April 7, 2022

expert reaction to study looking at risk of vision impairment and loss, and use of phosphodiesterase type 5 inhibitors

A study published in JAMA Ophthalmology looks at the risk of ocular adverse events associated with the use of phosphodiesterase 5 inhibitors, used to treat erectile dysfunction, in men in the US.

An MHRA spokesperson said:

“Sildenafil, tadalafil, vardenafil and avanafil are important medicines for the management of erectile dysfunction, and use of these products is known to be associated with a risk of visual problems.

“The study published in JAMA Ophthalmology confirms the known risk of certain visual problems which may rarely occur in regular users of these medicines.

“The product information for these medicines advises that if any sudden visual defect occurs, patients should stop taking the product and consult a doctor immediately.

“Patients can continue to use these products, but should discuss any concerns about their treatment with their doctor or pharmacist.”

Background

• Sildenafil, tadalafil, vardenafil and avanafil belong to a class of medicines called phosphodiesterase type 5 (PDE5) inhibitors, which help to relax blood vessels. They may also affect other types of phosphodiesterase in different parts of the body, especially the eye, and for this reason they are known to be associated with visual side-effects.
• The product information (for both healthcare professionals and patients) advises that in the event of any sudden visual defect, patients should stop taking the product and consult a physician immediately.
• These products are contraindicated (i.e. must not be used) in patients who have loss of vision in one eye because of non-arteritic anterior ischaemic optic neuropathy (NAION), where there is loss of blood flow to the optic nerve; sildenafil is also contraindicated in patients with known hereditary degenerative retinal disorders such as retinitis pigmentosa.
• The study by Etminan et al (2022) used data over a 14-year period from the PharMetrics Plus database (IQVIA) in Canada to look at the risk of serous retinal detachment (SRD), retinal vascular occlusion (RVO), and ischemic optic neuropathy (ION) in 213, 033 regular users of PDE5I.
• The authors found that the incidence of SRD was 3.8 cases per 10,000 person-years, for RVO, the incidence was 8.5 cases per 10,000 person-years and for ION the incidence was 3.2 cases per 10,000 person years. Although SRD is not a known side effect with any of these products, these findings are otherwise in agreement with the known frequency of ‘rare’ listed in product information (affecting fewer than 1 in 1000 but more than 1 in 10,000 patients) for RVO and ION for sildenafil and tadalafil, which are the two most widely-used PDE5 inhibitors in the UK.
• The risks of ION and SRD were significantly increased with regular use (compared to non-use), while the risk of RVO was increased (but not significantly so) and there was a trend to increased risk with increased number of prescriptions. The authors were not able to look at the risk for individual products.
• There is no reason for patients to be concerned by the findings of this study, which confirm what is already known about the risk of certain visual problems with these medicines.
• Sildenafil has been available without prescription since March 2018, but pharmacists must complete a checklist with the patient which includes a check that they have never had loss of vision because of damage to the optic nerve (such as non-arteritic anterior ischaemic optic neuropathy [NAION]) and do not have an inherited eye disease (such as retinitis pigmentosa).
• Regulators will consider whether the findings of this study may necessitate an update to product information, and will promptly communicate any new advice to patients and healthcare professionals.

Prof Ijeoma Uchegbu, Chair in Pharmaceutical Nanoscience, UCL, said:

“The identification of a slightly higher risk of side effects related to eyesight when one uses these drugs to treat erectile dysfunction is important, especially as failing eyesight is naturally associated with getting older.  The study used over 200,000 men (a very large cohort of patients) to study these effects and this is the most powerful part of the study as the causal relationship between the use of these erectile dysfunction drugs and impaired vision would not be easy to study with a smaller group of patients.  The authors state that men using these types of medicines to control their erectile dysfunction should err on the side of caution and get their sight tested regularly.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“I’m glad that the researchers don’t go further than saying that men who use these drugs should be told that there may be an association between using them and some serious eye conditions, and that they should tell their doctors if something seems to be going wrong with their vision.  (I’d say it would be sensible advice for any person, taking these drugs or not, to tell their doctors if something seems to be wrong with their vision, but then I’m in the UK where we don’t normally have to worry about big medical bills.)  But I’d hope that nobody is told by their doctors that taking these drugs for erectile dysfunction will definitely cause an increase in the risk of these eye conditions.  This study just can’t tell us whether or not that is true.

“The issue is the usual one with observational studies.  In this study, some men had regular prescriptions for the drugs in question, and some did not.  They were followed up for a time, about four years on average, and new cases of three serious eye conditions were recorded.  The researchers found that the conditions were more likely in men who had had regular prescription for a PDE5I drug, than in men who had no prescription for them at all.  The trouble is that there would be other differences between the men who did have regular prescriptions, and men who were not prescribed the drugs, apart from the prescriptions.  These other differences – so-called potential confounders – may have been the actual cause for the difference in the risk of the eye conditions, rather than the drugs.

“One way that this can happen – though there are many others – is that the men, who take the drugs regularly, presumably do so for some reason, probably because they have problems with their erections.  Those problems don’t just come from nowhere – in some cases at least, they could be caused by other conditions or diseases.  Those other conditions would probably be more common in the men who took the drugs regularly than in the other men.  If in fact some of those conditions also increase the risk of the serious eye conditions that were considered, then what’s causing the increased risk in the men regularly using the drugs could be these underlying conditions, and not the drugs at all.  I’m not saying that that’s definitely the explanation in this study – but it could be.

