A large retrospective database analysis comparing outcomes of intraoperative aberrometry with conventional...

To evaluate differences between the absolute prediction error using an intraoperative aberrometry (IA) device for intraocular lens (IOL) power determination versus the error that would have resulted if the surgeon's preoperative plan had been followed.

Multiple centers in the United States.

Retrospective analysis of data collected using an IA device.

The database information was limited according to predetermined inclusion/exclusion criteria. Primary endpoints included the difference between mean and median absolute prediction error with IA use versus preoperative calculation, and the percentage of cases were compared when the prediction error was 0.5 diopters (D) or less.

A total of 32 189 eyes were analyzed. The IA mean absolute prediction error was lower than the preoperative calculation, 0.30 D ± 0.26 (SD) versus 0.36 ± 0.32 D (P < .0001). The aberrometry absolute median prediction error was lower than the preoperative calculation, 0.24 D versus 0.29 D (P < .0001). There was a significantly greater percentage of eyes with an aberrometry absolute prediction error of 0.5 D or less than eyes with a preoperative absolute prediction error of 0.5 D or less (26 357 [81.9%] of 32 189 vs. 24 437 [75.9%] of 32 189, P < .0001). In addition, in those cases in which power of the IOL implanted was different than the preoperatively planned IOL power, significantly more eyes had an aberrometry absolute prediction error of 0.5 D or less (10 385 [81.3%] of 12 779 vs. 8794 [68.8%] of 12 779, P < .0001).

In a database of more than 30 000 eyes, calculations incorporating IA outperformed preoperative calculations. The difference was more pronounced in those cases in which the preoperatively planned IOL power was different than the power of the IOL implanted.