Early diagnosis in glaucoma

Introduction

The process of diagnosis involves the gathering of information about a patient, through historytaking, clinical examination, and diagnostic tests, to formulate the probability that a disease is either present or not present. At each stage of this process, a piece of information either increases or decreases the probability. This process is called Bayesian inference and clinicians use this in everyday practice, although usually subconsciously and not in a formalized, quantitative

Corresponding author. Tel.: +44 20 7566 2059; Fax: +44 20 7566 2059;

E-mail: david.garway-heath@moorfields.nhs.uk

manner — the clinician uses his or her experience to put pieces of information together to arrive at a conclusion.

Identification of the glaucomatous neuropathy, when the disease is moderately advanced, is not a difficult task. The typical features of diffuse with superimposed focal rim narrowing and retinal nerve fiber layer (RNFL) loss, and associated characteristic patterns of visual-field loss, are easily recognized. However, diagnosis of the early stages of glaucoma can be challenging, with equivocal features in the optic-nerve head (ONH), RNFL, or visual field. Additional diagnostic tests may aid the clinician in making a diagnosis of early glaucoma and much research effort has been directed to develop more sensitive tests to reliably identify early glaucomatous visual function loss

48

and imaging devices to identify the earliest signs of structural damage. Yet the role of these tests in the clinical routine is unclear. The main emphasis of this chapter will be on the interpretation and integration of test results in the diagnostic process.

To make appropriate use of diagnostic tests, and the results they produce, one should consider how early in the course of the disease it is necessary to make a definite diagnosis, and how much time and resources should be devoted to this task. The aim of glaucoma management is to prevent symptomatic vision loss during a patient’s lifetime. Patients at greatest risk of symptomatic vision loss are those in whom the diagnosis has been made late in the disease process (Chen, 2003; Forsman et al., 2007). As glaucoma is generally a slowly progressive disease, patients with very early or equivocal disease are not at great risk of symptomatic vision loss. Thus, in these cases of diagnostic uncertainty, an initial diagnosis of ‘‘suspect glaucoma’’ is probably sufficient. Such cases can be monitored (at a frequency dictated by the patients risk profile) and the diagnosis made when definite change (progression) is identified or the findings of the clinical examination or conventional tests become unequivocal.

The stage of disease that one aims to identify depends on the clinical situation. When screening, or case-finding, for glaucoma, the false–positive ‘‘identification’’ of cases needs to be limited to avoid over-burdening the health care system and causing unnecessary distress to individuals who are, in fact, healthy. The actual number of false positive cases identified is related to the prevalence of the condition (or prior probability that the disease is present) and the specificity (true negative, or 1 false positive rate) of the test being applied. The specificity of any test is lower (and the false positive rate is higher) when selecting a diagnostic criterion that is sensitive in earlier stages of disease. Therefore, one has to balance the stage that one aims to identify against the expected false–positive rate. When screening or case-finding in the general population, this may mean not attempting to identify the very earliest signs of glaucoma, but aiming to identify moderate (unequivocal) disease. However, in individuals already referred to secondary care, the probability

that glaucoma is present is already much higher than it is in the general population and the clinician can afford to aim to identify earlier stages of glaucoma without generating too many false– positive diagnoses.

History and examination

The purpose of history-taking is to establish the background probability that a particular condition is present. For glaucoma, this includes identifying the family history — glaucoma tends to run in families and a positive family history for glaucoma increases the probability that glaucoma is present. Other aspects of the patient history may raise the probability for glaucoma being present — age, ethnic background, myopic refractive error, previous ocular injury or surgery, previous steroid use, and other factors (Boland and Quigley, 2007). This background (or prior) probability becomes the context for the clinical examination, and the findings of the clinical examination are interpreted in the context of the patient history.

Important aspects of the clinical examination include intraocular pressure (IOP) measurement, gonioscopy, and ONH and RNFL examination.

The level of the IOP has a moderately strong effect on the probability that glaucoma is present, because although glaucoma can be present at all levels of IOP, the prevalence is much greater at higher IOP. This is illustrated by the findings of the Baltimore Eye Survey (Fig. 1). Similarly, the gonioscopy findings will influence the probability that glaucoma is present.

Clinical examination of the ONH and RNFL directly evaluates the eye for signs of glaucomatous damage. The signs, in the context of the history, may be so obvious that further tests are not needed to make the diagnosis. For instance, notching of the rim with adjacent wedge-shaped RNFL loss in an eye with an IOP of 28 mmHg is almost certain to have glaucoma. Additional testing is not required to make the diagnosis (although visual-field testing and imaging are required to document the extent of damage and to monitor for progression). However, lesser degrees of neural rim narrowing and uncertain