Amblyopia – the commonest vision abnormality of childhood – is characterized by a loss of visual acuity usually of one eye only. Treatment aims to promote function of the amblyopic eye and does this by restricting, usually through occlusion, the competitive advantage of the fellow eye. Recent experimental evidence demonstrates that the recovery of vision following early deprivation is facilitated by increasing visually evoked activity. An analogous approach in humans is to minimise image blur by correcting refractive error prior to treatment – a practice which may account for the poorly quantified improvements in visual acuity sometimes attributed to `spectacle adaptation'. Here we describe clinically significant gains in visual acuity obtained over a period of 4–24 weeks in a group of amblyopic children arising solely in response to the correction of refractive error. Consequences for the clinical management of refractive amblyopia are discussed.
Although the causal mechanisms have not been unequivocally established, most amblyopia is found in association with anisometropia, or with squint or, a combination of the two, acting so as to present an obstacle to visual development. Primate animal models of amblyopia have repeatedly shown the primary visual cortex (area V1) to be dysfunctional (Kiorpes and McKee, 1999). Functional imaging studies confirm processing abnormalities in area V1 in humans and hint at additional deficits in higher cortical areas (Anderson, 2002). In conspicuous contrast, to an examiner, the eye itself appears structurally normal.
Classic teaching dictates that optical correction (in practice the prescription of spectacles and contact lenses) can seldom on its own, eliminate the amblyopic eye deficit. The fellow eye is thought to have established a competitive advantage over the amblyopic eye (Wiesel and Hubel, 1965) which can only be overcome if visual input to the fellow eye is restricted or degraded. It is on this understanding that the mainstream treatments of amblyopia (occlusion and penalization) draw their empirical support.
However, recent work suggests that the recovery of vision following early experimentally induced deprivation depends on the absolute level of visually evoked activity (restoration of visual input) and not on a competitive activity-dependent process (Mitchell and Gingras, 1998). If this is indeed the case, then such a mechanism might underpin sporadic reports in humans and animals (Kivlin and Flynn, 1981; O'Dell et al., 1989; Clarke and Noel, 1990; Moseley et al., 1997), which hint that correcting the optical defects present in amblyopic eyes – sometimes referred to in a clinical setting as `spectacle adaptation' – may, in the long term, be therapeutic even where no attempt has been made to remove the competitive advantage of the fellow eye.
To determine whether, and to what extent, the correction of refractive error alone might reduce amblyopic visual deficits, we prescribed appropriate spectacle correction to a group of affected children and monitored their visual status at weekly intervals. No other treatment was administered during this period.
With written parental consent, 13 amblyopic children (age: 3.8–6.3 years) who had never previously worn spectacles or undergone any amblyopia treatment participated. These children had been referred to one of two paediatric ophthalmic outpatient clinics at either St Mary's or Hillingdon Hospital. No children were excluded on the basis of age although it was an entry requirement that they be capable of reading letters directly from an acuity chart or be able to match them using a hand-held key card. Diagnosis was confirmed by full ophthalmic examination including a cycloplegic refraction. When prescribing spectacles, the differential spherical error between the eyes was always preserved but for hypermetropic errors the full spherical correction was not always prescribed. Similarly, the correction of cylindrical errors of less than 0.50 D was at the clinicians' discretion. The visual acuity entry criterion was 0.1 logMAR (corrected) or worse in the eye with the poorest acuity. In 11 children, refractive error (aniso-, or isometropic) was the only amblyopic association whilst in the remaining two, squint was also present. A complete clinical description is provided in Table 1. Visual acuity was initially recorded upon first-time spectacle wear using ETDRS logMAR charts (Precision Vision®, La Salle, IL, USA) scored on a letter-by-letter basis (resolution=0.02 logMAR). Children were tested with the same chart throughout the course of the study. The (most) amblyopic eye was always tested first. Four further measurements were sought at consecutive weekly intervals. At this point, the protocol specified that children who had shown no improvement or whose acuity had declined should leave the study and revert to standard care. Beyond this period, weekly measurement continued until either resolution in the amblyopic eyes reached or exceeded 0.0 logMAR (normal visual acuity) or, was seen to stabilise (operationally defined as four inflexions in the plot of acuity against time or to remain unchanged over the same period).
