Comparison of Rates of Change Between Binocular and Monocular Visual Fields

Abstract

Purpose.: To compare rates of change between binocular and monocular visual fields.

Methods.: The study included 1264 visual fields from 62 normal-tension glaucoma patients with a minimum of nine pairs of visual fields for at least 5 years of follow-up. Integrated binocular visual fields (BVFs) were calculated from the two monocular visual fields using a binocular summation. Linear regression of mean deviation (MD) values was used to evaluate the rates of change of the BVFs and monocular visual fields. For each patient, the eye with the worse MD value at baseline was defined as the worse MD eye. The eye with the faster rate of change of monocular visual fields was defined as the faster-changing eye.

Results.: The mean age of subjects was 61.8 years at baseline, the mean number of paired visual field tests was 10.2, and the mean follow-up was 8.1 years. The mean rate of change in the BVFs (−0.10 dB/y) was significantly slower than that of the faster-changing eyes (−0.34 dB/y) and faster than that of the slower-changing eyes (−0.06 dB/y; P < 0.001 for both comparisons). Forty-five eyes (64.5%) among the worse MD eyes at baseline were identified as faster-changing eyes at last follow-up, and having a worse MD value at baseline was a risk factor for being the faster-changing eye (P = 0.025).

Conclusions.: The rate of change in BVFs was intermediate between the rates of the faster-changing and slower-changing eyes.

Introduction

The fundamental goal of glaucoma management is to prevent patients from developing visual impairment that produces disability in their daily lives. Evaluation of rates of visual field change is essential in assessing the risk of functional impairment. Rates of visual field progression can be highly variable among glaucoma patients.^1–4 Most patients progress relatively slowly, but others have aggressive disease with rapid deterioration that can eventually result in blindness or substantial impairment in the absence of appropriate interventions. Therefore, estimation of rates of change has a great impact on disease management. 

Although previous studies have evaluated rates of change in monocular visual fields for monitoring glaucoma,^2,3,5,6 very little is known about rates of change in binocular visual fields (BVF).^7,8 Recent studies have shown that functional losses measured by BVFs show a better relationship with patient-reported quality of vision compared with losses measured by monocular fields.^9,10 Therefore, evaluation of rates of change using BVFs could provide a better method for assessment of the risk of functional impairment in glaucoma.^7–12 Assessment of rates of BVF change has been limited by the fact that ‘true' BVF tests are not routinely performed in clinical practice with the Humphrey Visual Field Analyzer. However, several methods have been proposed to integrate results of monocular fields in order to predict the binocular field.^13,14 Among them, estimates of binocular sensitivity obtained using a binocular summation model and maximal sensitivity model have been shown to provide the best approximation to sensitivity values obtained by true binocular testing.¹⁴ 

Several studies have evaluated the relationship between these integrated BVFs and monocular visual fields.^14–17 The threshold sensitivities of each location of the BVFs are generally better than those of monocular visual fields,¹⁴ and in particular, the mean deviation (MD) of the BVF is better than that of the monocular visual field of the better eye.^15,16 Binocular visual fields have also demonstrated a good relationship with functional disability in glaucoma^{9–12,15,17,18} and, furthermore, the BVF is better correlated with quality of life than monocular visual fields.^11,12,15 However, no study has compared rates of change between the BVF and monocular visual fields in a longitudinal study. 

The purpose of this study was to evaluate rates of change in estimated BFVs and compare them with rates of change obtained from monocular visual fields in glaucoma patients. 

Methods

This retrospective observational study was approved by the Chung-Ang University Hospital institutional review board and adhered to the tenets of the Declaration of Helsinki. Chart reviews were performed for subjects who were longitudinally followed for more than 5 years for the management of glaucoma with complete clinical examination and several imaging and functional tests. 

We included 1264 visual fields of 62 subjects with a diagnosis of normal-tension glaucoma (NTG) in at least one eye at the baseline visit. Normal-tension glaucoma was classified as having repeatable abnormal visual field test results, defined as a pattern standard deviation (PSD) with P less than 0.05, and/or glaucoma hemifield test (GHT) results outside normal limits; having the appearance of glaucomatous optic discs, defined by the presence of neuroretinal rim thinning, excavation, or localized or diffuse retinal nerve fiber layer defects indicative of glaucoma; and IOP in the normal range (≤21 mm Hg). To be included, subjects had to have a best corrected visual acuity of 20/40 or better, spherical refraction within ± 10.0 diopter (D), and cylinder refraction within ± 3.0 D. For determination of the rate of visual field progression, a reasonable number of tests within at least 5 years are needed.^19,20 In particular, six visual field tests in the first 2 years are very useful for the detection of rapid progression.^21,22 Thus, eligible patients were required to have a minimum of nine paired visual field tests (visual field tests of both eyes on the same day) and a minimum follow-up period of 5 years. 

