Preoperative and Postoperative Clinical Factors in Predicting the Early Recurrence Risk of Intermittent Exotropia After Surgery

Abstract

To identify preoperative and postoperative early recurrence risk in intermittent exotropia (IXT) patients after surgery. Prospective clinical cohort study. We included 210 basic-type IXT patients who underwent either the bilateral rectus recession or unilateral recession and resection procedure and had complete follow-up until recurrence or for more than 24 months postoperatively. The primary outcome was early recurrence, defined as postoperative exodeviation over 11 prism diopters at any time beyond postoperative month 1 and within 24 months. Survival was estimated by the Kaplan-Meier method. Preoperative and postoperative clinical characteristics were collected from patients, and preoperative and postoperative Cox proportional hazards regression analyses were performed. Preoperative model was fit with 9 preoperative clinical factors (sex, onset age of exotropia, duration of disease, spherical equivalent of the more myopic eye, preoperative distant exodeviation, near stereoacuity, distant stereoacuity, near control, and distant control). Postoperative model was fit by adding 2 factors relevant to surgery (surgery type and immediate postoperative deviation). Corresponding nomograms were constructed and evaluated using the concordance indexes (C-indexes) and calibration curves. Decision curve analysis (DCA) was used to determine the clinical utility. The recurrence rate was 8.10% for 6 months, 11.90% for 12 months, 17.14% for 18 months, and 27.14% for 24 months after surgery. Younger age at onset, larger preoperative angle, and less immediate postoperative overcorrection were found to increase the risk for recurrence. Although onset age and age at surgery were strongly correlated in this study, age at surgery was not significantly associated with IXT recurrence. The C-indexes for the preoperative and postoperative nomograms were 0.66 (95% CI: 0.60-0.73) and 0.74 (95% CI: 0.68, 0.79), respectively. Calibration plots between predicted and actual observed 6-, 12-, 18-, and 24-month overall survival using the 2 nomograms revealed high consistency. The DCA indicated that both models yielded great clinical benefits. By relatively accurate weighing of each risk factor, the nomograms offer good prediction for early recurrence in IXT patients and may help clinicians and individual patients make appropriate intervention plans.