Findings from a right eye with glaucomatous optic neuropathy, advanced cupping, and normal results of standard automated perimetry (SAP). Note the inferior visual field defect on short-wavelength perimetry (SWAP) and frequency doubling technology (FDT), and corresponding loss of retinal nerve fiber layer superiorly on scanning laser polarimetry with variable corneal compensation (GDx-VCC) and optical coherence tomography (OCT). TSNIT indicates temporal (T, TEMP), superior (S, SUP), nasal (N, NAS), inferior (I, INF), temporal.
Scatterplot of short-wavelength perimetry mean deviation (SWAP MD) against vertical cup-disc ratio (CDR) stratified on the basis of the statistical probability of an optical coherence tomography retinal nerve fiber layer (OCT RNFL) thickness abnormality using the instrument's normative database. Eyes with increased CDR demonstrated significantly greater (R = 0.35, P = .01) visual field loss by means of SWAP and a higher prevalence of significant RNFL atrophy (small diamonds) by means of OCT.
Distribution of visual field abnormalities by means of short-wavelength perimetry (SWAP) and frequency doubling technology (FDT) in eyes with mild (cup-disc ratio [CDR], <0.4; n = 8), moderate (CDR, 0.4-0.7; n = 21) and advanced (CDR, >0.7; n = 14) cupping.
Distribution of retinal nerve fiber layer (RNFL) abnormality by means of scanning laser polarimetry with variable corneal compensation (GDx-VCC) and optical coherence tomography (OCT) in eyes with mild (cup-disc ratio [CDR], <0.4; n = 8), moderate (CDR, 0.4-0.7; n = 21), and advanced (CDR, >0.7; n = 14) cupping.
Receiver operating characteristic curves for the most sensitive structural and psychophysical measures. AUC indicates area under the curve; FDT, frequency doubling technology; INF, inferior; MD, mean deviation; OCT, optical coherence tomography; RNFL, retinal nerve fiber layer; PSD, pattern standard deviation; SUP, superior; SWAP, short-wavelength perimetry.
Bagga H, Feuer WJ, Greenfield DS. Detection of Psychophysical and Structural Injury in Eyes With Glaucomatous Optic Neuropathy and Normal Standard Automated Perimetry. Arch Ophthalmol. 2006;124(2):169-176. doi:10.1001/archopht.124.2.169
Copyright 2006 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2006
To compare the prevalence of structural and psychophysical abnormalities in normal eyes and eyes with glaucomatous optic neuropathy (GON) and normal standard automated perimetry (SAP).
Complete examination, SAP, short-wavelength automated perimetry (SWAP), frequency doubling technology (FDT), scanning laser polarimetry (GDx-VCC), and optical coherence tomography (OCT) of the peripapillary retinal nerve fiber layer (RNFL), optic disc, and macula were performed. Glaucomatous optic neuropathy was defined as cup-disc asymmetry between fellow eyes of greater than 0.2, rim thinning, notching, excavation, or RNFL defect. All eyes had normal SAP. Abnormal measurements on OCT, GDx-VCC, SWAP, and FDT were defined as those outside 95% normal limits. Eyes were stratified into 3 groups based on the OCT-generated vertical cup-disc ratio: mild, moderate, and advanced cupping (cup-disc ratio of <0.4, 0.4-0.7, and >0.7, respectively). Receiver operating characteristic curves were developed to assess sensitivity and specificity of structural and functional assessments.
Forty-seven eyes of 47 patients (25 with GON and 22 normal) were enrolled (mean ± SD age, 58 ± 16 years; range, 25-83 years). Compared with normal eyes, eyes with GON had significantly worse mean deviation and pattern standard deviation by means of SWAP and FDT (P = .02-.05); OCT-derived mean and superior and inferior RNFL thickness (P = .008, <.001, and .05, respectively); mean macular thickness (P = .01), rim volume, rim area, cup-disc ratio, and cup area (all P<.001); and GDx-VCC nerve fiber indicator and inferior average (P = .03). There was a significantly (P = .008, .002, .003, and .01, respectively) greater prevalence of abnormalities identified by SWAP, FDT, OCT and GDx-VCC in eyes with advanced cupping (43%, 43%, 57%, and 57%, respectively) compared with mild cupping (0%) and moderate cupping (9%, 5%, 19%, and 33%, respectively).
