Subjects included in the study

Descriptive characteristics of the subjects included in each group of the current study are summarized in Table 1. VT group was significantly older (p < 0.001) and had more subjects who underwent previous treatment for amblyopia in other clinics before their inclusion in the study (p < 0.001) than patching group. Also, there were differences between groups according to the type of amblyopia. VT group included more subjects with anisometropic amblyopia (24 out of 32) than the patching group (9 out of 20). There were no significant differences between groups in other variables, such as gender (p = 1.000), patching regimen (p = 0.258), spherical equivalent (SE) of fellow eye (p = 0.836) or SE of amblyopic eye (p = 0.910), and baseline VA of amblyopic (p = 0.090) or fellow eye (p = 0.066). However, it should be noted that in the VT group there were 1 subject with anomalous correspondence and 5 subjects who did not undergo patching properly despite the professional prescription. Furthermore, in relation to strabismic patients, in the VT group, there were 6 patients with strabismus, 1 had constant exotropia that became intermittent after optical correction, 2 patients had partially accommodative esotropia with residual microtropia, 1 patient underwent botulinum toxin treatment for esotropia before being included in the study, and 2 patients had accommodative esotropia that was adequately corrected with optical correction. In the patching group, out of the 11 strabismic patients, 3 patients had non-accommodative esotropia, 2 patients had constant exotropia, 5 patients had accommodative esotropia that was corrected with optical correction, and 1 patient had partially accommodative esotropia with residual microtropia after using glasses.

Table 1 Descriptive characteristics of patients in VT group and patching group. P-values with asterisk represent those corresponding to statistical significanceChanges in visual acuity

In Fig. 5, data on changes in VA from baseline to 6 months after initiating the treatment of both groups during treatment are displayed. In VT group, the VA of the amblyopic eye significantly improved from 0.24 ± 0.19 logMAR at baseline to 0.11 ± 0.15 (p < 0.001), 0.08 ± 0.19 (p = 0.004), and 0.05 ± 0.07 logMAR (p = 0.678) at 1, 3, and 6 months, respectively (Fig. 6). In patching group, the VA of the amblyopic eye significantly improved as well from 0.40 ± 0.32 logMAR at baseline to 0.23 ± 0.19 (p = 0.005), 0.14 ± 0.15 (p = 0.345), and 0.10 ± 0.13 logMAR (p = 0.018) at 1, 3, and 6 months, respectively. There were no statistically significant differences at baseline in the VA of the amblyopic eye between VT and patching groups (p = 0.090), with an improvement of this parameter with treatment in both groups, and no significant differences between them (p = 0.732) at the end of the follow-up. At the end of the follow-up, 36 patients (12 from the patching group and 24 from the VT group) reached a VA of 0.1 logMAR or better in the amblyopic eye.

Fig. 5

Changes in visual acuity in VT and patching groups during the follow-up in the amblyopic and fellow eye. An asterisk highlights which differences between groups reached statistical significance

Fig. 6

Distribution of changes in visual acuity in the amblyopic eye in VT and patching groups during the follow-up, remarking the criterion of amblyopia resolution

The total improvement of VA was 0.18 ± 0.16 logMAR in the VT group and 0.31 ± 0.35 logMAR in the patching group when comparing the last visit with the baseline exam. Although patching group showed on average a higher improvement, there were no statistically significant differences between groups (p = 0.317). This can be due to the non-significant but lower VA of the amblyopic eye in the patching group, having these patients more range of recovery. In addition, mean VA of the amblyopic eye in VT group reached the value of 0.1 logMAR VA or better after 3 months of treatment, while in the patching group this value was achieved after 6 months of treatment.

Regarding the changes in the VA of the fellow eye, there was a slight improvement during the treatment, but not reaching statistical significance in any of the two groups. Likewise, there were no statistically significant differences between groups in this parameter, except at 1-month follow-up, but the difference was not clinically relevant.

According to the amblyopia type, there were no statistically significant differences in VA at baseline between amblyopia subgroups. However, VT group showed slightly better VA than patching group during treatment for anisometropic and strabismic cases. Specifically, the VA was significantly better in VT group after 3 months of treatment for those cases with anisometropic amblyopia (p = 0.007). Nevertheless, both groups reached a similar mean magnitude of anisometropic amblyopia visual recovery at 6 months (patching group: 0.04 ± 0.06 logMAR; VT group: 0.03 ± 0.06 logMAR; p = 0.898). In strabismic amblyopia, the VA was significantly better at 1 month (p = 0.041) in the VT group, although both groups improved partially and obtained similar VAs during the rest of follow-up. From a clinical perspective, the anisometropic amblyopia was rehabilitated with both treatments, but the VA recovery was faster with VT. In strabismic amblyopia, both groups showed an improvement as well, but the mean VA values after therapy did not accomplish the amblyopia resolution criteria at 6 months in any of the two groups (patching group: 0.15 ± 0.14 logMAR; VT group 0.10 ± 0.08 logMAR; p = 0.699).

In addition, in the VT group, there were moderate correlations between the type of amblyopia and the VA at 1 month (r = 0.448, p = 0.010), 3 months (r = 0.593, p < 0.001), and 6 months (r = 0.535, p = 0.015), and therefore patients with anisometropic amblyopia reached better VA during treatment. However, there was no correlation between amblyopia type and baseline VA (r = 0.145, p = 0.428). On the other hand, patients in the patching group showed a moderate correlation between the type of amblyopia and baseline VA (r = 0.519, p = 0.019), but without correlation with the VA achieved during and after treatment (p > 0.05).

