ObjectiveTo identify factors associated with visual acuity regression following successful treatment of anisometropic amblyopia.Design and methodThis was a retrospective cohort study. Database records for 100 and 61 children with anisometropic amblyopia who met at least one criterion for successful treatment proposed by the Pediatric Eye Disease Investigator Group (PEDIG) and had at least 1 year of follow-up data available after the criterion was met were analyzed. The study sample was split into two groups, those who regressed within 1 year of successful treatment (no longer met any of the PEDIG criteria for successful treatment) and those who did not. A two-step analysis involving a least absolute shrinkage and selection operator (LASSO) regression and a logistic regression were used to identify predictor variables for increased risk of regression. A broad range of clinical, perceptual, and demographic variables were included in the analyses.ResultsSixty-eight (42.5%) children regressed within 1 year of successful treatment. Among the 27 predictor variables considered within the statistical modeling process, the three most important for predicting treatment regression were the extent of amblyopic eye visual acuity improvement, age at first hospital visit and sex. Specifically, lower risk of regression was associated with larger amblyopic eye visual acuity improvement with treatment, younger age at initiation of treatment and female sex.ConclusionPatients who received treatment at a younger age and responded well to treatment had a lower risk of treatment regression. This pattern of results suggests that early detection of amblyopia and strategies that enhance treatment adherence may reduce the risk of treatment regression. The higher risk of regression in boys than girls that we observed may reflect known sex differences in brain development and /or sex differences in environment within our sample of children from South China.