The Myopia Exchange
The Myopia Exchange
IMI definitions report paper with Prof Ian Flitcroft
Kate Gifford chats to Professor Ian Flitcroft about the International Myopia Institute (IMI) paper on Defining and Classifying Myopia, of which he was the lead author. Topics include why we need to define myopia for clinical, epidemiological and research reasons; grading systems for structural complications of myopia; why the refractive definition doesn't specifically require cycloplegia and the definition of pre-myopia.
Hello everyone. I'm here with Ian Flitcroft and this in the first in a series of what I hope will be engaging and important podcasts talking about the International Myopia Institute publications. They were published very recently, on the 28th of February(2019), and Ian was the lead author and the committee leader of the Definitions Report. Ian, I wanted to have a chat to you today about the Definitions Report- especially for our clinical audience- to get a concept of what you think the key take home messages will be. We're here tonight in Brisbane; there's a little bit of background noise for set up for a dinner, and Ian's going to be talking all about this to a bunch of Brisbanites shortly. This is something that is important for us to understand and to take on board right across the world- thanks very much Ian for being involved in this chat. I wanted to ask you first of all- why do you think we need to define myopia? What do you think are the important aspects that we need to understand in defining myopia?
Ian:There are two main reasons. One of which is to eliminate a lot of historical baggage with so many definitions and old theories which were wound into the terms. Part of that was cleaning things down to a simple set of terms,[which is] fundamental for the evidence base going forward, for meta-analysis and comparison. Other areas have standardized classifications for definition- for example diabetes and clinically significant macular edema. That definition has transcended through trial after trial[and been robust]. When you look at people's definitions of the higher end[of myopia], there's huge variations, and some of the analysis was looking at that variation. There's no absolute right or wrong answer around that, but we can coalesce around some consensus which allows for comparisons. People often get more bothered about the high[myopia definition] one, but statistically it's the lower[myopia definition] one that's more of a concern. Everyone says it's-0.50D. And there's a lot of reference in[the paper] to selection bias, which may be slightly obscure, but the problem is because of the way a distribution works- if you take all myopes at-0.50D, there's an awful lot in that group, some of those be sort of just slightly(myopic) emmetropes biologically. This is a question of a selection bias and trying to promote the concept of sensitivity analysis- whatever threshold you are using, if it is a different one to that, then throw in an analysis which compares different thresholds, so then the results will be there for meta-analysis. It will also allow us to look for biases in a sample. Particularly in epidemiological studies, if you're including-0.50D rather than-1.00D, you're looking at slightly different results, and that's an important distinction to standardize.
Kate:That's obviously important from a frequency and an epidemiological point of view. What about the question of high myopia? I want to digress just a bit. You're quite well known for making a very strong case that there's no physiological level of myopia, but the terminology of pathological myopia is something that was explored in the paper as well. Your paper published in 2012[demonstrated that] even with low levels of myopia there's increasing risk of pathology compared to the emmetrope, so therefore it's not fair to call less than 5D physiological and more than 5D pathological. How did you and your committee tackle that question of pathological/ high myopia and that delineation?
Ian:We got rid of'physiological' so that's definitely not included.
Kate:Okay, it doesn't exist guys. There's no such thing as'physiological' myopia!
Ian:And I lost the battle about'physiologic' with a whole bunch of retinal specialists who wanted a term. But the most important thing about that was the separation of'pathologic' as an umbrella term for the structural complications[of myopia]. If you look within that it doesn't actually refer to a refractive threshold- there are exceptions to that- or axial length. And'high myopia' is a purely refractive[definition]- a problem in the past is that people have used'pathologic' just to mean-6D without any problems. Someone who is-8D at[age] 25 may look pretty good[from a retinal health perspective]. Age 35, a bit thin... at 45 serious trouble, and at age 55 or 65 that eye has changed.[Defining on] the refraction or an axial length without understanding that biological process doesn't make any sense. So'pathological' myopia, now'pathologic myopia'- to have a single spelling- was the term the retinal people were quite keen to keep as an umbrella term. But it is clearly defined as a sort of container for a range of conditions; all of these conditions that have become very common.
