Consensus 8

8^th Consensus Meeting: Progression of Glaucoma

Paris, France, June 28, 2011

edited by R.N. Weinreb, D.F. Garway-Heath, C. Leung, J.G. Crowston and F.A. Medeiros
2011. 10 tables, 4 figures and 28 photos of with 1 in full color. Hardbound.
ISBN-10: 90 6299 231 5
ISBN-13: 978 90 6299 231 7
Published by: Kugler Publications.
Click here for more information on all publications in the Consensus series.

See meeting photos

Summary Consensus Points

• Visual function progression
• Structure
• Structure and function
• Risk factors
• Glaucoma and its impact on patient function

Section 1 – Visual function progression

1. Standard white-on-white automated perimetry (SAP), with a fixed testing matrix covering at least the central 24 degrees, is preferred for measuring progression in eyes with glaucomatous VF loss.
   Comment: more research is needed into the use of alternative measures of visual function (FDP, resolution perimetry, motion perimetry and others) to detect glaucomatous progression, before any of these can be considered alternatives to SAP for measuring progression.
   Comment: It is possible for glaucomatous optic neuropathy to progress structurally in the absence of functional progression and vice-versa.
2. Perform sufficient examinations to detect change.
   Comment: decisions on progression should not be made by comparing only the most recent field with the one before.
   Comment: suspected progression should be confirmed by repeating the field.

Baseline data collection (no previous VFs available) – first two years

3. In clinical practice, at least two reliable VFs is optimal in the first six months.
   Comment: In clinical scenarios, where the lifetime risk of visual disability is high, such as those who already have advanced damage, three baseline VFs may be necessary.
   Comment: A good baseline of reliable VFs is essential to be able to monitor for progression.
   Comment: Unless there are obvious learning effects, high false-positive errors, rim artifacts, or other obvious artifacts, examinations should not be removed from the analyses.
4. At least two further VFs should be performed within the next 18 months.
5. VF testing should be repeated sooner than scheduled if possible progression is identified on the basis of an ‘event’ analysis.
   Comment: In patients at risk of visual disability, performing six VFs in the first two years enables the clinician to rule out rapid progression (2 dB/year or worse) and establishes an ideal set of baseline data.
   Comment: the identification of possible progression may be on the basis of an ‘event’ criterion such as the Glaucoma Progression Analysis (in the Humphrey perimeter software) or ‘Nonparametric Progression Analysis’.
6. Establish a new baseline after a significant therapeutic intervention (e.g., surgery). Comment: the new baseline can be the last fields that defined the previous progression ‘event’.

Follow-up data collection (after the initial two years)

7. The frequency of follow-up VFs should be based on the risk of clinically significant progression (based on extent of damage and life expectancy).
8. In low and moderate risk patients, subsequent VF frequency should be one VF per year (unless there is a long follow-up) and, as a rule, repeated sooner if possible. Progression is identified on the basis of an ‘event’ analysis, or if other clinical observations are suggestive of possible progression or increased risk of progression.
   Comment: relevant clinical observations include structural progression (clinically noted or measured by imaging), a splinter hemorrhage, or inadequate IOP control.
9. In high risk patients, subsequent VF frequency should be two VFs per year and repeated sooner if possible progression is identified on the basis of an ‘event’ analysis, or if other clinical observations are suggestive of progression or increased risk of progression.
   Comment: following confirmed progression (by an ‘event’), the frequency of testing should be based on the estimated rate of progression, risk factors and other clinical indicators of progression, stage of disease and life expectancy.
   Comment: patients who have been stable for a long period, or who are progressing so slowly as to be at little risk for reaching disabling levels of field loss, and other clinical parameters indicate low risk of progression, may have VF testing less frequently than 1 VF per year.

Visual field progression may be analyzed by either ‘event-’ or ‘trend-’based methods

Event analysis: is change from baseline greater than a predefined threshold; the threshold is based on test retest variability (according to level of damage). Trend analysis: determines the rate of change over time; the significance is determined by the variability of the measurement and the magnitude of change.

