Variability in Refractive Measurements in Children with Autism Spectrum Disorder Under Cycloplegic and Non-Cycloplegic Conditions: Influence of Measurement Method

Purpose: To compare refractive measurements obtained with Retinomax K-plus 3 and Plusoptix A09 V in children with autism spectrum disorder (ASD) before and after cycloplegia and to assess inter-device agreement. Methods: This retrospective pilot study included 23 children with ASD (mean age 6.54 ± 2.83 years), of whom 14 underwent examination with two autorefractometers (Retinomax K-plus 3 and Plusoptix A09 V) before and after cycloplegia induced with 1% tropicamide. Direct inter-device comparisons were restricted to the 14 children with paired measurements and were performed using the Wilcoxon signed-rank test, with matched-pairs rank-biserial correlations reported as effect sizes. Agreement in spherical equivalent (SE) was assessed using Bland–Altman analysis. Cycloplegia-induced refractive shifts were compared between devices. Secondary comparisons with controls without documented neurodevelopmental disorders (n = 27) were restricted to Retinomax SE measurements and performed using analysis of covariance (ANCOVA), with age included as a continuous covariate. Results: Before cycloplegia, Retinomax showed significantly more myopic SE values than Plusoptix in both eyes (right eye [OD]: −1.21 ± 1.25 vs. 0.91 ± 0.67 D; left eye [OS]: −1.22 ± 1.15 vs. 0.77 ± 0.51 D; both p < 0.00417). Significant inter-device differences in SE persisted after cycloplegia (OD: 0.63 ± 0.94 vs. 2.00 ± 0.63 D; OS: 0.46 ± 0.96 vs. 1.98 ± 0.63 D; both p < 0.001). Bland–Altman analysis showed a systematic negative bias of Retinomax relative to Plusoptix, which decreased after cycloplegia but persisted (before: −2.12 D for OD and −1.99 D for OS; after: −1.37 D and −1.52 D, respectively). Cycloplegia-induced shifts were greater with Retinomax; however, the between-device differences were not statistically significant after Bonferroni correction. In secondary analyses, age-adjusted Retinomax SE remained significantly lower in the ASD group than in controls without documented neurodevelopmental disorders before and after cycloplegia (all p < 0.001). Conclusions: In this exploratory cohort of children with ASD, Retinomax measurements were associated with more myopic SE values than Plusoptix, particularly before cycloplegia. The inter-device bias decreased but persisted after cycloplegia. This reduction is compatible with a contribution of accommodation; however, the persistence of the bias after cycloplegia indicates that device-specific measurement characteristics should also be considered. Refractive measurements obtained with the two autorefractometers should therefore not be assumed to be directly interchangeable. These findings underscore the importance of cycloplegic control and careful interpretation of autorefractor measurements in children with ASD. Larger prospective studies are warranted.