Abstract
purpose. To analyze the sensitivity of various sensory tests adjusted for glaucomatous optic disc damage.
methods. In a cross-sectional study, the results of testing of 196 control subjects (age range, 18–69 years) and 308 patients with chronic open-angle glaucoma (age range, 18–70 years) were included. The perimetric mean defect (MD), a temporal contrast sensitivity test (TCS), a spatiotemporal contrast sensitivity test (STCS), the peak time of a blue-on-yellow visual evoked potential (BYVEP), and the amplitude of a pattern-reversal electroretinogram (PERG) were evaluated by a specific logistic regression model. This model included glaucomatous damage, quantified by neuroretinal rim area corrected for disc size, as a covariate of sensitivity.
results. Sensitivity of diagnostic tests increased for all procedures with increasing loss of neuroretinal rim area. With progressing optic disc damage, MD and STCS showed higher sensitivity than did TCS. BYVEP showed a higher sensitivity than PERG in all disease stages. In general, the psychophysical tests were more sensitive than the electrophysiological ones.
conclusions. The specific model used in this study was an appropriate tool to analyze the sensitivity of several sensory glaucoma tests in relation to disease stage. Moreover, tests that were more sensitive in early disease stages (TCS) and others that were more sensitive in more advanced stages (MD, STCS) were identified.
Chronic glaucomas are a heterogeneous group of eye diseases characterized by increased intraocular pressure (IOP) in many patients and by progressive damage to the optic disc and to the visual field. In these diseases, the IOP is an important risk factor, whereas optic disc changes and perimetric defects are definite signs of the disease and typically progress slowly over years or decades. An elevated IOP without any of the just mentioned signs of the disease is a state called ocular hypertension (OHT). Although it has been shown that IOP levels are important in the course of the disease and that IOP reduction slows down disease progression, the progression of OHT to the glaucomatous disease seems to be a relatively rare event, estimated to occur at a rate of approximately 1% to 2% per year.
1 2 3 If a conversion from OHT to glaucoma takes place, the optic disc or retinal nerve fiber layer changes often occur before the visual field defects
4 5 and thus may be considered useful signs for an early glaucoma diagnosis. Visual field loss, determined by conventional white-on-white perimetry, seems to be a late defect in the disease process and may occur only when a considerable number of optic nerve fibers have been lost.
6 Thus, perimetry may be considered a useful tool for the follow-up of patients with advanced stages of glaucoma. Apart from detecting rather advanced glaucomatous damage, perimetry is a time-consuming and fatiguing procedure burdened with high variability
7 of the patient’s performance and learning effects.
8 9 10 Therefore, during recent years, other sensory procedures, psychophysical as well as electrophysiological, have been devised that can indicate glaucomatous damage earlier and in a more reliable and quicker way than perimetry. The diagnostic value of such methods has been described previously.
11 12 13
In general, diagnostic tests become more sensitive with progressing damage. In addition, between the individual test types, there seems to be a great difference in which test performs best in which stage of the disease. The question of how the sensitivity of such tests changes during the disease process is important, because it may advise the ophthalmologist which of the experimental tests to use in different stages of glaucoma.
It was the purpose of the present investigation to answer this question by analyzing data from the Erlangen Glaucoma Registry, which was started in 1991 and has accumulated a considerable number of patients and control subjects who have been examined at regular intervals with various experimental sensory tests. Ideally, such an analysis should be undertaken by individual follow-ups. However, that is difficult because of the low rate of disease progression. In addition to the chronicity inherent in the disease, therapeutic interventions probably further delayed an advancement of the disease. For these reasons, the progression of the glaucomatous process was studied in cross-sectional analyses of patients who entered the glaucoma registry at different stages of glaucoma. In the present investigation, perimetry was treated equivalent to the other experimental sensory tests to compare them with each other. In several studies, a highly significant correlation (
P < 0.001,
r = −0.81) was shown between the neuroretinal rim (NRR) area and glaucoma stage according to optic disc morphometry.
14 15 This leaves the optic nerve head damage, measured as size of the neuroretinal rim (NRR) area, as the appropriate reference for scaling the glaucomatous damage.
