Saturday, November 7, 2026

1:30 PM - 1:45 PM

Ballroom 1 and 2

---

Overview

Rani Bahho

---

Details

Purpose: To determine the most accurate definition of pre-myopia for predicting future myopia and explore whether definitions should be age and ethnicity-specific. Method: Children aged 6 and 12 years (n=4118) from the Sydney Myopia Study were followed up after 5-6 years (n=2103, 51%). Refractive error was obtained by cycloplegic autorefraction (cyclopentolate 1%) with pre-myopia defined as a right eye SER of ≥-0.49 to ≤+0.75D (International Myopia Institute [IMI] definition), and myopia as ≤-0.50D. Using various thresholds (+0.25 to +1.50D), prediction of future myopia in pre-myopic children was explored using OR, ROC analysis, and Youden's index. Stratification by ethnicity allowed for comparison between European Caucasian (EC) and East Asian (EA) children. Results: Pre-myopia prevalence was 17.6% at 6 and 45.4% at 12 years of age. At follow-up, 36.7% (OR 13.4; p<.001) of pre-myopic children in the younger and 15.7% (OR 82.7; p<.001) in the older cohort had developed myopia. The IMI definition had an AUC of 0.73 and 0.74 in the younger and older cohort, respectively (p<.001). Accuracy was highest at ≤+1.00D (AUC 0.74) in the younger and ≤+0.25D (AUC 0.83) in the older cohort (p<.001). Pre-myopia of ≤+0.25D remained most accurate in older EC and EA children (p <.001). In younger children, a +1.00D threshold was most appropriate for EC (AUC 0.78, p<.001), and +0.75D for EA children (AUC=0.62, p=.033). Conclusion: The accuracy of myopia prediction in pre-myopic children was increased by using age-specific definitions, with a more hyperopic upper limit in younger children and a lower limit for older children.

---

Speaker