SEARCH

SEARCH BY CITATION

Keywords:

  • frequency-doubling perimetry;
  • glaucoma;
  • retinal nerve fibre layer;
  • visual field

Abstract.

  1. Top of page
  2. Abstract.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Purpose:  To study the topographic relationship of retinal sensitivity evaluated by frequency-doubling perimetry (FDT) in healthy subjects and patients with glaucoma.

Methods:  Hundred and thirty-four eyes from 134 subjects (72 patients with glaucoma and 62 healthy controls) were prospectively and consecutively selected. Only one eye of each subject was randomly selected for evaluation. All subjects underwent a full ophthalmic examination and a reliable FDT (full-threshold C-20-5 algorithm). Pearson correlation coefficients between threshold values within the same hemifield were calculated. Maps of related points were plotted according to these correlation coefficients.

Results:  In the control group, each FDT location strongly to moderately correlated with the other FDT locations in the same hemifield. In glaucoma subjects, only a few locations significantly correlated with other threshold values in the same hemifield. The strongest correlations were observed between neighbouring locations. The pairs of points with the strongest correlation corresponded to the inferior retinal regions. In general, perimetric maps showed the retinotopic distribution of ganglion cell axons in the retinal nerve fibre layer.

Conclusions:  The statistical correlations between the FDT threshold values in the same visual hemifield objectively highlight the structure–function relationship determined by the anatomic distribution of retinal nervous tissue. This structure is altered in patients with glaucoma.


Introduction

  1. Top of page
  2. Abstract.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Glaucoma is an optic neuropathy characterized by atrophy of the optic nerve because of the loss of retinal ganglion cells and their axons in the retina (Quigley 1999). Damage to the retinal nerve fibre layer (RNFL) is usually associated with corresponding visual field (VF) defects.

Frequency-doubling technology (FDT) perimetry was introduced in 1997 (Johnson & Samuels) and is based on a phenomenon described by Kelly (1966), who reported that the apparent spatial frequency of achromatic sinusoidal grating of low spatial frequency (<1 cycle/deg) appears to double when it undergoes counterphase flickering at a high temporal frequency (≥15 Hz). The low spatial frequency and high temporal frequency stimulus characteristics necessary for perception of the frequency-doubling illusion suggest that this phenomenon is mediated by magnocellular cells (Maddess & Henry 1992). This illusion is useful for detecting glaucomatous field loss (Quigley 1998; Burnstein et al. 2000; Casson et al. 2001) because it is these large-diameter ganglion cells that are lost in the course of the disease.

The applicability of FDT for glaucoma diagnosis was reported in 1981, and FDT has since proven to be useful in detecting glaucoma (Tyler 1981; Korth et al. 1989; Fogagnolo et al. 2008).

Morphologic and functional tests are now commonly used together to diagnose glaucoma and to follow up glaucomatous patients (Heeg & Jansonius 2009; Horn et al. 2011). Nevertheless, some studies (Landers et al. 2003; Ferreras et al. 2007a) have demonstrated that in many cases, FDT can detect VF defects even earlier than standard automated perimetry. On the one hand, therefore, this technology appears to be more sensitive for early glaucoma diagnosis than other VF tests, such as those using a white-on-white stimulus. On the other hand, however, the FDT stimulus has a lower dynamic range for useful follow-up of patients with glaucoma.

The aim of this study was to determine the relationship between FDT threshold values within the same hemifield in healthy individuals and in patients with glaucomatous optic neuropathy. Although the threshold sensitivity of the C-20-5 algorithm has been studied (Cello et al. 2000; Wadood et al. 2002), in the present study, we determined these measures using single test locations in the same hemifield in each study group. These findings could be useful for improving the interpretation of FDT VFs as a tool for detecting early glaucomatous defects.

Methods

  1. Top of page
  2. Abstract.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Subjects

The prospective study protocol was approved by the Clinical Research Ethics Committee of Aragón (CEICA), and written informed consent was obtained from all participants. The study design followed the tenets of the Declaration of Helsinki for biomedical research. Only one randomly chosen eye from each subject was included in the study.

Inclusion criteria were age between 18 and 80 years, best-corrected visual acuity of 0.5 or better (Snellen scale), refractive errors <3 spherical dioptres and two dioptres in cylinder (Chauhan & Johnson 1999; Anderson & Johnson 2002), transparent ocular media (nuclear colour/opalescence, cortical or posterior subcapsular lens opacity <1) according to the Lens Opacities Classification System (LOCS) III system (Chylack et al. 1993) and open anterior chamber angle.