The methods applied in the study adhered to the tenets of the Declaration of Helsinki for the use of human subjects in biomedical research. Informed consent was obtained from each subject before enrollment. Institutional Review Board and Ethics Committee approvals were achieved at baseline of the ongoing glaucoma study (the “Erlangen Glaucoma Registry”).
Excluded from the study were eyes with more than 12% false-positive and false-negative responses in visual field testing, and visual acuity of 0.6 or better was a requirement. Fixation was automatically controlled by the perimetric device (Octopus-501, G1, three-phases, 59 measure points; Interzeag, Schlieren, Switzerland). In addition, eyes without available results of NRR measurements had to be excluded in view of the study’s rationale. All subjects had clear optic media, and no disease was found by slit lamp examination, gonioscopy, and funduscopy.
Temporal Contrast Sensitivity Test (TCS).
Spatiotemporal Contrast Sensitivity Test (STCS).
Blue-on-Yellow Visually Evoked Potential (BYVEP).
Pattern-Reversal Electroretinogram (PERG).
Static Automatic Perimetry.
In our study, the mean perimetric defect (MD) quantified the extent of the visual field defect. Fixation losses are controlled through pupil videos taken by an integrated camera. If fixation loss occurs, an acoustic signal indicates that the measurement has to be repeated at the respective position.
18
For each patient, all tests were explained in detail and performed on the same day and in the same sequence (MD, STCS, TCS, PERG, and BYVEP). Learning tests were not performed. With the exception of perimetry with a duration of approximately 25 minutes per eye, the diagnostic procedures took approximately 5 to 10 minutes per eye.
For all five diagnostic tests, age dependency was adjusted for linear regression analysis in the control group. The measure for disease severity was the loss of the neuroretinal rim (LNRR) area—that is, the difference between the expected area of the NRR adjusted for optic disc size by a linear regression analysis in the control group. Each test was dichotomized by fixing the specificity in the control group to 80%. Descriptive analysis included means, standard deviations, medians, and ranges for continuous variables
(Tables 1 2 3) . In significance tests, the generalized estimating equations (GEE) method was used to account for the dependency arising from multiple measurements at the same subject.
24 25 The Bonferroni correction was used for multiple testing.
Because the specific logistic model used in the present investigation has not been applied before in the ophthalmology literature, it is described briefly: Given that an eye is actually glaucomatous (according to a diagnostic reference criterion), we might expect that the degree of optic nerve damage influences the result of a diagnostic test. Moreover, the degree of damage might influence the sensitivities of several tests differently. Both effects can be quantified and tested statistically using a specific logistic regression approach developed by Leisenring et al.
26 In this model, different tests are coded as dummy variables, and disease severity is a continuous covariate. Inclusion of interaction terms between diagnostic tests and disease severity gives a specific logistic model for each test type. These different models can be represented graphically as logistic curves, with sensitivity being the dependent and disease damage being the independent variable
(Fig. 1) . For a suitably chosen measure of disease damage (in our case the LNRR), the regression coefficients are the logarithms of increase in sensitivity per unit of increase in disease damage
(Table 4) . With this method, the sensitivities of all diagnostic procedures over the entire spectrum of disease damage can be evaluated simultaneously through one single model and its graphic representation
(Fig. 1) . An additional model including NPG as a dummy variable was built for detecting possible interactions between test performance and IOP
(Fig. 2) .
Statistical analyses were performed with commercially available software (SPSSWIN, release 11.0; SPSS Sciences, Chicago, IL). GEEs were analyzed using the MAREG program, which is available by the Department of Statisitics, Ludwig Maximillians University Munich, Munich, Germany, without cost to the user (www.stat.uni-muenchen.de/sfb386/).
In the present study, five different tests were evaluated, including automated white-on-white projection perimetry. There are some other reports in the literature of investigations that have used batteries of different psychophysical and electrophysiological tests to determine their usefulness in glaucoma diagnosis.
