CONTINUOUS LIGHT INCREMENT PERIMETRY (CLIP) STRATEGY COMPARED TO FULL THRESHOLD STRATEGY IN GLAUCOMA PATIENTS

B.K. WABBELS,1,2 S. DIEHM,1 K. ROHRSCHNEIDER1 and G. KOLLING1

1Department of Ophthalmology, University of Heidelberg, Heidelberg; 2Department of Paediatric Ophthalmology, Strabismology and Ophthalmogenetics, University of Regensburg, Regensburg; Germany

Abstract

Continuous light increment perimetry (CLIP) is an improved testing strategy for automated static perimetry designed to save test time and enhance patient compliance. Stimulus intensity is continuously increased, starting from a subthreshold intensity until recognition. The test is constantly modified according to patient performance. As CLIP showed good results in normal subjects in previous studies, we now compare CLIP to the standard 4/2-full threshold (4/2)-strategy in glaucoma patients. Thirty-two patients with glaucomatous visual field defects (mean sensitivities 2.8-21.1 dB), all with perimetric experience were tested with CLIP (three times) and 4/2 in a randomized fashion. Tests were performed at 55 test locations within the central 30° visual field (24-2 area) using the Twinfield perimeter. Average mean sensitivity was significantly higher for CLIP (13.2 dB) than for 4/2 (11.7 dB) (t test, p < 0.0001). Topographical analysis of gray scales showed similar scotoma detection: absolute scotomas and extension of scotoma were comparable for both strategies. The mean difference between CLIP and 4/2 at single test points (excluding the blind spot) was 3.8 dB, 75% of differences were less than 6 dB, 95% less than 13 dB. Large differences were mostly found at the borders of absolute scotoma. Mean test time was significantly shorter for CLIP (5.5 min) compared to 4/2 (8.8 min) (t test, p < 0.0001). Patient acceptance was better for CLIP than for 4/2. In conclusion, CLIP showed comparable results to 4/2 in a shorter testing time with good patient acceptance. Mean sensitivities were 1.5 dB higher than for 4/2, similar results were found previously in normal subjects. Other new strategies such as SITA or TOP are also able to shorten test time, but patient performance and acceptance were not generally evaluated.

Introduction

In automated static perimetry, thresholds of differential light sensitivity are generally estimated using staircase procedures. Disadvantages are not only the long test time, but also the fact that the test becomes more difficult at the end of the program as stimuli are presented near the threshold. Many new test algorithms have recently been

Address for correspondence: Bettina K. Wabbels, MD, University of Regensburg, Department of Paediatric Ophthalmology, Strabismology and Ophthalmogenetics, Franz Josef Strauss Allee 11, D-93053 Regensburg, Germany. Email: bettina.wabbels@klinik.uni-regensburg.de

Perimetry Update 2002/2003, pp. 135–145

Proceedings of the XVth International Perimetric Society Meeting, Stratford-upon-Avon, England, June 26–29, 2002

edited by David B. Henson and Michael Wall

© 2004 Kugler Publications, The Hague, The Netherlands

investigated to save test time without loosing reproducibility,1-8 but less attention has been placed on patient performance and compliance.

The continuous light increment perimetry (CLIP) strategy was developed to save test time and to increase patient compliance. Results of CLIP were promising in normal subjects,9,10 and therefore, this study was planned to confirm these results in glaucoma patients with regard to threshold, reproducibility, test time, and patient compliance.

Material and methods

Continuous light increment perimetry

In this study we used an improved threshold strategy known as CLIP.5,6 CLIP uses a modified ramp stimulus, where stimulus intensity is continuously raised, beginning at a subthreshold starting position until recognition. The central threshold is first tested with a staircase strategy. Then a luminance class is chosen according to the central threshold. Reaction times are estimated at eight test locations, two locations per visual field quadrant, using stimuli 5 dB brighter than the presumed threshold. The mean reaction time is calculated from these eight reaction times. If a stimulus is not detected initially, the same location is retested with maximum luminance in order to check for an absolute scotoma. In this case, the reaction time is not included.