Eyes with GON and normal results of SAP have significantly greater structural and psychophysical abnormalities than do normal eyes evaluated by means of OCT, GDx-VCC, SWAP, and FDT. Eyes with increased vertical cup-disc ratio are more likely to manifest such abnormalities on advanced diagnostic testing.
There is mounting evidence, based on longitudinal and histologic studies, that structural injury precedes visual field loss detectable by standard automated perimetry (SAP) in many eyes with early glaucomatous optic neuropathy (GON). Documentation of retinal nerve fiber layer (RNFL) atrophy in glaucomatous eyes was originally described by means of red-free photography.1 Quigley and colleagues2 demonstrated that 40% axonal loss may occur before any change in visual function is detected with perimetry. Sommer and colleagues3 reported that 60% of patients with ocular hypertension had evidence of RNFL loss that occurred up to 6 years before a detectable change in results of SAP; 88% of patients had evidence of RNFL loss at the time of visual field loss. Recently the Ocular Hypertension Treatment Study demonstrated that 55% of eyes that converted to glaucoma had isolated progressive optic disc structural damage without coexisting changes in visual function by means of SAP.4
Numerous reports have suggested that short-wavelength automated perimetry (SWAP) and frequency doubling technology (FDT) can detect glaucomatous damage earlier than SAP, based on their ability to target a specific subpopulation of retinal ganglion cells, thereby limiting redundancy.5- 11 There is evidence that FDT12 and SWAP6 abnormalities precede SAP visual field loss by as much as 4 and 5 years, respectively, in eyes with ocular hypertension and suspected glaucoma. Furthermore, both SWAP and FDT have been reported to demonstrate closer agreement with structural assessments of glaucomatous damage than SAP.13- 16 Recently, a longitudinal study involving 250 subjects with ocular hypertension demonstrated a greater baseline prevalence of SWAP deficits than SAP deficits; however, the incidences of new defects with SWAP and SAP were similar.17
Advances in posterior segment imaging technology provide objective, reproducible, and quantitative assessments of the peripapillary RNFL, optic disc, and macular thickness. Assessments of RNFL thickness by means of optical coherence tomography (OCT) and scanning laser polarimetry have been reported to demonstrate good correlation with SAP, high discriminating power for glaucoma detection, and close agreement with other imaging technologies.18- 21 Recent data22,23 have demonstrated that changes in macular thickness are detectable in glaucomatous eyes by OCT and are well correlated with changes in visual function and peripapillary RNFL atrophy. Optic disc analysis using OCT has been reported to demonstrate close agreement with manual disc assessment and confocal scanning laser tomography.24,25
Although there are reports that structural assessment of the RNFL by means of OCT and scanning laser polarimetry may discriminate normal eyes from eyes with ocular hypertension or suspected glaucoma,26- 28 conflicting evidence exists.29- 31 New evidence suggests that diffuse structural changes occur before the development of standard visual field loss in glaucomatous eyes with localized abnormalities on SAP.32 The purpose of the present investigation was to compare the prevalence of structural (RNFL, optic disc, and macula) and psychophysical abnormalities in normal eyes and eyes with GON and normal results of SAP, and to evaluate their relationship with the severity of glaucomatous cupping.
Normal and glaucomatous eyes meeting the eligibility criteria were enrolled in this prospective, cross-sectional study. Informed consent was obtained from all subjects by means of a consent form approved by the Institutional Review Board for Human Research of the University of Miami School of Medicine, Miami, Fla. All patients underwent complete ophthalmic examination including slitlamp biomicroscopy, gonioscopy, pachymetry, Goldmann applanation tonometry, dilated stereoscopic examination of the optic disc and fundus, SAP, SWAP, FDT, OCT, and scanning laser polarimetry with variable corneal compensation (GDx-VCC). One eye per subject was enrolled. If both eyes met the eligibility criteria, one eye was randomly selected.
