Factors affecting the benefit of glasses alone in treating childhood amblyopia: an analysis of PEDIG data

Abstract

Background: To evaluate factors associated with better outcomes from optical treatment alone in amblyopic children from 3 up to 7 years.

Methods: Data extracted from two studies with similar protocols, Amblyopic Treatment Studies 5 (n = 152) and 13 (n = 128) from the Pediatric Eye Disease Investigator Group database, were used to determine by regression analysis the factors associated with improvements in visual acuity in the amblyopic eye, inter-ocular visual acuity difference and stereoacuity. Input variables were aetiology of amblyopia (anisometropic, strabismic and combined-mechanism amblyopia), treatment compliance, visual acuity, interocular visual acuity difference, stereoacuity, tropia size at distance and near, age and refractive error at baseline.

Results: Despite the range of clinical factors considered, our models explain only a modest proportion of the variance in optical treatment outcomes. The better predictors of the degree of optical treatment success in amblyopic children are visual acuity of the amblyopic eye, interocular visual acuity difference, stereoacuity, treatment compliance and the amblyopic eye spherical-equivalent refractive error. While the aetiology of the amblyopia does not exert a major influence upon treatment outcome, combined-mechanism amblyopes experience the smallest improvement in visual acuity, tropia and stereoacuity and may need longer optical treatment periods.

Conclusions: While results identify the factors influencing optical treatment outcome in amblyopic children, clinicians will be unable to predict accurately the benefits of optical treatment in individual patients. Whether this is because relevant clinical or non-clinical factors (e.g. nature and volume of daily activities undertaken) influences the outcomes from optical treatment has not yet been identified and remains to be discovered.

Keywords: Amblyopia; Interocular difference; Optical treatment; Stereoacuity; Visual acuity.

Fig. 1

Selection criteria and participant numbers in the two Amblyopia Treatment Studies (ATS), ATS5 and ATS13. The studies are similar until the end of the OTA phase, and diverge thereafter. However, this analysis focuses on gains up to the end of the OTA phase and thus the results from ATS 5 and ATS 13 are combined. Total participants n = 280 [(n_ATS5= 152) + (n_ATS13= 128)]. Data sets for both studies are available on the PEDIG website [11]

Fig. 2

Distribution of initial (white) and final (gray) values of the outcome variables a) VAAE, b) IOD-VA and c) stereoacuity and changes in these three variables ((d)-(f)) in the three study groups (anisometropic (A), strabismic (S) and combined-mechanism (C) amblyopes). In panels d) to f), negative values correspond to improved performance following the OTA phase. The number of participants is n = 280 for a), b), d) and e); n = 235 for c) and n = 118 for f). Boxes represent the interquartile interval and whiskers cover the data at largest distance from the median that cannot be considered outliers. The narrowing of the box around the median (“notches”) define the 95% confidence interval of the median. Stereoacuity levels range from 1 (40”) to 7 (> 800”). Initial stereoacuity data for the anisometropic group were unavailable

Fig. 3

Changes between visits for VAAE (left), IOD-VA (centre) and stereoacuity (right). Not all patients had data for the intermediate visits: n = 148 for VAAE and IOD-VA. Stereoacuity is reported only for strabismic and mixed amblyopes. The y-axis in each plot shows the change in the clinical measure between the time points indicated on the x-axis. Negative values on the y-axis imply improved performance following the OTA phase. 8w: 8 weeks; 16w: 16 weeks; Final: time point when the OTA was judged to have come to an end

Fig. 4

Top: First and second principal components of the baseline variables used in the PCA analysis (IOD indicates the interocular difference for the indicated variable). Bottom: Values of the first (V₁) and second (V₂) principal components for the anisometropic (left), strabismic (center) and combined-mechanism (right) amblyopic samples. Blue squares: resolved amblyopia; Red circles: unresolved amblyopia. The percentage of explained variance appears between brackets