10. Both event and trend analyses are needed, largely for different time points in the follow-up during clinical care.
11. In general, event-based methods are used early in the follow-up, when few VFs are available for serial analysis.
    Comment: progression by an event criterion usually requires confirmation on at least two further occasions to be sufficiently sure that progression has truly occurred.
    Comment: confirmation of progression should usually be made on a separate occasion (patients have ‘off days’).
    Comment: When interpreting VF progression that is confirmed by an ‘event’ method, the clinician should look at:
    – the baseline fields, to ensure they are reliable and appropriate for the analysis
    – the estimated rate of progression and the confidence of the estimate;
    – the severity of the visual loss in terms of impending impairment;
    – the risk factors for progression.
12. In general, rate-based analyses are used later in the follow-up, when a greater number of VFs is available over a sufficient period of time to measure the rate of progression.
    Comment: a rate of progression in the first two years is a rough estimate (wide range of possible rates around the central estimate); in most patients it takes longer to obtain a reliable estimate of the rate of progression.
    Comment: trend (regression) analysis provides an estimate of the rate of progression and a measure of the reliability of the estimate; the reliability of the estimate is judged from the confidence limit.
    Comment: clinicians should consider other clinical measures of progression and risk of progression when interpreting this information (these data provide the ‘prior probability’ for progression).
13. When progression is identified, the clinician should ensure that the progression is consistent with glaucoma and not related to some other cause.

Measure the rate of visual field progression

14. Clinicians should aim to measure the rate of VF progression.
    Comment: Estimating the rate of progression is invaluable for guiding therapeutic decisions and estimating the likelihood of visual impairment during the patient’s lifetime.
15. In the absence of significant changes in therapy, the rate of progression of suitable global indices (MD or VFI, but not PSD or LV) is linear in treated glaucoma eyes, except at the most advanced stages.
16. As a linear model for progression is acceptable, trends may be extrapolated to predict future loss if there is no change in therapy, over appropriate intervals.
17. Both local and global metrics are needed for assessment of progression.
    Comment: Rates are most often measured on ‘global’ parameters, such as mean deviation, mean defect or visual field index. However, focal progression (such as paracentral) may be missed by a global index.
18. Total Deviation based methods are more sensitive to cataract than Pattern Deviation based methods. However, by eliminating or reducing the component of diffuse visual field loss, Pattern Deviation based methods may underestimate progression rates.
19. Use available software support.
    Comment: Subjective judgment of VF print-outs is unreliable and agreement among clinicians is poor. Statistical analysis, either in the perimeter software or stand-alone software, is advantageous to reliably identify and measure progressive VF change.

Pay attention to examination quality

20. Examinations of poor quality will likely lead to an erroneous assessment of progression.
    Comment: The most important factors to reduce test variability are a proper explanation of the test to the patient, appropriate instrument setup and 1:1 monitoring of the patient by a trained technician.
21. Do not rely automatically on the VF reliability indices.
    Comment: The VF reliability indices may be unreliable! The most useful index is the ‘False Positive’ rate; values greater than 15% likely represent a less reliable performance; values less than 15% do not guarantee reliability. The technician is the best judge to exam quality.
22. If unreliable tests require repeating, the patient should be carefully re-instructed.

Use the same threshold test

23. Clinicians should select their preferred perimetry technology, test pattern, and thresholding strategy for the baseline tests and stick with the same test throughout the follow up.
    Comment: any analysis of progression can only be performed if a compatible threshold algorithm and test pattern is used.
24. In advanced glaucoma, smaller angular size SAP testing grids, e.g., HFA 10-2 may be of value in a minority of patients.
    Comment: Kinetic perimetry and SAP with larger targets (e.g., size V) may also be useful.
    Comment: The advantages of a change in test pattern (e.g., from a 24-2 to a 10-2 grid) should also be weighed against the disadvantages for progression analysis by commercial software.