We included 55 paired visual field tests of age-matched healthy subjects from a database at Chung-Ang University Hospital. These subjects visited our clinic for ocular examination due to suspicious glaucomatous disc, occupational medical certificates, or physical screening programs. To be included, healthy subjects had to have a best-corrected visual acuity of 20/40 or better, spherical refraction within ± 10.0 D, and cylinder refraction within ± 3.0 D. Table 1 shows the comparison between the NTG and age-matched healthy population groups. 

Table 1

 

Comparison of Clinical Characteristics of NTG and Normal Control Groups

	NTG	Healthy Control	PValue*
Number of included patients	62	55
Age (y, mean ± SD; range)	61.8 ± 6.4 (48.9–72.8)	61.6 ± 6.4 (50.3–73.6)	0.735
Female/Male, n (%)	34 (54.8)/28 (45.2)	34 (61.8)/21 (38.2)	0.447
Spherical equivalent (D; range)	−1.79 ± 3.2 (−9.25 to 3.25)	−0.82 ± 2.7 (−6.25 to 4.00)	0.113
MD of the right VF, dB†	−2.87 ± 4.12 (−3.71, −2.13, −0.02)	−0.04 ± 0.91 (−0.75, 0.15, 0.67)	<0.01
MD of the left VF, dB†	−3.26 ± 5.36 (−3.99, −1.90, −0.33)	−0.42 ± 0.95 (−1.34, −0.38, 0.43)	<0.01

*  P value according to the Mann-Whitney U test.

†  Nonnormally distributed variables, mean ± SD (first quartile, median, third quartile).

Monocular and Binocular Visual Fields

Monocular standard automated perimetry was performed using the 30-2 Swedish Interactive Threshold Algorithm (SITA) standard of the Humphrey Visual Field Analyzer (Carl Zeiss Meditec, Inc., Dublin, CA, USA) at all visits during the follow-up period. Only reliable tests (≤20% fixation losses, ≤33% false negatives, and ≤33% false positives) were included. Visual fields were reviewed and excluded due to the presence of artifacts, such as lid or rim artifacts, fatigue effects, inattention or inappropriate fixation. Visual fields were also reviewed for the presence of abnormalities that could indicate diseases other than glaucoma, such as homonymous hemianopia. 

For the calculation of the estimated pointwise sensitivities of the BVFs, the raw threshold sensitivities of the right and left eyes were used. The 30-2 stimulus presentation pattern consists of 76 points within the central 30° in a 6° grid bracketing the horizontal and vertical meridians. Each point in one eye has a corresponding location in the opposite eye. For example, superior temporal points in the right eye correspond to superior nasal points in the left eye. Since a ‘true' BVF test has no blind spots, the four points corresponding to the blind spots were included in the analyses, although they did not have a spatial correspondent in the visual field of the fellow eye. Therefore, each individual BVF had a total of 76 overlapping points, as shown on Figure 1. 

Figure 1

 

Schematic representation of the left, right, and integrated BVF. Threshold sensitivities from right and left eyes were integrated using the summation method and four points (black squares in the BVF) corresponding to the blind spots of both eyes were included for the calculation of BVF.

The sensitivity for each point of the BVF was estimated using the binocular summation model, described by Nelson-Quigg et al.¹⁴ According to this model, the binocular sensitivity can be estimated using the following formula:  

 

where Sr and Sl are the monocular threshold sensitivities for corresponding visual field locations in the right and left eyes, respectively. In order to calculate the binocular sensitivity from this formula, light sensitivity had to be converted to a linear scale (apostilbs) and then converted back to a logarithmic scale (decibels).  

For calculation of monocular and BVF MDs, we applied the formula described by Anderson et al.²³:  

 

where L is the number of locations in the visual field, TDi is the total deviation in decibels from the age-matched healthy population at location i, and Si² is the variance of point i in the BVF of the age-matched healthy population. We used data from 55 paired visual fields of age-matched healthy subjects to calculate the TDi and the Si² of the monocular fields and BVFs.  

