Randomized clinical trial comparing streaming dichoptic films to patches for the treatment of amblyopia in children aged 3-7 years


The research protocol adhered to the principles of the Declaration of Helsinki, was approved by the University of Texas Southwestern Medical Center Institutional Review Board, and complied with the requirements of U.S. Health Insurance Portability and Accountability 1996. Written informed consent was obtained from the accompanying parent. The protocol is available at www.clinicaltrials.gov (ID: NCT03825107 [31/01/2019]). In addition, this study follows the recommendations proposed by the CONSORT statement.


Children aged 3 to 7 years who had been diagnosed by pediatric ophthalmologists as having amblyopia due to strabismus, anisometropia or both (combined mechanism) were eligible. The inclusion criteria were: (1) visual acuity of the amblyopic eye of 0.2 to 0.8 logMAR; (2) visual acuity in the other eye of −0.1 to 0.2 logMAR; (3) interocular visual acuity difference ≥ 0.2 logMAR; (4) corrected anisometropia or strabismus (

study design

This study was a randomized clinical trial (parallel group design). At the baseline visit, eligibility was checked and vision assessments were performed. Children were randomly assigned to watch contrast-rebalanced dichoptic animated films (experimental treatment) or to patch the other eye 2 h/day every day (standard amblyopia treatment) for 2 weeks. The 2-week time point was chosen for the primary outcome based on our pilot studies of contrast-rebalanced dichoptic films1.2 and our randomized treatment trial of contrast rebalanced dichoptic games8. Randomization was done via the website www.randomization.org with a coded distribution of a 1:1 ratio, with blocks of 12 allocations. Investigators only had access to the randomization assignment, which was sealed in a numbered envelope, after the child had been enrolled. At the 2-week primary outcome visit, children in the patch group switched to the dichoptic animation film group and all participants were asked to return for a secondary outcome visit vision assessment at 4 weeks. . Families had the option of continuing the dichoptic films for a total of 6 weeks (Fig. 1). All children returned for a secondary outcome visit at 4 weeks. After that, the children in the dichoptic films group had the option of continuing with the films for an additional 2 weeks and the children who were initially in the patch group had the option of continuing with the films for an additional 4 weeks (up to 6 weeks of films for both groups). Children who continued had follow-up visits with vision assessments every 2 weeks (14 ± 3 days).

Figure 1

Dichoptic animated films

Eighteen popular animated feature films were modified using custom MATLAB software to enable dichoptic presentation on a New Nintendo 3DS XL platform (Nintendo Corporation, Kyoto, Japan), which was loaned to the child for the duration of the study so that the films can be viewed at home. This Nintendo device allowed dichoptic viewing without the need for anaglyphic, polarized, or shutter glasses or a VR headset. The movies had the same design as in our recent prospective contrast-rebalanced movie study in which children came to the lab to watch movies on a passive 3D screen.1.2. In short, odd lines on the Nintendo 3D screen were visible with one eye, and even lines were visible with the other eye. A patterned image mask composed of irregularly shaped spots was multiplied by the images seen by the amblyopic eye, and the inverse patterned mask was multiplied by the images seen by the other eye, so that different parts of the display were seen by each eye. Film spots seen by the amblyopic eye still had high contrast (100%), while complementary drops seen by the other eye had reduced contrast. Because the blobs had Gaussian edges, the edges of the blobs overlapped and were seen by both eyes with different contrasts, preventing any meaningful rivalry. The shape and location of the blobs varied dynamically every 10 s. The contrast of the other eyes started at 20% for the first movie and increased by 10% for each following movie (i.e. 20%, 22%, 24%, 26%, 29%, …). The films were uploaded to our streaming service Amazon Web Services (AWS; Seattle, WA, USA) in two-week packages, each containing 6 films that preserved the contrast order of the other eyes from bottom to top .