The starting position at the other test locations is 5 dB dimmer than the presumed threshold (5 dB subthreshold), taking into account age and central threshold. The light intensity is continuously enhanced by 1 dB per reaction time. If the patient had not pressed the answer button after eight reaction times (8 dB luminance increments), stimulus intensity is then directly enhanced by 2 dB per reaction time for three steps. If the patient does not recognize the stimulus even then (in total 14 dB brighter than the starting position, 9 dB brighter than the presumed threshold), the stimulus intensity is enhanced by 4 dB per reaction time until recognition. Therefore stimulus luminance increases faster in deeper scotomata.

The final level of stimulus intensity before recognition is regarded as the threshold (therefore correcting the effect of reaction time). Absolute scotoma detected during reaction time testing are not retested. The test is constantly modified according to patient performance. For example, if the stimulus is seen within less than three reaction times from the initial luminance, it is retested starting from a 5 dB dimmer level. If a threshold differs by more than 10 dB from the quadrant mean threshold, it is automatically retested at the end.

Twinfield perimeter

CLIP is performed on a Twinfield perimeter (Oculus Inc., Wetzlar, Germany). The experimental design is shown in Figure 1. Stimuli are presented in a semi-transparent test bowl (radius 30 cm) using back-projection. An integrated infrared camera monitors eye movements and pupil size, with the results being presented on a computer monitor and processed by a personal computer. Head position can be corrected by clicking on the center of patient’s pupil on the monitor. Four red fixation marks are

Fig. 1. Experimental design: Twinfield perimeter with test bowl, computer monitor, personal computer, and printer.

located at 1° of eccentricity. Maximum luminance is 318 cd/m2, background luminance is 10 cd/m2.

The Twinfield perimeter can perform static and kinetic perimetry (manual or automatic) using Goldmann size III or I stimuli. Static perimetry test strategies include CLIP, 4/2, fast threshold (using information of neighboring test points), and suprathreshold tests. Test grids can either be chosen from a list or be self-generated.

Study design

Thirty-two glaucoma subjects (aged 37-82 years, mean age 60 years, 13 males/19 females) with glaucomatous visual field defects were tested using the Twinfield perimeter. All subjects were experienced in automated visual field testing and had a visual acuity of 0.3 or better. One eye per subject was chosen. The subjects performed the 4/2 full threshold program and the CLIP strategy three times in randomized order. All tests per subject were completed on one day with rest breaks as requested.

Test were performed using a 24-2 test grid (55 test locations within the central 30° of visual field; Fig. 2). Goldmann size III stimulus (0.43°) was used for all tests.

The first CLIP examination on each subject was discarded. Differences between the three groups (4/2 and CLIP 2 and 3) were analyzed by one-way ANOVA on a 1% level. If this was significant, the differences between the two groups were examined using Student’s t test.

Fig. 2. 24-2 test grid with 55 stimulus locations.

Fig. 3. Distribution of mean sensitivities (4/2 and CLIP strategy) with means and standard deviations.

Results

Mean sensitivities ranged from 3.3-18.6 dB for the 4/2 strategy, from 2.8-21.0 dB for CLIP 2 (second testing), and from 3.2-21.6 dB for CLIP 3 (third testing). CLIP showed higher average mean sensitivities than 4/2. Mean sensitivity was 13.2 dB for CLIP 2 and 3 and 11.7 dB for 4/2 (Fig. 3). The differences between 4/2 and CLIP were statistically significant (t test, p < 0.0001).

Topographical analysis of gray scales showed similar scotoma detection for both strategies in all subjects. Examples are shown in Figures 4, 5 and 6. Patients 9 (Fig. 4) and 16 (Fig. 5) are examples of good correlation between strategies. Patient 7 (Fig.

Fig. 4. Patient 9: Bjerrum scotoma with good correlation. a. 4/2 strategy; b. CLIP strategy with thresholds and gray scales.

6) had the worst correlation in this study. The deep relative scotoma in the lower part of the visual field is less pronounced with CLIP compared to 4/2.

We then calculated threshold differences between 4/2 and CLIP at single test points for all subjects. These distributions are shown in Figures 7a, b and c.

Absolute differences between CLIP (2 and 3) and 4/2 at single test points (excluding the blind spot) were 3.8 dB on average (median 3 dB). Seventy-five percent of absolute differences were less than 6 dB, 95% were less than 13 dB. Large differences were mostly located at the borders of absolute scotoma, where the test location is within the absolute scotoma with one strategy and within a normal area with the other. The comparison of absolute differences for CLIP 2 and 3 showed a very similar distribution. Absolute differences were 3.8 dB on average (median 2 dB), 75% of absolute differences were smaller than 5 dB, 95% were smaller than 13 dB. As in the comparison 4/2 with CLIP, large threshold differences were also located at the borders of absolute scotomas.

