THE CORRELATION BETWEEN CHANGE IN OPTIC DISC NEURORETINAL RIM AREA AND DIFFERENTIAL LIGHT SENSITIVITY

NICHOLAS G. STROUTHIDIS and DAVID F. GARWAY-HEATH

Glaucoma Research Unit, Moorfields Eye Hospital and Institute of Ophthalmology, London, UK

Introduction

It has often been suggested in primary open-angle glaucoma that structural damage to the optic disc characterized by neuroretinal rim loss occurs before functional change as recorded by automated perimetry.1-4 When visual field sensitivity is recorded in standard decibel (dB) scale, it has a curvilinear relationship to rim area in glaucoma.5-10 By this relationship, when the neuroretinal rim area is large and changes by a large degree, there will be a correspondingly small change to the visual field differential light sensitivity (DLS). Likewise, when the rim area is small and this decreases by a small amount, it will be reflected by a large change in the DLS. This observation suggests that the relationship between structure and function in glaucoma is a variable, changing one, depending on severity of disease. However, it is possible that functional loss is inappropriately measured using the dB scale, and that the observed curvilinear relationship may be a manifestation of its logarithmic nature.

The relationship between the physical (light) stimulus and its psychophysical perception in visual field testing is generally accepted to follow the Weber-Fechner law of subjective intensity. This states that a subject’s internal sense of magnitude is proportional to the logarithm of stimulus intensity. More recent work has suggested that subjective intensity is linearly related to the neural signal on which it depends, and this is supported by a series of combined psychophysical and neurophysiological studies of tactile texture perception.11 When DLS is converted from dB scale to a nonlogarithmic, linear scale (1/Lambert), it has been shown to have a linear relationship with rim area.8-10 There is also a suggestion that the 1/Lambert scale may be superior to the dB scale as a reflection of underlying ganglion cell function, and therefore a better measure of neuronal damage.9

It is possible that part of the reason for functional change being detected at a later stage than structural change could be due to the inappropriate logarithmic scaling of

Address for correspondence: Nicholas G. Strouthidis, MD, Glaucoma Research Unit, Moorfields Eye Hospital, City Road, London, EC1V 2PD, UK

Perimetry Update 2002/2003, pp. 317–327

Proceedings of the XVth International Perimetric Society Meeting, Stratford-upon-Avon, England, June 26–29, 2002

edited by David B. Henson and Michael Wall

© 2004 Kugler Publications, The Hague, The Netherlands

the visual field. The physiological relationship between rim area and DLS is not known for certain. If it were to be assumed that there is an equivalent rate of decay between both parameters, then a slope approaching a value of 1.0 would be observed when DLS is plotted against rim area. It would be expected that the dB scale manifests a smaller change in DLS than rim area in early glaucoma, and this would be reflected in a slope value of less than 1.00 (assuming the x-axis represents rim area and the y- axis represents DLS). Furthermore, the correlation should be non-linear. The slope value will more closely approximate a value of 1.00 in the 1/Lambert scale if there is an equivalence of measurements of rim area and visual field sensitivity. It may be possible to detect progression earlier using a linear (1/L) model.

The purpose of this study was to identify change in DLS in both dB and 1/Lambert scaling against change in neuroretinal rim area over time. Eyes at high risk of progression were selected to provide a cohort of subjects in whom both parameters would be changing. It was hoped that this may give an indication of the nature of the relationship between DLS and rim area (i.e., linear versus curvilinear).

Methods

Subjects

Subjects were recruited from the ocular hypertension clinics of the Glaucoma Research Unit at Moorfields Eye Hospital. There were two populations: the first was a cohort of patients with ocular hypertension who were deemed to have converted to glaucoma by pre-determined visual field criteria during the testing period. Ocular hypertension was defined as an intraocular pressure (IOP) of greater than 24 mmHg in the subject eye and normal visual field as tested by 24-2 Humphrey automated perimetry. Conversion to glaucoma was on the basis of AGIS visual field criteria (see below). The other population had an optic disc hemorrhage in the subject eye noted at some point during the testing period, in combination with either a diagnosis of glaucoma or raised IOP.

Testing

Visual field tests were performed with the 24-2 full threshold program using the Humphrey automatic perimeter. Tests were performed at four-monthly intervals and the data were analyzed using Progressor for Windows Version 3.0. Only visual field tests that coincided with HRT imaging on a test date were included in the study. Conversion criteria was satisfied when a glaucomatous defect was detected and then confirmed on a further two field tests performed one month apart.

Optic disc images were acquired using the HRT (Heidelberg Engineering, Dossenheim, Germany) at intervals of four months to one year. Three 10° images were acquired on each occasion. Images were exported to the HRT-II Explorer software (Version 1.10) for analysis. Images were rejected if the exported contour line did not correspond closely to the disc margin, or if there were differences in disc magnification. Recordings of the optic disc global rim area, as well as of the superotemporal, superonasal, inferotemporal and inferonasal rim area, were made (Fig. 1). A record of the initial cup shape measure was also made for each of these sectors.