Normal subjects had no history of ocular disease. All had intraocular pressure less than or equal to 21 mm Hg by Goldmann applanation tonometry, normal optic disc appearance based on clinical stereoscopic examination and review of stereo disc photography, and normal results of SAP. Absence of GON was defined as vertical cup-disc asymmetry less than 0.2, cup-disc ratio less than 0.6, and intact neuroretinal rim without peripapillary hemorrhages, notches, localized pallor, or RNFL defect. Normal visual field indexes were defined as a mean defect and corrected standard deviation within 95% normal limits and a glaucoma hemifield test result within normal limits.
Glaucomatous optic neuropathy was defined as either cup-disc asymmetry between fellow eyes of greater than 0.2, rim thinning, notching, excavation, or RNFL defect. The SAP was performed with the Humphrey Field Analyzer (Carl-Zeiss Meditec, Dublin, Calif) with a SITA Standard strategy, program 24-2 (Carl-Zeiss Meditec). All eyes had normal results of SAP, defined as a glaucoma hemifield test result within normal limits and pattern standard deviation (PSD) of probability greater than 5%. Eyes with visual acuity less than 20/40, peripapillary atrophy extending to 1.7 mm from the disc center, retinal disease, or unreliable or abnormal SAP results (>25% fixation losses and false-positive and false-negative rates) were excluded from this investigation.
The SWAP was performed with the Humphrey Field Analyzer with the use of a full threshold strategy, program 24-2 (software version 12.6; Carl-Zeiss Meditec). Abnormal SWAP measurements were defined as mean deviation (MD) or PSD outside 95% normal limits and glaucoma hemifield test result outside normal limits, or MD or PSD outside 95% normal limits and a cluster of 3 or more contiguous points with probability level of less than 5%, with 1 or more with probability less than the 1% level in the pattern deviation plot. The FDT was performed with the Humphrey Matrix System (Carl-Zeiss Meditec) by means of a maximum likelihood (zippy estimation of sequential testing, or ZEST) test strategy, program 24-2 (software version 2.02; Carl-Zeiss Meditec). Abnormal FDT measurements were defined as MD or PSD outside 95% normal limits and glaucoma hemifield test result outside normal limits.
All patients underwent OCT imaging (Stratus OCT; Carl-Zeiss Meditec) of the peripapillary RNFL, optic nerve head (ONH), and macula within 6 months of clinical examination with the use of version 3.0 software. The scan algorithms used in this study were fast RNFL (3 scans averaged to form a mean baseline), fast macula, and fast optic disc scans. Multiple scans were obtained per eye, and the single scan demonstrating the greatest signal-noise ratio was selected. Images with poor centration, poor focus, or a signal-noise ratio less than 33 dB were excluded. Imaging of the RNFL was performed with a circular scan 3.4 mm in diameter centered on the optic disc. A mean of 3 separate RNFL measurements (256 A-scans each) were used. Topographic measurements of the ONH using OCT were generated in an automated fashion by identifying the edge of the ONH as the termination of the retinal pigment epithelium–choriocapillaris layer and defining a parallel line 150 μm anteriorly.25 Images of the ONH were obtained with 6 radial scans (128 A-scans each) in a spokelike pattern, each 4 mm long, centered on the optic disc. Macular thickness measurements were generated with 6 radial lines (128 A-scans each), each 6 mm long, centered on the fovea, to generate a macular map 6 mm in diameter with quadrantic macular thickness in the central 12° (6 mm). Data from the central 0.5 mm of the map were excluded because the foveola is devoid of ganglion cells. The GDx-VCC imaging (Carl-Zeiss Meditec) was performed in a standardized fashion (software version 5.3.1) as previously described20,32,33 with a circular scan (3.2-mm diameter) centered on the optic disc. Images were excluded that were obtained during eye movement, as well as unfocused, poorly centered images, images with residual corneal birefringence of 13 nm or more (based on oral communication with Q. Zhou, PhD, of Carl-Zeiss Meditec, 2004), or images with a quality score grade less than 8.
