AIM:To investigate the predictive role of exodrift amount on long-term postoperative eye position regression in children with intermittent exotropia.METHODS:Prospective cohort study.Children with intermittent exotropia who underwent surgical treatment at the hospital between September 2023 and July 2025 were included. Based on the eye position regression status at 6 mo postoperatively, patients were divided into a regression group and a non-regression group. Clinical data of two groups were compared. The relationship between exodrift and visual function was analyzed. Multivariate linear regression was used to examine the association between exodrift and visual function. Logistic multivariable analysis identified factors influencing long-term postoperative eye position regression. A regression discontinuity design(RDD)was employed to evaluate the validity and robustness of the results. Interval likelihood ratios were analyzed to assess the relationship between exodrift and long-term postoperative eye position regression. Receiver operating characteristic(ROC)curves were used to evaluate the predictive value of exodrift for long-term postoperative eye position regression. Restricted cubic splines(RCS)were applied to analyze the dose-response relationship between exodrift and the risk of long-term postoperative eye position regression.RESULTS:This study initially included 133 children with intermittent exotropia. Among them, 5 cases were lost to follow-up. Eventually, 128 cases completed the follow-up(43 cases in the regression group and 85 cases in the non-regression group). The age of the children with intermittent exotropia ranged from 2 to 15 y(mean 9.89±2.61 y), with 82 boys and 46 girls. The age of the regression group was 9.82±2.64 y(26 boys and 17 girls). The age of the non-regression group was 9.96±2.58 y(56 boys and 29 girls). The two groups showed significant differences in Titmus near stereopsis, strabismus type, postoperative day 1 eye position, and postoperative day 1 exodrift(all P<0.05). There was no significant correlation between the amount of exodrift 1 d after the surgery and visual function(all P>0.05). Multiple linear regression analysis showed that there was no statistical significance between visual function and the amount of exodrift 1 d after surgery(all P>0.05). Titmus near stereopsis, convergence insufficiency type of strabismus, inward deviation(esophoria)at 1 d postoperatively, and an increased exodrift at 1 d postoperatively were identified as risk factors for postoperative eye position regression(all P<0.05). When exodrift after surgery was 9 PD at 1 d postoperatively, the risk of long-term eye position regression after surgery significantly decreased(P<0.001). When exodrift of the eye after surgery ≥14 PD at 1 d postoperatively, the probability of long-term eye position regression was relatively highest. When the cut-off value was 9 PD, the AUC of the predicted postoperative 1-day exodrift for predicting the long-term postoperative eye position regression was 0.809(95%CI: 0.749-0.869), indicating a relatively high predictive value. As the amount of exodrift increases 1 d after the surgery, the risk of long-term eye position regression also increases.CONCLUSION:The amount of exodrift 1 d after surgery is arisk factor for the long-term eye position regression in children with intermittent exotropia. The two are in a dose-response relationship. When the amount of exodrift 1 d after surgery is ≥14 PD, the risk of regression is the highest. The optimal cut-off value of 9 PD has a high predictive value for long-term eye position regression. Clinically, the amount of exodrift 1 d after surgery can be monitored to assess the risk of regression 6 mo after surgery, providing a reference basis for the clinical formulation of individualized follow-up plans.