Subjects with previous intraocular surgery; diabetes or other systemic diseases; history of ocular, congenital or neurologic disease; current use of a medication that affects VF sensitivity; or inability to perform any of the protocol tests were excluded.

All subjects underwent full ophthalmologic examination, which included clinical history, best-corrected visual acuity, biomicroscopy of the anterior segment, gonioscopy, Goldmann applanation tonometry, central corneal ultrasonic pachymetry (model DGH 500, Exton, PA, USA), funduscopic examination using a 78-dioptre lens, stereophotographs of the optic disc (Canon CF 60 UV; Canon Inc, Tokyo, Japan) and Humphrey FDT (Welch Allyn, Skaneateles Falls, NY, USA and Carl Zeiss Meditec, Dublin, CA, USA).

Participants were classified as normal (control group) if the intraocular pressure (IOP) was <21 mmHg and the optic nerve head appeared normal. The glaucoma group comprised patients with a baseline IOP of 21 mmHg or higher and glaucomatous optic nerve head morphology. Glaucomatous damage was defined as focal (localized notching) or diffuse neuroretinal rim narrowing with concentric enlargement of the optic cup, or both (Tuulonen & Airaksinen 1991).

Frequency-doubling perimetry (FDT)

Frequency-doubling perimetry utilizes a vertical sine-wave grating of low spatial frequency (0.25 cycles/deg) that undergoes counterphase flickering at a high temporal frequency (25 Hz). The contrast of the stimulus is modified for each location in the VF. The C-20-5 presentation pattern tests the central 20°. In this sense, 10 × 10° targets are used, 4 per quadrant within the central 20°, along with one smaller central target (5°-diameter circle) projected onto the macular region to measure contrast sensitivity. This test presents targets at a contrast level that 95% of healthy age-matched subjects would be expected to detect. The 17 tested points were numbered from the nasal-superior position to temporal-inferior position. Left eye data were converted to a right eye format in such a way that position one corresponded to the nasal-superior location, position four to the temporal-superior location, position 13 to the nasal-inferior location and position 16 to the temporal-inferior location (Fig. 1). Position 17 corresponded to the central location (fixation point).

image

Figure 1.  Grid of visual field points tested by the full-threshold C-20-5 algorithm of frequency-doubling perimetry. These locations were numbered as shown based on the side of the eye.

Download figure to PowerPoint

During the FDT examination, the targets were imaged at optical infinity; thus, patients wore their distance-corrected lenses while performing the test. Optimal near correction was not needed because large low-spatial-frequency targets are not significantly affected by refractive errors up to six dioptres (Johnson & Samuels 1997). All FDT tests were performed under low-illumination conditions. Fixation loss and false positive rate were <1/6, and the false negative rate was <1/3. If these rates were exceeded, the test was repeated until the examination was reliable.

Statistical analyses

Statistical analyses were performed with the Statistical Package for the Social Sciences (spss 17.0; SPSS Inc., Chicago, IL, USA) and MedCal 11.0.1. (MedCal Software, Mariakerke, Belgium) The size of both groups was determined by the Schlesselmann method (Schlesselmann 1982) for unpaired case–control studies.

Parametric tests were used because the distribution of the FDT results was determined to be normal (based on the Kolmogorov–Smirnov test). Pearson correlations between threshold values of FDT-tested points within each hemifield were calculated (superior and inferior hemifields were considered separately). Raw sensitivities measured at each test point are indicated in decibels (dB), which are tenths of a log unit (logarithmic scale). The dB levels in each location of the raw numeric plot were converted to a linear scale before calculating the correlations between them.

Results

  1. Top of page
  2. Abstract.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

The study included 134 subjects: 62 control eyes of 62 healthy individuals (mean age 59.65 ± 10.62 years) and 72 eyes of patients with glaucoma (mean age 61.55 ± 7.65 years). The baseline IOP and the cup/disc ratio, evaluated in stereophotographs, FDT mean deviation (MD) and FDT pattern standard deviation (PSD) differed significantly between groups (Table 1). Age, best-corrected visual acuity and central corneal thickness were not significantly different between the two groups.