Mean test times were 8.84 minutes (SD 0.61 min; range 7.97–10.2 min) for 4/2, 5.46 minutes (SD 0.61 min; range 4.25-6.79 min) for CLIP 2, and 5.57 minutes (SD 0.95 min; range 4.47-9.6 min) for CLIP 3, respectively (Fig. 8). Mean test time was significantly shorter for CLIP (examinations 2 and 3) compared to 4/2 (t test, p <

Fig. 5. Patient 16: advanced glaucomatous visual field defect with good correlation. a. 4/2 strategy; b. CLIP strategy with thresholds and gray scales.

0.0001). CLIP saved 38% of test time (range 21-53%). All subjects had shorter mean test times with CLIP (2 and 3) than with 4/2.

Mean response time was 448 msec (SD 53 msec). No correlation was found between test and response times.

At the end of the examination, subjects were asked about their favorite strategy. They were asked to give frank answers, as we wanted to know their subjective feelings about the advantages and disadvantages of both strategies without giving them a predefined choice. So no statistic evaluation can be performed on these data.

Most subjects preferred CLIP compared to 4/2. They complained that the 4/2 strategy put too much pressure on them and was too fast. Many people mentioned that they needed to concentrate more with 4/2, especially at the end of the examination, and that they were often not sure if there really was a stimulus or not. The minority of people who found CLIP more tiring than 4/2 (despite the shorter test time) all had very large scotomas.

Most persons preferred CLIP over 4/2 because it seemed to be more ‘at their pace’. They found it positive that stimuli were definitely seen at a certain luminance, even at the end of the examination. A few people told the examiner that if they were not sure whether there really was a stimulus, they just waited for the stimulus to become a little brighter.

All people liked CLIP because of the shorter test time.

Fig. 6. Patient 7: advanced glaucomatous visual field defect with moderate correlation. a. 4/2 strategy; b. CLIP strategy with thresholds and gray scales.

Discussion

In this study, CLIP showed similar scotoma detection compared to standard full threshold strategy (4/2), but in a significantly shorter test time. Mean sensitivities were 1.5 dB higher than with the 4/2 strategy, absolute differences at single test locations (and therefore reproducibility) were comparable between 4/2 and CLIP and between separate CLIP tests. Most patients preferred CLIP to 4/2, possibly due to the shorter testing time, but also because the strategy seemed easier to perform and less tiring in itself than the standard full strategy.

The results of this study confirm our findings in normal subjects.9,10 Mean sensitivity in normal subjects was about 2 dB higher in CLIP than 4/2, reproducibility of CLIP was significantly better compared to 4/2 and fast threshold strategy and as good as SITA Standard. In normal subjects, CLIP was able to save 58% (53–60%) of test time and was the shortest strategy in all subjects (compared to 4/2, fast threshold, and SITA Standard). All normal subjects preferred the CLIP strategy.

Studies concerning the so-called ramp stimulus, comparable to CLIP, were basically undertaken before 1970,11-15 when the staircase strategies appeared in automated perimetry.2,14,15 But the findings of Capris et al.16 suggested that continuous luminance changes can be a good alternative to staircase procedures.

a.

b.

c.

Fig. 7. Distribution of differences at single test locations between: a. 4/2 and CLIP 2; b. 4/2 and CLIP 3; and c. CLIP 2 and 3.

Fig. 8. Mean test time.

Newer studies indicate that the ramp stimulus tests different cell populations from pulse stimulation. The study by Okamoto et al.17 suggests that the y-system is excited by pulse stimuli, the x-system by ramp stimuli. Kani et al.18 found a smaller diameter of receptive fields for ramp stimuli compared to pulse stimulation, and responses were also different for those obtained by pulse stimulation. The density of receptive fields for ramp stimuli was similar to that of x-cells, for pulse stimuli parameters were similar to the y-cells.

