Comparison of a ray-tracing refractometer, autorefractor, and computerized videokeratography in measuring pseudophakic eyes.

Purpose: To test the accuracy and reliability of the first-generation prototype of Tracey-1, a ray-tracing refractometer.
Setting: Cullen Eye Institute, Baylor College of Medicine, Houston, Texas, USA.
Methods: The refractive spherical equivalent (SE) and astigmatism measured by Tracey-1, subjective manifest refraction (MR), and autorefraction (AR) (Nidek AR-3300) were compared in 58 pseudophakic eyes of 44 patients with a mean age of 69.95 years +/- 8.6 (SD). The astigmatic values obtained from Tracey-1 and computerized videokeratography were also compared. Astigmatism values were compared using power vector analysis.
Results: The mean SEs measured by MR, Tracey, and AR were -0.76 diopter (D) +/- 0.96 (SD) (range 1.00 to -3.00 D), 0.37 +/- 0.64 D (range 1.68 to -1.49 D), and -0.44 +/- 1.13 D (range 1.63 to -3.50 D), respectively. The Pearson correlation coefficients between MR and Tracey values and between MR and AR values were high (r = 0.895 and r = 0.875, respectively; both P < .0001). However, the mean difference between the Tracey and MR values was 1.13 +/- 0.48 D (range 2.45 to -0.06 D) (P < .001). The mean cylindrical differences between MR and Tracey and between MR and AR were 0.03 +/- 0.51 D (range 1.22 to -1.07 D) (P = .709) and -0.30 +/- 0.62 D (range 0.50 to -2.25 D) (P = .003), respectively. Power vector analysis revealed difference vectors of -0.16 x 60 between Tracey and MR and -0.22 x 87 between AR and MR.
Conclusions: Spherical and cylindrical refractive data obtained with the Tracey system correlated well with those derived from MR; however, there was a mean spherical error of approximately 1.10 D. Further work is required to refine the accuracy and range of this device.