optic nerve damage (after glaucoma), it’s still rare enough that there’s virtually no chance that the increased risk could have been detected in clinical trials for the drug, which typically enrol several hundred participants; one of the major meta-analyses of semaglutide for weight loss involved just 3613 individuals, from four randomised controlled trials.
This is of course true for most drugs. Because pharmaceutical companies carefully control the conditions of drug licensing trials in order to minimise adverse reactions (for example, by selecting relatively healthy participants who are on few to no other drugs), the full suite of adverse effects of any drug can only be detected once it is unleashed onto the general population.
This is called postmarketing surveillance, and is why drug regulators established pharmacovigilance mechanisms such as the FDA Adverse Event Reporting System (FAERS) and the TGA’s Database of Adverse Event Notifications (DAEN). But these systems rely upon medical practitioners and/or patients reporting potential adverse reactions to a drug. And since most doctors won’t see even a handful of cases of conditions that are very rare in their careers, they’re unlikely to suspect that a particular drug may have caused or contributed to such a disorder, and hence to report it.
Only a specialist referral service such as Mass Eye and Ear sees enough patients with rare eye disorders to notice a possible link between NAION and semaglutide, as a press release for the study explained:
“The impetus for the study occurred in the late summer of 2023 when Rizzo [principal investigator of the study, Joseph Rizzo, MD, director of the Neuro-Ophthalmology Service at Mass Eye and Ear], a resident (study co-author Seyedeh Maryam Zekavat, MD, PhD) and other Mass Eye and Ear neuro-ophthalmologists noticed a disturbing trend — three patients in their practice had been diagnosed with vision loss from this relatively uncommon optic nerve disease in just one week. The physicians observed all three were taking semaglutide.
This anecdotal recognition led the Mass Eye and Ear research team to run a backward-looking analysis of their patient population to see if they could identify a link between this disease and these drugs, which had been surging in popularity…
The researchers analyzed the records of more than 17,000 Mass Eye and Ear patients treated over the six years since Ozempic was released and divided the patients in[to] those who were diagnosed with either diabetes or overweight/obesity.
The researchers compared patients who had received prescriptions for semaglutide compared to those taking other diabetes or weight loss drugs. Then, they analyzed the rate of NAION diagnoses in the groups, which revealed the significant risk increases.”
Ozempic Linked to Increased Risk of Blinding Eye Condition
What were those significant risk increases? Compared to patients matched for age, sex, and comorbidities (hypertension, type 2 diabetes, obstructive sleep apnoea, obesity, hyperlipidemia, and coronary artery disease), patients with diabetes were 4.3 times more likely to be diagnosed with NAION if they were prescribed semaglutide than if they were prescribed non-GLP-1 agonist diabetic medications.
And patients with obesity were 7.7 times more likely to develop NAION if they were treated with semaglutide than if they took non-GLP-1 agonist weight-loss medications.
Advocates of GLP-1 receptor agonists will no doubt point out that a four-to-nearly-eight-fold increase in risk of a rare condition still results in that condition being very rare. That’s true, of course, but I would counter that the visual loss caused by NAION has no effective treatment. Is a heightened risk of incurable blindness an acceptable price to pay for losing weight and improving one’s diabetes bloodwork? And, given that 1.7 per cent of the US population was prescribed semaglutide in 2023 – that’s one out of every fifty-nine Americans – what are the economic and social implications of a sharp uptick in the risk of a condition that causes irreversible vision loss?
You won’t find blindness among the adverse reactions catalogued in my previous article, The high price of weight loss: GLP-1 agonist side-effects, because as of March 2024, when I published that article, no one except the Mass Eye and Ear investigators knew about it. And they hasten to point out that their study only establishes an association, not a causal relationship. Furthermore, as yet, there’s no known causal mechanism linking semaglutide to NAION; they speculate that “expression of the GLP-1 receptor in the human optic nerve26 and GLP-1 RA–induced enhanced sympathetic nervous system activity might influence optic nerve head perfusion and potentially increase the risk of NAION.” But, as they point out,
“As with any drug, … therapeutic benefits are inseparable from adverse effects.”
Risk of Nonarteritic Anterior Ischemic Optic Neuropathy in Patients Prescribed Semaglutide
That is, as I stressed in the subheader of Stop calling them ‘side effects’, you don’t get to pick and choose which ‘effects’ you’ll get from medicines.
I have absolutely no doubt that many more adverse effects of semaglutide and other GLP-1 receptor agonists will be identified in the years to come, and at least some drugs in this class will eventually either be pulled off the market by a drug regulator, or voluntarily withdrawn if the injury compensation payouts start to exceed the pharmaceutical companies’ profit.
