The Incremental Retinal Defocus Theory and its MeritThis shouldn’t be too relevant for most of you, but if you are interested, make sure you make it to the end.of this post. I thought I knew what IRDT is until I went through the papers by their authors - Chances are that one can't guess what IRDT is by its name.
IntroSince Hung and Ciuffreda published their initial paper on Incremental Retinal Defocus Theory (IRDT), they have been trying to convince their colleagues about the predictive power of their theory. Namely, that IRDT is consistent with the recent myopia findings on the effects of undercorrection and light.
To begin, what exactly is this Incremental Retinal Defocus Theory (IRDT)? It’s an unifying theory on myopia development, maintaining that the
decrease of retinal defocus, in a
time increment, reduces the release rate of retinal neuromodulators, which in turn reduces the proteoglycan synthesis rate, rendering scleral tissues more vulnerable, which leads to an increase of the axial growth rate.
In a similar but opposite manner, IRDT also maintains that the
increase of retinal defocus, in a time increment, will send a cascade of chemical signals which would ultimately retard the rate of axial growth.
Hung and Ciuffreda emphasizes that it’s the
change of retinal-image defocus area, rather than the magnitude of defocus itself, that triggers the different behaviours of retinal neuromodulators (e.g., dopamine), and the ensuing effects on axial length.
Without further ado, Let’s look at what Hung and Ciuffreda have to say on different optical treatments on myopia.
The alleged effect of full-correctionWith full correction, looking at optical infinity and 25cm do not result in substantial difference in retinal defocus. At optical infinity, although an individual might experience slight accommodation lead, the images at this distance are clear overall. Similarly, while at 25cm the same individual might experience a very slight accommodation lag, the presence of accommodation ensures that the retinal defocus at 25cm ultimately remains minimal.
That is, assuming that the person’s accommodation amplitude is at least 4D, then the person’s accommodation system, which enables adaptation to various visual stimuli, ensures that changes in retinal defocus from far to near is minimal. As a result, there is little change in the rate of axial growth.
ProblemsAll seems good, except that people wearing full prescription tends to see their diopter increasing year after year (until it stabilizes). Are they progressing at a genetically-determined default rate? No. It’s generally false that full prescription stabilizes myopia. In fact, anecdotal reports invariably suggest that most myopes who switch to an +2 undercorrection (for far and near) experience a slight initial reduction in myopia.
In fact, my experience with myopia shows that full prescription does exactly the opposite of stagnating myopia. My myopia only stopped when I stopped using minus lens for close work, while
simutaneously abstaining from focusing within 50cm. In brief, evidence suggests that the appeal to the so-called genetically-programmed progression rate is precisely an appeal to the “unquestioned wisdom from the past.”
The alleged effect of 0.75D undercorrectionAssuming that a person’s default distance is at optically infinity (i.e., -0.75D of accommodation stimulus), with the usual accommodation lead at far distance, the individual, with the accommodation “disabled” at far, now experiences a relatively significant retinal defocus at optical infinity (i.e, >= 6 meters).
However, during a close work session, the individual shifts from far to, say, 25cm, at which point the individual would experience a significant
reduction of retinal defocus (after the adjustment made by the accommodation system). Consequently, an increase in axial growth rate would ensue, as a result of the cumulated effect of regular far-to-near activities.
ProblemsI understand that these researchers were trying to find an approach that would explain the O’Leary et al. study, but I just have to say that the reasoning they cooked up is too much of a stretch (and this is not to even take into account the criticisms of that O’Leary study, which some believe either contains serious methodological flaws, or reveals no statistical significance between the full prescription group and the undercorrected group, after a correction on the data)
Dr. Alex Frauenfeld’s approach to myopia rehabilitation, which typically consists of giving a 1D undercorrection for far, and a 2D undercorrection for near, would probably make no optical sense according to IRDT. Firstly, the +1 undercorrection for far would “disable” the accommodation system, creating a significant amount of retinal defocus at far. Secondly, when a Frauenfeld patient focuses near slightly beyond their far point, there would be practically no retinal defocus.
What this means that the regular shifting from far to near represents a significant decrease of retinal defocus, and the magnitude of this decrease is even greater than that resulted from a typical 0.75D undercorrection. Consequently, one would expect a Frauenfeld patient to become even more myopic. In reality though, just the opposite is true. Frauenfeld patients generally fare very well with their myopia - Just the opposite of what Hung and Ciuffreda's reasoning would have predicted.
The alleged effect of strong plus lensWearing a strong plus in effect disables the accommodation system. As a result, a subject can no longer focus far and near equally well. Focusing far with plus lens then produces a significant increase of retinal defocus, which retards the axial growth rate over time.
ProblemsRead between the lines and you might see a double standard emerging

. There is an implicit assumption that the research animals don’t engage in far-to-near activities, and that’s the reason why there is an overall increase of retinal defocus. There are actually animal studies in which the focusing distance is controlled, so that animals only look at a particular distance most of the day.
Under those experimental settings, the increase of retinal defocus only happens in the initial days of the experiments. However, despite the lack of change in retinal defocus in the later days of the experiments, the myopia-inhibiting effect of plus lens still make itself present. IRDT would have predicted than little refractive change would occur in the later days of such an experiment.
Other predictionsSince children can accommodation fairly well, a small overcorrection would lead to little change in retinal defocus (after accommodation), hence the effect of overcorrection wouldn’t be significant. Similarly, since multifocals allows one to see both far and near equally well, IRDT predicts little change of myopia progression from the use of, say, bifocal.
ProblemsOvercorrection in animals and humans invariably worsen myopia (dubbed by the mainstream as the normal genetically-programmed myopia progression rate). Previous studies on multifocals, especially the bifocals, do display some degree of myopia-inhibiting effect. These effects were brushed off as being insignificant.
Alledged ideal prescriptionHung and Ciuffreda suggests a full prescription for far, and a weak undercorrection (+0.5 or +0.75 add) for near tasks, as their computer simulation suggests that doing so would minimize any change in retinal defocus, and hence result in little myopia progression.
ProblemsWith these prescriptions, the patients would still experience chronic close strain. The full prescription ensures that ciliary never relaxes, and the tiny amount of undercorrection in the prescription for near almost guarantee that reading will be done with a fair amount of close strain.
Final wordsThe Incremental Retinal Defocus Theory is not a theory about
minimal myopic defocus. Rather, it’s a theory about the
change of retinal defocus, in an increment of time. In fact, IRDT would predict that constant edge-of-blur will produce no change in retinal defocus over time, thus no change in myopia progression.
IRDT strikes me as being out of touch with the reality of myopes. For the sake of completeness though,
here is the link to one of Hung and Ciuffreda’s paper.