Table 1. . Subject details
We excluded from analysis one subject on grounds of non-attendance (from weeks 7–16). Six of the remaining twelve subjects missed occasional visits but never more than four consecutively. Mean (S.D.) initial corrected acuities in the amblyopic eye were 0.47 (0.26) logMAR. Final visual acuities were recorded from between 4 and 24 weeks at which time all subjects had demonstrated clinically significant improvements in amblyopic eye acuity (range: 0.1–0.5 log units=1–5 lines) (Figure 1). Eight of the twelve subjects now fell within an accepted range of normal acuity (0.0 logMAR or better) and hence there was no clinical indication to occlude these children.
Analysis of fellow eye data proved problematic given that in five subjects initial corrected acuities were actually poorer than their uncorrected acuities. Therefore, in the fellow eyes, start point acuities were defined as those recorded at that time at which each subject's corrected acuity was equivalent to, or exceeded, their initial uncorrected acuity (in practice this was never later than the third visit subsequent to the initial recording). Even on this basis, corrected fellow eye acuity in seven of the twelve subjects fell within the amblyopic range (≥0.1 logMAR) yielding an initial group mean (S.D.) of 0.13 (0.12) logMAR. By the time at which best acuity had been recorded in the amblyopic eye, mean (S.D.) fellow eye acuity had improved to 0.00 (0.05) logMAR.
One explanation for the acuity gains seen in the children is that of a simple practice effect arising out of the weekly test protocol. To examine this possibility we recruited a further six subjects (age: 4.1–6.8 years) who met the main study entry criteria. These underwent identical measurement procedures excepting that corrected acuity was recorded on only two occasions: immediately upon spectacle prescription and subsequently 7 weeks later. These subjects showed comparable gains (range: 0.08–0.60 log units) to those seen at around the same period in the main study suggesting that repeat testing did not significantly contribute to the observed improvements in visual resolution.
That the full benefits of refractive correction are not immediately manifest clearly rules out the simple, optical elimination of retinal image blur as a likely explanation for the gains observed; further, practice does not appear to contribute. Placebo effects could conceivably play a role but given the magnitude, universality and time-dependency are unlikely to be of significance. We propose that the gains in acuity we have observed in amblyopic children are mediated by the provision of images with their high spatial frequency content restored by the correction of refractive error. We suggest that this increase in the spatial complexity of the retinal image provides a stimulus sufficient to activate non-competitive, activity-dependent processes perhaps shaping cortical architecture in the manner discussed by Hübener and Bonhoeffer (1999). Very rapid improvement (within a period of hours and days) has been observed in the acuity of infants upon the removal of a congenital cataract and similarly attributed to a non-competitive process (Maurer et al., 1999).
Surprisingly, the improvement (0.28 log units) of one subject demonstrating both esotropia and anisometropia was not untypical of the remaining (straight-eyed) subjects. Clearly, if this improvement is to be attributed to the benefits of spectacle correction the `anisometropic component' of the subject's amblyopia must have been responsible, in the most part, for the subject's initial visual deficit.
The response to spectacle wear of the fellow eyes appeared to mimic that of the amblyopic eyes: gains in acuity occurring during the course of the study. In several cases this was predictable – high bilateral refractive errors associated with bilateral amblyopia – although in others this was not the case perhaps hinting at a small residual abnormality seen in the fellow eye when rigorously tested (e.g. see Woo and Irving, 1991).
We have shown that simply correcting the optical defect associated with an amblyopic eye can lead, over a period of several weeks, to substantial gains in visual acuity. As stated in the Introduction, this is not an original observation (see e.g. Clarke and Noel, 1990) but hitherto the magnitude, ubiquity and time-course of this phenomenon has not been established within an empirical framework. Further, the neurophysiological process which might underpin `spectacle adaptation' have not been addressed.
Finally, it is important to emphasise and reiterate the significant implications for clinical practice of a period of refractive correction; eliminating the need for conventional treatment in some cases, while reducing the amount required in others. In those children who do require further therapy, one might also predict improved compliance, for when the fellow eye is occluded or penalized, the available visual capacity in the amblyopic eye will be greater than had no period of spectacle wear been undertaken. Practitioners will already be aware that children find occlusion and penalization therapies invariably more unpleasant than spectacle wear.