| OCR Text |
Show Journal of Neuro- Ophthalmology 19( 2): 89- 99, 1999. © 1999 Lippincott Williams & Wilkins, Inc., Philadelphia Comparison of Threshold Visual Perimetry and Objective Pupil Perimetry in Clinical Patients Takeshi Yoshitomi, Takako Matsui, Akihiro Tanakadate, and Satoshi Ishikawa Objectives: In an attempt to measure the visual field objectively, we have performed pupil perimetry, by which the pupil light reflex is monitored in response to perimetric light stimuli. The purpose of this study was to ascertain whether pupil perimetry reveals defects similar to those revealed by standard threshold perimetry in patients with various diseases. Material and Methods: An infrared pupillometer was linked to an automated perimeter to record, at each perimetric location, 76 pupil contractions, which were comparable to the test locations of the Humphrey Field Analyzer ( HFA 30- 2 program; Humphrey, San Leandro, CA). One hundred eighteen patients with various diseases were investigated. Results: Ninety- one patients ( 77.1%) maintained a pupil area large enough ( more than 10 mm2 in area) to respond adequately to focal light stimuli throughout the test. The correlation between the pupil field and the threshold visual field was subjectively judged to be good in most cases. However, pupil perimetry showed less damage than that seen in threshold perimetry in six of nine patients who had Leber's hereditary optic neuropathy ( LHON). Conclusions: Pupil perimetry is a good method for measuring the visual field objectively and has potential for clinical use in most of the cases. Key Words: Humphrey Field Analyzer- Leber's hereditary optic neuropathy- Objective perimetry- Pupil perimetry- Visual field. Automated threshold perimeters of various kind have been developed recently and play a major role in the evaluation of patients with visual field loss ( 1,2). These perimeters enable ophthalmologists to find earlier visual field loss or to analyze visual field change using a computer. However, these perimeters measure visual field subjectively and require patient cooperation, which is sometimes limited. In an attempt to measure the visual field objectively, we performed pupil perimetry, to monitor the pupillary constriction in response to perimetric Manuscript received July 29, 1998; accepted February 10, 1999. From the Department of Ophthalmology ( T. Y., T. M., S. I.), Kitasato University School of Medicine; and the Department of Clinical Engineering ( A. T.), Kitasato University, School of Allied Health Sciences, Kitasato, Kanagawa, Japan. Address correspondence and reprint requests to Takeshi Yoshitomi, M. D., Department of Ophthalmology, Wakayama Medical College, 7- Bancho 27, Wakayama, 640- 8155, Japan. light stimuli. The concept of pupil perimetry, which uses the light reflex to monitor the visual function objectively, is not new. Pupillary hemiakinesia, the absence of pupillary response in the blind hemifield of a visual field, was first noted in 1880 ( 3,4). The first attempt to measure the visual field using pupil perimetry was undertaken by Harms ( 5), and attempts were subsequently made by many investigators ( 6- 9). However, these investigations concentrated testing along the horizontal meridian and also required much effort to analyze the pupil movement without a computer. With the development of automated perimetry and computed pupillogra-phy, these laborious, time- consuming analyses have been streamlined, making automated pupil perimetry available for clinical use ( 10,11). Kardon et al.( 10) reported automated pupil perimetry by linking an infrared pupillometer to a Humphrey Field Analyzer ( HFA; Humphrey, San Leandro, CA), which seemed to be useful for assessing the visual field objectively. However, the correlation between pupil field and the threshold field is not always high. This may be because some diseases affect the pupillary input differently than the visual input. In an attempt to clarify these questions, we have also developed pupil perimetry. A preliminary report of this method has been made ( 12). This report compared the results of standard automated perimetry in normal subjects and in patients with various diseases using pupil perimetry and standard threshold