| OCR Text |
Show ORIGINAL CONTRIBUTION Laser Pointer Visual Field Screening Michael S. Lee, MD, Laura J. Balcer, MD, MSC. E., Nicholas J. Volpe, MD, Grant T. Liu, MD, Gui S. Ying, MS, and Steven L. Galetta, MD Background: Sensitivity of confrontation visual field ( CVF) screening is low unless defects are significant. We compared the sensitivity of laser pointer visual field screen-ing ( LVF) with conventional CVF for identifying eyes with abnormal automated perimetry. Methods: Ninety consecutive patients presenting for HVF prospectively underwent a masked comparison of CVF and LVF testing ( 175 eyes) from April to May 2000. LVF was performed using a laser pointer target projected onto a tangent screen. Points were tested in random fashion on ei-ther side of the vertical and horizontal meridians, near cen-tral fixation, around the blind spot, and in each quadrant. Single and double simultaneous finger counting was used to test CVF. Results: LVF demonstrated significantly greater sensitiv-ity as compared with CVF ( 73% versus 31%, P = 0.001) in identifying field defects found on HVF. Specificities for LVF and CVF were 82% and 99%, respectively. The aver-age testing times per eye were 0.5 minute for CVF, 1.5 min-utes for LVF, and 8.0 minutes for HVF. Conclusions: In this cohort, laser visual field testing was significantly more sensitive than confrontation testing. It may represent an effective, time- efficient tool for visual field screening. ( J Neuro- Ophthalmol 2003; 00: 260- 263) Computerized perimetry, such as Humphrey visual field ( HVF) testing, has become the gold standard in visual field assessment. However, increased time, office space, and staff make it impractical to test every patient. Confron-tation visual fields ( CVF) represent a quick and simple screening test but miss many field defects. Several authors have compared CVF to formal perim-etry and found it relatively insensitive. 1- 3 One report found 40% sensitivity when compared with HVF. 1 An-other report concluded that " confrontation visual field test-ing is relatively insensitive unless a moderate to dense de-fect is present, and as such is a poor screening test". 2 A tangent screen technique using a projection light as the tar-get has been described, 4 but this technique has not been formally compared with established forms of visual field testing. The primary purpose of this study was to compare the sensitivities of laser pointer visual field testing ( LVF) and conventional CVF testing in the detection of visual field abnormalities found on HVF. 5 METHODS After obtaining institutional review board approval, we prospectively screened all patients presenting for HVF from April 2000 to June 2000 with CVF and LVF testing at the Philadelphia Veterans Administration ( VA) Medical Center. At the time of screening, the examiner was masked to previous diagnoses or visual field defects. Informed con-sent was obtained from each participant. Three experienced, masked graders evaluated each HVF and LVF by consensus. Using all information from HVF ( corrected pattern standard deviation, total and pattern plots, mean deviation), each field was judged normal or ab-normal. Abnormal HVFs were categorized as one of 14 monocular field defects based on Optic Neuritis Treatment Trial ( ONTT) criteria. 6 Degree of abnormality was graded as minimal ( MD better than − 6.00 dB), moderate ( MD worse than − 6.01 dB but better than − 20.0 dB), or severe ( MD worse than − 20.01 dB). CVFs were considered abnormal if there was any error or subjectively perceived scotoma. CVF was performed with the patient and examiner seated 2/ 3 meter apart. 1- 3 After occluding one eye and From Cole Eye Institute, Cleveland Clinic Foundation ( MSL), Cleve-land, Ohio, and the Departments of Ophthalmology ( LJB, NJV, GTL, GSY, SLG) and Neurology ( LJB, GTL, SLG) at the University of Penn-sylvania School of Medicine, and Philadelphia Veterans Administration Medical Center ( NJV), Philadelphia, Pennsylvania. Reprints: Steven L. Galetta, M. D., 3 East Gates, 3400 Spruce Street, Philadelphia, PA 19104. E- mail: galetta@ mail. med. upenn. edu Supported by the Heed Foundation, Cleveland, OH ( MSL) and by NIH Grant EY 00351 ( LJB). The authors have no proprietary interest in the publication of this study. Presented in part at the North American Neuro- Ophthalmology Soci-ety Annual Meeting, Palm Springs, CA, February 2001 and the Associa-tion for Research in Vision and Ophthalmology, Ft. Lauderdale, FL, May 2001. 