Clinical trials

25. Event analyses aim to identify a statistically significant difference between study arms and not necessarily a clinically significant difference.
    Comment: As glaucoma is a chronic progressive disease and progression is generally linear, small amounts of progression that reach statistical significance become larger, clinically significant amounts of progression if there is no additional therapy.
26. Rate analyses of VF indices are an appropriate statistical approach to identify differences between treatment groups.
    Comment: Rate analysis methods have been used often in trials for other chronic progressive diseases, such as dementia.
27. Difference in the progression ‘event’ criterion applied in the various clinical trials limits comparison of the incidence of progression determined in those trials.
    Comment: Comparison of groups in different clinical trials is also hampered by mismatch of subjects with regard to stage of glaucoma, quality of visual field exams, and other traits.

Research needs

1. The development of ‘event’ criteria for progression based on individual patient test-retest variability.
2. There is a need to compare event-based endpoints and rate of progression outcomes in a data set with data acquired with appropriate frequency and test intervals with respect to clinical trials.
3. Further research is needed into the added value of smaller angular size test grids, and different size stimuli, e.g., size V, in advanced glaucoma.
4. Determine appropriate dynamic ranges of stimulus contrasts for size III, and develop new stimuli with larger dynamic ranges of appropriate stimulus contrasts.
5. Improve the interface between perimetrist and device, and between patient and device.
6. Identify, or develop, stimulus types (e.g., FDT) and test algorithms which provide optimal information content for progression analysis in children and adults who have difficulty performing a reliable SAP test.
7. Develop alternate methods for selecting stimulus locations in order to avoid extensive testing of blind areas and to focus on areas of interest.
8. Further assess the benefits of using prior threshold as a starting point in a follow-up test (or if threshold is < 0 dB previously, confirmation at that point that a 0 dB stimulus is not seen is sufficient).
9. Determine the optimal frequency and timing of tests for individual patients.
10. Use of good mathematical modeling.
11. Develop better approaches to identify learning effects.
12. Identify the appropriate test and frequency of testing for patients with progressive glaucomatous optic neuropathy and SAP within normal limits.

Section 2 – Structure

2.1 Technologies for measurement of optic disc and retinal nerve fiber layer (RNFL) parameters

1. Serial optic disc stereo-photography and RNFL photography are valuable and enduring methods for monitoring structural progression.
   Comment: Stereoscopic clinical examination of optic disc and RNFL may be useful to detect change in comparison with a baseline photograph.
   Comment: Subjective estimates of cup/disc ratio only detect large changes in cupping and are insufficient for monitoring structural changes.
2. Color fundus photography is the preferred imaging modality to identify disc hemorrhages and parapapillary atrophy.
   Comment: Disc hemorrhages and beta-zone PPA are known risk factors for glaucoma progression.
3. Changes in beta-zone parapapillary atrophy can signal glaucoma progression.
   Comment: Methods for evaluating changes in PPA require further validation and include fundus photography, CLSO, and SDOCT.
4. Several imaging instruments, including confocal scanning laser ophthalmoscopy, scanning laser polarimetry, and optical coherence tomography objectively provide reproducible measurements and quantitative assessment of the optic disc and RNFL change.
   Comment: The detection of glaucoma progression by comparing sketches or descriptions of cup disc ratio in the clinical chart is generally not suitable for an early detection of progression and may be replaced by imaging techniques and/or optic disc photography.
   Comment: Imaging instruments provide progression detection analyses that can determine whether change is greater than the measurement variability of an individual eye.
5. There are several structural components of longitudinal change detection that likely contribute to the variability of measurements.
   Comment: These include variation in clinical disc margin visibility, intersession variation and accuracy of segmentation algorithms, variation in vascular blood volume and reference plane anatomy, and longitudinal image registration.
6. Image quality can influence our ability to detect structural change.
   Comment: Automated quality indices vary by instrument and are often proprietary with little information available about how they are constructed.
   Comment: Poor quality images can lead to either false positive or false negative results.
   Comment: For patient management decisions, clinicians should review the quality of images included in glaucomatous progression assessment.
7. More than one good quality baseline image facilitates progression analysis.
   Comment: Some instruments automatically acquire several baseline images during one imaging session.

