In a polycentric study, the influence of some of the most common varia
bles capable of influencing the reliability of prekeratotomy pachymetr
y has been evaluated. The fluctuation in a sequence of measurements fr
om the same point on the cornea, with the patient in different positio
ns (orthostatic and clinostatic), during progressive drying of the lac
rimal film and before and after corneal marking have been evaluated. T
hree operators, working randomized, used the same computerized pachyme
ter (DGH1000) to make the measurements. The data obtained in this way
were statistically analyzed by the Statgraphics personal computer prog
ram. Some interesting and useful considerations about the more correct
methods to run the test and its real accuracy have emerged.