Kate:In talking about the question of pathologic myopia, your paper deals with grading scales and in particular myopic macular degeneration or myopic maculopathy. We've got a little bit of ice cube noise in the background!(Caterers setting up!) So hopefully you guys can still hear us.
Ian:Just to prove we're in Brisbane.
Kate:Yes! We need ice cubes here! So do we have an effective grading scale for myopic pathology? Do we have it for particular conditions? What should clinicians be aware of in terms of understanding these structural complications of myopia?
Ian:In terms of the ideal outcome, the terminology was a bit of a compromise because of previous WHO[World Health Organization] publications. The META-PM(Meta-Analysis for Pathologic Myopia) Study Group is a large consensus group which defines the standards now. Within that, the condition[of myopic maculopathy] is quite close to an international standard now. Photographic grading systems are just that- we have the LOCS grading system for cataract; it doesn't include visual acuity, which you'd consider an important clinical measure. The whole point of photographic standards is they allow for interventions to be standardized and compared. It wasn't designed to be a comprehensive clinical tool.[In addition,} there are a whole lot of other imaging techniques which are becoming essential- high definition OCT is providing a huge amount of information on choroidal structure, and also imaging the RPE layers- we have far more effective clinical tools. That's for clinical management,[while the grading scales] provide a useful set of definitions. There are three thresholds,[the first] which is from a WHO point of view. One definition, which is super simple, was designed for a RAAB study- the Rapid Assessment of Avoidable Blindness-[data collected] with a retinoscope and a direct ophthalmoscope, a long way from a plug[power point]. This is a limited functional assessment which is[basic], but[important]. The next grade up is META-PM[grading scale for myopic maculopathy], which is a photographic study; good for natural history studies and intervention studies. Over and above that,[in the IMI paper] there are set of definitions about some of the conditions which help to standardize terminology, but how that would be monitored would be part of a study design. You don't need to define ahead of time exactly how you measure a condition.
Kate:Your[IMI] paper makes a clear remit for the fact that we need to assess for these issues that we have some classification and grading systems- the META-PM analysis in particular you've mentioned- but that they don't necessarily bring in all aspects of all of the imaging that we have available at this point. What are the other non macular complications of myopia that were key mentions in your report?
Ian:The nuance is trying to bring in a wide range of conditions[such as] retinoschisis and tractional macular hole within a single umbrella- MTM, or myopic tractional maculopathy. In relation to[optic] disc changes, we're used to[seeing] these funny looking myopic discs and association with glaucoma, but at this higher end[of myopia] the structure of the optic nerve and supporting sclera is so abnormal- an optic neuropathy which is not correlated at all with intraocular pressure. This isn't LTG[low tension glaucoma]; this is where the posterior structural changes are so severe that they're directly affecting optic nerve function.
Kate:What's it been called? Myopia-associated glaucoma-like optic neuropathy. Can we make an acronym out of that? MAGLOC-N?(laughs)
Ian:Yeah, no(laughs)
Kate:Can we talk for a moment about retinal detachment? We generally believe that there's a higher risk of retinal detachment with higher levels of myopia but because retinal detachment occurs[frequently] in non-myopic situations, how did the committee deal with that as part of the continuum of myopia pathology?
Ian:It was in the same way that cataract is a complication that occurs in people with myopia and in people that don't have it- there's more frequency in people with higher[myopia]. Similarly with retinal detachment- even though there are distinct features[in myopia, such as] early onsets, and more challenging surgery, it's not a uniquely myopic complication. The thought process is that if the complication is a direct consequence[of myopia], and is uniquely myopic, that it falls within that category[of myopia-associated pathology]. The risk profile[is higher] for the very high myopes- they're a reasonably rare beast presently- and there's still plenty of hyperopes with detachments or traumatic PVD's to swing the numbers. It logically sits in the same bracket as a cataract complication, just that the risk is much steeper.
Kate:I'm interested to ask you about whether the definition[of myopia] includes cycloplegia or not. This was an important factor in the[writing of the IMI] Clinical Management Guidelines. Not all optometrists out there have access to cycloplegia and so from a clinical management point of view, this is important[to understand]. From a clinical trials or research point of view, we might expect that cycloplegia is a requirement. I'm interested to hear how you and your committee decided on the definition of cycloplegia or not as part of that refractive definition.