“It’s possible for researchers to try to take into account potential confounders, on which they have data, by making statistical adjustments.  These researchers did that for four confounders (hypertension, smoking, diabetes and coronary artery disease) for all the eye conditions, and in addition another (sleep apnea) for just one of the eye conditions.  This makes it less likely that differences in these factors are the cause of the differences in the risk of the eye conditions.  But it’s never possible to be sure that one has adjusted for everything relevant, and you can’t adjust for potential confounders on which you don’t have data.  So the question of cause and effect does still remain open.

“One particular issue here is, I’d say, that the researchers did not adjust for the men’s ages.  There wasn’t a difference between the men who got one of the eye conditions (the ‘cases’) and those who didn’t (the ‘controls’) in terms of age, because the controls were chosen to match the cases in age.  But that does not take away the possibility that age is a confounder.  A BMJ paper¹ on this point says, “However, matching in a case-control study does not control for confounding by the matching factors;” and explains that the matching can even introduce confounding that wasn’t there in the population being studied.  And there could be other confounders too.

“Another point about the study is that, though the total cohort of men that they considered looks impressively large – they had data on over 200,000 men who used the drugs, and many more who didn’t – what really determines the precision of the estimates is the number of men who had one of the serious eye conditions, and that was considerably smaller.  (Size does matter for medical studies, but not always in the most obvious way.)  In men like those that were studied, these eye conditions are not common.  In 10,000 men like those studied, the researchers report that about 15 or 16 would get one of the eye conditions in a year.  So, though the researchers estimated that the risk of having one of the conditions was somewhere between 40% and 140% higher in men prescribed one of the drugs regularly than in men who were not prescribed them, that’s a relative measure of the risk, and in absolute terms the increase in risk in men who take one of the drugs regularly is not large.

“The reason why there’s a lot of statistical uncertainty in the increased risk is, again, because the number that really counts here is the number of diagnoses of the eye conditions, not the number of men studied in total, and relatively few were diagnosed with the eye conditions.  And I should point out again that the study can’t tie down cause and effect. It might or might not be the drugs that are the cause of the difference in risk.”

1 Neil Pearce, ‘Analysis of matched case-control studies’, BMJ 2016, https://doi.org/10.1136/bmj.i969

Prof Linda Sharples, Professor of Medical Statistics, London School of Hygiene & Tropical Medicine, said:

“This is an interesting use of routinely collected data to assess associations between PDEI5 treatments and three measures of vision impairment.  The cohort study tells us that these vision impairments are fairly rare, occurring in 1.55 patients per thousand in this population.

“The analysis used good selection of the two comparison groups and matched cases and controls for age and time of follow up.  Three diagnosed comorbidities were adjusted for, hypertension, diabetes and coronary artery disease.  The resulting IRR (incidence rate ratio, which means the probability of having a visual impairment if treated divided by the probability of being having an impairment if not treated) was between 1.44 for RVO (retinal vascular occlusion), the most common of the three visual impairments to 2.58 for SRD (serous retinal detachment).  The IRR for the composite of these three was 1.85.  This equates to an additional 1.3 patients per thousand people who are treated with these drugs being diagnosed with visual impairment.

“Because there is no clear biological mechanism for the relationships, we cannot be sure that all confounding has been adjusted for, so that a causal relationship cannot be claimed.  It is difficult to completely adjust for confounding in all observational studies and it is likely to be in play here.  However, the dose response relationship and the consistency of the IRRs means that further study might be worthwhile.”

Prof Robert MacLaren, Professor of Ophthalmology, University of Oxford, said:

“Phosphodiesterase inhibitors (e.g. Viagra) are an effective treatment for men with erectile dysfunction.  The research in this paper implies that there might be a slightly increased risk of blood vessel problems in the eye in regular users of this medication.  However, it must be remembered that the conditions that lead to erectile dysfunction, such as obesity, diabetes, and high blood pressure also affect the ocular blood vessels.  Hence this may simply be an association of two conditions, rather than one causing the other.  Only a randomised placebo-controlled clinical trial can determine if phosphodiesterase inhibitor drugs cause ocular problems.  This would be difficult, because most men prescribed the placebo would know about it and might find an excuse to discretely visit their local pharmacy.  Overall, the relative risks appear to be very low and there is no cause for concern.”

Dr Dipender Gill, NIHR Clinical Lecturer in Clinical Pharmacology and Therapeutics, St George’s, University of London, said:

“The is an observational study in American men that identifies a positive association between use of phosphodiesterase 5 inhibitors and risk of adverse events affecting the eye, namely serous retinal detachment, retinal vascular occlusion and ischemic optic neuropathy.  While the authors are correct in their assertion that the findings warrant further pharmacovigilance regarding such possible harmful effects of phosphodiesterase 5 inhibitors, we can’t say from this study alone that the drug class caused these adverse events.  Specifically, in such studies it is difficult to unravel identified associations from confounding factors.  As always, medications should be taken in consultation with a physician, and only when their benefits outweigh their potential risks.”

‘Risk of Ocular Adverse Events Associated With Use of Phosphodiesterase 5 Inhibitors in Men in the US’ by Mahyar Etminan et al. was published in JAMA Ophthalmology at 16:00 UK time on Thursday 7 April 2022.

DOI: 10.1001/jamaophthalmol.2022.0663

Declared interests

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  My quote above is in my capacity as an independent professional statistician.”

Prof Robert MacLaren: “No conflicts of interest.”

Dr Dipender Gill: “I am a clinical pharmacologist based at Imperial College London. I am employed part-time by Novo Nordisk for work unrelated to the current study.”

No others received.