Regarding the effect size of the change in VA with each type of treatment, the Wilcoxon effect size was r = 0.53 in the VT group and r = 0.44 in the patching group at 3 months. However, at the end of the follow-up, the effect size was slightly higher in VT group (r = 0.48) than in patching group (r = 0.54).

Changes in contrast sensitivity

An enhancement in CS was observed in the amblyopic eye of the VT group (Fig. 7). This improvement was significant for all spatial frequencies at 1 month and continued improving until 3 months, although in this visit the change was only significant for 0.5 (p = 0.003), 1 (p = 0.009), 2 (p = 0.028), and 4 (p = 0.049) cycles/°. In the last visit, after 6 months, CS significantly increased as well in 0.5, 1, 2, and 8 cycles/°, but decreased in 4 (p = 0.037) and 16 (p = 0.673) cycles/°. Nonetheless, there was significant improvement in 0.5, 1, 2, 8, and 16 cycles/° from the baseline until the end of follow-up, with mean changes of 7.00 ± 5.57 for 0.5 cycles/° (p = 0.001), 25.94 ± 25.64 for 1 cycles/° (p = 0.001), 42.81 ± 39.02 for 2 cycles/° (p < 0.001), 33.31 ± 54.52 for 4 cycles/° (p = 0.004), 20.40 ± 22.50 for 8 cycles/° (p = 0.002), and 1.47 ± 2.85 for 16 cycles/° (p = 0.020), respectively.

Fig. 7

Contrast sensitivity of patients in the VT group during follow-up. Data extracted from the software VisionaryTool. The Y axis shows the contrast sensitivity based on Weber’s contrast, and the X axis shows the spatial frequencies from low (0.5 cycles/°) to high (16 cycles/°)

Binocular function score

In the VT group, binocularity experienced a relevant change, improving significantly after the direct stimulation of fusion and stereopsis with dichoptic therapy (Fig. 8). Before treatment, 12.5% patients had suppression of the amblyopic eye, 31.2% subjects had simultaneous vision, and 56.2% subjects showed stereoacuity between 40 and 2000 arc sec. After 1 month of treatment, there was a significant improvement (p = 0.002), and as consequence, 81.2% subjects presented stereoacuity, 18.7% subjects had simultaneous vision, and any subject had suppression. At 3 and 6 months of follow-up, results were stable and there were no significant changes in BF.

Fig. 8

Changes in binocular function of patients in the VT group during follow-up. Percentage of patients with suppression (red), simultaneous vision (yellow), and stereopsis (green) from baseline to 6-month follow-up. After 1 month of treatment, no patients had suppression and more patients reached stereopsis

Compliance and duration of the treatment with VT

The treatment varied among subjects, performing all subjects at least 3 months of VT, with 75% of patients reaching a VA of 0.1 logMAR or better in this period. The mean time of VT at home performed was 7.7 ± 2.8 h in the first month, 7.0 ± 3.1 h in the second month, and 5.8 ± 3.5 h in the third month. Consequently, comparing the VT prescribed with the VT performed at home, compliance was 72.4% in the first month, 68.9% in the second month, and 56.8% in the third month. Compliance was poorly correlated with the VA at 3 months in the VT group (r = 0.383, p = 0.036).

Correlations

VA showed no significant correlation with CS in VT group (p > 0.05). In VT group, baseline VA was strongly correlated with the total VA improvement (r =  − 0.770, p < 0.001), and VA at 1 month (r = 0.625, p < 0.001), poorly correlated with VA at 3 months (r = 0.386, p = 0.032) and no correlated with VA at 6 months (r = 0.107, p = 0.654). In the patching group, baseline VA was strongly correlated with the total VA improvement (r =  − 0.847, p < 0.001), but no with the VA reached during follow-up.

Baseline BF was significantly correlated with BF (r = 0.567, p = 0.001) and VA (r = 0.469, p = 0.007) at 1 month, BF (r = 0.586, p = 0.001) and VA (p = 0.574, p = 001) at 3 months, and BF at 6 months (r = 0.719, p < 0.001). Besides this, BF was also correlated with amblyopia type at baseline (r = 0.552, p = 0.001) and at 6 months (r = 0.563, p = 0.010). Age and gender were not found to be correlated with VA improvement in VT group. However, in patching group, there was a significant correlation between age and total VA improvement (r = 0.542, p = 0.037). In addition, previous treatment was correlated with total VA improvement in VT group (r = 0.512, p = 0.021), but not in the patching group (p > 0.05).

Multiple linear regression analysis

Multiple linear regression analysis confirmed that a prediction equation of the visual improvement achievable at 1 month with the combined treatment of VT and patching could be obtained (p < 0.001, R2 = 0.747, Durbin–Watson: 2.153):

ΔVA = 0.100 − 0.060*(Amblyopia Type: 1, anisometropic; 2, strabismic) + 0.732*(Baseline VA) − 0.28*(interocular difference VA).

The normality of the unstandardized residuals distribution (p = 0.131) and the absence of outliers confirmed the homoscedasticity of this model. Similarly, no multicollinearity was detected in the model (VIF between 1.114 and 1.560).

Similarly, a prediction equation of the visual improvement achievable at 3 months with the combined treatment was obtained (p < 0.001, R2 = 0.850, Durbin–Watson: 2.620):

At 3 months: ΔVA = 0.161 − 0.093*(Amblyopia Type: 1, anisometropic; 2, strabismic) + 0.925*(Baseline VA) − 0.045*(interocular difference VA).

The normality of the unstandardized residuals distribution (p = 0.200) and the absence of outliers confirmed the homoscedasticity of this model. Similarly, no multicollinearity was detected in the model (VIF between 1.105 and 1.589).