Table 1.   Demographic and clinical characteristics of the study population.
 Control group (n = 62)Glaucoma group (n = 72)p-value*
Mean (SD)Mean (SD)
  1. SD = standard deviation; BCVA = best-corrected visual acuity; IOP = baseline intraocular pressure (without treatment); CD = vertical cup-to-disc ratio in stereophotographs; MD = mean deviation; PSD = pattern standard deviation; FDT = Frequency-doubling perimetry.

  2. * Student’s t-test between control and patients with glaucoma.

Age (years)59.65 (10.62)61.55 (7.65)0.226
BCVA (Snellen)0.90 (0.10)0.83 (0.13)0.069
IOP (mmHg)14.57 (2.28)24.31 (3.89)<0.001
CD0.30 (0.15)0.73 (0.18)<0.001
Pachymetry (μm)549.42 (29.43)544.81 (36.21)0.423
MD of FDT (dB)−0.75 (2.59)−3.73 (4.49)<0.001
PSD of FDT3.81 (0.72)5.76 (2.92)<0.001

The correlation analyses of threshold values obtained with FDT are shown in Tables 2 and 3. To simplify the interpretation of these results, two correlation maps were drawn, one for the control group (Fig. 2) and other for the glaucoma group (Fig. 3). The FDT C-20-5 algorithm analyses a typical 4 × 4 grid of 16 locations that are each 10 × 10° and a foveal circular grid location of 5° (fixation point, Fig. 1). To generate the maps shown in Figs 2 and 3, each of the 16 locations was overlaid with a 4 × 4 FDT C-20-5 grid that shows the degree of correlation between each location with that of the other locations. Correlations with the fixation point were omitted. For interpretation of these maps, the following colour code was used: the location being evaluated was marked in black, the location with the strongest correlation was coloured dark blue, other locations with strong-to-moderate correlation were coloured light blue, and the remaining locations are coloured white. The degree of correlation was categorized as strong when r ≥ 0.6, moderate when 0.6 > r ≥ 0.4 and mild when r < 0.4.