perimetry. SUBJECTS AND METHODS Subjects Eleven normal volunteers ( mean age, 32.7 ± 9.9) were recruited and examined in our clinic before testing and had normal visual field threshold testing with the HFA to ensure that they were free of ocular disease, excepting refractive error. One hundred and eighteen patients with various diseases also participated in this study ( Table 1). Miotics, such as pilocarpine were not administered to these patients for at least a year before testing and mydriatics were not administered at least a month before measurement. Visual fields were measured first by the HFA 30- 2 program. After providing informed consent, the volunteers were asked to participate in pupil perim- 89 90 T. YOSHITOMIET AL. TABLE 1. Patients tested by pupil perimetry Diagnosis Glaucoma Postgeniculate lesion Retinitis Pigmentosa Functional field loss LHON Pituitary tumor AION Others Total Number of patients 50 11 3 9 10 5 20 10 118 Male/ female 25/ 25 8/ 3 2/ 1 5/ 4 10/ 0 1/ 4 12/ 8 6/ 4 69/ 49 Mean age ± S. D. 59.7 ± 11.8 58.5 ± 18.0 62.0 ± 8.5 28.2 ± 8.5 34.7 ± 9.7 44.2 ± 20.7 40.0 ± 17.7 46.2 ± 20.1 50.0 ± 18.0 AION, anterior ischemic optic neuropathy; LHON, Leber hereditary optic neuropathy. Others included toxic optic neuropathy, retinal vein occulusion, retinal artery occlusion, aged macular degeneration, multiple evanescent white- dot syndrome, and visual fields abnormalities of unknown origin. etry testing within a week after the threshold visual field test. Pupil Perimetry Figure 1 shows the general hardware arrangement for our pupil perimetry, which is somewhat similar to that reported by Kardon et al. ( 10) An infrared video camera ( C2400- 07; Hamamatsu Photonics, Hamamatsu, Japan) was linked to an automated perimeter ( AP- 3000; Kowa, Japan) at the fixation- monitoring scope to record direct pupil response. An infrared light source was projected into the bowl so that its reflection onto the eye provided enough illumination of the eye for the infrared video v camera. An area analyzer ( C- 3163; Hamamatsu Photonics, Hamamatsu, Japan) constantly counted the number of pixels in the pupillary area ( 60 times/ seconds) and was relayed to the computer ( PC- 9800; NEC, Tokyo, Japan). The video image was obtained from a 40 ( horizontal) x 30- mm ( vertical) window area, with a resolution of 2560 pixels x 1936 pixels. A bowl background of 31.5 apostilb ( asb) was reduced to 5 asb, because high- intensity background caused the pupil to become too small, which limited its dynamic range of movement. Stimulus intensity was kept at 1000 asb, and a size V stimulus ( 1.7°) was used with a stimulus duration of 0.2 seconds. As we have reported previously, the blind spot could not be plotted by size V stimulation. Although it is possible to plot the blind spot with a size III stimulus, we found that three of seven normal eyes had no pupillary response to a size III stimulus at 30° from fixation ( 12). We therefore concluded that some intraocular scattering occurs with a size V stimulus and that scotomas smaller than the blind spot cannot be detected with a size V stimulus. We tested 76 perimetric locations that were comparable to that of HFA 30- 2 with a stimulus interval of 3.0 seconds. Single measurement at each location was performed using a custom software program ( Kowa). A software algorithm was also developed to detect artifact due to blinks or fixation loss caused by eye movement. In such cases, the software ignored this response and sent a signal to the perimeter to repeat the stimulus at the same perimetric location. Testing time was approximately 4 Automated Perimetry Video Monitor PM- 953T • • • • • • C3160 ( M3154) Area analyzer Trigger Signal Printer FIG. 1. General hardware arrangement for pupil perimetry. J Neuro- Ophthalmol, Vol. 19, No. 2, 1999 AUTOMATED PUPIL PERIMETRY IN CLINICAL PATIENTS 91 3.18 180 3.78 3.38 150 3.81 132 3.87 151 125 A 3.71 3* 0 184 3.87 3.98 2.98 2.77 141 2.81 148 173 178 4.09 4.41 4.08 144 2* 1 ISO 100 131 189 4.13 841 4* 9 180 139 182 128 18S 4.19 4.14 4.48 4* 1 4.98 4.05 109 143 104 104 121 179 424 4* 4 4.42 420 188 115 2.01 138 131 427 4.78 184 188 149 108 170 4.07 171 190 187 182 133 3.78 3.45 2.98 1.07 140 1.92 207 1* 9 1.88 1.10 1.37 1.82 1.59 B 