Statistical Analysis

Clinical and demographic variables were tested for normality using the Anderson-Darling test. Descriptive statistics included the mean and SD for normally distributed variables and the median, first quartile and third quartile for nonnormally distributed variables. Nonnormally distributed variables between two related samples were compared using the Wilcoxon rank sum test. To compare the characteristics of the NTG group and age-matched healthy group, the Mann-Whitney U test was used. To evaluate MD values at baseline as a possible risk factor for being the faster-changing eye, Pearson's χ² test was used. 

To assess rates of visual field change, ordinary least-squares linear regression of monocular and BVF MD values over time (dB/y) was performed for each eye and patient, respectively. The beta coefficient (slope) of the regression equation was considered to be the best estimate of the rate of change. For each patient, the eye with the better MD value at baseline was defined as the better MD eye and the other eye was defined as the worse MD eye. The eye with the slower rate of change (more positive slope) in its monocular visual field was defined as the slower-changing eye, and the eye with the faster rate of change (more negative slope) was defined as the faster-changing eye. 

Statistical analyses were performed using SPSS software version 19.0 (PASW, ver. 19.0; SPSS, Inc., Chicago, IL, USA). The alpha level (type I error) was set at 0.05. 

Results

The study included 1264 visual fields (632 paired visual fields) from 62 patients. The mean number of paired visual fields per patient was 10.2 ± 1.2, and the mean follow-up time was 8.1 ± 2.1 years. The average MD of all 62 BVFs was −1.35 dB and it was significantly better than that of the individual right and left eyes (−2.87 dB and −3.26 dB, respectively, P < 0.001). The average MD of the BVFs was significantly better than that of the worse MD eyes (−1.35 dB vs. −4.67 dB, P < 0.001) at baseline. In addition, the average MD of the BVF was better than that of the better MD eye even though there was no statistical significance (−1.35 dB vs. −1.47 dB, P = 0.611) at baseline. Table 2 shows the clinical and demographic characteristics of the study patients. 

Table 2

 

Demographic and Clinical Characteristics of Study Patients

Number of included patients	62
Age at baseline, y* (range)	61.8 ± 6.4 (48.9–72.8)
Visual field (VF)
Total number of VFs (number of pairs)	1264 (632)
Mean number of paired VFs per patient (range)*	10.2 ± 1.2 (9–14)
Right VF MD at baseline, dB†	−2.87 ± 4.12 (−3.71, −2.13, −0.02)
Left VF MD at baseline, dB†	−3.26 ± 5.36 (−3.99, −1.90, −0.33)
Better VF MD at baseline, dB†	−1.47 ± 2.60 (−3.21, −0.92, 0.22)
Worse VF MD at baseline, dB†	−4.67 ± 5.82 (−6.19, −3.09, −0.81)
BVF MD at baseline, dB†	−1.35 ± 2.48 (−2.83, −1.15, 0.46)
Follow-up, y (range)†	8.1 ± 2.1 (5.1–12.2)

 VF, visual field.

*  Normally distributed variables, represented by mean ± SD.

†  Nonnormally distributed variables, represented by mean ± SD (first quartile, median, third quartile).

The mean rate of change in the BVFs was −0.10 dB/y (± 0.22), and the distribution of progression rates of the BVFs ranged from −0.86 to +0.29 dB/y with a negatively skewed deviation (Fig. 2). Table 3 summarizes rates of change in the BVFs and monocular visual fields from 62 patients. The rate of change in the BVFs was significantly faster than the rate of change in the slower-changing eyes (−0.10 dB/y vs. −0.06 dB/y, P < 0.001), and significantly slower than of the rate of change of the faster-changing eyes (−0.10 dB/y vs. −0.34 dB/y, P < 0.001). A boxplot of the slopes for the BVFs, the slower-, and the faster-changing eyes demonstrated that the rate of change in the BVFs appeared to be between that of the slower-changing eyes and the faster-changing eyes, although it is close to the slower-changing eyes (Fig. 3). 

Figure 2

 

Rates of change in integrated binocular visual fields expressed in decibels/year. The distribution shows a negative tail among the faster-changing eyes.

Figure 3

 

Boxplot for slopes of binocular and monocular visual fields. The slope of the integrated binocular visual field was intermediate between those of the slower- and faster-changing eye visual fields.