Ian:This definition has to apply in all sorts of contexts. If you look at epidemiology papers, survey results from adult populations, some of them will be cycloplegic and some won't be. In clinical practice if someone's refracted somebody and haven't used cycloplegia, are they not myopic? Obviously they are myopic. If there's a[strict] definition and there's a remuneration issue in relation to myopia in glasses, for example with insurance,[this could be a problem]. So if it's to define myopia in a practical and a research setting, saying'when accommodation is relaxed' includes cycloplegia, but also means that you can't do[the refraction incorrectly].
Kate:That's an important factor- the definition of myopia[in this IMI paper] is'when ocular accommodation is relaxed'. That allows for normal clinical techniques which occur in practice as well as cycloplegia.
Ian:You don't want to have definition which is so pure that it's only applicable for research. This is a question of where that sort of detail is best placed- within a research protocol or a definition? A research protocol- that's the place to describe a technique. Omitting a reference[in the definition] to the fact that[accommodation] should be relaxed would be remiss in the sense that there's a minimum standard to ensure that you are measuring[accurately]. When young adults or kids are involved, then cycloplegia is a strongly recommended intervention. For adult population surveys, the clinical techniques have evolved over hundreds of years to be able to measure[refraction accurately].
Kate:Absolutely. On some level we've got to be trusted to do the job right! But it's also important to have definitions that encompass a variety of different levels of evidence.
Ian:If you're using a definition within a study then it's almost inevitable, if the age profile indicates, that[cycloplegia would be used]. The idea[in setting a definition] was to be technique agnostic. Sometimes terminology like'objective' are thrown in, but if you do a subjective is that not a myopic measurement?
Kate:Obviously objective techniques could be very important from that research study point of view, but if we've got an enormous amount of epidemiological data, that that might not be as pure in terms of being able to garner[objective instead of subjective refraction] information.
Ian:There's no reason to include that[objective definition]. Where the need is there, it should be included. One of the'get out' clauses[with respect to] the paper is that whenever a small number of people try to put something together, no matter how much of a broader health station there is, there are people who are going to disagree with you! I'm just making an appeal that where nonstandard terms are used, they're clearly defined. Oftentimes, terms are just thrown in there and to the person who uses them they have meaning.[The problem is] where the same term used in multiple ways-'pathological' myopia as a refractive threshold was exactly that. It was used for sort of all sorts of things. There are papers where-4D in a child is termed'pathological', just because it was unusual. It becomes unhelpful.
Kate:It[-4D] may be unusual in a four year old, but not necessarily unusual in a population of 12 year olds.
Ian:The key separation was to entirely divorce that'pathologic' concept from a refractive definition. Refractive thresholds are quantitative- clear and simple.
Kate:So from your analysis of the research and in writing this paper, was it appropriate at all to make an axial length delineation of'pathologic'? I know you said earlier that'pathologic' could be independent of refraction and independent of axial length. We've talked about how it was defined by refraction. What do you say about definition by axial length?
Ian:Whatever threshold you try to[use], there are examples and counter examples- you'll see staphylomas in short eyes. That's a function of how thin the posterior sclera started, and genetic factors.[Dr} Kyoko Ohno-Matsui was a real driving force in that. She took over the clinic in Tokyo[Medical and] Dental[University] from Takashi Tokoro, who had recorded and monitored these high myopes over decades- the most extraordinary database. Most of it is not published- it was published in one pretty hard to find book, which I have and will lend to nobody! So she[Dr Ohno-Matsui] is sitting on this enormous historical database of information. She was a real source of information about all those exceptions. She was the one who was most clear cut that axial length obviously is a correlate,[but] we can define this[myopic pathology] as a set of structural complications. They're defined by those structural changes.
Kate:Yes, and that could be independent of axial length. So obviously we've still got that risk factor there, but we do have to be concerned about these structural complications regardless of axial length and more so as we get into higher levels of myopia.