205 1.00 127 227 1.88 1.39 1* 8 1.53 1* 9 1.78 1.03 123 1.74 142 2.01 1.81 154 1.07 1.59 1.50 1* 7 1* 7 1.92 0* 1 2.07 1.74 1.71 1.58 1* 8 2.08 1.40 1.89 1.48 1.42 1.15 1.83 1.00 1.78 1.79 2* 8 1* 2 1* 2 1* 8 1.97 1.95 0* 3 0* 3 1* 2 1.40 109 0* 3 1.88 i 0* 8 1.18 1.48 1.10 118 1.78 k, 2 L 108 1.12 2.09 128 1.84 1.03 8^ S88$ 8$ nr , \ :~ n : ] m Central area Middle area | Peripheral area 1. H ± 1.3B Uddl* i rH Ptrfphtnil m Mefdl* * rt* FXIptittM * n * FIG. 2. A: Distribution of average amplitude of pupil contraction ( in square millimeters) from nine normal subjects within the 30° field. Each location was tested once. Data were obtained from the right eye in each subject. B: Distribution of average differences between first and second measurement ( 30- minute interval). C: Seventy- six locations were divided into three areas: central, middle, and peripheral. D: Distribution of average amplitude of pupil contraction ( in square millimeters) in the central, middle, and peripheral areas. E: Distribution of average differences between first and second measurement in the central, middle, and peripheral areas. J Neuro- Ophthalmol, Vol. 19, No. 2, 1999 92 T. YOSHITOMI ET AL. tfumphmy^ ieldjAnalyzer Pupil Perimetry CO II. t w s c t. t fl. l i- A udi liiisi Mil1 1,4 t . l 3, » a. i ! . ' • 1- 4 li :. FIG. 3. A 76- year- old woman with normal- tension glaucoma. Results of threshold perimetry ( Humphrey Field Analyzer; Humphrey, San Leandro, CA; top) are compared with the results of pupil perimetry ( bottom). In pupil perimetry, each gray scale represents the range of the amplitude of contraction in square millimeters, shown at the bottom. minutes ( 3.8- 5.0). Refractive correction was not made, because lens correction alters the size of the pupil in the monitor. We checked all patients' status throughout the measurement to maintain their concentration on the fixation point. The visual field obtained by pupil perimetry was expressed by mapping the change of the pupil area from baseline to the trough of the pupillary light reflex at each target location within the field. We did not analyze latency data because preliminary investigation suggested that the latency time was as variable as amplitude of contraction when measuring the patients with small pupils or with dark irises because of the noise. Gray- scale maps ( with 11 scales) of the pupil fields were also displayed using a custom program in which the location with the maximum amplitude of pupil constriction of each eye was defined as 10 ( white) and that with no constriction was defined as 0 ( black). J Neuro- Ophthalmol, Vol. 19, No. 2, 1999 AUTOMATED PUPIL PERIMETRY IN CLINICAL PATIENTS 93 Humphrey Field Analyzer HK* W H Pupil Perimetry OMiV IVMX fl, » i i : ra! I . I ( txt sfM. a \ m mm II s - 1 i . i M i t S. J .1.11 4.4 F. Ti 1,1 !- » 2 4.4 FIG. 4. A 72- year- old man with open- angle glaucoma. Results of threshold perimetry ( Humphrey Field Analyzer; Humphrey, San Leandro, CA; top) are compared with the results of pupil perimetry ( bottom). RESULTS Patients and normal subjects found pupil perimetry easier than threshold perimetry because it required them only to concentrate on the fixation point, without having to respond if they saw the target. One normal subject had pupils that were too small ( less than 10 mm2 in area and 3.57 mm in diameter) at the beginning of the test to detect reliable pupil constrictions in response to focal perimetric stimuli and was excluded from the study. Among 118 patients with various diseases, 91 patients ( 77.1%) maintained a pupil area large enough ( more than 10 mm2 in area) to respond to focal light stimuli throughout the test without having to exclude any test locations. Fifteen patients had pupils larger than 10 mm2 initially; however, their pupils gradually became smaller during testing, which made the results of the last half of the testing unreliable. Eight patients had pupils less than 10 mm2 throughout the measurement. Small pupils in these patients may be partly caused by convergence on the fixation target. Four patients could not control their blinks. J Neuro- Ophthalmol, Vol. 19, No. 2, 1999 94 T. YOSHITOMIET AL. Humphrey Field Analyzer iiiiiiKiJSS& M. : : i | : : [ ! r i : : ! j j ! i l j ; ^ * ,_ ; v : i i i : i ; i i : " : : : ! " i i " l ! i':!!! ii: il:-': i::": i: iLr |!; ii| i- l.::, irli: » !