Purpose The Humphrey Matrix (FDT2) may be more sensitive in detecting glaucomatous visual field loss than SITA standard automated perimetry (SAP) performed on the Humphrey Field Analyzer (HFA). Therefore, FDT may be a good candidate to determine disease progression in patients with glaucoma. Our aim was to test the hypothesis that automated perimetry using the FDT2 would be equal to, or more effective than, HFA SITA-Standard, in identifying glaucomatous progression.
Methods One hundred and twenty patients with glaucoma were tested twice at baseline and every 6 months for 4 years with HFA SITA-Standard and FDT2. FDT2 values were standardised to HFA SAP values. We used pointwise linear regression (PLR) over the full data series to identify glaucomatous progression and generated an array of results using three different criteria: (1) three or more clustered test locations progressing, (2) three or more non-clustered test locations progressing and (3) total number of progressing test locations. We compared HFA SAP and FDT2 for the number of locations signalled by the PLR detection algorithm.
Results Regardless of the criteria, HFA SAP with SITA-Standard testing detected visual field progression at a higher rate than the FDT2 overall (P<0.001).
Conclusion HFA SAP identifies glaucomatous visual field progression at a rate at least as high if not higher than FDT2.
- visual testing
- glaucoma progression
This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Statistics from Altmetric.com
Frequency doubling perimetry V.2 (FDT2) or the Humphrey Matrix has been shown to be a sensitive method to detect glaucomatous and neurologic visual field damage.1–8 The 10° frequency doubling stimulus used in FDT V.1 (FDT1) has been compared with standard automated perimetry (HFA SAP; HFA, Humphrey Field Analyzer) to detect visual field progression.9 The progression was usually identified in different patients and outcomes varied considerably with the criteria. Linear regression revealed more defects using HFA SAP but glaucoma change probability analysis showed more progression with FDT1.9
It is well known that with HFA SAP, variability increases exponentially with decreasing sensitivity.10 11 FDT2, possibly related to its larger stimulus size and test strategy (ZEST—Zippy Estimation of Sequential Thresholds) used by FDT2 shows little increase in variability with decreasing sensitivity.11 12 This makes FDT2 an excellent potential candidate for detecting visual field progression in glaucoma. However, both variability and effect size (rate of change) must be considered.
Using FDT2, Redmond and coworkers13 studied 64 subjects with primary open angle glaucoma and 36 controls. Subjects were tested every 6 months with FDT2 and HFA SAP using the 24–2 pattern on the same day. All subjects had at least five perimetric examinations and median mean deviation (MD) in the glaucoma subjects was −2.6 dB. Using permutation of pointwise linear regression analysis (PoPLR) and total deviation values to determine progression, they found statistically significant deterioration in 17% of subjects with FDT2 and in 34% of subjects with HFA SAP. In healthy observers, significant progression was found in 8% with FDT2% and 31% with HFA SAP. Redmond et al concluded there was no evidence that FDT2 was more sensitive than HFA SAP for determination of visual field progression in patients with glaucoma.
Hu et al14 compared the performance of HFA SITA-Standard and FDT2 perimetry in monitoring glaucoma progression in 113 eyes of subjects with primary open-angle glaucoma with three criteria (conservative, moderate and liberal). The conservative criterion was the presence of at least three adjacent test locations in the same hemifield signalled as changing with at least one non-edge location. The moderate criterion was the same as the conservative but having any three test locations changing. The liberal criterion was the presence of any two test locations signalled as changing with at least one non-edge location. Hu et al14 found no statistically significant difference in the proportion of progressing and improving eyes across tests using the conservative criterion. Fewer eyes showed improvement on HFA SAP compared with FDT2 using the moderate criterion and FDT2 detected less progressing eyes than HFA SAP using the liberal criterion. FDT2 did not appear to have significant benefits over HFA SAP in monitoring glaucoma progression but the two tests often detected progression in different glaucoma eyes according to their findings.
Liu et al15 compared the detection of visual field progression between HFA SITA Standard and FDT2 testing in 179 patient eyes with glaucoma using PLR. In their work, a test location was considered progressing when the rate of change of visual sensitivity was ≤ −1 dB/year for non-edge and ≤−2 dB/year for edge locations. Three criteria were used to define progression in an eye: ≥3 adjacent non-edge test locations (conservative), any three locations (moderate) and any two locations (liberal) progressed. They found 6.1% and 3.9% progressed with the conservative criteria, 14.5% and 5.6% of eyes progressed with the moderate criteria and 20.1% and 11.7% of eyes progressed with the liberal criteria by FDT2 and HFA SAP, respectively. Taking all test locations into consideration (total, 54×179 locations), FDT2 detected more progressing locations (176) than HFA SAP (103, P<0.001). The rate of change of visual field MD was significantly faster for FDT2 (all with P<0.001). No eyes showed progression in the normal group using the conservative and the moderate criteria. They concluded FDT2 detected more progressing eyes than HFA SAP at a comparable level of specificity.
