Variability in the Calibration Error of the Goldmann Applanation Tonometer

Abstract

To investigate the variability in the measured calibration error with continued usage of Goldmann applanation tonometers (GATs) having unacceptable calibration error. The study included 132 slit-lamp mounted Goldmann tonometers (Model AT 900 C/M; Haag-Streit, Switzerland). A single observer twice checked a randomly selected set of 25 instruments on 2 consecutive days to determine the intraobserver agreement in the measurement of GAT calibration error. The same observer prospectively checked all the instruments between 8 and 9 AM on any given day at all testing levels namely 0, 20, and 60 mm Hg and rechecked the faulty instruments (calibration error more than ± 2 mm Hg at any testing level) 2 times more on the same day between 12 noon and 1 PM and 4 and 5 PM. The single measures intraclass correlation coefficients for the intraobserver agreement at the 20 mm Hg testing level were 0.78 for positive error and 0.83 for negative error. Twenty-eight (21%) instruments were faulty at any testing level. Nineteen (14%) were faulty at the clinically most important 20 mm Hg testing level. The maximum observed variability in the positive and negative calibration error at any testing level was +4 and -23 mm Hg, respectively. Fifteen (53%) faulty instruments had high variability (≥ 2 mm Hg) in the calibration error at any testing level. The calibration error of faulty GATs can frequently have a high variability. One should avoid estimating the true intraocular pressure from a faulty GAT by instinct.