Table 2.   Pearson’s correlation coefficients (r) and significance levels for two-tailed test (p) between the frequency-doubling technology (FDT) threshold values in eyes of healthy subjects.
 Location 1Location 2Location 3Location 4Location 5Location 6Location 7Location 8Location 9Location 10Location 11Location 12Location 13Location 14Location 15Location 16
Location 1r 0.4610.5360.5550.8220.5130.5720.796        
p <0.001<0.001<0.001<0.001<0.001<0.001<0.001        
Location 2r0.461 0.6030.6040.4840.7820.5630.403        
p<0.001 <0.001<0.001<0.001<0.001<0.0010.002        
Location 3r0.5360.603 0.5750.6170.6780.6090.420        
p<0.001<0.001 <0.001<0.001<0.001<0.0010.001        
Location 4r0.5550.6040.575 0.5910.6420.6270.364        
p<0.001<0.001<0.001 <0.001<0.001<0.0010.005        
Location 5r0.8220.4840.6170.591 0.6540.6980.804        
p<0.001<0.001<0.001<0.001 <0.001<0.001<0.001        
Location 6r0.5130.7820.6780.6420.654 0.7550.369        
p<0.001<0.001<0.001<0.001<0.001 <0.0010.004        
Location 7r0.5720.5630.6090.6270.6980.755 0.520        
p<0.001<0.001<0.001<0.001<0.001<0.001 <0.001        
Location 8r0.7960.4030.4200.3640.8040.3690.520         
p<0.0010.0020.0010.005<0.0010.004<0.001         
Location 9r         0.7160.7860.6250.4420.6620.6490.440
p         <0.001<0.001<0.0010.001<0.001<0.0010.001
Location 10r        0.716 0.7530.5770.4720.6050.5720.496
p        <0.001 <0.001<0.001<0.001<0.001<0.001<0.001
Location 11r        0.7860.753 0.6160.5280.6660.6290.587
p        <0.001<0.001 <0.001<0.001<0.001<0.001<0.001
Location 12r        0.6250.5770.616 0.6120.5980.5130.444
p        <0.001<0.001<0.001 <0.001<0.001<0.001<0.001
Location 13r        0.4420.4720.5280.612 0.6770.4390.429
p        0.001<0.001<0.001<0.001 <0.0010.0010.001
Location 14r        0.6620.6050.6660.5980.677 0.7680.613
p        <0.001<0.001<0.001<0.001<0.001 <0.001<0.001
Location 15r        0.6490.5720.6290.5130.4390.768 0.715
p        <0.001<0.001<0.001<0.0010.001<0.001 <0.001
Location 16r        0.4400.4960.5870.4440.4290.6130.715 
p        0.001<0.001<0.001<0.0010.001<0.001<0.001 
Central locationr0.5100.4200.5890.4880.6950.6770.5800.4500.6730.6630.6920.7130.5700.7180.5420.426
p<0.0010.001<0.001<0.001<0.001<0.001<0.001<0.001<0.001<0.001<0.001<0.001<0.001<0.001<0.0010.001
Table 3.   Pearson’s correlation coefficients (r) and significance levels for two-tailed test (p) between the frequency-doubling technology (FDT) threshold values in eyes of patients with glaucoma. Values significant at the 5 % level are marked in bold.
  Location 1Location 2Location 3Location 4Location 5Location 6Location 7Location 8Location 9Location 10Location 11Location 12Location 13Location 14Location 15Location 16
Location 1r −0.031−0.0200.2760.481−0.0460.2400.153        
p 0.8050.8750.026<0.0010.7170.0540.225        
Location 2r−0.031 0.967−0.042−0.0690.9690.024−0.029        
p0.805 <0.0010.7410.584<0.0010.8480.820        
Location 3r−0.0200.967 −0.023−0.0400.9970.070−0.041        
p0.875<0.001 0.8570.749<0.0010.5810.748        
Location 4r0.276−0.042−0.023 0.563−0.0630.6230.580        
p0.0260.7410.857 <0.0010.618<0.001<0.001        
Location 5r0.481−0.069−0.0400.563 −0.0720.3110.558        
p<0.0010.5840.749<0.001 0.5680.012<0.001        
Location 6r−0.0460.9690.997−0.063−0.072 0.005−0.061        
p0.717<0.001<0.0010.6180.568 0.9660.632        
Location 7r0.2400.0240.0700.6230.3110.005 0.444        
p0.0540.8480.581<0.0010.0120.966 <0.001        
Location 8r0.153−0.029−0.0410.5800.558−0.0610.444         
p0.2250.8200.748<0.001<0.0010.632<0.001         
Location 9r         0.2030.2430.1160.2210.6670.7110.301
p         0.1040.0510.3570.076<0.001<0.0010.015
Location 10r        0.203 0.4460.0600.0560.2210.2100.565
p        0.104 <0.0010.6360.6600.0760.092<0.001
Location 11r        0.2430.446 0.0890.7100.3490.5580.567
p        0.051<0.001 0.480<0.0010.004<0.001<0.001
Location 12r        0.1160.0600.089 0.0350.2060.1460.614
p        0.3570.6360.480 0.7800.0990.246<0.001
Location 13r        0.2210.0560.7100.035 0.1820.4530.122
p        0.0760.660<0.0010.780 0.146<0.0010.334
Location 14r        0.6670.2210.3490.2060.182 0.6480.433
p        <0.0010.0760.0040.0990.146 <0.001<0.001
Location 15r        0.7110.2100.5580.1460.4530.648 0.394
p        <0.0010.092<0.0010.246<0.001<0.001 0.001
Location 16r        0.3010.5650.5670.6140.1220.4330.394 
p        0.015<0.001<0.001<0.0010.334<0.0010.001 
Central locationr0.370−0.008−0.0310.5800.518−0.0540.4620.6740.5790.5130.5580.1560.2800.6400.6050.546
p0.0020.9500.809<0.001<0.0010.672<0.001<0.001<0.001<0.001<0.0010.2140.024<0.001<0.001<0.001
image

Figure 2.  Correlation map between frequency-doubling perimetry threshold values in the control group. The point being evaluated appears in black, the location with the strongest correlation is coloured dark blue, other locations with strong or moderate correlations are coloured light blue, and the remaining locations are coloured white.

Download figure to PowerPoint

image

Figure 3.  Correlation map between frequency-doubling perimetry threshold values in patients with glaucoma. The point being evaluated appears in black, the location with the strongest correlation is coloured dark blue, other locations with strong or moderate correlations are coloured light blue, and the remaining locations are coloured white.

Download figure to PowerPoint

In the control group, each FDT location strongly to moderately correlated with the other FDT locations within the same hemifield. In the glaucoma group, only a few locations correlated significantly with other threshold values in the same hemifield, and the correlations were strong to moderate.

