Interocular Axial Length and Corneal Power Differences as Predictors of Postoperative Refractive Outcomes after...

To determine whether differences between eyes in axial length (AL) and corneal power (K) on optical biometry are predictive of refractive outcomes.

Retrospective cohort study.

A total of 729 patients (1458 eyes) who underwent bilateral phacoemulsification at TLC (Mississauga, Ontario, Canada) from September 2013 to August 2015.

We compared the proportion of patients having >0.5 diopters (D) of refractive error from target stratified by interocular axial length differences (IALDs) and interocular K differences (IKDs) between eyes as measured by optical biometry (IOL-Master, Carl Zeiss Meditec, Oberkochen, Germany). Analysis was repeated for 0.25 D or 1.0 D targets and for patients with uncorrected visual acuity (UCVA) >0.3 logarithm of the minimum angle of resolution (logMAR) postoperatively.

Proportions, odds ratios (ORs), and corresponding 95% confidence intervals (CIs) were computed using generalized estimating equations to account for within-patient correlation.

Some 79.1% of eyes were ≤0.5 D of refractive target, 47.0% were ≤0.25 D, and 97.2% were ≤1.0 D. The OR of having a refractive outcome >0.5 D from target for IALD cutoff of 0.2 mm was 1.4 (1.1-1.8), of 0.3 mm was 1.6 (1.2-2.1), and of 0.4 mm was 1.8 (1.3-2.5). This translates to 70.0% (63.5-75.7) within target for IALD of ≥0.4 mm versus 80.7% (78.4-82.9) for <0.4 mm. For a given patient with IALD, the chance of being off target was similar for the shorter and longer eye. Eyes outside of target were twice as likely to be <-0.5 D than >0.5 D. Interocular K difference was largely not associated with prediction error, yet larger IKD-flat, steep, and average were associated with increased odds of UCVA >0.3 logMAR postoperatively.

Interocular axial length difference of as little as ≥0.2 mm is associated with a higher chance of >0.5 D of refractive error from target and worse UCVA. Interocular K difference was not associated with worse refractive error from target, although a difference of ≥0.4 D was associated with worse UCVA. These cutoffs should be considered in preoperative planning and discussion with patients. Future study is required to assess whether repeating measurements, using adjunctive measurement devices, or attempting to separate true differences from artifact based on preoperative refractive characteristics reduces residual refractive error.