Comparison of keratometry and videokeratography after penetrating keratoplasty.

BACKGROUND As new methods for corneal curvature measurement have evolved, users of videokeratscopes need to know the practical limitations of these instruments. We assessed agreement between keratometry and videokeratography in measuring highly astigmatic corneas. METHODS Two independent examiners made three keratometric and videokeratographic measurements on each of 33 corneas after penetrating keratoplasty. The non-orthogonal keratometric readings obtained with a Zeiss 10 SL/O keratometer (Carl Zeiss Ltd.) were compared to the non-orthogonal simK readings (maxK, minK) calculated by the algorithms of a TMS-1 videokeratoscope (Tomey). Measurement agreement was evaluated for steep and flat meridian power and location, and astigmatism magnitude (D). RESULTS A systematic bias of the TMS-1 in measuring steeper than keratometry for the steep meridian was demonstrated (95% confidence interval: -0.34 to -1.20 D). The limits of agreement (d - 2SD to d + 2SD) between the two instruments were found to be unacceptable for clinical purposes in measuring steep meridian power (-3.17 to +1.63 D), flat meridian power (-4.92 to +4.48 D) and astigmatism magnitude (-5.84 to +4.87 D). Clinically acceptable differences were observed in identification of steep and flat meridian location. CONCLUSIONS The Zeiss 10 SL/O keratometer and the TMS-1 videokeratoscope showed poor measurement agreement for irregular corneal surfaces, despite the good correlation previously shown between keratometry and videokeratography in calibrated spheres and regular corneas. The TMS-1 showed a systematic bias, measuring a greater power in the steeper meridian than the Zeiss 10 SL/O keratometer. It is suggested that the two instruments cannot be used interchangeably in comparing the curvature of corneas after penetrating keratoplasty.