[Reproducibility of the results of two non-contact tonometers. Comparison with the Goldmann applanation tonometer].

Two non-contact tonometers (Keeler "Pulsair" and Reichert NCT II) were compared regarding the reproducibility of their readings mutually and with referring literature about Goldmann applanation tonometry. Differences of maximum 2 mmHg were found using Pulsair tonometer/NCT II in 61%/83% for two subsequent readings, in 73%/89% for subsequent mean values of 3, in 86%/95% for subsequent mean values of 6 and in 97% each for subsequent mean values of 9 readings. The confidence intervals of 90% were +/- 5.87 mmHg/+/- 3.38 mmHg for subsequent readings, +/- 3.75 mmHg/+/- 2.29 mmHg for mean values of 3, +/- 2.63 mmHg/+/- 2.00 mmHg for mean values of 6 and +/- 2.13 mmHg/+/- 1.61 mmHg for mean values of 9 readings, signifying the high scatter of Pulsair measurements; reaching the NCT II's reproducibility, which is comparable to that of the Goldmann readings of an experienced observer, requires mean values of a higher number of Pulsair readings. The discreet pressure lowering effect occurring with repeated NCT II applications could not be observed during Pulsair measurements. Perhaps the Pulsair instrument might be interesting for scientific questions, but it is not recommendable for routine diagnostic.