Objective analysis of toric intraocular lens rotation and centration.

PURPOSE To assess the repeatability of an objective image-analysis technique to determine intraocular lens (IOL) rotation and centration. SETTING Six European ophthalmology clinics. METHODS Eyes with an Akreos AO aspheric IOL with orientation marks were imaged. A masked observer rated image quality. The axis of rotation was determined from a line bisecting the IOL orientation marks. This was normalized for rotation of the eye between visits using the axis bisecting 2 consistent conjunctival vessels or iris features. The center of ovals overlaid to circumscribe the IOL optic edge and the pupil or limbus were compared to determine IOL centration. Intrasession repeatability was assessed in a subgroup of eyes and the variability of repeated analysis examined. RESULTS The SD of intrasession repeatability was +/-0.79 degrees for IOL rotational stability, +/-0.10 mm for horizontal centration, and +/-0.10 mm for vertical centration. The SD of repeated-analysis variability of the same image was +/-0.70 degrees, +/-0.20 mm, and +/-0.31 mm, respectively. The mean eye rotation (absolute) between visits was 2.23 degrees +/- 1.84 (SD) (10% >5 degrees rotation) using 1 set of consistent conjunctival vessels or iris features and 2.03 +/- 1.66 degrees (7%>5 degrees rotation) using the mean of 2 sets (P = .13). Poorer image quality resulted in larger apparent absolute IOL rotation (r = -0.45, P<.001). CONCLUSIONS Objective analysis of digital retroillumination images allowed sensitive assessment of IOL rotation and centration stability. Eye rotation between images can lead to significant errors if not taken into account. Image quality is important to analysis accuracy.