With the CLIP strategy, mean sensitivity was higher than 4/2 full threshold strategy in normal subjects9,10 and glaucoma patients. Analyzing the strategy, in CLIP, both local adaptation (Troxler, 18042; Cibis and Monjé, 195419) and summation phenomena20,21 could affect the results. As mean sensitivities are higher in CLIP, it can be presumed that this could be more an effect of temporal summation. But other studies showed that summation is only strong for a stimulus duration of up to 100 msec, and the influence on stimuli longer than 200 msec, and especially longer than 500 msec, is small.2,21 In CLIP, the stimuli are longer (one step per reaction time in a mean of 443 msec) and subthreshold until recognition. Mean sensitivities were higher despite the fact that some subjects waited until the stimulus was a little brighter, if they were not sure. Takashima et al.22 noted similar findings for ramp stimuli in normal subjects, especially in the central areas of the visual field. Their hypothesis was that it could be differences between the x- and y-systems.

Evaluation of mean sensitivity in studies concerning other new strategies such as Fastpac, SITA and TOP was inconsistent. In some studies, there were also higher mean sensitivities compared to standard full threshold, as we found with CLIP,5,6,23-25 while other studies did not find any differences.4,26,27 It was presumed that higher sensitivities could be due to a shorter test time, and therefore to less fatigue effect.24,30

Most new test strategies are designed, as is the case with CLIP, to save test time compared to the 4/2 strategy. Test time is shortened by between 30 and 80%,3-6,23-29 depending on the strategy and the subjects included. Test time is generally shorter and savings greater in normal subjects, as is found with CLIP and for SITA and Fastpac.28,29

Patient acceptance was not generally evaluated in studies about new strategies. Perhaps it was presumed that a shorter test time would be sufficient. But nowadays, not only test time should be kept in mind, but also patient compliance.

In this study we found that subjects with minimal to moderate visual field loss preferred CLIP. Subjects with severely damaged visual fields generally found both CLIP and 4/2 hard to perform. This can be explained since one of the advantages of CLIP is feedback on performance and the fact that the stimulus luminance increases until the patient can see it. Compared to 4/2, a larger proportion of stimuli are seen in the end.

Future modifications could include taking into account previous results and thresholds of neigboring locations. Such modifications will help make CLIP even more acceptable to patients due to a further reduction in test time.

Further studies will show whether CLIP performs as well in other diseases, e.g., ocular hypertension. The fact that CLIP is easy to perform would also help in examining patients with less experience or those who, for some other reason, do not cooperate during automated static perimetry.

Conclusions

CLIP is a new strategy in automated visual field testing which has been designed to save test time and enhance patient compliance. In glaucoma patients, CLIP was able to reduce test time by 38% when compared to the full threshold strategy, with good reproducibility. Patient acceptance was excellent.

References

1.Johnson CA, Chauhan BC, Shapiro LR: Properties of staircase procedures for estimating thresholds in automated perimetry. Invest Ophthalmol Vis Sci 33:2966-2974, 1992

2.Flammer J: Chapter I. Theoretische Grundlagen. In: Gloor B (ed) Perimetrie mit besonderer Berücksichtigung der automatischen Perimetrie, pp 34-59. Stuttgart: Enke 1993

3.Weber J: Eine neue Strategie für die automatisierte statische Perimetrie. Fortschr Ophthalmol 87:3741, 1990

4.Flanagan JG, Wild JM, Trope GE: Evaluation of FASTPAC, a new strategy for threshold estimation with the Humphrey Field Analyzer in a glaucomatous population. Ophthalmology 100:949-954, 1993

5.Bengtsson B, Heijl A, Olsson J: Evaluation of a new threshold visual field strategy, SITA, in normal subjects. Acta Ophthalmol Scand 76:165-169, 1998

6.Bengtsson B, Heijl A: SITA fast, a new rapid perimetric threshold test: description of methods and evaluation in patients with manifest and suspect glaucoma. Acta Ophthalmol Scand 76:431-437, 1998

7.Gonzalez de la Rosa M, Martinez A, Sanchez M: Accuracy of the tendency oriented perimetry (TOP) in the Octopus 1-2-3 perimeter. In: Wall M, Heijl A (eds) Perimetry Update 1996/97, pp 119-123. Amsterdam/New York: Kugler Publ 1997