In summary, there are currently conflicting reports on the ability of FDT2 to detect visual field progression in glaucoma. To further investigate this issue, we tested 120 patients with glaucoma and 60 ocular healthy controls twice at baseline and every 6 months for 4 years with HFA SAP and FDT2 for their relative abilities to detect glaucomatous progression. Our null hypothesis was that due to its high sensitivity to detect defects and its lack of a rise in variability with decreasing sensitivity, FDT2 would be equal to, or more effective than HFA SAP, in identifying glaucomatous progression. Conversely, HFA SAP may have better performance than FDT2.
Materials and methods
Sixty normal subjects and 120 patients with glaucoma were tested at baseline and again at a separate sitting within 1–8 weeks and then every 6 months for 4 years with SITA-Standard size III, Full Threshold size V, FDT2 and motion perimetry (only the results of HFA SAP and FDT2 in subjects with glaucoma are presented here). All subjects gave informed consent to participate in the study. We followed the tenets of the Declaration of Helsinki. The normals were volunteers, paid in accordance with the Institutional Review Board, who answered advertisements inviting them to participate in research.
The patients with glaucoma were invited from the glaucoma clinic at the University of Iowa Department of Ophthalmology and Visual Sciences if they met entry criteria. They were enrolled if they had glaucomatous optic disc changes in conjunction with abnormal SAP (glaucomatous visual field defects, that is, three or more adjacent abnormal test locations in a clinically suspicious area at the P<0.05 level or two adjacent locations abnormal with at least one at the P<0.01 level). In addition, MD was in the range of 0 to −20 dB on HFA SAP. We included patients with primary, secondary or normal tension glaucoma. The patients did not have another disease affecting vision and were capable of performing HFA SAP and FDT2 as well as being willing to return for follow-up visits. Patients were excluded if they had cataract causing visual acuity of worse than 20/30, pupil size less than 2.5 mm, age less than 19 years or pregnancy. If both eyes qualified for the study, one eye was chosen at random as the study eye.
For HFA SAP, we used automated perimetry with SITA-Standard program 24–2 of the HFA (Carl Zeiss Meditec, Dublin, California, USA). We followed the manufacturer’s recommendations for using corrective lenses. Care was taken to prevent lens rim artefacts. The subjects had testing in one eye, chosen at random, but the same eye was used for all tests. An eye patch was used to occlude the fellow eye. All visual field examinations met the following reliability criteria: fixation losses <20%16 or normal gaze tracking, false positive rate <10%17 and false negative rate <33%. The four tests were administered in a random order with at least a 5 min rest break between the testing sessions.
FDT2 or Humphrey Matrix Frequency Doubling Technology Perimeter, Welch Allyn (Skaneateles, New York, USA) was performed either before or after conventional perimetry testing in random order with at least a 5 min rest period between examinations to reduce the effect of fatigue. Testing was performed in a dim room using the Humphrey Matrix device.18
Since FDT2 and HFA SAP have different scales, we standardised the FDT2 values using the framework reported by Fredette et al.19 Fredette et al studied 50 patients with glaucoma and three normal subjects and defined a range within which FDT2 sensitivity values and those of the HFA were comparable and the range within which a one-to-one correspondence is plausible. The algorithm we used in order to build the equivalence between the ranges for plausible comparison between FDT and HFA is as follows:
FDT values less than 4 dB are converted to 0 dB.
FDT values between 6 and 7 dB are converted to values between 6 and 9 dB.
FDT values between 10 and 14 dB are converted to values between 22 and 25 dB.
FDT values between 18 and 22 dB are converted to values between 29 and 30 dB.
FDT values between 22 and 32 dB are converted to values between 30 and 34 dB.
FDT values above 35 dB remain unchanged.
We then used PLR over the full data series to determine progression. Since there is no consensus for choosing slopes, P values and clusters of progressing locations, we generated arrays of results using three different criteria: (1) C1: three or more clustered test locations progressing, (2) C2: three or more non-clustered test locations progressing and (3) C3: total number of progressing test locations. Our primary outcomes (total number of progressing test locations) were compared among these three PLR arrays using analysis of variance (ANOVA), with factor group (FDT2, HFA); factor slope (10 levels) and factor p-value (10 levels).
We compared the level of slope and P values from the arrays that strike statistical significance between FDT2 and HFA SAP based on factorial analysis of variance.
The average age of the patients with glaucoma was 64.9 years (SD, 9.5 year; range, 38–81 years); their average MD was −6.7 dB with SD of 4.4 dB; median of −5.7 dB. Seventy-five of the patients with glaucoma were women and 45 were men. Further assessment regarding the temporal change in MD can be found in table 1. The rates of progression are found in figure 1.