But the emergence of this link between semaglutide and NAION prompted me to delve into the file in which I store articles to discuss in future posts, for examples of adverse effects linked to other widely-used drugs, that have only emerged after years (in some cases, decades) of use, by tens of thousands of people. I’m going to summarise each of these cases, and as you read them, I would like you to keep in mind the following question:
What do each of these cases, of a widely-used drug eventually being found to be linked to increased risk of a disease or adverse outcome, have in common?
OK? Here goes:
Example #1: Prolonged use of certain progestogerone-like drugs is linked to increased risk of the most common primary brain tumour in adults
Synthetic versions of the sex hormone progesterone – known as progestogens – are taken by hundred of millions of women worldwide, mostly for contraceptive purposes but also for treatment of endometriosis, heavy and painful menstrual periods, premenstrual syndrome and perimenopausal symptoms.
In a study published in the BMJ in March 2024, researchers used data from the French National Health Data System to identify 18 061 women living in France who had surgery to remove a meningioma between 1 January 2009 and 31 December 2018. Each case with surgically-excised meningioma was matched with five controls who had the same year of birth and area of residence, and the researchers compared use of a range of progestogens in cases vs controls.
Meningiomas comprise 40 per cent of primary tumours of the central nervous system; they are mostly benign (noncancerous) and slow-growing, but they may require surgical removal if they begin to compress adjacent brain tissue. The surgery to excise them carries certain hazards; in particular, seizures. In fact, 28 per cent of the cases in this study were still using antiseizure medications a full three years after their meningiomas were removed.
The authors were spurred to explore the relationship between meningioma and the use of progestogens that are commonly taken for contraception and menstrual issues, because in 2022, they had already identified an increased risk of surgically-excised meningioma associated with the use of three other progestogens (cyproterone acetate, chlormadinone acetate and nomegestrol acetate), that are rated as more potent than the ones studied in this latest paper.
They found that women who used promegestone, medrogestone, or injectable medroxyprogesterone acetate for one year or more, had a 2.7-fold, 3.5-fold and 5.6-fold increased risk of intracranial meningioma requiring surgery, respectively. The most potent progestogens – chlormadinone acetate, nomegestrol acetate and cyproterone acetate, were associated with a 3.9-fold, 4.9-fold and 19.2-fold higher risk of surgically-removed meningioma.
Of the three less-potent progestogens found to be linked to heightened risk of meningioma, injectable medroxyprogesterone is the most widely-used (and in fact, promegestone was withdrawn from the market in 2020, and had only ever been licensed in France). As a long-acting contraceptive whose effectiveness in preventing pregnancy is not reliant on women remembering (or even wanting) to take a daily dose, it is extensively used in women who are, to put it delicately, not at the top of the socioeconomic ladder:
“Worldwide, in 2019, 3.9% of women of childbearing age were using injectable contraception (medroxyprogesterone), that is, 74 million users, but figures vary widely between world regions (from 1.8% in high income countries to 8.7% in low income countries).41 This method of contraception is the most widely used in Indonesia (13 million women),42 Ethiopia (4.6 million women), and South Africa (3.6 million women).41 In the USA, medroxyprogesterone acetate is used in more than 2 million prescriptions in 2020 and more than one of five sexually active American women report having used injected medroxyprogesterone acetate (150 mg/3 mL) in their lifetime.4344 Injectable contraceptives are much less widely used in Europe (3.1% of women of childbearing age in the UK and 0.2% in France41)… In addition, medroxyprogesterone acetate used as an injected contraceptive is known to be prescribed to specific populations, especially people with mental illnesses.50“
Use of progestogens and the risk of intracranial meningioma: national case-control study
As I discussed in Hormonal contraception and women’s mental health, the risk of depression was found to be 90 per cent higher in women using a medroxyprogesterone acetate injection than in those using no form of hormonal contraception, but hey, who even cares if brown, black, poor, or ‘mentally ill’ women get depressed, anyway? (And yes, for those who have trouble detecting it in a written text, that was sarcasm.)
The authors noted with concern that
“Results showed that progestin related meningiomas tend to occur more frequently at the skull base and that surgery for lesions in this location is much more challenging.”
Use of progestogens and the risk of intracranial meningioma: national case-control study
What do think the chances of successful surgery for meningioma removal are going to be, if you’re a poor woman in Indonesia, South Africa or Ethiopia? (Or for that matter, a poor woman in the US?) And how about your chances of obtaining antiseizure medication at an affordable price? Yeah, good luck with that.
In summary, several drugs that are widely used as long-acting contraceptives (especially in poor women), and for menstrual disorders and menopausal symptoms, increase the risk of a type of brain tumour that may necessitate surgical removal, with a high risk of residual postoperative seizures.
