Outcomes for second-eye IOLs may improve by considering first-eye prediction errors

The authors investigated the relationship between first- and second-eye prediction errors (PEs) among patients who underwent bilateral sequential phacoemulsification cataract surgery with the same IOL model. They found, by studying a database of 2,129 patients, that a 50 percent correction factor (CF) reduces second-eye PE when either first-eye optimized PE is within ±1.50 D or when nonoptimized IOL constants are used. The correlation is weaker when interocular corneal power differences are greater than 0.60 D.

The authors say the results provide strong statistical evidence to justify the use of a 50 percent correction factor in routine clinical practice for bilateral use of the Bausch & Lomb LI61AO Sofport IOL model, which was used in the study, when the arrangements for follow-up and scheduling of second-eye surgery allow this to take place without undue excess cost. However, the study’s findings do not support the practice of simultaneous bilateral cataract surgery for patient convenience because refractive outcomes can be improved for the second eye if the interval between surgeries is delayed for at least 30 days so that the refractive outcome of first-eye surgery can stabilize.

The authors used data on consecutive cataract operations performed at one National Health Service hospital in the United Kingdom during a 56-month period. They analyzed retrospectively calculated PEs (Hoffer Q, Holladay 1, SRK/T) for any association between paired eyes. A range of CFs derived from the first-eye PE were applied to the second-eye PE using optimized and nonoptimized IOL constants.

They found that prediction errors of paired eyes were correlated. Interocular corneal power differences exceeding 0.60 D were associated with a weaker correlation but interocular axial length differences did not affect the correlation. Applying a 50 percent CF to second eyes of patients with a first-eye PE between ±0.50 and ±1.50 D improved refractive outcomes from 30 percent, 56 percent and 92 percent to 42 percent, 75 percent and 96 percent within ±0.25 D, ±0.50 D and ±1.00 D, respectively, and reduced the mean absolute error (MAE) from 0.49 to 0.37 D (P < 0.0001). For first-eye PE below ±0.50 D, a 50 percent CF reduced the MAE from 0.32 to 0.30 D (P < 0.00001). A 50 percent CF also reduced second-eye MAE for eyes with nonoptimized IOL constants.

These results demonstrate that refractive outcomes after cataract surgery for each eye of a patient are associated. The study also demonstrates with very high statistical confidence that a 50 percent correction brings about significant improvements in MAE for the second eye and the improvement in theoretical PE reduction is proportional to the first-eye PE. For second eyes whose first eyes had PEs of more than ±0.50 D, the improvements using a 50 percent correction factor were statistically significant and the magnitude of improvement also seemed to be clinically significant. For second eyes whose first eyes had PEs of ±0.50 D or less and ±0.25 D or less, the improvement in MAE with a 50 percent correction factor was significant, but the magnitude of improvement was less compared with higher PE groups.