We created matrices of different combinations of slopes and P values for FDT2 and HFA SAP subjects with glaucoma and normals (tables 2 and 3). We were able to establish, for C1, C2 and C3, an overall statistically significant difference between groups, that is, HFA SAP versus FDT2 (P<0.001), a statistically significant difference between FDT2 and HFA SAP at each of the levels of slope (P<0.001) and a statistically significant difference among these groups at each level of P value (P<0.05)—with HFA SAP yielding higher detections on average compared with FDT2. Regardless of the criteria, the higher detection of abnormalities using HFA SAP for glaucomatous progression was maintained. Since some of the criteria would not likely be clinically useful, we repeated the analysis using a subset of reasonable clinical criteria (see shaded area in tables). This subgroup analyses showed differences favouring HFA SAP at a statistical significance (P<0.001).
There was modest overlap of progression detection between the tests with some criteria but poor overlap with others. To highlight the differences in the results from the two methods, we subtracted the number of progressing locations by criteria found with FDT2 from HFA SAP. Figure 2 shows HFA SAP testing signals at least as many progressing locations as FDT2. For the commonly used criteria of a PLR slope of −1 dB/year and P values of <0.01 and <0.05, HFA SAP outperformed FDT2 for all conditions (tables 2 and 3 and figure 2). The largest differences per subject are found with the criteria of at least three non-clustered locations progressing.
To see if the results were similar in early glaucoma, 34/112 people met the criterion of MD greater than or equal to −4 dB. The results from this subgroup for all three criteria were similar to the full cohort results (online supplementary tables 1 and 2).
Supplementary file 1
We found HFA SAP testing identified glaucomatous visual field progression more frequently than testing with FDT2 over a 4-year period in patients with glaucoma with moderate visual field loss based on criteria C1, C2 and C3. This was true for all three conditions we tested with the largest differences per subject found with the three non-clustered locations. Therefore, our null hypothesis has been rejected across our range of slope and P value examined. This suggests that HFA SAP may have better performance for detecting glaucomatous progression compared with FDT2.
Since there is no single criterion that has been agreed on for identifying glaucomatous progression, we supply a matrix of criteria from which a clinician can survey given the clinical question (see tables 2 and 3 and figure 2). The slope and P value chosen should depend on the clinical question being asked—whether one is more concerned with sensitivity or whether specificity is more critical. This issue has been studied20 and while we have shown several optimal criteria that balance sensitivity and specificity, the choice depends on the specific clinical or research setting.
FDT2 was designed to use the frequency doubling effect to preferentially stimulate the M retinal ganglion cells. An early hypothesis was that since M and My cells (if they exist in humans) are sparse, they would likely have little redundancy if tested, possibly allowing for the early detection and progression of visual field defects. However, Swanson et al21 found the converse—HFA SAP may be more effective in preferentially stimulating the M cells than using the frequency doubling effect. The frequency doubling effect is most likely mediated by the combined activity of several types of retinal ganglion cells and more central visual mechanisms and does not appear to have special properties that would isolate a ganglion cell type.21 22
A number of publications have concluded that using the FDT2 stimulus could find visual field defects earlier than HFA SAP,23–25 while other reports show similar rates of glaucomatous detection.2 26 However, these studies are difficult to analyse because the database of normals for the two tests were composed of different subjects and used different criteria. For example, the database of normals provided with the FDT2 required normal HFA SAP results to be included. Therefore, poorly performing perimetry subjects at the lower tail of the normal distribution may have been truncated by using these criteria. Also, tests that do not have stable reproducibility characteristics may appear sensitive to detect disease but their apparent superiority may be explained by their high fluctuation of test results. A high retest variability would hinder detection of glaucomatous progression.
Several studies have investigated the ability of FDT2 perimetry to detect visual field progression compared with HFA SAP. Redmond and coworkers13 using PoPLR found HFA SAP detected progression in about twice as many subjects as FDT2 but far more normal subjects were identified as progressing with HFA SAP. Hu et al found no benefit of FDT2 over HFA SAP in monitoring glaucoma progression. And Liu and coworkers15 reported FDT2 detected significantly more progressing locations than HFA SAP. Our results suggest that across the spectrum of detection, with different combinations of PLR slopes and P values, HFA SAP identifies glaucomatous visual field progression at a rate at least as high if not higher than FDT2.
We are very appreciative of the cooperation of the University of Iowa, Department of Ophthalmology and Visual Sciences Glaucoma service.
Contributors All authors made substantial contributions to the design, analysis and writing of this manuscript.
Funding This study was supported by a VA Merit Review Grant RX000140-03 and by an unrestricted grant to the Department of Ophthalmology from Research to Prevent Blindness, New York, New York, USA.
Competing interests None declared.
Patient consent Obtained.
Ethics approval The visual testing protocol was approved by the University of Iowa Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.