8.Gonzalez de la Rosa M, Gonzalez-Hermandez M, Abraldes M, Azuara-Blanco A: Quantification of topographic correlation of threshold values in glaucomatous visual fields. Invest Ophthalmol Vis Sci 41(Suppl):85, 2000

9.Wabbels BK, Burk ROW, Kolling G: CLIP: an improved strategy in automated visual field testing. In: Perimetry Update 2000/2001, pp 177-186. The Hague: Kugler Publ 2001

10.Wabbels BK, Kolling G: Continuous light increment perimetry (CLIP strategy) as a new method measuring thresholds in automated static perimetry. Ophthalmologe 97(Suppl):149, 2000

11.Sloan LL: Instruments and technics for the clinical testing of light sense. Arch Ophthalmol 22:233251, 1939

12.Dannheim E: Statische Perimetrie im binokularen Sehakt. Orthoptik Pleoptik 1:14-21, 1973

13.Anderson DA: Perimetry With and Without Automation. St Louis, MO/Washington, DC: Mosby Co 1987

14.Spahr J: Optimization of the presentation pattern in automated static perimetry. Vision Res 15:12751281, 1975

15.Bebié H, Fankhauser F, Spahr J: Static perimetry: strategies. Acta Ophthalmol Scand 54:325-338, 1976

16.Capris P, Spinelli G, Zingirian M: Comparing continuous and stepwise luminance variation in static campimetry using the Grignolo-Tagliasco-Zingirian projection campimeter. Int Ophthalmol 8:5558, 1985

17.Okamoto O, Mimura O, Kani K, Inui T: Characteristics of two systems of human vision using fundus perimetry. Doc Ophthalmol Proc Ser 49:483-488, 1987

18.Kani K, Takashima M, Nagata S, Mimura O, Takubo K: An analysis of human visual receptive fields using the perimetric method. In: Lakshminarayanan V (ed) Basic and Clinical Applications of Vision Science, pp 283-286. The Hague: Kluwer Academic Publ 1997

19.Schober H VI: Die Adaptation. In: Schober H (ed) Das Sehen. Leipzig: Fachbuchverlag 1954

20.Krakau CE: Temporal summation and perimetry. Ophthalmic Res 21:49-55, 1989

21.Funkhouser A, Fankhauser F: Temporal summation measurements with the Octopus 1-2-3-perimeter. German J Ophthalmol 3:120-128, 1994

22.Takashima M, Nagata S, Tsukada H, Obata N, Kani K: Ramp stimulation perimetry testing the X- system. In: Mills RP, Wall M (eds) Perimetry Update 1994/95, pp 149-157. Amsterdam: Kugler Publ 1995

23.Shirato S, Inoue R, Fukushima K, Suzuki Y: Clinical evaluation of SITA: a new family of perimetric testing strategies. Graefe’s Arch Clin Exp Ophthalmol 237:29-34, 1999

24.Morales J, Weitzman ML, Gonzalez de la Rosa M: Comparison between tendency-oriented perimetry (TOP) and Octopus threshold perimetry. Ophthalmology 107:134-142, 2000

25.Sharma AK, Goldberg I, Graham SL, Mohsin M: Comparison of the Humphrey Swedish interactive threshold algorithm (SITA) and full threshold strategies. J Glaucoma 9:20-27, 2000

26.Schaumberger M, Schäfer B, Lachenmayr BJ: Glaucomatous visual fields: FASTPAC versus full threshold strategy of the Humphrey Field Analyzer. Invest Ophthalmol Vis Sci 37:1390-1397, 1995

27.Maeda H, Nakaura M, Negi A: New perimetric threshold test algorithm with dynamic strategy and tendency oriented perimetry (TOP) in glaucomatous eyes. Eye 14:747-751, 2000

28.Remky A, Arend O: Klinische Erfahrungen mit dem ‘Swedish interactive threshold algorithm’ (SITA). Klin Mbl Augenheilk 216:143-147, 2000

29.Roggen X, Herman K, Van Malderen L, Devos M, Spileers W: Different strategies for Humphrey automated perimetry: FASTPAC, SITA standard and SITA fast in normal subjects and glaucoma patients. Bull Soc Belg Ophtalmol 279:23-33, 2001

30.Hudson C, Wild JM, O’Neill EC: Fatigue effects during a single session of automated static threshold perimetry. Invest Ophthalmol Vis Sci 35:268-280, 1994