Example #2: Diabetics who use statins have greater deterioration in their diabetic condition than those who don’t take statins
Back in 2017, I wrote The futility of risk factor management: Cholesterol-lowering drugs cause diabetes, which discussed a meta-analysis of observational studies that found that cholesterol-lowering statin drugs raise the risk of developing type 2 diabetes by 44 per cent. As I wrote in that article,
“Having diabetes more than doubles the risk of having a heart attack overall, and more than quadruples it in women. For Asians and South Asians who become diabetic, the increased risk of having a heart attack is even more dramatic than for people of European descent.”
The futility of risk factor management: Cholesterol-lowering drugs cause diabetes
Diabetics are prescribed statins in order to reduce their risk of cardiovascular complications, such as heart attacks. But a retrospective cohort study of patients covered by the US Department of Veterans Affairs, which involved 83 022 matched pairs of statin users and nonusers who had all been diagnosed with diabetes, found that patients who used statins had nearly 40 per cent higher odds of greater progression of their diabetes than patients who did not. That is, statin users were more likely to be on multiple glucose-lowering drugs, to be prescribed insulin (which is pretty much the end of the road if you’re a type 2 diabetic), to have significant hyperglycaemia (high blood sugar), and to suffer acute glycaemic complications such as ketoacidosis.
Furthermore, there was a dose-response relationship between statin use and diabetes progression:
“The odds of diabetes progression among statin users vs nonusers were 1.83, 1.55, and 1.45 for high-, moderate-, and low-intensity cholesterol lowering, respectively.”
Association of Statin Therapy Initiation With Diabetes Progression: A Retrospective Matched-Cohort Study
The mechanism of action by which statins cause and exacerbate diabetes is well-known: insulin resistance. And insulin resistance, in turn, drives several of the pathological mechanisms underlying cardiovascular disease – you know, the cardiovascular disease that statins are supposed to help prevent:
“Several RCTs,3,4,43 Mendelian randomization studies,44 observational studies,5 and animal studies45,46 have demonstrated that statin use increases insulin resistance. The association of statin use with diabetes progression may be explained by its effect on insulin resistance (eDiscussion in the Supplement). Insulin resistance has been shown to increase the risk of diabetic complications,10,47 endothelial dysfunction, inflammation, and increased platelet reactivity.48–52“
Association of Statin Therapy Initiation With Diabetes Progression: A Retrospective Matched-Cohort Study
The number needed to harm was only 13; that is, for every 13 diabetic patients who took a statin, one experienced worsening diabetic control, with all its attendant harms.
In summary, statins, which are prescribed to diabetics in order to protect them against the cardiovascular disease that their diabetic condition makes them more susceptible to, cause insulin resistance which not only exacerbates diabetes, but also contributes to multiple pathogenic pathways of cardiovascular disease – by far the leading cause of death in diabetics.
Example #3: Long-term use of drugs for attention-deficit/hyperactivity disorder (ADHD) is linked to increased risk of cardiovascular disease
Using nationwide health registers in Sweden, researchers identified 278 027 individuals aged between 6 and 64 who had been diagnosed with attention-deficit/hyperactivity disorder (ADHD) and/or received a prescription for an ADHD medication for the first time, between 1 January 2007 and 31 December 2020. Average follow-up time was 14 years. Out of this cohort, they found 10 394 individuals who had been diagnosed with cardiovascular disease of any type, during this period (cases). They then matched each case with five controls from the same cohort.
They found that a longer cumulative duration of ADHD medication was associated with an increased risk of cardiovascular disease (CVD) compared with nonuse, in both young and older people. (CVD is the leading cause of death in Sweden.) The risk of developing CVD increased rapidly for the first three continuous years on ADHD medication, and then stabilised. Taken as a whole, each one-year increase in the use of ADHD medication was associated with a four per cent increased risk of CVD.
Furthermore, higher dosage regimes were associated with a higher risk of CVD. This dose-response relationship is a powerful indicator of the biological plausibility of the link between ADHD medication and CVD risk.
Among specific types of CVD, hypertension (high blood pressure) and arterial disease had the strongest association with long-term use of ADHD medication. For example, compared with nonusers of ADHD medications, those who took them for three to five years had 72 per higher risk of becoming hypertensive, and 65 per higher risk of being diagnosed with arterial disease. The association between CVD and cumulative use of ADHD drugs was strongest for the stimulant medications such as methylphenidate, amphetamine, dexamphetamine and lisdexamfetamine, while the nonstimulant atomoxetine was only associated with increased risk in the first year of use.
In summary, prolonged use of medications prescribed for ADHD, particularly stimulant drugs, increases the risk of cardiovascular disease, particularly hypertension and arterial disease, in people diagnosed with ADHD.
