Article Text


Biometric aspects and comparison with published papers
  1. Turnstrasse 24, 66953 Pimasens, Germany

    Statistics from

    Editor,—In their discussion on central corneal thickness determined with optical coherence tomography in glaucoma in the BJO, Bechmann et al 1 mention the results of Ehlerset al.2 and compare them with the results of Whitacre et al 3without regarding generally accepted principles of interpretation.4 Some biometrical considerations will be found in the following.

    In the paper by Ehlers et al in figure 4 the correlation coefficient between the correction value and corneal thickness is 0.768 at n = 29. In the comparable figure 2 of Whitacreet al no correlation coefficient is given at n = 15. This coefficient was calculated by us after digitalising the data points. It equals 0.51. According to Klemm,4 (page 97) the estimate of regression is extremely unreliable and thus useless at r <‖0.6‖. The data of Ehlers et al, therefore, are much more convincing than the data of Whitacre et al.This fact does not reduce the merit of Whitacre, who brought the problem of corneal influence on tonometry to our notice.

    It escaped the attention of Bechmann et althat figure 4 of Ehlers et al 2and figure 2 of Whitacre et al 3differ fundamentally from figure 2 in the paper by Wolfset al.5 Furthermore, regarding the results of the Rotterdam study, the ordinate of figure 4 of Ehlerset al shows the correction value according to corneal thickness, and in figure 2 of Whitacreet al the ordinate shows the measurement error according to corneal thickness. These two ordinates (Ehlerset al and Whitacre et al) differ by sign and show the result of subtraction of intracamerally measured IOP and applanation tonometry values. The ordinate in figure 2 of the Rotterdam study, however, shows the results of applanation tonometry. This is a fundamental difference that absolutely forbids a comparison. The Rotterdam study does not provide a correlation coefficient of the data shown in figure 2, which may be interpreted as a cloud of points. We have similar data and have calculated the coefficient of correlation r= 0.17. Therefore, in this case it may be concluded that the estimate of regression is playing with figures only (Klemm,4 page 97).

    In summary, the data of Ehlers et alcurrently show the association of measurement error and corneal thickness in the most convincing way. Although Bechmannet al have (erroneously) seen a small influence of central corneal thickness in IOP measurement in the literature they attribute an important part to corneal thickness in the diagnosis and understanding of various types of glaucoma. It can be concluded from the context that the authors treat corneal thickness as a new quantity in the diagnosis of glaucoma, comparable with optic disc parameters. They have nicely shown different values of corneal thickness in the various types of glaucoma. However, they do not believe that corneal thickness influences applanation tonometry. Therefore, they have to explain their findings in a more complicated way. The psychologist and philosopher Watzlawick6 (page 67) states that we prefer declaring undeniable facts (which are inconsistent with our explanation) to be untrue or unreal instead of fitting our explanation to these facts. The application of biometric knowledge in judging the data of Whitacre et al 3 and a reinterpretation of figure 2 of the Rotterdam study5 may fit the opinion of the authors to the most likely explanation7-9 that corneal thickness influences the results in applanation tonometry to a clinically relevant degree, and that recommends the application of OCT in the diagnosis of glaucoma if available.


    View Abstract

    Request permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.