Age‐Related Clinical Agreement Between Noncycloplegic Autorefraction and Subjective Refraction: A Power‐Vector Analysis

Purpose: To assess the agreement between noncycloplegic autorefraction and subjective refraction across age groups and refractive error types in a large real-world clinical study using clinically meaningful thresholds for spherical and astigmatic components. Methods: This retrospective study analyzed 827 eyes from patients aged 7-85 years examined between May 2023 and January 2025. Objective refraction was measured using a KR-1 autorefractor (Topcon, Tokyo, Japan), followed by standard subjective refinement. Spherical equivalent (SE), cylinder power (CP), and power vectors (J0, J45) were calculated using Fourier vector analysis. Agreement was evaluated using absolute differences and predefined clinical criteria (SE ± 0.50 D; CP ± 0.25 D; axis ≤ 10° when CP ≥ 0.50 D). Comparisons were performed across three age groups (7-21, 22-49, ≥ 50 years) and refractive subgroups. Results: SE agreement was high across all ages, with 90.8% of eyes within ±0.50 D. J0 demonstrated a progressive shift from with-the-rule to against-the-rule astigmatism with age, whereas J45 remained near zero. CP differences were generally small but slightly larger in older adults. Axis deviations > 10° occurred in 27.2% of eyes, predominantly when CP was low (mean 0.51 ± 0.42 D), suggesting limited visual impact in most cases. Conclusions: Noncycloplegic autorefraction has been demonstrated to exhibit a high degree of agreement with subjective refraction for SE in both adolescents and adults. The astigmatic shifts that occurred with age were minimal and had only a limited effect on clinical outcomes. Subjective refinement remains a critical component in the precise determination of axes, particularly in cases of hyperopia, mixed astigmatism, and in eyes of advanced age. Future studies that include hyperopic children and young patients requiring cycloplegia would be valuable in extending these findings to a broader range of pediatric populations.