| i| : i; Lii;< iifc& i: j:: iiii Pupil Perimetry OKAVSCflJ. |" fl. O n > I .' VI ( , t VI 8tA* StALt t . l nsflii 1.3 » . l I. I 9!! H = SrS I , T a, I l. u FIG. 5. A 68- year- old man with homonymous visual field loss due to a postgeniculate lesion. Top: Humphrey Field Analyzer ( Humphrey, San Leandro, CA). Bottom: Pupil perimetry. Normal Subjects The distribution of the average amplitude of pupil contraction ( in square millimeters) from 10 normal subjects within the 30° field is shown in Figure 2A. The same 76 locations were tested that were used in HFA 30- 2 and each location was tested once. The same test was repeated once more after a 30- minute interval. The distribution of average differences between the first and second measurements is shown in Figure 2B. Seventy- six locations were divided into three areas ( Fig. 2C). A significant difference was found in contraction amplitude among the three areas ( repeated- measures analysis of variance; n = 10; P = 0.0002; Fig. 2D). However, the difference in variability was not statistically significant among the areas ( Fig. 2E). J Neuro- Ophthalmol, Vol. 19, No. 2, 1999 AUTOMATED PUPIL PERIMETRY IN CLINICAL PATIENTS 95 Humphrey Field Analyzer Pupil Perimetry C* JW SC*. E 0- » O. B mm mm MAT ( CALI M i r 4.1 1,4 EiiEffi t. i » T i • M M 4.4 S. i t. a « . T 1,4 FIG. 6. A 53- year- old woman with retinitis pigmentosa. Top: Humphrey Field Analyzer ( Humphrey, San Leandro, CA). Bottom: Pupil perimetry. Patients With Various Diseases In two patients with dense field loss due to glaucoma, both pupil perimetry and threshold perimetry were performed by HFA ( Figs 3 and 4). These patients had received no therapy that would affect the pupil reaction, such as miotic agents or surgery. The field seems similar between the two methods when subjectively compared. However, not every patient showed good correlation between results of pupil perimetry and light threshold perimetry. Among 38 patients with glaucoma who successfully completed pupil perimetry, more damage was shown in 4 by pupil perimetry than by threshold perimetry. All four of these patients had minimal field damage, and variability of the pupil J Neuro- Ophthalmol, Vol. 19, No. 2, 1999 96 T. YOSHITOMI ET AL. Goldmann Perimeter Pupil Perimetry flIWV SCALE 5I> ••.', SfJILE mm mm *. i g. t » .* T. I I 1 » . « j . a • .* 4.) o . - in R FIG. 7. A 39- year- old man with functional field loss. Goldmann's perimetry showed a spiral kinetic field ( top); pupil perimetry showed a normal field ( bottom). response may have exaggerated the pupil perimetry field damage. One patient showed less damage with pupil perimetry than with threshold perimetry and had large and very reactive pupils. A patient with homonymous visual field loss due to a postgeniculate lesion is shown in Figure 5. Pupil response was decreased in the area of field loss in all nine patients tested including one patient who was tested 5 days after a stroke. All of these patients had neither midbrain area nor optic tract damage, according to magnetic resonance scan. The visual field in a patient with retinitis pigmentosa is shown in Figure 6. The fields obtained from both methods were very similar. The diagnosis of functional field loss was considered to be the most useful made by pupil perimetry. Such a case is shown in Figure 7 in a subject with a spiral kinetic field and a normal pupil field, which confirmed our clinical impression. The results of visual field testing using both methods in a patient with LHON are shown in Figure 8. The threshold perimetry showed marked visual field loss. However, pupil perimetry showed smaller damage than that seen with threshold perimetry. This J Neuro- Ophthalmol, Vol. 19, No. 2, 1999 AUTOMATED PUPIL PERIMETRY IN CLINICAL PATIENTS 97 Pupil Perimetry GRAY bl/ i. t, H H M na ^ n. • J. I 1.