|Year : 2012 | Volume
| Issue : 2 | Page : 97-102
Comparison between Humphrey Field Analyzer and Micro Perimeter 1 in normal and glaucoma subjects
Vineet Ratra, Dhanashree Ratra, Muneeswar Gupta, K Vaitheeswaran
Medical Research Foundation, Sankara Nethralaya, Chennai, India
|Date of Web Publication||4-Aug-2012|
Senior Consultant and Incharge, Navasuja Sankara Nethralaya, 73, Venkata Krishna Road, RA Puram, Chennai-600028
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Purpose: To determine the correlation between fundus perimetry with Micro Perimeter 1 (MP1) and conventional automated static threshold perimetry using the Humphrey Field Analyzer (HFA) in healthy individuals and in subjects with glaucoma.
Materials and Methods: In this study, we enrolled 45 eyes with glaucoma and 21 eyes of age-matched, healthy individuals. All subjects underwent complete ophthalmic examination. Differential light sensitivity was measured at 21 corresponding points in a rectangular test grid in both MP1 and HFA. Similar examination settings were used with Goldmann III stimulus, stimulus presentation time of 200 ms, and white background illumination (1.27 cd/m 2 ). Statistical analysis was done with the SPSS 14 using linear regression and independent t-test.
Results: The mean light thresholds of 21 matching points in control group with MP1 and HFA were 14.97 ± 2.64 dB and 30.90 ± 2.08 dB, respectively. In subjects with glaucoma, the mean values were MP1: 11.73 ± 4.36 dB and HFA: 27.96 ± 5.41 dB. Mean difference of light thresholds among the two instruments was 15.86 ± 3.25 dB in normal subjects (P < 0.001) and 16.22 ± 2.77 dB in glaucoma subjects (P < 0.001). Pearson correlation analysis of the HFA and MP1 results for each test point location in both cases and control subjects showed significant positive correlation (controls, r = 0.439, P = 0.047; glaucoma subjects, r = 0.812, P < 0.001). There was no difference between nasal and temporal points but a slight vertical asymmetry was observed with MP1.
Conclusion: There are significant and reproducible differences in the differential light threshold in MP1 and HFA in both normal and glaucoma subjects. We found a correction factor of 17.271 for comparison of MP1 with HFA. MP1 appeared to be more sensitive in predicting loss in glaucoma.
Keywords: Correction factor, glaucoma, glaucoma, Humphrey Field Analyzer, Micro Perimeter 1, retinal sensitivity, sensitivity, static threshold perimetry
|How to cite this article:|
Ratra V, Ratra D, Gupta M, Vaitheeswaran K. Comparison between Humphrey Field Analyzer and Micro Perimeter 1 in normal and glaucoma subjects. Oman J Ophthalmol 2012;5:97-102
|How to cite this URL:|
Ratra V, Ratra D, Gupta M, Vaitheeswaran K. Comparison between Humphrey Field Analyzer and Micro Perimeter 1 in normal and glaucoma subjects. Oman J Ophthalmol [serial online] 2012 [cited 2019 Jan 19];5:97-102. Available from: http://www.ojoonline.org/text.asp?2012/5/2/97/99372
| Introduction|| |
Microperimetry is a procedure in which retinal sensitivity is assessed while the fundus of the eye is directly examined, enabling exact correlation between macular pathology and corresponding functional defect. Recent reports have documented the use of microperimetry in quantifying macular sensitivity in various macular pathologies and they have also shown the usefulness of this investigational tool in assessing the effectiveness of a particular treatment modality in restoring macular function. ,, In the Micro Perimeter 1 (MP1) (Nidek Instruments Inc, Padova, Italy) the software has been further advanced to perform automated static threshold fundus perimetry and automated kinetic fundus perimetry. The MP1 offers distinct advantages over conventional automated perimetry. The fixation analysis and surveillance is far superior with autotracking of the eye movements and correction for loss of fixation, making it an ideal tool for visual field analysis in patients with poor fixation. Static threshold perimetry with MP1 can be performed using customized testing parameters that allow the examiner to select the pattern, intensity, form, and number of stimuli being projected onto the retina. Microperimetry has been compared with Octopus 101 conventional static perimetry by Springer et al. where they found significant correlation between the two instruments at each test point. However, the difference varied depending on the stimulus location in their study. It has also been compared with Humphrey Field Analyzer (HFA) by Ozturk et al.  They found MP1 correlated well with HFA in detecting glaucomatous field defect. It is, however, important to establish normative data and reference values for a direct comparison between the instruments. Our study was undertaken to determine the correction factor for comparison of MP1 with HFA.