Ian:Yes, on axial length- the paper on axial and refractive risk factors of uncorrectable visual loss in European people- the Rotterdam study- was kind of the way to go, where you can see that risk profile.
Kate:Ok, that was 26 millimeters people, 26 millimeters.[The cut off for increased risk of uncorrected vision impairment- Tideman et al 2016]
Ian:We've mentioned some of this today about the individual risk and population risk-in terms of the inflection points on where individual risk comes in, regarding refraction,-6D is more relevant. The Brien Holden paper(Holden et al 2016 on global projections of myopia and high myopia)[defined high myopia] on-5D, which will produce a higher level of high myopes on that threshold. It's not that hard to adjust to what that is, but it's not insubstantial. The logic to that paper- being a global paper on a global setting- was that a high proportion of myopes haven't got access to visual correction. So if you are-5D, you reach the WHO definition of legal blindness, uncorrected. There is a perfectly good rationale for that. That's a situation whereby if you're publishing papers around that, you can publish a sensitivity analysis. For that paper, it's probably worth some point doing a sensitivity analysis showing what the differences are at the different[thresholds of high myopia- 5D or 6D]- that'd be very interesting. If you want to put a paper's[definition of high myopia] at 5D and you published it with a sensitivity analysis- it's not hard to do it once the data is all there- it provides a lot more understanding about the question, and also automatically enhances the value of the analysis. I don't think there's any great sex appeal in sensitivity analysis, though!
Kate:A scientific necessity. So this is the projection of global myopia hitting 50% by 2050 that we're discussing, which uses the high myopia threshold of-5D.(Holden et al 2016)
Ian:That's okay, that's fine- because the 50%[projection] was at the low threshold, and that's all[incidence of] myopia. The only impact is on the high myope numbers. The[incidence of] myopia[quoted in this paper] was everybody more myopic than-0.50D, so that[high myopia threshold] doesn't matter. The headline figure is unchanged. It's the absolute numbers for the high myopes which will be slightly different, but they've also published an association between refractive and risk model, and it's related to the continuous refractive error. If you go back to the risk level, it all comes out in the wash.
Kate:More myopia, more risk.
Ian:The core numbers just of high myopia will be different from that[if 5D is used as the threshold], but there's enough analysis there in relation to risk.
Kate:Okay. I wanted to ask you about pre-myopia. In the Myopia Profile Facebook group, there was some heated debate recently about whether pre-myopia was a true entity or whether it's something made up by the evangelists; and it wasn't actually a thing; and if we call every child a pre-myope, then they're satisfying the definition. But, I was really interested to see in your paper that pre-myopia has achieved a definition- it's a distinct entity and not an evangelistic fiction as such. Can you tell us a bit about pre-myopia?
Ian:Yes, well it's now an evangelistic fact! For anyone who doesn't believe it, I would just say one word-[the] Orinda[study]. In longitudinal surveys of when you look at what's happening to eye growth,[they're] showing that three years prior to becoming myopic, the people who end up being myopic start to show a different, accelerated growth pattern. Normally growth is continuing and decelerating. It's an acceleration of that which happens before myopia. And that's not just in[the] Orinda[study].
Kate:Yes. It's been found in the CLEERE[Collaborative Longitudinal Evaluation of Ethnicity and Refractive Error] study as well- the fastest progression of myopia is the year before starting; the year before onset.
Ian:Did I get it the wrong way around? Ah, I think I did.(laughs)
Kate:Well, yeah... look, it's been proven guys. It's been proven. Pre-myopia is a thing!(laughs)
Ian:Thank you Kate, yes it is! There's something that Jane Gwiazda[found]...[there are] all sorts of papers showing these longitudinal studies- when you follow kids through, the fastest rate of change is often in the year before they[become] myopic. There's a biological trigger which produces that change. Now the proof of the pudding is in the eating! For pre-myopia to be a valid concept within that definition, it sounds a bit woolly- it's where there are enough risk factors to produce probability of myopia.
Kate:The key risk factor is refractive- being not as hyperopic as a child should be[for their age]. We also have other risk factors in terms of visual environment that we understand, and family history as well, that are risk factors in pre-myopia.