* 1.4 *, I mm mm EIEIII •=•• ::;::: 1.4 J. i t. W 1.4 1.1 T. I FIG. 8. The results of visual field testing using threshold perimetry ( top) and pupil perimetry ( bottom) in a 30- year old man with Leber's hereditary optic neuropathy. phenomenon was seen in six of nine patients with LHON. DISCUSSION According to the results in this study, pupil perimetry is advantageous in less cooperative or unreliable patients. It is also useful in aged paretic patients who have trouble pressing a button in threshold perimetry. Moreover, it is less stressful for the patients. The patients have only to concentrate on the fixation point without having to respond when they see the target; therefore, fixation loss did not seem to be as much of a problem in pupil perimetry as it is in threshold perimetry. Moreover, differential diagnosis of functional field loss is the most useful in pupil perimetry because this method is objective. However, not every patient is suitable for pupil perimetry. We tested 118 patients by pupil perimetry and failed to obtain reliable results in 27 patients ( 22.9%). The main problem in this measurement was the size of the pupil. Patients with very small pupils ( i. e., less than 10 mm2 in area) or whose pupils gradually constricted during measurements were not suitable for pupil perimetry. Automated threshold perimetry also has its limitations. J Neuro- Ophthalmol, Vol. 19, No. 2, 1999 98 T. YOSHITOMI ET AL. The reliability of a subject's automated perimetric test result is generally assessed by reliability index ( i. e., fixation loss, false negative rate, and false- positive rate). It has been reported that results in 45% of subjects with glaucoma ( first- time measurement) are considered unreliable when screened with reliability criteria ( 13). Therefore, a 22.9% failure rate in pupil perimetry in this study should not be considered unacceptable. We have studied nine patients with homonymous visual field loss due to a postgeniculate lesion. Because it is generally considered that the light reflex is purely a midbrain pathway, visual field defects with supragenicu-late lesions should not be detected by this method. In fact, Wernicke theorized that the light reflex in the blind part of the visual field could differentiate between pre-geniculate and postgeniculate lesions ( 14). However, pupil responses were decreased in the area of field loss in all cases in the present study. Moreover, all these patients had no midbrain area damage, according to magnetic resonance images. Similar results have been found by many investigators ( 5,15- 17). A number of hypotheses have been offered to explain these findings. One possible explanation is that transsyn-aptic degeneration may take place as described by van Buren ( 18). This could occur retrograde across the lateral geniculate synapse or could spread to the pretectal area, originating in the visual cortex. However, this explanation is unlikely, because we had one patient who was tested only 5 days after a stroke ( which is too early to develop transsynaptic degeneration) and still had decreased pupil responses in the area of field loss. Another explanation of an accessory pathway would require reconsideration of the basic anatomy of the light reflex, which Cibis et al. ( 17) concluded was unlikely. However, a cortical pathway to the pretectal area has been reported in cats ( 19). Therefore, it may be possible that some aspect of the pupillary light reflex may be mediated through a cortical pathway, at least under the stimulus conditions of pupil perimetry. As we have shown in this study, the subjective correlation between the pupil field and the threshold visual field was good in most cases. However, our pupil perimetry showed less damage than did threshold perimetry in six of nine patients with LHON. Wakakura et al. ( 20) showed that LHON is characterized by well- preserved afferent fibers for pupillary light response. Moreover, Kardon et al. ( 10) also showed disparities between the two methods in patients with optic neuritis. It is therefore concluded that some diseases damage pupillary input differently than they damage visual input. We must remember that the pupil field is pathophysiologically different from the threshold field, even though these two methods provide similar fields in most cases. Pupil perimetry measures the visual field by means of an evoked light reflex instead of a patient's subjective response. Many attempts have been made to use pupil perimetry in patients ( 10,12,21); however, there are no apparent data from