One of the most flawed studies of the year gained major media attention. Yet there are great teaching points for those interested in critical appraisal
"The success of Western society has resulted in widespread obesity—and obesity-related diseases, such as diabetes, high blood pressure, heart failure and vascular disease. Obesity is clearly not a good thing for human health. The statistics on obesity—especially in children—are staggering."
Not sure that "success" is the right word there... The destruction of our food supply is a total failure for everyone except BigAg cashflow, and the congressmen who are lobbied by them
So far as I can tell, just as, per Ed Livingston, they don't control for the severity of T2D, they don't control for the severity of obesity. The propensity matching doesn't match for BMI.
This Is a fantastic analysis and much appreciated.
I am actually a neuro ophthalmologist and I also practice at a large academic medical center. I see many, many patients with NAION. I know the authors of this study fairly well, and I would like to be charitable: They do at least acknowledge that they are not claiming a causal relationship So this is likely done in good faith. But this study is a mess.
It is important to understand that NAION is a purely clinical diagnosis. We have no specific biomarker. The confidence in the diagnosis is dependent upon clinical factors including age, subacute painless onset, presence of disc swelling presentation (which is required ), resolution of disc swelling spontaneously in 6-8 weeks, and no or partial improvement over time. Although diabetes and hypertension are risk factors, they are often not present in many cases. The primary risk factor is structural crowding of the optic discs which is an anatomic feature and non-modifiable. The exact mechanism is not understood and still much debated but the theory is that structural crowding of axons of the optic nerve head in small optic discs result in relative hypoperfusion, and then a variety of factors (age, microvascular disease, impaired local and systemic auto regulation, possibly yet to be discovered genetic factors) converge to cause reduced perfusion to a vulnerable optic nerve head just long enough to cause ischemia. It is difficult to determine cup disc ratio in a swollen optic nerve so we use the other eye has a reference which is generally reliable but not perfect. This is presumably the reason that this disease primarily occurs and Caucasian, and is much less common in African Americans, since the size of the optic disc various significantly between these groups.
So even though the charts were reviewed, I still think there would be some question about the confidence in the diagnosis. You would really want to prospectively collect and follow patients over time to be certain they truly had NAION. They also did not even speculate about mechanism. Based on what we know about the semaglutides, and our current understanding of NAION, it is hard to see how there would be a true causal relationship.
I already thought that this was a highly flawed study but it is nice to have a definitive analysis since I am already receiving questions from my patients about it (also from my colleagues in Ophthalmology and Neurology!). The sample size alone is a huge issue, and it is nice to learn more about collider bias. I think I understood it intuitively but I have not seen that particular term before. I think studies like this are not necessarily dangerous but they can really cloud the waters since the statistical jargon is dense and most doctors may take it for granted that they are legitimate and justify the conclusions.
But I am disappointed that your title wasn't "...but EYE doubt it!".
At our local medical school we have an endocrinologist/bariatrician who has treated over 1000 patients with Ozempic and has yet to see even one case of ischemic optic neuropathy. And your patient going blind isn't something you are likely to miss. So something is wrong with that study, I think.
Well done thoughtful post. To your last point about the neurosurgeon only seeing brain bleed complications of anticoagulation though I would suggest that such a myopic view is bidirectional. The absolute risk reduction of stroke with anticoagulation is often overstated and the decision making is nuanced, especially in the frail elderly who are not included in the clinical trials.
Reflecting back on the many studies on anticoagulants for stroke reduction, they generally show a consistent pattern---very small differences in strokes between those on anticoagulants vs. placebo; these practically insignificant differences are then subjected to the magic of statistics and said to be "statistically significant." Most of the time, the small differences are completely countered by the complications of bleeding.
I lost an 80 year old friend to a brain bleed that was thought to be a result of Eliquis at what must have been too high a dose. I have other friends worrying about that med because of bleeding. One just opted for a Watchman because he was afraid of a brain bleed. (I showed him Dr Mandrola's article on it but he went ahead anyway. I guess no one likes a buttinsky.)
Correct me if I'm wrong, but are studies like these not at least partially the result of the endless pressure on medical professionals at all levels at all times to build better resumes by publishing publishing publishing? And so wouldn't competition for ink space drive competitive people into areas with the most buzz and fresh funding, seeking the most sensational take-away possible? Seems that the incentives are driving less and less medical researchers towards scientific truth building and ever more toward personal brand building. Scientific endeavor in all disciplines seems trapped in this, no?
Is it immortal time bias? Both groups have eye problems, and the exposed group got the GLP. But the eye problem started, likely, before their exposure. Isn't it more like a misclassification? If GLPs are not causal, why would we see a signal?
"The success of Western society has resulted in widespread obesity—and obesity-related diseases, such as diabetes, high blood pressure, heart failure and vascular disease. Obesity is clearly not a good thing for human health. The statistics on obesity—especially in children—are staggering."