Calculations

Progressor produces a sensitivity output in dB scale for each of the points tested in the 24-2 test (Fig. 2) for each visual field test. The dB scale for each point on each test date can be converted to 1/Lambert scale by 1/L = 10(dB sensitivity/10).

Mean 1/Lambert DLS values were calculated for each visual field sector. Sector mean dB values were calculated as follows: mean dB = log(mean 1/Lambert) x 10.

At each time point, the rim area and field sensitivity were expressed as a proportion of the starting value. Proportion of initial DLS was plotted against proportion of initial rim area. Plots were analyzed for each subject for global, superotemporal, inferotemporal, superonasal and inferonasal rim area, and their respective corresponding field sectors. The sensitivity is plotted in both the dB scale and in 1/Lambert scale with a significant linear correlation taken as an r2 value of 0.1 or greater with a corresponding p value of 0.05 or less.

A positive slope value indicates that both the visual field sensitivity and rim area are changing in the same direction. A negative value would indicate one parameter changing in the opposite direction to the other, for example, the DLS value increasing and the rim area decreasing. A slope value of 1.00 occurs when both DLS and the rim area decrease at the same proportional rate. The same would be true if both values improved at the same rate, but this should not be observed in progressive glaucoma. A slope value of less than 1.00 indicates that there is a slower visual field change compared to rim area. A slope value of greater than 1.00 suggests a faster change to the visual field compared to the rim area.

By analyzing the visual field sensitivity and the rim area in this manner, we are examining the relationship between the two in individual eyes at isolated time points. Therefore, this does not make assumptions about the pattern of rim or field change (i.e., whether it is linear, curvilinear or episodic) over time.

The hypothesis that stage of disease may modify the relationship between change

Fig. 2. Visual field template.

in rim area and change in field was tested by plotting the rim/field slope values against starting cup shape measure, starting rim area and starting visual field sensitivity.

Results

Seventeen eyes of 17 subjects were studied, with nine coming from the ocular hypertension conversion group and eight from the optic disc hemorrhage group (seven of these went on to develop glaucomatous disc changes).

The subjects’ demographics are summarized in Table 1.

Table 2 shows the subjects in whom there was a significant (p less than 0.05) decline in rim area, 1/Lambert and dB over time. The negative slope value in all cases is a reflection of the general decrease in value of rim area and DLS over time in these subjects. In fact, this was observed in all cases, although only eight achieved levels of significance for rim area and 1/Lambert and nine for decibel light sensitivity.

The overall results for the global field and disc parameters in 1/Lambert scale are depicted in Table 3 and for the decibel scale in Table 4. All subjects with significant correlations (r2 greater than 0.1 and p value less than 0.05) are shown and are ranked in decreasing magnitude of slope value. In the dB scale, four of 18 subjects have a significant correlation and five in the 1/Lambert scale. The median slope value for 1/ Lambert was 2.45 (range, 1.09-3.52); median slope was much shallower in dB scale at 0.5 (range, 0.42-1.29).

Taking each disc segment tested individually, most significant correlations were found in the inferotemporal segment (Table 5). Overall, the inferotemporal disc had

Table 2. Decline of rim area and field sensitivity over time

Table 5. Significant correlations between change in segmental disc rim area and change in field sensitivity

and successful treatment of IOP (Graph 10). As both DLS and rim area change in the same direction, the corresponding plots show positive slopes (Graphs 11 and 12). Once again, the slope is far steeper on the 1/Lambert scale compared to the dB scale, suggesting relatively greater change in 1/Lambert field sensitivity compared to the rim area.

Discussion

There is a wide variation in the slope values observed in this study, although only a small number achieved statistical significance. Such wide variation may indicate that the relationship may vary between individual eyes, and may be modified by stage of disease. We hypothesized that, if both visual field sensitivity and rim area decayed at the same rate, then a slope value of 1.00 would be observed.

A slope value different from 1.0 may be explained by disassociation between changes in disc structure and visual field. It is possible that some functional changes may not necessarily be accompanied by structural changes to the disc. This might arise if there was ganglion cell dysfunction without ganglion cell death in the presence of high IOP. It is also possible that there may be changes to the disc structure which do not necessarily result in functional change, such as conformational changes to the supporting tissue without axonal loss. An alternative explanation is that the changes to the disc and visual field were disassociated in time, with the disc changes occurring at a different time point from the corresponding change to the field.

The rim conformation at the start of testing may also influence the rim/field change slope value. The results from this particular study were unable to demonstrate a significant relationship between starting cup shape measure and slope value. We were also unable to show a significant relationship between starting rim area and slope, as well as starting visual field sensitivity and slope.