patients with various diseases. We have observed that the pupil field and threshold field are notably different in patients with LHON and in patients with optic neuritis. It seems that both of these conditions damage pupillary input differently than they damage visual input. However, we have observed that subjective visual fields and pupil response fields are similar in patients with glaucoma, homonymous visual field loss, and retinitis pigmentosa. Although pupil perimetry provides a field similar to the threshold field in most cases, we must be cautious in our interpretation of the exceptions, because a different response system is under study. Moreover, it is possible that the visual field obtained from pupil perimetry may contain a component influenced by a possible cortical pathway in addition to the well- known midbrain reflex pathway. In conclusion, although there are potential problems in pupil perimetry, as in other forms of perimetry, it is a good method for measuring the visual field objectively and is potentially useful for clinical testing. Acknowledgment: Supported by the Japanese Ministry of Environment. REFERENCES 1. Harrington DO, Drake MV. Computerized perimeters. In: Klein EA, ed. The visual fields. Text and atlas of clinical perimetry. St. Louis: C. V. Mosby, 1990: 35- 56. 2. Heijl A, Krakau CET. An automatic static perimeter, design and pilot study. Acta Ophthalmol 1975; 53: 293- 310. 3. Heddaeus E. Klinishe Studien Uber die Beziehungen zwischen Pu-pillarreaction und Sehvermogen. Halle- Wittenberg: Diss Halle- Wittenberg, PlOtzsch Buchdruckerei, 1880: 65. 4. Wilbrand H. Uber Hemianopsie und ihr Verhaltnis zur topischen Diagnose der Gehirnkrankenheiten. Berlin: A. Hirschwald, 1881: 214. 5. Harms H. Grundlagen, Methodik und Bedeutung der Pupillenpe-rimetrie fur die Physiologie und Pathologie des Sehorgans. Graefes Arch Klin Exp Ophthalmol 1949; 149: 1- 68. 6. Bresky R, Charles S. Pupil motor perimetry. Am J Ophthalmol 1969; 68: 108- 12. 7. Narasaki S, Kawai K, Kubota S, Noguchi J. Videopupillographc perimetry and its clinical application. Jpn J Ophthalmol 1974; 18: 253- 74. 8. Aoyama T. Pupillographic perimetry ( Japanese). Nippon Ganka Gakkai Zasshi 1975; 79: 1247- 55. 9. Aoyama T. Pupillographic perimetry: clinical application ( Japanese). Nippon Ganka Gakkai Zasshi 1977; 81: 1527- 256. 10. Kardon R, Kirkali P, Thompson H. Automated pupil perimetry: pupil field mapping in patients and normal subjects. Ophthalmology 1991; 98: 485- 96. 11. Fankhauser F II, Flammer J. Puptrak 1.0: A new semiautomated system for pupillometry with the Octopus perimeter: a preliminary report. Doc Ophthalmol 1990; 73: 235- 48. 12. Yoshitomi T, Matsui T, Mukuno K, Ishikawa S. Objective visual field measurement using pupil perimetry ( Japanese). Nippon Ganka Gakkai Zasshi 1996; 100: 825- 31. 13. Katz J, Sommer A. Reliability indexes of automated perimetric tests. Arch Ophthalmol 1988; 106: 1252^ t. 14. Wernicke G. Uber hemianopische Pupillenreaction. ForstschrMed 1883; 1: 49- 53. 15. Walker CB. Topical diagnostic value of the hemiopic pupillary reaction and the Willbrand hemianoptic prisma phenomenon. JAMA 1913; 61: 1152- 6. 16. Yoshitomi M. The examination of pupillary light reflex with a / Neuro- Ophthalmol, Vol. 19, No. 2, 1999 AUTOMATED PUPIL PERIMETRY IN CLINICAL PATIENTS 99 specific pupilloperimeter, particularly hemianopic pupillary paralysis. Nippon Ganka Gakkai Zasshi 1955; 59: 858- 73. 17. Cibis, GW, Campos EC, Aulhorn E. Pupillary hemiakinesia in suprageniculate lesions. Arch Ophthalmol 1975; 93: 1322- 7. 18. Van Buren JM. Trans- synaptic retrograde degeneration in the visual system of primates. J Neurol Neurosurg Psychiatry 1963; 26: 402- 9. 19. Distler C, Hoffmann K. The pupillary light reflex in normal and innate microstrabismic cats: II. Retinal and cortical input to the nucleus pretectalis olivaris. Vis Neurosci 1989; 3: 139- 53. 20. Wakakura M, Yokoe J. Evidence for preserved direct pupillary light response in Leber's hereditary optic neuropathy. Br J Ophthalmol 1995; 79: 442- 6. 21. Donahue SP, Moore P, Kardon RH. Automated pupil perimetry in amblyopia. Generalized depression in the involved eye. Ophthalmology 1997; 104: 2161- 7. J Neuro- Ophthalmol, Vol. 19, No. 2, 1999 |