Abnormal RNFL measurements using OCT were defined as mean or quadrantic thickness values outside 95% normal limits, based on the instrument's normative database, that were confirmed on at least 2 of 3 repeat scans. Abnormal GDx-VCC measurements were defined as a retardation value outside 95% normal limits and a cluster of 3 or more contiguous superpixels outside 95% normal limits, with 1 outside 99% normal limits, based on the instrument's normative database, that was confirmed on at least 2 of 3 repeat scans. Eyes were stratified into 3 groups on the basis of the OCT-generated vertical cup-disc ratio (CDR): mild cupping (CDR, <0.4), moderate cupping (CDR, 0.4-0.7), and advanced cupping (CDR, >0.7). Structural and functional characteristics were assessed separately in each group.
Statistical analysis was performed with JMP (SAS Institute Inc, Cary, NC) and SPSS software (SPSS Inc, Chicago, Ill). Analysis of variance and Yates corrected χ2 were used to compare different measures among the groups. Forward stepwise logistic regression analysis was performed to identify structural and functional variables associated with GON. Receiver operating characteristic curves and the areas under the curve were quantified for each measure and compared by the method described by Hanley and McNeil.34P≤.05 was considered statistically significant.
Forty-seven eyes of 47 patients (25 with glaucoma and 22 normal subjects) were enrolled (mean ± SD age, 58 ± 16 years; range, 25-83 years). Average ± SD visual field MD on SAP was −0.2 ± 1.2 dB (normal eyes) and −0.3 ± 1.6 dB (eyes with GON). Three (12%) of the 25 eyes with GON had an intraocular pressure greater than 21 mm Hg at the time of enrollment into the study. Although the mean central corneal thickness (CCT) in normal and glaucomatous eyes was similar, a nearly significantly (P = .08) higher prevalence of eyes with CCT less than 545 μm was observed in eyes with GON (13 [52%]) compared with normal eyes (5 [23%]). There were no other differences between the groups (Table 1) with regard to their demographic and clinical characteristics.
Figure 1 illustrates findings from a right eye with GON, advanced optic disc cupping, and a normal visual field on SAP. Note the inferior visual field defect seen with SWAP and FDT, and corresponding RNFL loss superiorly with GDx-VCC and OCT. Table 2 gives the structural and functional assessments in normal eyes (n = 22) and eyes with GON (n = 25). Significant (P = .02, .05, .04, and .05, respectively) differences in SWAP MD, SWAP PSD, FDT MD, and FDT PSD were observed between eyes of similarly aged normal subjects and patients with GON. Significant differences in all ONH topographic measures (P<.001) except disc area (P = .21), OCT-derived RNFL thickness (P = .008), macular thickness (P = .01), and nerve fiber indicator (P = .03) were observed between the 2 groups.
Table 3 illustrates the magnitude of visual field loss and RNFL atrophy in eyes with GON by means of advanced structural and functional testing. Among the 25 eyes with GON, 8 (32%) had abnormal results of SWAP, 7 (28%) had abnormal results of FDT, 12 (48%) had abnormal RNFL by OCT, and 15 (60%) had abnormal RNFL by GDx-VCC. Eyes with GON demonstrated a higher proportion of structural abnormalities (18 [72%]) than functional abnormalities (10 [40%]). Figure 2 illustrates a scatterplot of SWAP MD against vertical CDR stratified by the probability of an RNFL thickness abnormality by OCT according to the instrument's normative database. Eyes with increased CDR demonstrated significantly greater (R = 0.35, P = .01) visual field loss according to SWAP and a higher prevalence of significant RNFL atrophy (small diamonds, Figure 2) with OCT.
Table 4 illustrates differences in clinical characteristics and structural and functional assessments between eyes with mild (vertical CDR, <0.4), moderate (vertical CDR, 0.4-0.7), and advanced (vertical CDR, >0.7) cupping. Eyes with advanced cupping had significantly worse mean FDT MD (P = .04), OCT-derived ONH topography measures (P≤.001 to .004), RNFL thickness (P = .02), and macular thickness (P = .03). The prevalence of visual field abnormalities by SWAP and FDT (Figure 3) and structural abnormalities by GDx-VCC and OCT (Figure 4) were assessed in eyes with mild, moderate, and advanced cupping. A significantly (P = .008, .002, .003, and .01, respectively) greater prevalence of abnormalities by SWAP, FDT, OCT, and GDx-VCC was identified in eyes with advanced cupping (43%, 43%, 57%, and 57%, respectively) than in eyes with mild (0%) and moderate (9%, 5%, 19%, and 33%, respectively) cupping.
