Statistics from Altmetric.com
Editor,—In his interesting article on the seven million people blind from glaucoma worldwide, Quigley highlights the need for better detection of cases with moderate and severe glaucoma damage and hence the greatest risk of blindness.1 In his editorial, Rait correctly states that the costs of standard methods of glaucoma screening, including computerised perimetry, are prohibitive in most developing countries so that glaucoma remains undiagnosed until very late in its course.2 A simple and inexpensive visual field test has been designed3specifically for the detection of moderate and severe glaucomatous visual field loss in situations precluding other methods, but Rait seems to dismiss this technique as ‘less reliable’, quoting an article4 that is not representative of the published literature.
The oculokinetic perimetry (OKP) glaucoma test, to which Rait referred, is a handheld card, with 26 numbered fixation targets arranged in a spiral around a central, black spot. The patient is asked to look at each number for 1 second and to state whether or not the spot is seen. If any number is associated with disappearance of the spot, the test result is abnormal. Wishart reported only a 60.5% sensitivity and 61.5% specificity when 56 patients were examined by a technician in hospital.4 In contrast, Greve and Chisholm in Canada reported 95% and 93% concordance rates in 584 sector for sector comparisons with the Friedmann analyser and 994 sector comparisons with the Humphrey analyser respectively.5 The 26 point OKP test took under 1 minute in normal eyes, which was about a third of the time taken with the other methods. Sponsel and colleagues from Prevent Blindness America reported that the 26 point test detected 92% of 48 patients with moderate to severe visual field loss, with a false positive rate of 3.6%.6 Statistically, these results were not significantly different from those obtained with the Henson visual field analyser. Christoffersen et al reported a 94% specificity, when 185 patients were examined by medical secretaries in a general practice in Norway, mentioning two cases detected by the secretaries with OKP but missed by general practitioners with the Bjerrum tangent screen.7
The OKP chart used in these studies had a single, black stimulus to make the cost low enough for worldwide distribution free of charge by a pharmaceutical company. This device (now called the ‘multifixation campimeter’) has been improved, now incorporating a disc, which is dialled by the examiner so as to make a 3 mm grey spot of the desired contrast (10% or 25%) appear briefly in a central window. Twenty points in the central 28° field are examined, with two extra points in the normal blind spot to check reliability. Compared with the version it supersedes, this test is easier (for both patient and examiner) and more ‘foolproof’, with interim results of an ongoing evaluation indicating greater sensitivity and specificity (C H Fenerty and B E Damato, unpublished data).
Independent evaluations of the new multifixation campimeter are required, including studies performed under normal working conditions.
Editor,—Damato and Fenerty are to be congratulated on providing a new and improved multifixation campimeter, ‘easier’ and ‘more foolproof’ than its predecessor. I have no doubt that 95%/93% sector concordance is possible with the original device1 but it can only detect moderate to severe visual field loss with any reliability.
I understand that the original Damato stimulus is roughly equivalent to a size III Humphrey stimulus of 19dB (Patella M, Zeiss–Humphrey Instruments, personal communication). This calculation is based on the contrast ratio between the white card and the black ink, plus a Goldmann type conversion for size difference. Such a stimulus value leads to a very specific but not very sensitive screening level when applied to a population at large. Obviously it is useful to have a specific but not very sensitive test to detect moderately severe glaucoma but it is not ideal. It is hoped that the sensitivity of disease detection has been improved in the new version and ultimately field trials on an appropriate population should confirm or deny this.
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.