advertisement
How do we monitor our glaucoma patients in the next millennium?
Report on the Gullstrand Foundation Meeting, April 17, 1999, Uppsala, Sweden
Erik L. Greve
Paul Kaufman pointed out that the lifetime of technology is becoming increasingly shorter. He wondered whether sophisticated technology would ever replace critical judgments. He offered as major questions: How important is it to detect glaucomatous damage earlier than we do now? And: How much more time and effort do we want to invest in detecting progression sooner? Glaucoma is a slowly progressive disease. We have time! In fact, it may be that progression of the visual field, as measured by automated perimetry, goes unnoticed except in the late stages (see also, Editors Selection on Management Quality). After all, we are interested in quality of vision, not in minuscule changes of light sensitivity.
Anders Heijl discussed the problems associated with the progression of visual field defects. He illustrated that there may be no more than seven to eight steps between normal visual fields and blindness. Major problems are: learning, variability and cataract. The learning effect is variable, individual and underestimated. It may conceal progression. The learning effect is more pronounced in the peripheral field and more in glaucoma than in normals. Variability is higher in pathological locations than in normal locations. In defect positions, the threshold may reach any value by chance alone. Therefore, deterioration of normal locations has more significance than in pathological locations. The possibility of detecting progression is dependent on 'testpoint status' (defect or normal), eccentricity, and general status of the visual field. In contrast to MD, PSD is not disturbed by cataract. PSD is only suitable in the early stages.
Fifty-four percent of glaucoma patients have cataract. Cataract usually causes diffuse loss. Diffuse loss caused by glaucoma is rare. Programs using linear regression and MD will find cataract but not glaucoma. Estimating diffuse loss improves the detection of progression.
In a point-by-point analysis (as opposed to global indices), progression may be evaluated compared to a population base or to regression analysis. Any analysis needs more than five consecutive visual fields. Significant progression should involve at least five positions. Note: If five visual fields are needed, visual field examination once a year may take five years! More visual fields would improve the determination of progression. However, this poses an economical problem, yet to be solved.
Chris Johnson stated that we need better means to establish progression, particularly for the evaluation of treatments (whether in daily management or in drug research). "It is important to refine progression." Judgment of progression may be based on:
1. Pseudo-arbitrary criteria, i.e., expert opinion without a mathematical basis;
2. event analysis, i.e., event compared to baseline;
3. trend analysis, e.g., linear regression.
Chris Johnson pointed out that, in all major longitudinal studies, different criteria for progression were used (AGIS, CITGS, EMCT, ALTS, etc.) In the ocular hypertension treatment study (OHTS), only 14% of the defects - as found by the definition used in this study - were confirmed. Solutions to this problem may be found in improving the signal-to-noise ratio:
1. smarter statistical analysis. e.g., fuzzy logic;
2. boost signal more than noise, e.g., SWAP;
3. reduce noise = variability, e.g., larger target, such as in ring perimetry.
In longitudinal studies, ring perimetry could detect progression one and a half years earlier than white-on-white perimetry. Note: the question arises as to whether indeed the quality of life, real quality of vision, of glaucoma patients benefits from refined progression analysis. Will the patient notice the difference? This will be further discussed under optic nerve head analysis.
A few points on electrophysiological methods:
1. patient cooperation is still needed;
2. variability is great;
3. signal-to-noise ratio is a major problem;
4. lack of standardization.
Johanni Airaksinen opened the discussion on optic nerve head analysis. He pointed out that 50% of optic nerve head progression is linear, 25% is episodic, and 25% curvilinear. Glaucoma patients may loose 3% of their axons (36,000) each year, normals only 0.23%. He wondered whether modern technology is able to detect this loss.
Balwantray Chauhan explained confocal laser tomography (as implemented by the HRT) and his experience with the longitudinal follow-up of glaucoma patients and normal subjects. He felt that the mass detection of glaucoma, using optic nerve head data, will be impossible, because of the wide variation of normal discs. Disc area may vary by a factor of 7 (0.6-5.7 mm2), rim area by a factor of 6. The sensitivity and specificity of CD ratio is as bad as that of IOP. Chauhan uses a program with the HRT that analyses 6×64 points of the measured area of 25×256 pixels. The variability per point is established. Pathological change is compared to normal variability at each point. Trend over time is measured in at least three consecutive plots. The method provides variability plots, difference plots, and probability plots.
In a longitudinal study, he showed that a large proportion of patients (on treatment) progressed on HRT, but not on perimetry. It is not known whether detection of progression is simply earlier, or whether the rate of progression is also faster (unlikely). The significance of such findings for clinical practice needs to be established. In relation to earlier statements concerning refinement of the progression of visual defects, it may be interesting to monitor treatment on the basis of ONH analysis. As the goal of a treatment is to prevent further deterioration of the visual field, perimetry or other visual function tests will always be necessary. The combination of improved quantitative ONH analysis and refined visual function follow-up may considerably improve the management of glaucoma patients.
Erik L. Greve
The Gullstrand Foundation was set up on the initiation of the Department of Ophthalmology of the University of Uppsala, in 1996. The Foundation invites European ophthalmologists interested in glaucoma to a one-day meeting. Its goal is to become a forum for information and lively discussions on current problems important for our understanding and management of patients with glaucoma. The Gullstrand Foundation arranges an annual meeting with a small number of invited speakers from different parts of the world. Each meeting covers a special aspect of glaucoma. Ophthalmologists renowned for their research in glaucoma, as well as distinguished researchers outside the field of glaucoma, are invited to cover different aspects of the topic. With this mix of speakers, the discussion is stimulating and will perhaps provoke some new ideas. Ample time is assigned for discussion, and the participation of the audience is of major importance for the success of the meeting.