Discussion

  1. Top of page
  2. Abstract.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Studies of the threshold sensitivity of the VF are widely used for the detection and follow-up of glaucomatous damage. Many different types of perimeters and numerous perimetric techniques are used.

Frequency-doubling perimetry perimetry offers the potential advantage of reducing the testing time, thereby decreasing the patient’s loss of concentration and fatigue, which can decrease the reproducibility and accuracy of the tests (Wild et al. 1991).

Here, we found significant correlations between the threshold values of nearby locations and between the threshold values of very distant points that are close to the blind spot (optic nerve). Our findings suggest high functional dependence between distant VF points, according to the retinal ganglion cell anatomy. Similar results were reported by Gonzalez de la Rosa et al. (2002), who quantified the interpoint correlations of threshold values within the VF in patients with glaucoma using the Octopus 32 program. Unlike Gonzalez de la Rosa et al. (2002), however, we performed an independent analysis of the superior and inferior hemifields. Because of the asymmetry in the normal distribution of the neuroretinal rim [inferior-superior-nasal-temporal (ISNT) rule], an asymmetric map with respect to the horizontal line would be expected, and even more so in patients with glaucoma. The ganglion cell fibres in the retina are arranged in fascicles, which run parallel up to the vicinity of the optic disc (Fitzgibbon & Taylor 1996; FitzGibbon 1997). This circumferential organization is maintained in such a way that the area surrounding the optic disc and adjacent axons are located in the same sector. Within the fascicles, the fibres do not have a strict radial organization. As the fascicles of fibres pass into the optic nerve head, there is some change in the position, disrupting the retinal circumferential organization. The poor radial organization of the fibres in the retinal fascicles persists as the fibres pass through the intraocular segment of the nerve. Only a few fibres on the temporal side can cross to the opposite hemisphere. So, based on this fact, we considered it more appropriate to perform an independent analysis for each hemifield.

The maps that we developed (Figs 2 and 3) represent the functional interrelationship and provide a perimetric model for FDT. Furthermore, these charts suggest the circumferential distribution of the points with related retinal sensitivity in relation to the retinotopic distribution of the ganglion cell axons in the RNFL. Taking into consideration that the FDT analyses wide locations, however, the typical curved distribution of the ganglion cell axons was not always evident.

The strongest correlations were observed between neighbouring locations. Several cell physiology studies have suggested the existence of secondary cellular neurodegeneration in glaucoma neuropathy. Subsequent to the primary neuronal damage associated with optic neuropathies such as glaucoma, retinal ganglion cells that survive the primary insult are injured by toxic effects of the primary degenerating neurons (Schori et al. 2001; Kaushik et al. 2003; Brao-Osuna et al. 2007). This effect is attributed to the release of excitatory amino acids, and glutamate release in particular is strongly implicated in this secondary retinal ganglion cell degeneration described in glaucoma. This could explain why the strongest correlations were observed between the closest FDT locations.

In the same way, the pairs of points that showed the strongest correlations were 3–6, 2–6 and 2–3. Alterations in papillary morphology are a parameter with a high predictive value in glaucomatous patients (Mardin et al. 1999). Glaucoma is characterized by a progressive narrowing of the neuroretinal rim that is usually more pronounced in the inferior and superior poles and is especially marked in the inferior sector (Tuulonen & Airaksinen 1991; Jonas et al. 1993). This is consistent with the fact that the thickness of the RNFL is greater and more susceptible to glaucomatous damage in these areas around the optic nerve (ISNT rule). Taking into consideration that the visual image focused on the retina is inverted top to bottom and reversed right to left, we expected that the first VF defects in glaucoma would be observed in the superior regions, such as FDT locations 2, 3 and 6, areas with the highest correlations in our study.

In our study, the normal control group had no previous VF testing experience, although most of the patients with glaucoma had previously undergone standard automated perimetries. All control subjects had FDT values within normal limits (Ferreras et al. 2007b) and did not undergo a second perimetry when the first one was reliable, because training should always go in the direction of improvement. Nevertheless, because of a learning effect, the normal control subjects might show a slight improvement in the FDT values if the test were to be repeated.

To summarize, the statistical correlations between the FDT threshold values in the same visual hemifield objectively highlight the structure–function relationship determined by the specific distribution of the retinal nervous tissue. The quantification of interpoint correlations could be useful for interpreting VF tests and optimizing the use of FDT to detect glaucomatous defects.

References

  1. Top of page
  2. Abstract.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References