Optimization of biometry for intraocular lens implantation using the Zeiss IOLMaster

Purpose: To compare the accuracy of biometry using conventional A-scan ultrasonography and partial coherence interferometry, and to improve the accuracy of biometry by sequential audit of postoperative refractive error. Methods: The study was performed in three phases. In phase 1, 20 consecutive patients undergoing routine phaeoemulsification underwent biometry using both A-scan ultrasonography and the Zeiss IOLMaster (ZIOLM). A single experienced optometrist refracted all patients 2 weeks after surgery. The errors between expected and achieved refraction were calculated and compared between the two methods. In phases 2 and 3, a further 22 and 20 patients, respectively, were recruited and only the ZIOLM was used for biometry. The manufacturer's suggested A-constant was refined and the error between expected and achieved refraction was calculated. Results: In phase 1, the median unexpected error for the ZIOLM was + 0.63 (interquartile range + 0.368 to + 1.015) and for A-scan biometry was - 0.24 (interquartile range - 1.335 to + 0.802). In phase 1 65% of patients' postoperative refractions were found to be within 1.0 D of emmetropia using the ZIOLM (55% using A-scan biometry). Refinements to the A-constant improved this to 95% by phase 3. Conclusion: An error was identified in IOL power estimation with the ZIOLM, when using the manufacturer's recommended A-constant (recommended and previously optimized ultrasound A-constant 118.0; ZIOLM optimized A-constant 118.6). Serial modifications to the A-constant were successful in reducing the unexpected error to well within the tolerance limits of published international standards.