From this study, we can conclude that there was a wide variation of visual field sensitivity change compared to rim area change over time. The median slope value in 1/Lambert scale was 2.45, steeper than the hypothesized value of 1.00. This means that the 1/Lambert scaling has a tendency to predict an apparent faster change to the visual field compared to the rim area. In contrast, the dB scale had a median slope

value of 0.50, and the slope values were shallower than 1.00. This suggests a slower change to the visual field compared to the rim area. The median slope value with the closest approximation to 1.00 was the inferotemporal segment, as scaled in 1/Lambert (0.92). This was also the disc segment with the greatest number of significant correlations.

There are some issues regarding the image analysis which may have affected our study. Contour line misalignment, image size change, and image quality concerns resulted in 16% of images being excluded from the analysis. Of these factors, misalignment was the most frequent, and this is already being addressed in the continuing development of the Explorer software. In addition to contour line misalignment, images were also discarded, but to a much lesser degree, due to variation in image magnification. The reference plane is set at the fixed standard reference plane in the current Explorer software. It is likely that, in the subjects tested, the height of the disc rim varies with disease progression and, as such, a fixed reference plane may not be optimal. An experimental reference plane tailored for each optic nerve head, such that it is located entirely below the optic nerve head, at a level where the variability is least and at a fixed distance from the optic nerve head topographical z-axis, may yield more uniform, predictable, and less variable reference plane data than the standard HRT reference plane.12 Another criticism of our study is that the method for calculating global and sector DLS using the Progressor output does not take into account the influence of eccentricity upon local retinal sensitivity.

Recent longitudinal studies have shown that disc changes are observed more often than visual field changes in patients with glaucoma.13 These findings suggest that, with current methods of testing, the disc structure may be a more reliable indicator of disease progression than the visual field. It is unclear what is the most appropriate scaling for DLS from the comparison to rim change in this study, and work on a larger series is underway to establish the optimal scaling of DLS in evaluating the relationship between structure and function.

Acknowledgment

NGS is supported by a Friends of Moorfields Research Fellowship and by an unrestricted grant from Heidelberg Engineering.

References

1.Sommer A, Pollack I, Maumenee AE: Optic disc parameters and onset of glaucomatous field loss. I. Methods and progressive changes in disc morphology. Arch Ophthalmol 97:1444-1448, 1979

2.Motolko M, Drance SM: Features of the optic disc in preglaucomatous eyes. Arch Ophthalmol 99:1992-1995, 1981

3.Pederson JE, Anderson DR: The mode of progressive disc cupping in ocular hypertension and glaucoma. Arch Ophthalmol 98:490-495, 1980

4.Airaksinen PJ, Tuulonen A, Alanko HI: Rate and pattern of neuroretinal rim area decrease in ocular hypertension and glaucoma. Arch Ophthalmol 110:206-210, 1992

5.Airasinen PJ, Drance SM, Douglas GR, Schulzer M: Neuroretinal rim areas and visual field indices in glaucoma. Am J Ophthalmol 99:107-110, 1985

6.Funk J, Bornscheuer C, Grehn F: Neuroretinal rim area and visual field in glaucoma. Graefe’s Arch Clin Exp Ophthalmol 226:431-434, 1988

7.Nanba K, Iwata I: Correlation of optic disc changes and visual field defects in glaucoma. In: Mills RP (ed) Perimetry Update 1992/1993, pp 165-169. Amsterdam/New York: Kugler Publ 1992

8.Garway-Heath DF, Viswanathan A, Westcott M, Kamal D, Fitzke FW, Hitchings RA: Relationship between perimetric light sensitivity and optic disc neuroretinal rim area. In: Wall M, Wild JM (eds) Perimetry Update 1998/1999, pp 381-389. The Hague: Kugler Publ 1998

9.Garway-Heath DF, Caprioli J, Fitzke FW, Hitchings RA: Scaling the hill of vision: the physiological relationship between light sensitivity and ganglion cell numbers. Invest Ophthalmol Vis Sci 41:17741782, 2000

10.Garway-Heath DF, Holder GE, Fitzke FW, Hitchings RA: Relationship between electophysiological, psychophysical, and anatomical measurements in glaucoma. Invest Ophthalmol Vis Sci 43:22132220, 2002

11.Johnson KO, Hsiao SS, Yoshioka T: Neural coding and the basic law of psychophysics. Neuroscientist 8(2):111-121, 2002

12.Poinoosawmy D, Tan J, Hitchings RA: Reference plane definition and its influence on optic nerve head rim area measurement. Invest Ophthalmol Vis Sci 42:S132, 2001

13.Chauhan BC, McCormick TA, Nicolela MT, LeBlanc RP: Optic disc and visual field changes in a prospective longitudinal study of patients with glaucoma. Arch Ophthalmol 119:1492-1499, 2001