Receiver operating characteristic curves were generated to identify the structural and functional measures with the highest discriminating power to separate normal eyes from eyes with GON. Table 5 illustrates the area under the receiver operating characteristic curve for various structural and functional assessments. Figure 5 illustrates the receiver operating characteristic curves for various structural and psychophysical measures (OCT superior RNFL thickness, SWAP PSD, FDT MD, and GDx-VCC inferior average). By means of a forward stepwise logistic regression model, only OCT superior RNFL thickness (P = .009) and SWAP MD (P = .007) were statistically significant; CCT was significant (P = .03) if included as a dichotomous variable.
Although clinical stereoscopic examination of the optic disc and RNFL, in addition to SAP, are currently the standard methods for diagnosis and monitoring of GON, evidence exists that new psychophysical tests have superior sensitivity.5,6,35- 37 Recent studies suggest that structural changes are detectable by advanced imaging technology before the development of an SAP abnormality.26,32 Recent prospective randomized trials38 have demonstrated that eyes with advanced glaucomatous cupping and field loss are more likely to progress. Although unproved, it is likely that early detection and treatment of GON may reduce the incidence of blindness from glaucoma.
This study demonstrates that, compared with eyes of similarly aged normal controls, eyes with GON and normal SAP results have a high prevalence of RNFL atrophy detectable by GDx-VCC or OCT (48%-60%), and abnormal visual function assessed by SWAP or FDT (28%-32%). Although there are reports suggesting early detection of glaucomatous injury in eyes with preperimetric GON by means of advanced imaging and psychophysical testing, limited data exist regarding the characteristics of such eyes by a comprehensive diagnostic approach. A recent study by Medeiros and coworkers39 reported greater discriminating power with OCT, GDx-VCC, and Heidelberg Retina Tomograph (HRT); however, the study population involved eyes with SAP deficits (average ± SD MD, −4.9 ± 3.9 dB). The use of FDT and scanning laser polarimetry with fixed corneal compensation each identified glaucomatous damage in 25% of eyes with preperimetric glaucoma.40 Others41 have suggested that scanning laser polarimetry with fixed compensation is more sensitive than FDT in predicting visual field loss by SAP in eyes with preperimetric glaucoma. In a recent study, the sensitivity of RNFL thickness measurements by OCT was reported to be 71% at a fixed specificity of 90% in eyes with preperimetric GON.26 Mok and colleagues16 also reported good correlation between RNFL thickness abnormalities determined by OCT and SWAP. It should be recognized that, within a given eye, results of ancillary diagnostic testing may disagree. For example, Figure 1 demonstrates good concordance between the superiorly localized RNFL atrophy and inferiorly localized selective field loss, yet the defects on SWAP and FDT are in different quadrants.
Reduced CCT has been identified as a risk factor for conversion of ocular hypertension to glaucoma42 and is associated with abnormal results of SWAP and FDT in ocular hypertension.43,44 A CCT less than 545 μm has been demonstrated to be a significant longitudinal predictor of standard visual field loss in eyes with preperimetric GON.45 In this cross-sectional analysis, although there were no significant differences in the mean CCT values between eyes with GON and normal eyes, we found a higher prevalence of CCT less than 545 μm in eyes with GON. This may be explained by our relatively limited sample size and the variance of CCT measurements among the general population. In contrast, although both the Barbados Eye Study Group46 and the Early Manifest Glaucoma Trial38 found higher intraocular pressure to be a significant risk factor for glaucoma progression, they did not find a relationship between intraocular pressure and CCT. Additional longitudinal studies with increased sample size may be necessary to further investigate this observation.