Ian:If you look at the latest CREAM[Consortium for Refractive Error and Myopia- genetic links with myopia] paper, which was primarily about different visual mechanisms- it's a great paper. Buried in that, there was one[key] figure. Gene prediction for myopia is rubbish because it only explains a tiny amount of the variance. If you look at it as a decile- the highest decile, the top 10%- there's a risk profile and look at the risk of high myopia within that group, it's enormous. Being able to predict accurately across a population is not going to work. But that's not what we care about. In pre-myopia, we're not trying to treat everybody. The idea is to get the really bad ones into treatment as soon as you can. Actually if[the definition and management of] pre-myopia is only used for the top 10 to 15 percent...
Kate:At least we're going to capture more aren't we?
Ian:There's an intervention- and there are trials now looking at pharmacological pre-prevention- outdoor time is more effective pre- than post-[myopia onset]. If nothing else, the moment they cross the line[into myopia], you can do something about it. The other thing I'm looking at, at the moment, is the use of centile charts to see where kids'[axial length measurements] are within a population, to see if they're changing, when kids are growing[normally].
Kate:I think that's fascinating and that's going to be the next clinical frontier for us, because we understand what these growth charts are like for their height and for other measures of pediatric growth, but we're yet to really achieve clear pictures of what that's like for the eye. I think that could really revolutionize picking up the pre-myope and myopia progression.
Ian:There is a bucket of large population based observational cohorts.
Kate:So we need to merge it. Someone's got to merge them all together- Ian that sounds like a job for you!
Ian:I'm actually doing this right now! Caroline Klaver and Cathy Williams who are involved in the[CREAM] Consortium pulled in some of the Rotterdam study, the UK data, and some of the old European data in relation to centile charts- where you are on the refractile spectrum- that's pretty easy to pull together.(Tideman et al 2018) There are a couple of great European longitudinal cohorts-[for example,] Kathryn Saunders' NICER(Northern Ireland Childhood Errors of Refraction) study. There are buckets of data there- the data is available to convert it all to[highly] predictable centile charts. Because they are longitudinal studies, you can tell who ended up becoming myopic, in terms of risk prediction.
Kate:[In summary] pre-myopia is a thing; we have some interventions available. Your paper doesn't cover interventions- that's a whole extra paper altogether. But it's definitely something that we need to be able to detect, as you've said- to look at the children who are likely to progress and end up with those higher levels of myopia, and the higher associations with pathology.
Ian:If the parents are myopic and the kid is close to[the refractive threshold] then you are duty bound to monitor them closely- do that, do that. You can believe that pre-myopia is complete rubbish. Then if they become myopic...
Kate:Then you can believe in it!
Ian:You don't have to believe in it! But if you just know the fact that some kids at risk and accept the fact that young emmetropes might become myopic, particularly if their parents are myopic, offer them monitoring.
Kate:Absolutely. Are there any final key points or key take homes that you'd really like clinicians to take away from your paper? Aside from reading it! Because it's a brilliant read and I think it's actually the shortest of the IMI papers- it's brilliant and succinct- you've done a very good job in that respect! Is there anything else that you'd like to convey to our listeners, particularly from the clinical point of view? We've talked about understanding definitions; the importance of that from a[research] study point of view, but also from a clinical point of view. We've talked about pre-myopia; we've talked about structural complications of myopia.
Ian:The one request is in relation to the terminology- try to avoid all the modifying terms.
Kate:Yes.'Physiological'[myopia]- not a thing.
Ian:Yes.'School age','school correlational','juvenile onset'- all of these terms[exist], because European kids become myopic later; Chinese children earlier; are they different conditions? Just think very hard before you put any other word in front of'myopia' that's not'high'. It's either myopia, or it's high myopia. And if there's any other words, adjectives in there...
Kate:... you might be getting creative.
Ian:Just take it out! If you really have to put it in for really good reason- define it and justify it.
Kate:Absolutely. I think that's a great place to finish. Thank you so much for your time Ian- we need to get ourselves sorted now for this presentation. We've had bread rolls put in front of us so we can have a snack beforehand! Thank you very much and I look forward to hearing more of what you've got to tell everyone tonight.