Optimizing precision in toric lens selection by combining keratometry techniques

To increase precision in toric intraocular lens selection by reducing the frequency of outliers that arise from technology-dependent variability during the preoperative assessment for routine cataract surgery.

Mean preoperative values for absolute sphere, amount of astigmatism, and steepest cylindrical axis were obtained for 87 eyes (54 patients) each with a manual keratometer and four automated keratometers. The mean sphere, amount of astigmatism, and steepest cylindrical axis across five technologies for each eye were defined as the meld sphere, meld astigmatism, and meld axis, respectively. Each technology was evaluated against the meld by Bland–Altman analysis, Student’s paired t test, and correlation coefficients. Further comparison between individual technologies and the meld quantified the number of outlier measurements each technology produced.

The number of outliers between individual keratometers and the meld differed with specific measurement of sphere, axis, or amount of astigmatism. Although statistical analysis using Bland–Altman plots, correlation coefficients, and paired t tests suggested insignificant difference from meld measurements for each parameter, precision-guided analysis presented more clinically significant outliers. The number of outliers can be reduced for sphere (range: 2%-46% to 1%-6%), astigmatism (range: 6%-23% to 0%-2%), and axis outliers (range: 15%-27% to 3%-6%) by averaging measurements from automated and manual keratometers.

Although multiple keratometry technologies produced similar, average measurements, the authors found a disturbing number of outliers that may be overlooked when employing a single technology. Measurement errors can be dramatically reduced by averaging measurements from manual keratometry with any automated technology to make toric lens selection more precise.