260 J Neuro- Ophthalmol, Vol. 23, No. 4, 2003 fixating on the examiner's nose, the patient subjectively identified any parts of the examiner's face that appeared to be dim or " missing." Single and double simultaneous finger counting was used in each quadrant to assess peripheral fields. LVF was performed using a commercially purchased laser pointer ( Beta Electronics, OH). The red laser beam had a wavelength of 633- 680 nm ( Class IIIa Laser) with less than 5 mW of output power. The same room, 1- m tan-gent screen, and lighting conditions ( lights on) were used for CVF and LVF testing. The examiner stood beside the patient ( outside the patient's field of view), seated 1mfrom the tangent screen ( Fig. 1). The laser was calibrated each time with a spot drawn on the wall measuring 1.75 mm di-ameter ( visual angle 0.1°). Occluding one eye, the patient fixated on the tangent screen central button and verbally signaled when the laser target was noted. Initially, the laser target was presented in the periphery and was brought slowly toward fixation to assess for constriction in the four quadrants. Static points were tested in random fashion on either side of the vertical and horizontal meridians, near central fixation, around the blind spot, and in each quadrant ( Fig. 2). 4 The blind spot was repeatedly tested to assess fixation losses. Scotomas were recorded and rechecked for reproducibility. Automated perimetry was performed prior to appla-nation tonometry using a Humphrey Visual Field Analyzer ( Humphrey Instruments, San Leandro, CA) with standard testing conditions: 31.5 asb white background, a size III white stimulus, and either 24- 2 FastPac or 24- 2 SITA Stan-dard protocol. Eyes with more than 33% fixation losses, false negatives, or false positive values were excluded for unreliable visual field parameters, as in current glaucoma trials. 7 Statistical analyses were performed using SAS 8.0 ( SAS Institute, Inc., Cary, NC) and Stata 7.0 ( StataCorp, College Station, TX). Sensitivities and specificities for LVF and CVF were compared using alternating logistic re-gression options for GEE estimation, adjusting for inter- eye and inter- test correlations. 8 RESULTS One hundred seventy- five eyes of 90 patients were evaluated. Twenty- five eyes were excluded because of un-reliable HVF parameters ( including both eyes of 6 pa-tients); 150 eyes of 84 patients were included in the final analysis. Twenty- five eyes underwent SITA threshold testing and 125 eyes had the FastPac program. No patient was excluded for fixation losses during LVF. There were 82 men and 2 women ( approximating the Philadelphia VA Medical Center patient population). The mean age was 66 ± 12 years. Diagnoses included glaucoma ( n = 63, 42%), glaucoma suspect ( n = 72, 48%), anterior ische-mic optic neuropathy ( n = 5, 3%), hydroxychloroquine screening ( n = 5, 3%), stroke ( n = 3, 2%), and optic neuritis ( n = 2, 1%). HVFs were categorized by expert opinion as normal in 91 eyes ( 61%) and abnormal in 59 ( 39%). LVF was ab-normal in 59 eyes ( 39%), while CVF was abnormal in 19 ( 13%) ( Table 1A). The sensitivity of LVF, 73%, was sig-nificantly greater than that of CVF, 31% ( P = 0.001, ad-justed by inter- eye, inter- test correlation) ( Table 1B). Sen-sitivities were 40% for LVF vs 10% for CVF with minimal HVF defects ( n = 20), 91% for LVF vs 33% for CVF with moderate HVF defects ( n = 33), and 83% for LVF vs 83% for CVF with severe HVF defects ( n = 6). Average testing times were 0.5 minutes for CVF, 1.5 minutes for LVF, and 8.0 minutes for HVF. FIGURE 1. Laser pointer field testing using a 1- m tangent screen. FIGURE 2. Protocol for points tested with the laser pointer. Laser Pointer Visual Fields J Neuro- Ophthalmol, Vol. 23, No. 4, 2003 261 For the eight ONTT field defect categories, the over-all agreement between LVF and HVF was 60% ( Table 2). Eight of the 25 excluded eyes had abnormal CVF and each had an abnormal LVF as well. Four other excluded eyes had abnormal LVF with normal CVF. DISCUSSION Confrontation techniques are widely used to screen for visual field defects as part of the neuro- ophthalmologic examination. Defects may be asymptomatic, particularly when they spare central vision, and may be difficult to iden-tify by CVF techniques unless they are of moderate to se-vere density. 1- 3 Ideally, screening tests should be highly sensitive and specific, but a trade- off is usually required. If the subsequent diagnostic evaluation involves a test with minimal cost or risk, such as HVF, then screening should have high sensitivity. 9 We have demonstrated that LVF can be used in the clinical setting with significantly greater sensitivity than confrontation as a screening tool. This LVF sensitivity increases if the field defects are of moderate density (− 6.01 dB< MD<− 20.0 dB). Intuitively, testing more spots could increase sensitivity, but time of testing would increase. Adding colored objects to confrontation could increase its sensitivity, but time of testing would increase. We believe that the majority of patients examined as outpatients by ophthalmologists, neurologists, and neuro- ophthalmolo-gists receive CVF based on finger counting alone. Thus, we chose not to include colored objects in our experimental CVF protocol. Because we did not select patients according to par-ticular diagnosis, but rather by prevalence, most of our patients had glaucoma or were glaucoma suspects ( 90%). Because visual field defects in neurologic disorders are typically denser, 2 LVF and CVF may have had greater sensitivities in this population of patients. Further testing of this laser technique is warranted in a group of patients with other neuro- ophthalmic conditions. Nonetheless, glaucoma is a reasonable disorder to