Table 3

 

Rates of Change in Integrated Binocular and Monocular Visual Fields From 62 Patients

	Slope, dB/y, mean ± SD	First, Second, Third Quartiles	PValue*
Integrated binocular VF	−0.10 ± 0.22	−0.19, −0.06, 0.05
Slower-changing eye	−0.06 ± 0.21	−0.14, −0.11, 0.11	<0.001
Faster-changing eye	−0.34 ± 0.56	−0.55, −0.18, −0.03	<0.001

*  For pairwise comparisons between the slope of the binocular visual field and the slope of the slower- or faster-changing eye according to the Wilcoxon rank sum test.

Forty (64.5%) of the worse MD eyes at baseline were identified as the faster-changing eye at last follow-up. Being classified as the worse MD eye at baseline was a risk factor for being the faster-changing eye (Pearson's χ² test, P = 0.025). However, in 35.5%, the better MD eye at baseline was identified as the faster-changing eye at last follow-up. 

The field series of one patient is shown in Figure 4 as an example. At baseline, his left eye was the better MD eye with an MD of −2.80 dB; however, the rate of change of the left eye identified it as the faster-changing eye with a slope of −0.29 dB/y, and the right eye was identified as the slower-changing eye with a slope of −0.12 dB/y. The rate of change of the BVF was −0.18 dB/y, a value between the slower and faster-changing eyes. 

Figure 4

 

Visual field series from one patient. The slope of the slower-changing eye was −0.12 dB/y and that of the faster-changing eye was −0.29 dB/y. The slope of the integrated binocular visual field was −0.19 dB/y, which was between the slower- and faster-changing eyes. The left eye, which had the better MD at baseline, was identified as the faster-changing eye at last follow-up.

Discussion

In this study, we determined rates of change in binocular and monocular visual fields of NTG patients who had a minimum nine pairs of visual fields with at least 5 years of follow-up. The rate of change in the integrated binocular visual fields was significantly faster than that of the slower-changing eyes and significantly slower than the rate of change of the faster-changing eyes. 

Understanding the relationship between rates of change in BVFs and monocular visual fields is an important aspect of management of glaucoma. We normally use both eyes at the same time and the fellow eye can compensate for the loss of visual function of the other. The BVF combines the sensitivities of both visual fields and, therefore, is likely to be a better representation of the patient's experience of the external world. In fact, there is some evidence that BVF damage is more strongly correlated with impairment in daily activities than monocular visual field damage^11,12,15 and 5-year forecasted visual field index (VFI) values using BVFs were more confident than those of monocular visual fields.⁸ Accordingly, the rate of binocular change is potentially a better tool than monocular change for predicting deterioration in quality of life. 

We used the estimated BVF from monocular visual fields instead of ‘true' binocular visual fields. Although ‘true' BVFs, such as the Easterman test in Humphrey Visual Field Analyzer and Octopus 900 perimeter (HAAG-STREIT AG, Koeniz, Switzerland), can be obtained using specific perimetric strategies, they are not routinely performed in clinical practice due to several limitations, including the absence of direct monitoring of fixation of both eyes and the binary test algorithm (pass or fail) in the Humphrey Visual Field Analyzer.^13,15 Among several different methods previously proposed for construction of the BVF from monocular fields, we used the summation method, as it has a superior correlation with the ‘true' binocular visual field exam compared with other approaches.¹⁴ The binocular summation or probability summation is calculated by the square root of the summed squares of the two monocular sensitivity values^14,17,24,25; therefore, it is essentially a composite of both eyes' threshold sensitivities. Many previous studies support binocular viewing as superior to monocular.^25–28 Values of binocular sensitivities by the binocular summation model are better than those of monocular sensitivities. Therefore, assessment of rates of change from the BVF constructed using the summation model is considered a valid method. Although the methodology used in our study to determine the BVF requires some calculations, it can be easily added into standard software used in visual-field instruments in order to provide clinicians with information about the BVF and rate of change in the BVF, which may help determine the risk of functional impairment. There are two other procedures that have been used for probability summation: a multiplicative probability summation²⁹ and the fourth-root summation of the two sensitivities.^30,31 The general effect of these models is the amplification of the activities of all relevant neural inputs. It would be useful to determine whether these procedures yield more favorable results than the current findings. 

