Non-Cycloplegic Devices Fall Short in Detecting Hyperopia, Study Shows

A meta-analysis published in Advances in Ophthalmology Practice and Research shows that non-cycloplegic autorefraction and photoscreening significantly underdetect hyperopia, with mean differences of -0.65D to -0.78D compared with cycloplegic methods. The bias was greatest in younger children and with the use of photoscreeners. Despite high testability, authors warn that rapid methods cannot replace cycloplegic evaluation for identifying amblyogenic refractive errors and ensuring accurate pediatric vision screening. Photo: Getty Images. Click image to enlarge.Timely detection and correction of refractive errors in children are integral in the prevention of amblyopia and strabismus. Cycloplegic retinoscopy is widely accepted as the most accurate method of measuring refractive errors; however, technological innovations have introduced handheld autorefractors and photoscreeners to provide quicker results in school and community screening programs. Some studies have found that these rapid methods may underestimate hyperopia, and cycloplegia itself is often met with reluctance in older children and adults due to its side effects. The authors of a new study published in the journal Advances in Ophthalmology Practice and Research conducted a systematic review and meta-analysis comparing cycloplegic and non-cycloplegic methods to determine which provides the best results.Data was gathered from 24 studies across 22 countries. The findings showed that, compared with cycloplegic reference methods, non-cycloplegic autorefractors and photoscreeners consistently underestimated refractive error, showing pooled mean differences of -0.65D and -0.78D, respectively. On more specific comparisons, the authors found the following:In non-cycloplegic autorefractors vs. cycloplegic retinoscopy (10 studies), the pooled mean difference was -0.81D.In non-cycloplegic vs. cycloplegic autorefraction (11 studies), the overall mean difference was -0.58D. In children aged seven to 12 years, the pooled mean difference was -0.73D, whereas in adults over 20 years it was -0.41D, indicating that the bias decreases with age but persists into young adulthood.For non-cycloplegic versus cycloplegic photoscreeners (four studies), the pooled mean difference was -0.85D.The pooled mean difference was -0.28D, indicating a trend where non-cycloplegic retinoscopy yields slightly more myopic (or less hyperopic) values than cycloplegic retinoscopy, although this did not reach statistical significance.In the comparison between non-cycloplegic subjective refraction and cycloplegic autorefraction (six studies), the pooled mean difference was 0.10D, suggesting a slight yet statistically significant bias toward more hyperopic (or less myopic) values with non-cycloplegic subjective refraction.For cycloplegic autorefractor vs. subjective refraction (five studies), the pooled mean difference was -0.20D.Testability rates for modern autorefractors were consistently high (above 95%), which the authors wrote “underscores their value in field and community-based screening where cooperation is limited.” However, the recurring finding is the systematic underestimation of hyperopia, they continued. Mean differences between non-cycloplegic autorefraction and cycloplegic retinoscopy ranged from -0.30D to -1.50D, with the largest discrepancies in hyperopic children or when photoscreening devices were used. The systematic bias of non-cycloplegic autorefractors and photoscreeners can not be fully corrected by adjusting cutoffs or statistical models, the authors wrote in the journal.“Cycloplegia remains indispensable for accurate detection of hyperopia and prevention of missed amblyogenic errors, as reinforced by recent systematic reviews,” they noted. The authors continued, “Variability across studies reflects age, device, examiner experience and prevalence, but the underlying myopic bias is universal.”The researchers noted various limitations, including high heterogeneity due to age, refractive profiles, device technologies and cycloplegic protocols. “Most primary studies used observational designs, with risks of selection and reporting bias,” they wrote. “Lack of standardized definitions, masking and subgroup data (e.g., developmental delay, high astigmatism) further limit interpretation. Additionally, subgroup analyses by race or country could not be performed because most studies did not provide stratified data by ethnicity or geographic region.” The inclusion of some outdated devices also reduced generalizability.Future studies should give special attention to underrepresented populations, such as infants, individuals with strabismus/amblyopia and people of different ethnicities, the authors concluded. “Clinical protocols must ensure that all positive or borderline screening cases undergo cycloplegic confirmation, with programmatic pathways to guarantee follow-up,” they wrote. “Device selection should consider local needs and performance, while education is needed to ensure evidence-based use of refractive technologies in pediatric care.”Click here for the journal source. Roque A, Nunes AF, Nascimento H, Martinez-Perez C. Instrument-based, non-cycloplegic versus cycloplegic refraction in pediatric and young adult populations (≤25 years): A systematic review and meta-analysis. Advances in Ophthalmology Practice and Research. November 14, 2025. [Epub ahead of print]. This article was developed by the editorial staff in conjunction with experts in the field. In the process, AI may have been among the editorial tools used to meet the goals of human editors, who approved all content.