2.2 Reproducibility of digital imaging instruments

1. Measurement variability influences the ability of any device to detect progression.
   Comment: There is a wide range of reproducibility estimates in the literature for SLP, CSLO, and OCT. Although studies of comparisons of instruments within the same patient populations are limited, these techniques likely provide data of similar reproducibility.
   Comment: Overall, SDOCT has better reproducibility than TDOCT.
2. There is a lack of consensus in the literature as to whether reproducibility changes across disease severity and this may vary across measured anatomic structures and techniques.

2.3 How to detect and measure structural change?

1. Event and trend based analyses are both useful for change detection.
   Comment: These analyses do not always concur.
2. It is important to estimate the rate of structural progression for clinical management decisions.
   Comment: The rates of change obtained from measurements from optic disc, RNFL and macular parameters may vary from each other.
3. Quantitative assessment of optic disc and retinal nerve fibre layer (RNFL) with imaging instruments is useful and complementary for change detection.
   Comment: Data are limited on whether macular measurements may be useful for change detection.
4. Differences in technologies and scan protocols could influence the detection of progression even when the same structure is measured.
5. There is no clear consensus on which instruments or parameters are optimal to detect structural progression. As technologies evolve, new instruments and parameters which are clinically useful will emerge.

2.4. How to define clinically significant structural change?

1. Interpretation of statistically significant change should take into account test-retest variability and knowledge on the magnitude of age-related change in healthy individuals.
2. Knowledge of age-related change in healthy individuals should preferably come from actual longitudinal data and not extrapolation from cross-sectional data.
3. A statistically significant change in a structural parameter such as rim area or nerve fiber layer thickness is a relevant change, however, it may not be clinically meaningful. The latter also should take into account the age and stage of the disease as well as an assessment of risk factors present.
   Comment: Currently, we have the tools to measure statistically significant change, however, to date we do not know how to fully assess the clinical importance of this change.

2.5 Issues in clinical practice

1. The optimal frequency of imaging tests is unknown.
   Comment: It depends on the severity of the disease and on the expected speed of progression.
2. In longitudinal studies investigating optic disc and RNFL progression in glaucoma, imaging tests have been performed once a year to three times a year.
3. The same structural measures (e.g. RNFL thickness) obtained with different instruments from the same manufacturer or the same technology from different instrument manufacturers (i.e., spectral domain OCT) are not necessarily interchangeable for progression assessment.
4. Structural assessment of change is a valid method for detection of glaucomatous progression in a clinical trial.
   Comment: structural change has been shown to be predictive of future functional loss in glaucoma.

Section 3 – Structure and function

1. Both optic nerve structure and function should be evaluated for detection of glaucomatous progression.
2. Currently, no specific test can be regarded as the perfect reference standard for detection of glaucomatous structural and/or functional progression.
3. Progression detected by functional means will not always be corroborated using structural tests, and vice-versa.
   Comment: This is due to the imperfect nature of testing analysis, individual variability, and the structure-function relationship.
4. The use of standard automated perimetry as the sole method for detection of change may result in failure to detect or underestimate progression in eyes with early glaucomatous damage.
   Comment: In glaucoma suspect or ocular hypertensive eyes with initially normal achromatic perimetry, a change in optic nerve structure (e.g., optic topography, retinal nerve fiber layer, optic disc hemorrhage, or parapapillary atrophy) may occur before perimetric change.
5. In general, detection of progression is more difficult in eyes with advanced disease.
   Comment: In eyes with advanced visual field damage, alternative perimetric strategies (i.e., larger stimulus, macular strategies, kinetic perimetry, etc.) may need to be employed.
6. A statistically significant change in structure and/or function (which takes age and variability into account) is not always clinically relevant.
   Comment: Its clinical relevance for patient management must take into account other risk factors and lifetime risk of visual disability.
7. Progressive structural changes are often but not always predictive of future development or progression of functional deficits in glaucoma.
   Comment: The predictive strength depends on the method used to assess structural/functional change.
8. Corroboration of glaucomatous progression through the use of more than one test may provide more effective and more rapid detection of glaucomatous progression than repeated confirmation of change using a single modality.
   Comment: Examples of corroborative change include structure-function (e.g., a structural change of the optic nerve and a spatially consistent functional change).
9. In order to increase the likelihood of detecting progression, test results should be of sufficient quality and appropriate quantity to provide meaningful information.
   Comment: While adjunctive testing can help clinical decision making, the use of multiple modalities of testing, at the expense of quality and appropriate frequency and quantity, should be avoided.
10. Life expectancy should be considered when evaluating the clinical relevance of a structural and/or functional change in glaucoma.
11. Structural and/or functional testing should be conducted throughout the duration of the disease.