We found that vertical CDR was associated with glaucomatous damage in eyes with preperimetric GON. Our data demonstrate that eyes with a vertical CDR greater than 0.7 had a higher prevalence of visual field loss and RNFL loss than did eyes with a vertical CDR of 0.7 or less. Johnson and colleagues47 found similar results, showing a significant association between vertical CDR and abnormal SWAP in eyes with ocular hypertension. Furthermore, the Ocular Hypertension Treatment Study reported a 32% increase in risk of conversion to primary open-angle glaucoma for every increase of 0.1 in vertical CDR.42 On the basis of these observations, clinicians should have a heightened level of suspicion regarding the presence of an abnormal psychophysical or structural diagnostic test in eyes with increased vertical CDR and normal results of SAP.
Our study has limitations that warrant consideration. The optic disc severity categorization was based on OCT assessment of vertical CDR, which may bias statistical associations with other OCT-derived measures such as RNFL, neural rim, and macular thickness. On the other hand, clinical assessment of vertical CDR is subjective, poorly reproducible, and imprecise; thus, categorization by clinical assessment would have introduced considerable bias. Studies have demonstrated that OCT-derived measurements of optic disc topography are highly reproducible and well correlated with topographic measurements derived by scanning laser tomography.25,48 Other limitations include small sample size and the possibility that our strict entry criteria for normal subjects may have produced a “supernormal” population. Although other studies20,22,30,31,49 have used similar criteria for normal subjects, exclusion of eyes with CDR of 0.6 or greater may have contributed to the 100% specificity observed in Table 3, which contrasts with clinical experience and other published reports. Finally, although structural abnormalities were repeatable on 2 of 3 OCT and GDx-VCC imaging studies, abnormalities on SWAP and FDT were not consistently confirmed with repeat testing. Although future studies are needed to validate these results, evidence suggests that FDT perimetry exhibits significantly less variability than SAP,50 especially within regions of visual field sensitivity loss, suggesting that fewer confirmatory tests may be necessary.
Several important considerations should be noted. No single diagnostic test has perfect sensitivity or specificity; therefore, clinical decisions should not be based on isolated test results. In the present study, because of the large number of significance tests conducted, there may be some α inflation and the statistical significance of the P values could be overstated; however, we have not used Bonferroni correction or a similar adjustment procedure to avoid inflation of β errors. The results of this study are not necessarily generalizable to other populations. This study concerns individuals with glaucoma not yet detectable by SAP; no conclusions can be drawn about individuals who may have visual field defects detectable by SAP but not evident by SWAP, FDT, OCT, or GDx-VCC. Finally, as demonstrated in other studies involving eyes with preperimetric glaucoma,45,51 it is possible that some patients in this study will not develop SAP deficits and that false-positive results may exist. Greater longitudinal follow-up is necessary to assess the rate of SAP conversion among the glaucomatous eyes in the current investigation.
In conclusion, eyes with GON and normal results of SAP have significantly greater structural and psychophysical abnormalities by OCT, GDx-VCC, SWAP, and FDT than do normal eyes. Vertical CDR is associated with structural and psychophysical abnormalities; eyes with CDR greater than 0.7 have a higher prevalence of abnormal findings on SWAP, FDT, and RNFL atrophy than eyes with vertical CDR of 0.7 or less when results of SAP are normal.
Correspondence: David S. Greenfield, MD, Bascom Palmer Eye Institute, 7108 Fairway Dr, Suite 340, Palm Beach Gardens, FL 33418 (firstname.lastname@example.org).
Submitted for Publication: June 9, 2004; final revision received March 21, 2005; accepted March 21, 2005.
Financial Disclosure: Dr Greenfield has received research support from and has served as a consultant for Carl-Zeiss Meditec and Laser Diagnostic Technologies.
Funding/Support: This study was supported in part by the Maltz Family Endowment for Glaucoma Research, Cleveland, Ohio; a grant from Barney Donnelley, Palm Beach, Fla; The Kessel Foundation, Bergenfield, NJ; grants R01-EY08684 and P30-EY014801 from the National Institutes of Health, Bethesda, Md; and an unrestricted grant from Research to Prevent Blindness, New York, NY.
Previous Presentation: This study was presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology; April 29, 2004; Fort Lauderdale, Fla.