Discriminant analysis of congenital esotropia surgery. Predictor variables for short- and long-term outcomes.

In a retrospective study, a computer-based stepwise discriminant analysis was used to create a biostatistical model of the results of surgery in congenital esotropia. One hundred fifty-seven patients who had bilateral medial rectus muscle recession as an initial operative procedure with at least 6 months postoperative follow-up were studied. The outcome groups for the discriminant analysis were based on the ocular deviation 6 weeks after the first surgery and at the patient's last visit, or the presence or absence of stereopsis at the last visit. With respect to the two predominant ocular deviation groups (Esotropia or Success) 6 weeks after surgery, 11 independent variables were analyzed; those determined to be prognostic of the outcome grouping were preoperative deviation, millimeters of medial rectus muscle recession, and interval between the onset of strabismus and surgery. With respect to ocular deviation at last visit, 105 patients were followed for over 2 years and were last examined when over 5 years of age; of the 12 independent variables analyzed, only postoperative deviation at 6 weeks was predictive. With respect to stereopsis, 94 patients had similar follow-up; of the 12 independent variables analyzed, preoperative deviation and age at surgery were prognostic. The results indicate that a smaller preoperative deviation and earlier surgical intervention increase the probability of achieving some degree of stereopsis in patients with congenital esotropia.