| Materials and Methods|| |
Forty-five eyes of 34 subjects with primary open angle glaucoma and 21 eyes of 12 healthy volunteers were enrolled in this study. The study was approved by the institutional review board. Informed consent was obtained from all the participants. Subjects with glaucoma were diagnosed based on intraocular pressure, optic disc changes, and visual field changes. An elevated intraocular pressure (IOP) of ≥21 mm Hg by Goldmann appalnation tonometer, glaucomatous cupping on slit lamp biomicroscopic examination with 90-diopter lens with corresponding visual field defects on HFA 24-2 were the defining criteria for diagnosing glaucoma. Controls were patients without any history of ocular or systemic diseases, other than refractive error who agreed to participate in the study. They had normal IOP, normal disc and normal visual fields, and no family history of glaucoma. Those patients with visual acuity less than 20/60, media opacities, lens haze, macular diseases including drusen or pigment alterations, systemic disorders, on medications that can affect visual function and history of trauma were excluded from the study. All study subjects underwent complete ophthalmic examination including best-corrected visual acuity, slit lamp biomicroscopy, intraocular pressure measurement, gonioscopy, fundoscopy, perimetry with HFA, and MP1. All subjects were given training of these two tests before the actual examination.
The testing protocol which included Goldmann III stimulus with presentation time of 200 ms, white stimulus color, and a white background illumination was same in both HFA and MP1. The MP1 tested 21 locations in a rectangular pattern from the fovea. HFA tested 56 locations in a rhomb-shaped six-degree grid. Both tests were performed with pupillary dilation.
The MP1 (Nidek Instruments Inc.) takes a high-quality digital fundus image using an infrared fundus camera with a 45-degree field of view. Perimetry is performed using a liquid crystal display controlled by special software (MP1 1.4.1. SP1, Navis). The stimuli are projected on the area of interest in the fundus. A computer controlled system tracks the eye movements (25 times/s) during examination. The system automatically adapts the projected stimuli to the fundus shifts. If the fundus is not aligned at a particular time, the MP1 automatically recalibrates and realigns before resuming the test. The perimetry and fixation pattern maps are then superimposed on the color fundus image allowing simultaneous anatomical and functional correlation.
For the purpose of this study, we used manually designed testing stimuli pattern which contained 21 testing locations in rectangular grid pattern (separate test patterns for right eye and left eye) with Goldmann III stimulus size (Ø = 26 arc/min), with attenuation of stimulus intensity at 16 dB. The stimuli were projected with presentation time of 200 ms, and white background was used with illumination of 1.27 cd/m 2 .
Fixation characteristics were measured according to Springer et al. where eyes with >75% fixation points located within the central 2° are classified as "stable fixation." If <75% fixation points are located within the 2° but >75% fixation points are located within the 4° it is classified as "relatively unstable fixation". And it is labeled "unstable fixation" if <75% fixation points are located within the 4° diameter circle.
Humphrey Field Analyzer
Conventional static threshold perimetry was performed with the HFA in all subjects. In the HFA, the 24-2 program tested 56 locations in a rhomb-shaped grid. Fixation behavior and false negatives were monitored during the examination and only tests with fewer than 30% false-positive answers were considered reliable and included in the study. Tests with fixation losses of greater than 20% and false-positive and false-negative responses greater than 30% were considered as unreliable.