test because the field defects may be subtle and are frequently undetectable with confron-tation. Notably, HVF, the gold standard for visual field defects used in this study and many others, was originally developed for the detection of glaucomatous visual field defects. One limitation of this study is that LVF was per-formed immediately prior to CVF. Therefore, the examiner knew if the LVF was abnormal prior to performing CVF. This order of screening may falsely raise the sensitivity of CVF, perhaps biasing the difference in sensitivities toward the null. In spite of this, the difference in sensitivities was highly significant ( P = 0.001), making it an even more com-pelling observation. On the other hand, the observer could have been biased in the opposite direction by believing that TABLE 1B. Sensitivity and specificity of laser visual field ( LVF) and confrontation visual field ( CVF) testing to defects found with the Humphrey visual field analyzer Sensitivity Specificity Estimate 95% CI* Estimate 95% CI* LVF 0.73 0.59- 0.81 0.82 0.77- 0.95 CVF 0.31 0.17- 0.38 0.99 0.92- 1.00 * 95% Confidence interval adjusted by intereye, intertest correlation. TABLE 2. Graders' interpretations of Humphrey visual field ( HVF) and laser visual field ( LVF) Field interpretation No. of eyes (%) HVF LVF Normal 91 ( 61%) 91 ( 61%) Arcuate scotoma 21 ( 14%) 33 ( 22%) Double arcuate 2 ( 1%) 7 ( 5%) Altitudinal 12 ( 8%) 4 ( 3%) Enlarged blind spot 2 ( 1%) 4 ( 3%) Central scotoma 2 ( 1%) 1 ( 1%) Peripheral rim 5 ( 3%) 2 ( 1%) Nasal step 15 ( 10%) 8 ( 5%) TABLE 1A. Numbers of patients with normal and abnormal laser visual field ( LVF) results and confrontation visual field ( CVF) results for a given Humphrey visual field ( HVF) result LVF HVF result CVF HVF result Normal Abnormal Total Normal Abnormal Total Normal 71 20 91 Normal 79 52 131 Abnormal 9 50 59 Abnormal 1 18 19 Total 80 70 150 Total 80 70 150 J Neuro- Ophthalmol, Vol. 23, No. 4, 2003 Lee et al 262 © 2003 Lippincott Williams & Wilkins the laser was a superior detecting instrument. We tried to limit this bias by accepting any wrong answer as an abnor-mal CVF. Another limitation is that the sample population is predominantly male. There is no evidence, however, that females would perform any differently on visual field test-ing. Finally, LVF used only one size and one luminance ( suprathreshold to any HVF stimulus). Varying the size and intensity of light could have raised the sensitivity of LVF. As recognized above, this would increase testing time. In clinical practice, a black felt tangent screen can be placed on the wall behind the examining table or chair. To perform LVF, the patient can face the screen, and the pocket laser pointer can be quickly used to screen the visual field. A bedside consultation without the benefit of a tangent screen could be performed on a wall or ceiling with a central fixation target. This would provide a unique opportunity to screen patients who are unable to leave their hospital room for formal perimetry. Finally, outpatients who are un-able to fit into or remain in a Humphrey perimeter would benefit from the increased sensitivity of LVF over CVF testing. In this prospective study, we have demonstrated that LVF testing, performed using a commercially available la-ser pointer projected onto a tangent screen, is significantly more sensitive than confrontation visual field testing with fingers in screening for HVF visual field defects in this cohort. Although HVF and other perimetric techniques remain the standard for documenting visual field defects in patients with suspected afferent visual pathway dis-ease, LVF may represent a more practical yet ade-quately sensitive method for office and bedside visual field screening. REFERENCES 1. Johnson LN, Baloh FG. The accuracy of confrontation visual field test in comparison with automated perimetry. J Natl Med Assoc. 1991; 83: 895- 898. 2. Shahinfar S, Johnson LN, Madsen RW. Confrontation visual field loss as a function of decibel sensitivity loss on automated static pe-rimetry. Ophthalmology. 1995; 102: 872- 877. 3. Trobe JD, Acosta PC, Krischer JP, et al. Confrontation visual field techniques in the detection of anterior visual pathway lesions. Ann Neurol. 1981; 10: 28- 34. 4. Farris BK. The Basics of Neuro- Ophthalmology. St. Louis: Mosby Yearbooks; 1991: 54. 5. Johnson CA. Standardizing the measurement of visual fields for clinical research. Ophthalmology. 1996; 103: 186- 189. 6. Keltner JL, Johnson CA, Spurr JO, et al. Visual field profile of optic neuritis: one- year follow- up in the optic neuritis treatment trial. Arch Ophthalmol. 1994; 112: 946- 953. 7. Gordon MO, Kass MA. The ocular hypertension treatment study: design and baseline description of the participants. Arch Ophthal-mol. 1999; 117: 573- 583. 8. Qaqish BF, Liang KY. Marginal models for correlated binary re-sponses with multiple classes and multiple levels of nesting. Bio-metrics. 1992; 48: 939- 950. 9. Hennekens CH, Buring JE. Epidemiology in Medicine. Boston: Little, Brown and Co; 1987: 327- 345. Laser Pointer Visual Fields J Neuro- Ophthalmol, Vol. 23, No. 4, 2003 263 |