27 28 29 30 31 The discussion of their findings in comparison with the present results is difficult because different stimulus paradigms were used and heterogeneous groups of patients and patients with suspected glaucoma in different stages were examined. None of the studies included disease severity itself into the analyses to rank the various tests under explicit consideration of their differing performance in different glaucoma stages. In the present cross-sectional analysis, the degree of the glaucomatous damage was graded on a continuous scale of the NRR area of the optic disc, and the concomitant performance in different sensory tests can be judged
(Figs. 1 2) .
Figures 1 and 2 of the present study suggest that in early stages of glaucoma, in POAG and NPG, the TCS is the most sensitive of all. The STCS and the perimetric MD seem to have rather low value in early glaucoma diagnosis, and both take a similar course as the glaucomatous neuropathy increases
(Figs. 1 2) .
The sensitivity of perimetry is comparably lower in the beginning of the disease but becomes very important in late stages. However, we are aware of the fact that the MD ignores spatial clustering and dispersion of local damages. Especially, early glaucomas might be indicated by a few clusters of damage that do not shift the entire MD into the pathologic range. Thus, it is not the sensitivity of the entire perimetric measurement we are analyzing. In addition, more elaborated perimetric techniques have been developed recently that are able to detect optic neuropathy in glaucomatous eyes with normal standard visual fields.
29 The curve of the STCS has a similar shape as that of perimetry. Although the test uses a temporally alternating pattern the stimulus is localized like perimetric stimuli and probes an area known to have a high susceptibility for glaucomatous damage.
32
In our study, the electrophysiological procedures in general showed lower sensitivity than the psychophysical ones. The BYVEP, which takes a course rather similar to the temporal contrast-sensitivity curve, is our most sensitive electrophysiological test in early glaucoma stages and ranks directly behind the temporal contrast sensitivity test there.
Several PERG studies regarded the PERG as an early diagnostic test.
27 28 33 34 35 36 Figure 1 , however, does not confirm these findings. In contrast, our results show a comparably poor sensitivity of PERG in early stages that cannot be compensated through the relatively steep course of the curve. In fact, Arden
30 states that ERG is not a screening technique, although many psychophysical tests appear promising. On the contrary, Graham et al.
28 and Drance et al.
27 used a series of different functional tests. In their studies, PERG alterations were found the most sensitive indicators of glaucomatous damage.
A conclusive comparison between studies concerning the sensitivity of PERG would be possible, if the relationship between stage of the disease and sensitivity had been quantified by the authors. In the work of Weinstein et al.,
35 individual data are given both for disease severity in terms of cup-to-disc ratio and results of the PERG (P1/N1). We have reanalyzed these data and found a U-shaped dependency of sensitivity on cup-to-disc ratio—implying that this ratio may not have been the only relevant indicator for disease severity in that study.
Stratification for NPG and POAG led to different performance in the case of the three psychophysical procedures
(Fig. 2) , whereas the electrophysiological tests performed nearly identically in both subgroups. Although the damage in POAG is rather diffuse, NPG seems to be characterized by local, more heterogeneous defects of the NRR
37 38 that may be detected more reliably through the electrophysiological procedures than through the psychophysical ones used in this study. In perimetry, for example, the corrected loss variance (CLV) is greater in eyes with NPG than in those with POAG and equal LNRR area, and the curve for MD would be shifted up with additional consideration of the CLV. Similar effects might occur for the other two psychophysical procedures.
One of the study’s advantages lies in the considerable sample size available from the Erlangen Glaucoma Registry. However, we could not use diagnostic procedures that have been developed most recently. Although somewhat striking differences were shown between the five tests according to optic disc damage, the sensitivity of every test was rather poor for early disease. However, the method we have introduced may be of general interest in vision research concerning the mechanisms of various electrophysiological and psychophysical tests in the course of disease progression.
The results of the present study indicate that the diagnostic sensitivity of a particular sensory test depends on the degree of the glaucomatous optic neuropathy. With progressing disease damage, sensitivity increases for all tests, but the curves describing the increase in sensitivity with increasing neuropathy are not all the same shape but show different steepness and may show crossover. This behavior may partly explain some of the discrepant results described in the literature.