Not sure that "success" is the right word there... The destruction of our food supply is a total failure for everyone except BigAg cashflow, and the congressmen who are lobbied by them
"There are ways to study drug safety. But you need large databases and careful controls. "
But you don`t/Sarcasm....its all about saving $ and Time to get shite products to market regardless of of efficiency or safety...
The FDA should just classify Ozempic as a vaccine, and get it on the childhood schedule. Then they wouldn't be required to study safety...
So far as I can tell, just as, per Ed Livingston, they don't control for the severity of T2D, they don't control for the severity of obesity. The propensity matching doesn't match for BMI.
Do you know what causes eye problems? Diabetes.
This Is a fantastic analysis and much appreciated.
I am actually a neuro ophthalmologist and I also practice at a large academic medical center. I see many, many patients with NAION. I know the authors of this study fairly well, and I would like to be charitable: They do at least acknowledge that they are not claiming a causal relationship So this is likely done in good faith. But this study is a mess.
It is important to understand that NAION is a purely clinical diagnosis. We have no specific biomarker. The confidence in the diagnosis is dependent upon clinical factors including age, subacute painless onset, presence of disc swelling presentation (which is required ), resolution of disc swelling spontaneously in 6-8 weeks, and no or partial improvement over time. Although diabetes and hypertension are risk factors, they are often not present in many cases. The primary risk factor is structural crowding of the optic discs which is an anatomic feature and non-modifiable. The exact mechanism is not understood and still much debated but the theory is that structural crowding of axons of the optic nerve head in small optic discs result in relative hypoperfusion, and then a variety of factors (age, microvascular disease, impaired local and systemic auto regulation, possibly yet to be discovered genetic factors) converge to cause reduced perfusion to a vulnerable optic nerve head just long enough to cause ischemia. It is difficult to determine cup disc ratio in a swollen optic nerve so we use the other eye has a reference which is generally reliable but not perfect. This is presumably the reason that this disease primarily occurs and Caucasian, and is much less common in African Americans, since the size of the optic disc various significantly between these groups.
So even though the charts were reviewed, I still think there would be some question about the confidence in the diagnosis. You would really want to prospectively collect and follow patients over time to be certain they truly had NAION. They also did not even speculate about mechanism. Based on what we know about the semaglutides, and our current understanding of NAION, it is hard to see how there would be a true causal relationship.
I already thought that this was a highly flawed study but it is nice to have a definitive analysis since I am already receiving questions from my patients about it (also from my colleagues in Ophthalmology and Neurology!). The sample size alone is a huge issue, and it is nice to learn more about collider bias. I think I understood it intuitively but I have not seen that particular term before. I think studies like this are not necessarily dangerous but they can really cloud the waters since the statistical jargon is dense and most doctors may take it for granted that they are legitimate and justify the conclusions.
But I am disappointed that your title wasn't "...but EYE doubt it!".
Has anyone noted an uptick in appendicitis since this stuff became popular? There must be some interesting data there.
At our local medical school we have an endocrinologist/bariatrician who has treated over 1000 patients with Ozempic and has yet to see even one case of ischemic optic neuropathy. And your patient going blind isn't something you are likely to miss. So something is wrong with that study, I think.
Amazing analysis! Thank you
Well done thoughtful post. To your last point about the neurosurgeon only seeing brain bleed complications of anticoagulation though I would suggest that such a myopic view is bidirectional. The absolute risk reduction of stroke with anticoagulation is often overstated and the decision making is nuanced, especially in the frail elderly who are not included in the clinical trials.
Reflecting back on the many studies on anticoagulants for stroke reduction, they generally show a consistent pattern---very small differences in strokes between those on anticoagulants vs. placebo; these practically insignificant differences are then subjected to the magic of statistics and said to be "statistically significant." Most of the time, the small differences are completely countered by the complications of bleeding.
I lost an 80 year old friend to a brain bleed that was thought to be a result of Eliquis at what must have been too high a dose. I have other friends worrying about that med because of bleeding. One just opted for a Watchman because he was afraid of a brain bleed. (I showed him Dr Mandrola's article on it but he went ahead anyway. I guess no one likes a buttinsky.)
Industry fluffer
Correct me if I'm wrong, but are studies like these not at least partially the result of the endless pressure on medical professionals at all levels at all times to build better resumes by publishing publishing publishing? And so wouldn't competition for ink space drive competitive people into areas with the most buzz and fresh funding, seeking the most sensational take-away possible? Seems that the incentives are driving less and less medical researchers towards scientific truth building and ever more toward personal brand building. Scientific endeavor in all disciplines seems trapped in this, no?
Is it immortal time bias? Both groups have eye problems, and the exposed group got the GLP. But the eye problem started, likely, before their exposure. Isn't it more like a misclassification? If GLPs are not causal, why would we see a signal?