Study protocol

Children assigned to the dichoptic film arm were asked to watch 3 films per week (approximately 4.5 h per week) during each 2-week period, based on the 4–5 h per week used in our previous binocular treatment studies contrast-adjusted and randomized treatment trials1,2,8,9,10. When they visited, we provided a link to their Nintendo loaner for all 6 movies scheduled to be viewed over the next two weeks. The names of the films were numbered on the Nintendo device, in order of eye contrast, and the family was told that the films should be viewed in that order at home with the date recorded on a sheet of paper newspaper. The diary sheet contained images of the title frames of each of the 6 films with space next to each for the parent to write the date it was watched. The contrast of the other eyes in films 1 to 6 increased by 20 to 32% during the first 2 weeks, films 7 to 12 increased by 35 to 57% during weeks 3 and 4 and, finally, films 13 to 18 increased from 63% to 100% if children participated in optional weeks 5 and 6. Children assigned to the patching arm received patches and were asked to patch for two hours a day for two weeks. During the 2-week visit, the children lent a Nintendo and started watching movies, starting with 20% eye contrast for the next two weeks, and had the option of completing up to 6 weeks of movies .

Vision assessments

Monocular congested best-corrected visual acuity (BCVA) was tested using the Amblyopia Treatment Study HOTV protocol for children 3-6 years of age (N=37)11,12,13 and the E-ETDRS protocol for 7 year olds (N = 16)14,15,16. A few of the younger children who were unable to complete the ATS-HOTV were tested with Allen Pictures on the Mentor Baylor II Visual Acuity Tester (BVAT-II; Mentor Ophthalmics, Inc., Norwell, MA, United States; N=7). Whichever visual acuity test was used for baseline assessment, it was also used at all follow-up visits. Stereoacuity was assessed with the Randot Preschool stereoacuity test and the Stereo Butterfly test [Stereo Optical Inc, Chicago, IL]17; stereoacuity was converted to log arcsec for analyses; zero stereoacuity was arbitrarily assigned a value of 4.0. The extent of suppression was assessed using the Worth-4 point test at 7 distances (3 m, 2 m, 1 m, 0.67 m, 0.5 m, 0.33 m and 0. 16m)18; the farthest distance at which the child reported 4 points was converted to the scotoma suppression area in logarithmic degrees. Depth of suppression using a dichoptic eye diagram that identifies the non-preferred eye/preferred eye contrast ratio (i.e., balance point) at which the child can overcome suppression and point out the letters presented to each eye with equal probability (Contrast Balance Index; CBI)19.20.

Sample size and statistical analysis

The primary outcome measure was the change in BCVA of the amblyopic eye at the 2-week visit. The analysis was conducted with a modified intention-to-treat approach, limited to participants who completed the 2-week visit within the predefined analysis window (± 3 days after baseline/randomization visit) and no imputation for missing data. A t-test was used to compare the change in visual acuity in the two treatment groups. Based on our recent prospective study of contrast-rebalanced dichoptic films as a treatment for amblyopia1.2, we anticipated a 0.15 ± 0.10 logMAR (7.5 letters) improvement in visual acuity at the 2-week primary outcome visit for children assigned to the dichoptic film arm. For children assigned to the patch group, we anticipated an improvement of 0.07 ± 0.10 logMAR (3.5 letters) at 2 weeks based on the results of our previous study of the patch versus a dichoptic game at rebalanced contrast as a treatment for amblyopia8. With these expected means, a sample size of 28 per group (56 total) would provide 85% power to report that the two groups have significantly different means, using a two-tailed analysis p-value less than 0.0521. We recruited and randomized 65 children (32 to 33 per group) to account for potential loss to follow-up.

As a secondary analysis of BCVA at the 2-week primary outcome visit, one-sample t-tests were performed to determine whether improvement in BCVA of the amblyopic eye at the 2-week visit was significant in each group. In addition, confidence intervals (95%) were calculated to determine whether there were significant improvements in stereoacuity, depth of suppression, or extent of suppression at the primary 2-week outcome for each group. Exploratory analyzes were conducted to determine whether the improvement in visual acuity at the 2-week visit was associated with the child’s baseline visual acuity, baseline stereoacuity, or response to conventional treatment. with glasses and patch before registration. Response to conventional treatment was categorized as “responded to treatment” if improvement in visual acuity ≥ 0.1 logMAR occurred with glasses and patch on two or more visits prior to enrollment. Children were classified as having had “no prior response” to conventional treatment if there was no improvement in visual acuity over two visits separated by at least 8 weeks.


About Author

Comments are closed.