Analysis of perimetric results
Differential threshold values were compared for 21 matching points [Figure 1] in a rectangular grid covering an area of 27 × 18 degrees. Data from left eyes were converted into right eyes for concordance of blind spot. MP1 results were inverted from top to bottom as well to correspond to the HFA. Central 21 points of HFA were considered for analysis. For the 21 matching locations, the mean difference between both instruments and its standard deviation were calculated. Statistical analysis was done using the Statistical Package for the Social Sciences (SPSS, version 14.0). Differential light thresholds were compared using the paired t-test and ROC curve. The two machines were compared using the intraclass correlation coefficient (ICC) and the Bland-Altman plots.
|Figure 1: (a). Microperimetric test pattern of 21 testing location used in this study, (b). Test grid of 24-2 Humphrey Field Analyzer, central 20 test points and foveal threshold considered for study analysis are shown inside the box|
Click here to view
| Results|| |
There were 15 males and 8 females (45 eyes) in the glaucoma group and 9 males, 2 females (21 eyes) in the control group. The power of the study was 96%. (α error = 0.05%). The mean age was 37.6 ± 14.6 years (range: 14-61 years) in control group and 46.2 ± 13.50 years (range: 14-67 years) in cases with glaucoma (P = 0.06). The mean best-corrected visual acuity using LogMAR was 0.15 ± 0.24 in controls and 0.2 ± 0.5 in the glaucoma group (P = 0.5).
The mean light thresholds with the standard deviation at each point with both MP1 as well as HFA are shown in [Table 1],[Table 2],[Table 3] and [Table 4]. In the control group, the mean light threshold using MP1 was found to be 14.97 ± 2.64 dB and with HFA it was 30.90 ± 2.08 dB. In subjects with glaucoma the mean light thresholds were MP1: 11.73 ± 4.36 dB and HFA: 27.96 ± 5.41 dB.
|Table 1: Microperimeter 1: Mean and SD of differential light threshold (dB) in glaucoma subjects|
Click here to view
|Table 2: Microperimeter1: Mean and SD of differential light threshold (dB) in controls|
Click here to view
|Table 3: HFA: Mean and SD of differential light threshold (dB) in glaucoma subjects|
Click here to view
|Table 4: HFA: Mean and SD of differential light threshold (dB) in controls|
Click here to view
The mean difference of light thresholds among the MP1 and HFA was 15.86 ± 3.25 dB in normal subjects and 16.22 ± 2.77 dB in subjects with glaucoma. The difference was marginally more in the glaucoma subjects than the control group. This difference between both instruments was highly significant (P < 0.001) in both the groups.
A vertical asymmetry was observed with marginally higher values in the inferior field with MP1 in glaucoma subjects and the control group. However, it did not reach statistical significance. No difference was seen between the nasal and the temporal points.
The difference in mean threshold values between glaucoma patients and control subjects was significant in all quadrants [Table 5].
In contrast to HFA, the microperimeter enabled direct assessment of fixation characteristics. The fixation characteristics are given in [Table 6].
Pearson correlation analysis of the HFA and MP1 results for each test point location in both cases and control subjects showed significant positive correlation (controls, r = 0.439, P = 0.047; glaucoma subjects, r = 0.812, P < 0.001). Using linear regression we found that with every 1dB change in MP1 value in controls, the HFA value changed by 0.345 (95% CI 0.006-0.685, P = 0.047). And for every 1dB change in MP1 value in glaucoma subjects, the HFA value changed by 1.006 (95% CI 0.784-1.229, P < 0.001).
The ROC curves obtained for both HFA and MP1 after averaging the values are displayed in [Figure 2]. The area under curve (AUC) for MP1 was found to be 0.735 while the AUC for HFA was 0.67. The MP1 appears to be more sensitive in predicting loss in glaucoma with a sensitivity of 75% while HFA showed sensitivity of 62%.
We found an overall correction factor of 17.271. The linear regression analysis showed significant correlation at each test location between the MP1 and HFA (r = 0.911, 95% CI 0.735-1.087, P < 0.001). Pearson correlation for all test locations in both the groups showed r = 0.791 and r2 = 62.6%. The HFA value can be predicted in 62.6% patients using this correction factor. (See Annexure 1 for correction factor calculation.)
