Use of fellow eye data in the calculation of intraocular lens power for the second eye

Purpose: To test the hypothesis that the refractive outcome of the first eye can be used to improve the intraocular lens (IOL) power calculation for the second eye.

Design: Retrospective, consecutive case series in a university clinic setting.

Participants: One thousand two hundred thirty-five patients who underwent sequential bilateral IOL implantation using the same type of IOL for both eyes.

Methods: The refractive outcome was analyzed in retrospect and the IOL power calculation was reviewed using 3 different methods: (1) the old Sanders-Retzlaff-Kraff (SRK) II formula, (2) the newer SRK/T formula, and (3) the more recent Olsen formula. In a subgroup of 232 eyes, the actual postoperative anterior chamber depth (ACD) was measured using the Lenstar LS900 laser biometer (Haag-Streit AG, Koeniz, Switzerland).

Main outcome measures: The error in diopters (D) of the predicted refraction in the spectacle plane.

Results: The correlation coefficients between the prediction errors for the right and left eyes were 0.56, 0.38, and 0.27 for the SRK II, SRK/T, and Olsen formula, respectively (P<0.001). Based on the observed correlation, the prediction of the second eye could be corrected according to the regression formula: Rx(cor) = Rx(exp)+β×Px(err), where Rx(cor) and Rx(exp) are the corrected and the uncorrected refractive prediction, respectively, Px(err) is the observed error of the first eye, and β is a formula-specific regression coefficient. Using this formula, the mean absolute error (MAE) was found to decrease from 0.56 D to 0.46 D, from 0.47 D to 0.41 D, and from 0.42 D to 0.40 D with the SRK II (P<0.001), the SRK/T (P<0.001), and the Olsen formula (P<0.05), respectively. Substituting the measured postoperative ACD of the first eye for the estimated ACD for the second eye in the Olsen formula was found to have a similar accuracy as the empirical regression formula.

Conclusions: Depending on the IOL power calculation formula, the prediction error of the first eye may be used to improve the prediction for the second eye. The reason for this effect seems to be the formula-dependent inaccuracy in the prediction of the IOL position (ACD) after surgery.