Our findings suggest that the specific approach used in this study may be a standard technique in the evaluation of diagnostic tests in glaucoma. It provides an analytical tool to compare performance and appropriateness of various diagnostic measurements, depending on disease severity. It may advise the ophthalmologist which of several tests to use in which stage of glaucoma; and, furthermore, it may improve the comparison of results obtained from different populations in different studies.
Supported by Deutsche Forschungsgemeinschaft (SFB 539), Bonn, Germany.
Submitted for publication April 29, 2002; revised February 18, 2003; accepted March 11, 2003.
Commercial relationships policy: N.
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “
advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Corresponding author: Andrea Stroux, Department of Medical Informatics, Biometry and Epidemiology, Free University Berlin, Hindenburgdamm 30, 12200 Berlin, Germany;
[email protected].
| Gender | Age (y) |
POAG (201 subjects, 385 eyes) | 119 Male | 49.5 ± 11.8 |
| 82 Female | [51; 18–70] |
NPG (107 subjects, 201 eyes) | 34 Male | 53.5 ± 10.3 |
| 73 Female | [55; 23–70] |
Control (196 subjects, 383 eyes) | 107 Male | 44.0 ± 13.2 |
| 89 Female | [45; 18–69] |
Table 2. Eye-Specific Characteristics
Table 2. Eye-Specific Characteristics
| IOP (mm Hg) | Neuroretinal Rim Area (mm2) | Loss of Neuroretinal Rim Area (mm2) | Mean Perimetric Defect (dB) |
Glaucoma (n = 586 eyes) | 26.6 ± 8.0 | 1.11 ± 0.40 | +0.44 ± 0.39 | +4.33 ± 4.79 |
| [25.0; 13.0–80.0] | [1.10; 0.13–2.66] | [+0.44;−0.87–+1.38] | [+2.80;−1.6–+25.5] |
Control (n = 383 eyes) | 17.7 ± 2.7 | 1.66 ± 0.33 | 0.00 ± 0.32* | +1.01 ± 1.27 |
| [18.0; 10.0–25.0] | [1.63; 1.07–2.96] | [−0.08;−1.43–+0.67] | [+1.00; −2.3–+4.1] |
Table 3. Results of Diagnostic Procedures
Table 3. Results of Diagnostic Procedures
| TCS (log %−1) | STCS (log %−1) | BYVEP (ms) | PERG (μV) |
Glaucoma (n = 586 eyes) | 1.26 ± 0.27 | 1.64 ± 0.48 | 126.1 ± 14.1 | 3.03 ± 1.12 |
| [1.31; 0.34–1.83] | [1.79; 0.00–2.20] | [122.7; 103.4–186.2] | [2.92; 0.21–7.41] |
Control (n = 383 eyes) | 1.48 ± 0.18 | 1.93 ± 0.13 | 117.1 ± 9.7 | 3.55 ± 1.12 |
| [1.48; 1.03–2.15] | [1.93; 1.41–2.29] | [116.0; 99.1–181.4] | [3.49; 0.00–7.43] |
Table 4. Results of the Logistic Regression Model
Table 4. Results of the Logistic Regression Model
| MD | TCS | STCS | BYVEP | PERG |
MD | 61% (56–66) | 0.08 | 0.67 | 0.001* | < 0.0001* |
| 1.61 (1.41–1.85) | | | | |
TCS | < 0.0001* | 56% (50–61) | 0.18 | 0.04, † | < 0.0001* |
| | 1.23 (1.11–1.37) | | | |
STCS | 0.31 | 0.001* | 60% (54–65) | 0.002* | < 0.0001* |
| | | 1.52 (1.34–1.72) | | |
BYVEP | 0.04, † | 0.20 | 0.20 | 49% (43–54) | 0.002* |
| | | | 1.37 (1.20–1.56) | |
PERG | 0.20 | 0.03, † | 0.56 | 0.47 | 38% (33–43) |
| | | | | 1.45 (1.31–1.61) |
Heijl, A, Leske, MC, Bengtsson, B, Hyman, L, Bengtsson, B, Hussein, M. (2002) Reduction of intraocular pressure and glaucoma progression: results from the early manifest glaucoma trial Arch Ophthalmol 120,1268-1279
[CrossRef] [PubMed]Kitazawa, Y, Hrie, T, Aoki, S, Suzuki, M, Nishioka, K. (1977) Untreated ocular hypertension: a long-term prospective study Arch Ophthalmol 95,1180-1184