Agreement between the two machines MP1 and HFA was studied using the ICC and the Bland-Altman plots for normal subjects and glaucoma subjects separately as well as for all study participants overall. [Table 7] reveals that both MP1 and HFA had high degrees of ICC (>0.87, P < 0.0001) in all study participants overall and in glaucoma subjects, but moderate degree ICC in normal subjects (0.598, P = 0.024).
Bland-Altman bias plots of the average of the MP1 and HFA measurements and the difference between them are depicted in [Figure 3], [Figure 4] and [Figure 5]. The bias plot for each of the variables showed excellent agreement with 95% limits of agreement being in acceptably narrow range in case of all the parameters. The mean bias in overall subjects was -16.14 dB (SD = 2.96, 95% limits of agreement = -21.94 dB to -10.34 dB; actual % outside specification limits = 4.5%) ; bias in normal subjects was -15.93 dB (SD = 2.55 dB, 95% limits of agreement = -20.92 to -10.94; actual % outside specification limits 4.8%) and bias in glaucoma subjects was -16.23 dB (SD = 3.15 dB, 95% limits of agreement = -22.42 dB to -10.05 dB; actual % outside specification limits 2.2%). These results thus demonstrate an acceptably low bias with excellent agreement between the MP1 and HFA values.
|Figure 3: Bland– Altman plot of MP1 and HFA in overall subjects (average vs. difference)|
Click here to view
|Figure 4: Bland– Altman plot of MP1 and HFA in normal subjects (average vs. difference)|
Click here to view
|Figure 5: Bland– Altman plot of MP1 and HFA in glaucoma subjects (average vs. difference)|
Click here to view
| Discussion|| |
It is important to establish normative data and provide reference values for direct comparison between two instruments. This study provides the normative data for Indian population for MP1. The mean differential light threshold of MP1 for normal subjects in our study was 14.97 ± 2.64 dB. Springer et al. and Ozturk et al. have reported similar values in normal population. The differential light threshold in normal persons using the MP1 was reported to be 17.79 ± 1.09 dB by Ozturk et al. and 15.5 ± 0.8 dB by Springer et al.
The mean differential light threshold at any given point was always lower in MP1 than in HFA. Similar finding has been reported by Ozturk et al. They reported a mean value of 17.79 ± 1.09 dB in MP 1 compared with 29.15 ± 2.38 dB with HFA. Our study shows comparable results with values of 14.97 ± 2.64 dB with MP1 and 30.90 ± 2.08 dB with HFA in normals. Springer et al. compared the light threshold of MP1 with the Octopus and the values were 15.5 ± 0.8 dB and 30.2 ± 1.2 dB, respectively. They attributed this difference to the higher background illumination than claimed by the manufacturer in MP1 causing lower contrast. The difference in testing strategy between the two methods may likewise account for the difference in threshold values because the MP1 takes the last seen threshold as final threshold whereas the Octopus calculates the difference between the last seen and last unseen stimulus as final threshold.
Among the patients with glaucoma we found similar reproducible difference in the differential mean light thresholds between MP1 and HFA which was statistically significant. We found an overall correction factor of 17.271 while comparing MP1 and HFA values. Using this correction factor it is possible to predict the anticipated HFA values from MP1 values and vice versa. Springer et al. found a correction factor of 11.4-18.3 dB on comparing with Octopus 101. In their study, the correction factor differed according to stimulus location. The inferior field showed considerably lower threshold values. We, however, did not find a significant difference in the light thresholds in different quadrants. The threshold values were marginally higher in the inferior field in our study which is in agreement with previously published studies. ,
This study also indicates that MP1 appears to be more sensitive in predicting glaucomatous field loss. Using microperimetry Orzalesi et al. showed presence of reduced sensitivity corresponding to localized areas of retinal nerve fiber layer defect in eyes which had a normal standard central 30 degree visual field test. Miglior  also reported that with microperimetry early loss of retinal sensitivity could be detected in eyes with normal standard visual field examination. Using the scanning laser ophthalmoscope microperimetry, Lima et al. were able to pick up subtle paracentral functional defects in patients with normal standard automated perimetry. In contrast, Ozturk et al. did not find any advantage of MP1 over HFA in the evaluation of glaucomatous visual field defects in macular area. They, however, did demonstrate perimetric defects even in the absence of anatomic defects in eyes with primary open angle glaucoma. Thus, eyes believed to be normal on the basis of standard automated perimetry may be found to have functional defects when more accurate tests are used for evaluation.
