Abstract Background The purpose of this study was to characterize the longitudinal refractive outcomes of orbital decompression in thyroid eye disease (TED), compare refractive trajectories across decompression strategies, and identify the anatomical correlates and baseline predictors of postoperative refractive changes. Methods This retrospective longitudinal study included 57 eyes of 46 patients with moderate-to-severe TED who underwent orbital decompression. Comprehensive ophthalmic evaluations including eyelid position parameters, refractive and axial measurements, and corneal biomechanical and tomographic assessments were performed preoperatively and during postoperative follow-up. Postoperative observations were analyzed using generalized estimating equations (GEE) models with restricted cubic splines, with time treated as a continuous variable to allow flexible modeling of temporal patterns. Multivariate GEE models were used to compare decompression strategies, evaluate associations between postoperative ocular biometric changes and postoperative refractive changes, and identify baseline predictors of clinically significant postoperative myopic drift. Results Orbital decompression was associated with a biphasic pattern of refractive change characterized by an initial myopic shift, followed by partial hyperopic recovery over time ( P < 0.001), with the greatest myopic shift occurring at approximately 2–3 months postoperatively. The refractive trajectories differed significantly across decompression strategies. Compared with one-wall decompression, two-wall decompression demonstrated a greater magnitude and more rapid progression of refractive change, as well as an increased risk of clinically significant myopic drift, independent of decompression extent. Standardized GEE analyses identified postoperative changes in measured axial length as the strongest correlate of postoperative refractive change ( P < 0.001). Additional associations were observed for changes in exophthalmos, eyelid position, corneal biomechanics, and corneal tomographic parameters. Several baseline ocular biometric parameters were independently associated with susceptibility to clinically significant postoperative myopic drift. Conclusions Orbital decompression induces dynamic postoperative refractive changes that vary according to decompression strategy and individual ocular characteristics. Postoperative refractive changes appeared to peak at approximately 2–3 months and partially recovered thereafter, with refractive status gradually approaching relative stabilization over time, suggesting that definitive refractive correction may be better deferred until postoperative stabilization. These findings provide clinically relevant insights into postoperative refractive behavior and may support individualized surgical planning and perioperative management for TED.
Yu et al. (Wed,) studied this question.