Example #4: Antibiotic use before the age of two is associated with increased risk of chronic health and behavioural conditions, ranging from allergies to obesity to ADHD
In a population-based cohort study, researchers compared the number of antibiotic prescriptions received by children born in Olmsted County, Minnesota, before the children turned two, to the childhood-onset medical diagnoses they received. They found that:
“Early antibiotic exposure was associated with an increased risk of childhood-onset asthma, allergic rhinitis, atopic dermatitis, celiac disease, overweight, obesity, and attention deficit hyperactivity disorder.”
Association of Infant Antibiotic Exposure With Childhood Health Outcomes
Furthermore, the more antibiotic prescriptions the children received, the higher the likelihood of multiple diagnoses.
Previously-published studies have found associations between early-life antibiotic use and individual conditions, as I’ve discussed in Acid suppressors and antibiotics increase allergic disease, Of bugs and brains – how your gut microbiome affects mental health, and Early-life antibiotic use makes kids fatter. However, this research project was far more ambitious. It utilised the Rochester Epidemiology Project medical records-linkage system, which links and archives the medical and prescription records of nearly all persons residing in Olmsted County, Minnesota (the location of the world-famous Mayo Clinic), to identify all diagnostic codes assigned to children born between 1 January 2003 and 31 December 2011, during any health care visit after they turned two years of age, and all outpatient prescriptions for antibiotics.
Astonishingly, 70 per cent of the 14 572 children were prescribed at least one antibiotic between birth and their second birthday, with most receiving multiple antibiotics. The majority of these prescriptions were written for conditions for which there is little to no evidence that antibiotics are beneficial, or for which they are not recommended, such as viral infections, fever of unknown origin, non-infectious gastroenteritis and nonsuppurative otitis media.
After adjusting for confounders such as birth weight, birth method (caesarean vs vaginal birth), maternal smoking and antibiotic use during pregnancy, they found that:
“Both girls and boys prescribed at least 1 course of antibiotics had a higher cumulative incidence of asthma, allergic rhinitis, overweight, and ADHD (Figure 1). Girls exposed to antibiotics were uniquely more susceptible to atopic dermatitis and celiac disease, whereas boys were uniquely more susceptible to obesity.”
Association of Infant Antibiotic Exposure With Childhood Health Outcomes
The relationship between the number of courses of antibiotics prescribed, the type of antibiotic, the timing of antibiotic exposure, and the risk of particular diagnoses, was somewhat complicated:
“Among children who received 1 or 2 prescriptions, only girls were at a significantly higher risk of developing asthma and celiac disease than those unexposed (Table 2). By contrast, receiving 3 to 4 prescriptions was associated with a higher incidence of asthma, atopic dermatitis, and overweight in both sexes; ADHD and celiac disease in girls; and obesity in boys. Both girls and boys who received 5 or more prescriptions had a significantly higher risk of developing asthma, allergic rhinitis, overweight, obesity, and ADHD, and girls also were at a higher risk of celiac disease…
Exposure to cephalosporins was associated with an increased risk of the highest number of conditions and, uniquely, autism and food allergies (Table 3). Penicillins were associated with an increased risk of asthma and overweight in both sexes, celiac disease and ADHD in girls, and obesity in boys, whereas they were associated with a reduced risk of autism in girls. Sulfonamides only increased the risk of overweight in boys. Macrolides were associated with an increased risk of asthma and overweight in both sexes and allergic rhinitis and obesity in boys, but were associated with a reduced risk of atopic dermatitis in girls and learning disability in boys…
Antibiotics prescribed before 6 months of age were significantly associated with the risk of atopic dermatitis, overweight, and ADHD in both sexes; asthma, allergic rhinitis, and obesity in girls; and food allergy in boys (Supplemental Table 5, available online at http://www.mayoclinicproceedings.org). Antibiotics prescribed between 6 and 12 months were most strongly associated with the incidence of asthma, allergic rhinitis, overweight, and obesity in both sexes. Antibiotics prescribed from 12 to 24 months were associated with a significantly increased risk of asthma and overweight in both sexes, ADHD in girls, and allergic rhinitis in boys.”
Association of Infant Antibiotic Exposure With Childhood Health Outcomes
While the causal mechanisms linking antibiotic use with physical and mental health conditions are not fully understood, the authors infer that they are related to perturbations to microbiome composition during key developmental periods.
In summary, taking an antibiotic in infancy and early childhood increases the risk of developing immunological, metabolic, and neurobehavioral health conditions, with multiple prescriptions associated with higher risk.
OK, have you been reflecting on what the four case examples that I’ve just described, and the newly-identified association between semaglutide and NAION, have in common? I want to read your responses to this question before I share my thoughts on it, so please leave a comment below, and in the next post, I’ll tell you what I believe is the bedrock philosophical assumption underlying these five instances of delayed recognition of harms associated with widely-used medications.
Leave A Response