The MP1 has a few limitations. In MP1 the last seen threshold is taken as final threshold. Although the examiner can define the initial threshold value, there is no provision for adaptive test strategy. The instrument tests the same level of luminance at all test locations before moving onto the next level. The examiner cannot add additional stimuli once the test begins, thus possibly precluding exact delineation of the scotoma.
This study indicates that the MP1 provides reproducible threshold values with a systematic difference with HFA, in both normal individuals and glaucoma subjects. We found a correction factor of 17.271. Using this factor the values of HFA can be accurately predicted from MP1 values and vice versa in 62.6% patients. We also established normative data for the MP1 for Indian population. The MP1 appeared to be more sensitive in predicting loss in glaucoma with a sensitivity of 75% compared with 62% with HFA.
The MP1 provides a new vista in automated fundus perimetry. It shows high sensitivity in detecting field defects in contrast to the conventional perimetry which shows limited precision, repeatability, and low sensitivity to small scotoma especially in the presence of low vision. By enabling exact real-time image tracking and alignment, MP1 also provides a quantitative analysis of fixation in terms of site and stability. The combined results of morphologic and functional testing may lead to new treatment strategies in glaucoma.
| References|| |
|1.||Rohrschneider K, Bültmann S, Springer C. Use of fundus perimetry (microperimetry) to quantify macular sensitivity. Prog Retin Eye Res 2008;27:536-48. |
|2.||Yodoi Y, Tsujikawa A, Kameda T, Otani A, Tamura H, Mandai M, et al. Central retinal sensitivity measured with the Micro Perimeter 1 after photodynamic therapy for polypoidal choroidal vasculopathy. Am J Ophthalmol 2007;143:984-94. |
|3.||Ozdemir H, Karacorlu SA, Senturk F, Karacorlu M, Uysal O. Assessment of macular function by microperimetry in unilateral resolved central serous chorioretinopathy. Eye (Lond) Eye 2008;22:204-8. |
|4.||Springer C, Bültmann S, Völcker HE, Rohrschneider K. Fundus perimetry with Micro Perimeter 1 in normal individuals: Comparison with conventional threshold perimetry. Ophthalmology 2005;112:848-54. |
|5.||Oztürk F, Yavas GF, Küsbeci T, Ermis SS. A comparison among Humphrey field analyzer, microperimetry and Heidelberg retina tomograph in the evaluation of macula in primary open angle glaucoma. J Glaucoma 2008;17:118-21. |
|6.||Rohrschneider K, Becker M, Schumacher N, Fendrich T, Völcker HE. Normal values for fundus perimetry with the scanning laser ophthalmoscope. Am J Ophthalmol 1998;126:52-8. |
|7.||Orzalesi N, Miglior S, Lonati C, Rosetti L. Microperimetry of localized retinal nerve fiber layer defects. Vision Res 1998;38:763-71. |
|8.||Miglior S. Microperimetry and glaucoma. Acta Ophthalmol Scand Suppl 2002;236:19. |
|9.||Lima VC, Prata TS, De Moraes CG, Kim J, Seiple W, Rosen RB, et al. A comparison between microperimetry and standard achromatic perimetry of the central visual field in eyes with glaucomatous paracentral visual-field defects. Br J Ophthalmol 2010;94:64-7. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7]
|This article has been cited by|
||Improving detection of mild loss of retinal light increment sensitivity at the posterior pole with the Microperimeter MP1
| ||Bowl, W. and Lorenz, B. and J√§ger, M. and Friedburg, C. |
| ||Investigative Ophthalmology and Visual Science. 2013; 54(7): 4666-4674 |
||Canadian Journal of Ophthalmology
| ||Kulkarni, S.V., Coupland, S.G., Stitt, D.M., Brownstein, J.J., Damji, K.F. |
| ||Efficacy of SLO-Microperimetry and Humphrey for evaluating macular sensitivity changes in advanced glaucoma. 2013; 48(5): 406-412 |