Per-operative retinoscopy as a predictor of final post-operative refraction.
PURPOSE: To assess the accuracy of streak retinoscopy performed at the end of cataract surgery as a predictor of final post-operative error. METHOD: Retinoscopy was performed on 68 patients as they lay on the operating table after routine cataract extraction and intraocular lens implantation. In each case the predicted post-operative refraction by biometry and the retinoscopy at the end of the operation were compared with the 6 week post-operative subjective refraction. RESULTS: The retinoscopy had a mean difference of 0.6 D (standard deviation of 0.5 D). The post-operative refraction predicted by biometric measurements had a mean difference of 1.6 D (standard deviation 0.6 D). When corrected for systematic error, 8% of patients were found to have an error of greater than 2 D as predicted by pre-operative biometry. Prediction by retinoscopy made no error greater than 2 D. The accuracy in the retinoscopic prediction of post-operative refraction was significantly better than the biometry using the F-test (p = 0.001). CONCLUSION: Retinoscopy at the end of cataract surgery may be a valuable tool to alert the surgeon to an unexpected refractive error. This would enable immediate intraocular lens exchange, if required.