The average MD of the BVF was better than that of the better MD eye even though there was no statistical significance. This result was very similar to that of a large-scale, cross-sectional study of 7543 subjects by Arora et al.¹⁷ who used a binocular summation method for the BVF. Additionally, our BVF MD result is similar to that of a cross-sectional study by Asaoka et al.¹⁶ even though they used different binocular integration method (maximal sensitivity) and different MD calculation method for binocular and monocular visual fields. However, to our knowledge, there are no previously published reports that compare rates of change between BVFs and monocular visual fields. 

The rate of change in the BVFs (−0.10 dB/y) was intermediate between those of the faster-changing eyes (−0.34 dB/y) and the slower-changing eyes (−0.06 dB/y). The MD of the BVFs was significantly better than those of individual eyes, and it was also better than those of the better MD eyes and the worse MD eyes in a cross-sectional analysis. However, this did not imply that the rate of change in the BVFs would be better than those of individual eyes when evaluated longitudinally. The rate of change in the BVFs was intermediate between rates of change for the individual eyes. Therefore, we should cautiously avoid the expectation that the rate of change in the BVF will be better than that of the better eye (or the slower-changing eye). The rate of change in the BVF provides useful information for determining the risk of functional deterioration in glaucoma.⁷ Therefore, it is important to consider not only the MDs of the BVFs, but also their progression rates for successful treatment of glaucoma, especially when potentially harmful treatment is considered for the faster-changing eye. 

The rate of change in the natural history of NTG without treatment was reported to be between −0.2 and −2.0 dB in the Early Manifest Glaucoma Trial Group³ and −0.36 dB/y in the Collaborative Normal-Tension Glaucoma Study Group.² In the current study, the rate of change in the fast progressing NTG eyes with treatment was −0.34 dB/y. This slope is consistent with earlier proposed reports,^2,3 and is better than other recent reports regarding the progression of Asian NTG patients.^32–34 This means that our study evaluated only patients with early glaucomatous visual field loss. 

Approximately 65% of the eyes classified as having worse MD at baseline were identified as the faster-changing eye at last follow-up. The worse MD eye at baseline (that is, the more progressed eye on detection) was a risk factor for faster progression. Our results are consistent with previous trials,^35,36 which indicated a higher risk of progression in eyes with worse MD values. However, recent studies^6,37 have reported conflicting results that there is no difference in progression rates depending on MD values, and that even eyes with worse baseline MD are associated with slower progression. These conflicting results imply that it is important to monitor both better and worse eyes with an equal amount of vigilance for treatment of glaucoma and maintenance of quality of life, despite controversy regarding whether the worse MD eye is at increased risk of progression. 

Our study had several limitations. We included patients with NTG in at least one eye at baseline. Consequently, the opposite, better eye had wide variance from being normal to having NTG at baseline. However, this approach was necessary in order to have an initial evaluation of the rate of change in BVFs in a population that resembles the rate of change found in the general population. Also, this study evaluated only patients with early glaucomatous visual field loss, and not patients with moderate or advanced visual field loss. The rate of change is probably more important for patients with moderate and advanced visual field loss from glaucoma. Therefore, attention is needed in the analysis of these study results. Another limitation is that we did not consider the influence of cataracts, which could be relevant because this study included patients with a mean age of 61.8 years at baseline and patients were followed for 5 or more years. The MD is more susceptible to the influence of media opacities than other global visual indices used to evaluate progression, such as the VFI.³⁸ However, the use of the VFI is not straightforward, as it takes into consideration probability values and eccentricity weighting factors. Also, VFI exhibits both a ceiling effect³⁹ (a substantial proportion of early glaucomatous visual fields with MDs better than −5 dB have a value of 100%) and a floor effect⁴⁰ (advanced glaucomatous visual fields with MDs close to −20 dB shift from a pattern deviation plot to a total deviation plot). The other limitation of our study was that we assumed a linear rate of BVF change over time. Several studies have suggested that functional changes do not follow a linear course during the natural course of the disease,^41–43 which might be related to the logarithmic scaling (decibel) of visual field sensitivity data. Nevertheless, the assumption of linear change is probably a reasonable one for short and medium follow-up periods, as performed in clinical practice. 

In conclusion, our results demonstrated that the rate of change in BVFs was significantly faster than that in the visual fields of the slower-changing eyes and slower than that in the visual fields of the faster-changing eyes. We expect that our findings will have significant implications for studies related to quality of life, monitoring of glaucoma, and providing therapeutic guidance. 

Acknowledgments

Disclosure: Y.S. Chun, None; J.-H. Shin, None; I.K. Park, None