Keratometry obtained by corneal mapping versus the IOLMaster in the prediction of postoperative refraction in routine cataract surgery

Le résumé fourni par la source

BACKGROUND: To establish whether simulated keratometry values obtained by corneal mapping (videokeratography) would provide a superior refractive outcome to those obtained by Zeiss IOLMaster (partial coherence interferometry) in routine cataract surgery. DESIGN: Prospective, non-randomized, single-surgeon study set at the The Royal United Hospital, Bath, UK, District General Hospital. PARTICIPANTS: Thirty-three patients undergoing routine cataract surgery in the absence of significant ocular comorbidity. METHODS: Conventional biometry was recorded using the Zeiss IOLMaster. Postoperative refraction was calculated using the SRK/T formula and the most appropriate power of lens implanted. Preoperative keratometry values were also obtained using Humphrey Instruments Atlas Version A6 corneal mapping. MAIN OUTCOME MEASURES: Achieved refraction was compared with predicted refraction for the two methods of keratometry after the A-constants were optimized to obtain a mean arithmetic error of zero dioptres for each device. RESULTS: The mean absolute prediction error was 0.39 dioptres (standard deviation 0.29) for IOLMaster and 0.48 dioptres (standard deviation 0.31) for corneal mapping (P = 0.0015). Keratometry readings between the devices were highly correlated by Spearman correlation (0.97). The Bland-Altman plot demonstrated close agreement between keratometers, with a bias of 0.0079 dioptres and 95% limits of agreement of -0.48-0.49 dioptres. CONCLUSIONS: The IOLMaster was superior to Humphrey Atlas A6 corneal mapping in the prediction of postoperative refraction. This difference could not have been predicted from the keratometry readings alone. When comparing biometry devices, close agreement between readings should not be considered a substitute for actual postoperative refraction data.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.