[CrossRef] [PubMed]Lundberg, L, Wettrell, K, Linner, E. (1987) Ocular hypertension: a prospective twenty-year follow-up study Acta Ophthalmol 65,705-708
Pederson, JE, Anderson, DR. (1980) The mode of progressive disc cupping in ocular hypertension and glaucoma Arch Ophthalmol 98,490-495
[CrossRef] [PubMed]Sommer, A, Katz, J, Quigley, HA, et al (1991) Clinically detectable nerve fiber atrophy precedes the onset of glaucomatous field loss Arch Ophthalmol 109,77-83
[CrossRef] [PubMed]Kerrigan-Baumrind, LA, Quigley, HA, Pease, ME, Kerrigan, DF, Mitchell, RS. (2000) Number of ganglion cells in glaucoma eyes compared with threshold visual field tests in the same persons Invest Ophthalmol Vis Sci 41,741-748
[PubMed]Werner, EV, Petrig, B, Krupin, T, Bishop, KI. (1989) Variability of automated visual fields in clinically stable glaucoma patients Invest Ophthalmol Vis Sci 30,1083-1089
[PubMed]Wood, JM, Wild, JM, Hussey, MK, Crews, SJ. (1987) Serial examination of the normal visual field using Octopus automated projection perimetry: evidence for a learning effect Acta Ophthalmol 65,326-333
Wild, JM, Dengler-Harles, M, Searle, AET, O’Neill, EC, Crews, SJ. (1989) The influence of the learning effect on automated perimetry in patients with suspected glaucoma Acta Ophthalmol 67,537-545
Heijl, A, Bengtsson, B. (1996) The effect of perimetric experience in patients with glaucoma Arch Ophthalmol 114,19-22
[CrossRef] [PubMed]Korth, M. (1997) The value of electrophysiological testing in glaucomatous diseases J Glaucoma 6,331-323
[PubMed]Martus, P, Korth, M, Horn, F, Jünemann, A, Wisse, M, Jonas, JB. (1998) A multivariate sensory model in glaucoma diagnosis Invest Ophthalmol Vis Sci 39,1567-1574
[PubMed]Martus, P, Jünemann, A, Wisse, M, et al (2000) Multivariate approach for quantification of morphologic and functional damage in glaucoma Invest Ophthalmol Vis Sci 41,1229-1238
[PubMed]Jonas, JB, Gusek, GC, Naumann, GOH. (1988) Optic disc morphometry in chronic primary open-angle glaucoma: I. Morphometric intrapapillary characteristics Graefes Arch Clin Exp Ophthalmol 226,522-530
[CrossRef] [PubMed]Jonas, JB, Gusek, GC, Naumann, GOH. (1988) Optic disc morphometry in chronic primary open-angle glaucoma: II. Correlation of the intrapapillary parameters to visual field indices Graefes Arch Clin Exp Ophthalmol 226,531-538
[CrossRef] [PubMed]Jonas, JB, Fernandez, MC, Naumann, GOH. (1992) Glaucomatous parapapillary chorioretinal atrophy: occurrence and correlations Arch Ophthalmol 110,214-222
[CrossRef] [PubMed]Littmann, H. (1982) Zur Bestimmung der wahren Grösse eines Objekts auf dem Hintergrund des lebenden Auges [Determination of the real size of an object on the fundus of the living eye] Klin Monatsbl Augenheilkd 180,286-289
[CrossRef] [PubMed]Flammer, J, Drance, SM, Augustiny, L, et al (1985) Quantification of glaucomatous visual field defects with automated perimetry Invest Ophthalmol Vis Sci 26,176-181
[PubMed]Horn, FK, Jonas, JB, Korth, M, et al (1997) The full field flicker test in early diagnosis of chronic open-angle glaucoma Am J Ophthalmol 123,313-319
[CrossRef] [PubMed]Owsley, C, Sekuler, R, Siemsen, D. (1983) Contrast sensitivity throughout adulthood Vision Res 23,689-699