Section 4 – Risk factors

1. Risk factors for glaucoma progression should be ascertained in all patients with glaucoma or suspected of being at increased risk of glaucoma.
2. Clinical risk factor assessment in glaucoma serves two roles. It provides (a) prognostic information; and (b) a basis for disease management.
   Comment: While proof of causality is desirable, the pragmatic nature of clinical medicine allows the use of risk factors of varying evidence quality and even clinical signs to be used in clinical management.
3. The use of risk factors in clinical management should take into account: (a) the strength of the risk factor for disease progression; and (b) the practicality and potential harm of reducing that risk factor.
4. Ocular hypertension is itself a strong risk factor for glaucoma, with rates of progression depending on the presence or absence of other risk factors.
   Comment: Accounting for these risk factors is critical to clinical decision making in the management of OHT patients.
   Comment: Risk factor assessment in OHT helps determine an individual’s need for IOP lowering medication and also informs on the frequency of follow up.
5. Risk calculators provide a means for quantifying risk of glaucoma progression in appropriate individuals with similar baseline characteristics to those present in the study.
   Comment: The utility of these risk calculators in clinical practice still needs to be determined.
6. Higher mean IOP is a strong risk factor for glaucoma progression.
   Comment: More studies are needed to evaluate the role of other IOP parameters as risk factors for glaucoma progression.
7. A thinner central cornea is a risk factor for progression in patients with higher baseline IOP.
8. The presence of pseudo-exfoliation syndrome is an independent risk factor for progression.
9. The presence of a disc haemorrhage, older age, and lower ocular perfusion pressure are risk factors for progression. Comment: The relationship between low blood pressure and risk of progression is complex.
10. While estimates of risk of progression for individual patients based on completed large clinical trials are available, the use of such estimates varies considerably in clinical practice.
11. There is greater information available regarding the importance of risk factors for progression from early to moderate disease than from moderate to severe disease.
    Comment: Few adequately powered studies have prospectively assessed the risk factors for blindness from glaucomatous disease.
12. The relative importance of risk factors for progression may vary depending upon the stage of glaucomatous disease.
    Comment: Some risk factors that do not appear to be important predictors of progression from early to moderate glaucoma may be relatively more important in predicting progression from moderate to severe disease and vice versa.
13. Studies that longitudinally assess risk factors for functional vision loss and blindness from glaucomatous disease are needed.

Section 5 – Glaucoma and its impact on patient function

1. Standard measures for assessing glaucoma include measures of optic nerve structure and function including cup/disc ratios, thickness of the retinal nerve fiber layer and ganglion cell layer, white on white visual fields, blue on yellow visual fields, and intraocular pressure. While these measures provide an assessment of the eye, they are surrogates for how the patient is functioning. Both PROs and functional tests provide important information in addition to standard tests on the impact of glaucoma on the patient.
2. It was previously believed that only advanced glaucoma damage has an impact on the patient ability to function. However, more recent cross-sectional clinic-based and population-based studies have demonstrated that early glaucomatous visual field loss has an impact on the patients’ ability to function as assessed by patient reported outcome measures and functional tests.
3. Future studies are needed to explore the relationship between PROs and functional measures and glaucoma progression.
4.  Numerous instruments and tests have been used for assessing PROs and functional measures in research settings. However, there is no consensus on a single PRO or functional measure (or set of PROs or functional measures) for clinical practice. There is a need to create simpler PROs and functional tests which can easily be reproduced in a wide variety of settings.