H J Simonsz1, M J C Eijkemans. 1. Department of Ophthalmology, Erasmus Medical Center, Rotterdam. simonsz@compuserve.com
Abstract
UNLABELLED: In the Early vs. Late Infantile Strabismus Surgery Study (ELISSS), 13.5% of children operated at 20 months vs. 3.9% of those operated at age 4 had gross binocular vision (Titmus Housefly). Reoperation rates were 28.7% in the former vs. 24.6% in the latter group and, although all were eligible for surgery at baseline at 11 SD 3.7 months, 8% in the early group vs. 20% in the late group were never operated, mostly because their angle decreased spontaneously. We assessed the predictive value of age, angle, and refraction in these matters. METHODS: The ELISSS reoperation rates were first compared with those found in nine series of consecutive cases in nine university clinics operated during one particular year, between 6 and 23 years previously. Logistic regression was used to estimate the effect of postoperative angle and clinic on the chance of reoperation. Secondly, a meta-regression analysis was done of these and other reported reoperation rates. The mean age at operation and the mean duration of follow-up were regressed on the logistically transformed reported reoperation rates. Finally, to estimate the chance of spontaneous decrease of the angle without surgery, a random-effects model was fitted on the 6-monthly orthoptic measurements of angle and refraction in the ELISSS that antedated surgery, loss to follow-up, or final examination. In the random-effects model (see online-only supplement link or visit, www.simonsz.net), for ELISSS patients the random effect was defined as the deviation of the average angle, the fixed effect. A vector was defined based on age and spherical equivalent of the patient. The variance around the prediction consisted of uncertainty in the estimations, random effects, and residuals. RESULTS: In the retrospective study, 204 patients who had been first operated between 6 and 23 years previously were eligible. A reoperation had been performed in 32 (19.3%) of the remaining 166 children who were 4.33 SD 1.35 years old at first surgery. The reoperation rate was 7.3% for those with a postoperative angle of -4° to +4° (N = 82), 25% for postoperative divergence > 5°, and 29% for postoperative convergence 10° to 14°. Strabismologists overestimated the reoperation rates at double. In the meta-regression analysis, 12 studies were included. Reoperation rates were between 60% and 80% for children first operated around age 1 and approximately 25% for children operated around age 4 (best fit: -0.221 Ln [age in months] + 1.1069; R(2) = 0.5725). Finally, in the predictions of random-effects model, a small angle at age 1 and hyperopia of approximately +4 increased the chance of spontaneous decrease of the angle into a microstrabismus. DISCUSSION: The benefit of early surgery for gross binocular vision is balanced by a higher reoperation rate and an occasional child being operated that would have had a spontaneous decrease into a microstrabismus without surgery. The fact that, in the ELISSS, hyperopia was associated with a decrease of the angle underscores the benefit of early refractive correction.
UNLABELLED: In the Early vs. Late Infantile Strabismus Surgery Study (ELISSS), 13.5% of children operated at 20 months vs. 3.9% of those operated at age 4 had gross binocular vision (Titmus Housefly). Reoperation rates were 28.7% in the former vs. 24.6% in the latter group and, although all were eligible for surgery at baseline at 11 SD 3.7 months, 8% in the early group vs. 20% in the late group were never operated, mostly because their angle decreased spontaneously. We assessed the predictive value of age, angle, and refraction in these matters. METHODS: The ELISSS reoperation rates were first compared with those found in nine series of consecutive cases in nine university clinics operated during one particular year, between 6 and 23 years previously. Logistic regression was used to estimate the effect of postoperative angle and clinic on the chance of reoperation. Secondly, a meta-regression analysis was done of these and other reported reoperation rates. The mean age at operation and the mean duration of follow-up were regressed on the logistically transformed reported reoperation rates. Finally, to estimate the chance of spontaneous decrease of the angle without surgery, a random-effects model was fitted on the 6-monthly orthoptic measurements of angle and refraction in the ELISSS that antedated surgery, loss to follow-up, or final examination. In the random-effects model (see online-only supplement link or visit, www.simonsz.net), for ELISSS patients the random effect was defined as the deviation of the average angle, the fixed effect. A vector was defined based on age and spherical equivalent of the patient. The variance around the prediction consisted of uncertainty in the estimations, random effects, and residuals. RESULTS: In the retrospective study, 204 patients who had been first operated between 6 and 23 years previously were eligible. A reoperation had been performed in 32 (19.3%) of the remaining 166 children who were 4.33 SD 1.35 years old at first surgery. The reoperation rate was 7.3% for those with a postoperative angle of -4° to +4° (N = 82), 25% for postoperative divergence > 5°, and 29% for postoperative convergence 10° to 14°. Strabismologists overestimated the reoperation rates at double. In the meta-regression analysis, 12 studies were included. Reoperation rates were between 60% and 80% for children first operated around age 1 and approximately 25% for children operated around age 4 (best fit: -0.221 Ln [age in months] + 1.1069; R(2) = 0.5725). Finally, in the predictions of random-effects model, a small angle at age 1 and hyperopia of approximately +4 increased the chance of spontaneous decrease of the angle into a microstrabismus. DISCUSSION: The benefit of early surgery for gross binocular vision is balanced by a higher reoperation rate and an occasional child being operated that would have had a spontaneous decrease into a microstrabismus without surgery. The fact that, in the ELISSS, hyperopia was associated with a decrease of the angle underscores the benefit of early refractive correction.