[CrossRef] [PubMed]Horn, FK, Martus, P, Korth, M. (1995) Comparison of temporal and spatio-temporal contrast-sensitivity tests in normal subjects and glaucoma patients Ger J Ophthalmol 4,97-102
[PubMed]Korth, M, Nguyen, NX, Jünemann, A, et al (1994) VEP test of the blue-sensitive pathway in glaucoma Invest Ophthalmol Vis Sci 35,2599-2610
[PubMed]Korth, M, Horn, FK, Storck, B, et al (1989) The pattern-evoked electro-retinogram (PERG): age-related alterations and changes in glaucoma Graefes Arch Clin Exp Ophthalmol 227,123-130
[CrossRef] [PubMed]Katz, J, Zeger, S, Liang, KY. (1994) Appropriate statistical methods to account for similarities in binary outcomes between fellow eyes Invest Ophthalmol Vis Sci 35,2461-2465
[PubMed]Liang, KY, Zeger, S. (1986) Longitudinal data analysis using generalized linear models Biometrika 73,13-22
[CrossRef] Leisenring, W, Pepe, MS, Longton, G. (1997) A marginal regression modelling framework for evaluating medical diagnostic tests Stat Med 16,1263-1281
[CrossRef] [PubMed]Drance, SM, Airaksinen, PJ, Price, M, Schulzer, M, Douglas, GR, Tansley, B. (1987) The use of psychophysical, structural, and electrodiagnostic parameters to identify glaucomatous damage Graefes Arch Clin Exp Ophthalmol 225,365-368
[CrossRef] [PubMed]Graham, SL, Drance, SM, Chauhan, BC, et al (1996) Comparison of psychophysical and electrophysiological testing in early glaucoma Invest Ophthalmol Vis Sci 47,2651-2662
Bayer, AU, Maag, KP, Erb, C. (2002) Detection of optic neuropathy in glaucomatous eyes with normal standard visual fields using a test battery of short-wavelength automated perimetry and pattern electroretinography Ophthalmology 109,1350-1361
[CrossRef] [PubMed]Arden, GB. (1993) Comparisons of new psychophysics and perimetry with electrophysiological techniques in the diagnosis of glaucoma Curr Opin Ophthalmol 4,14-21
[PubMed]Holopigian, K, Seiple, W, Mayron, C, Koty, R, Lorenzo, M. (1990) Electrophysiological and psychophysical flicker sensitivity in patients with primary open-angle glaucoma and ocular hypertension Invest Ophthalmol Vis Sci 31,1863-1868
[PubMed]Aulhorn, E, Karmeyer, H. (1976) Frequency distribution in early glaucomatous visual field defects Doc Ophthalmol Proc Ser 14,75-83
Wanger, P, Persson, HE. (1985) Pattern-reversal electroretinogram and ocular hypertension Doc Ophthalmol 61,27-31
[CrossRef] [PubMed]Trick, G. (1986) PRRP abnormalities in glaucoma and ocular hypertension Invest Ophthalmol Vis Sci 27,1730-1736
[PubMed]Weinstein, GW, Arden, GB, Hitchings, RA, Ryan, S, Calthorpe, CM, Oodm, JV. (1988) The pattern electroretinogram (PERG) in ocular hypertension and glaucoma Arch Ophthalmol 106,923-928
[CrossRef] [PubMed]O’Donaghue, E, Arden, GB, O’Sullivan, F, et al (1992) The pattern electroretinogram in glaucoma and ocular hypertension Br J Ophthalmol 76,387-394
[CrossRef] [PubMed]Quigley, HA, Sanchez, RM, Dunkelberger, GR, L’Hernault, NL, Baginski, TA. (1987) Chronic glaucoma selectively damages large optic nerve fibers Invest Ophthalmol Vis Sci 28,913-920
[PubMed]Yamazaki, Y, Koide, C, Miyazawa, T, Kuwagaki, N, Yamada, H. (1991) Comparison of retinal nerve fiber layer in high- and normal tension glaucoma Graefes Arch Clin Exp Ophthalmol 229,517-520
[CrossRef] [PubMed]