COMPARISON OF DIFFERENT METHODS FOR DETECTING GLAUCOMATOUS VISUAL FIELD PROGRESSION

EIJA VESTI,1,2 BALWANTRAY C. CHAUHAN3 and CHRIS A. JOHNSON2

1Helsinki University Eye Hospital, Helsinki, Finland; 2Discoveries in Sight, Devers Eye Institute, Portland, OR, USA; 3Department of Ophthalmology, Dalhousie University, Halifax, Nova Scotia, Canada

Abstract

Purpose

To compare the performance characteristics of seven criteria for analyzing glaucomatous visual field progression, using a combination of real patient data and computer simulation.

Methods

Initial and final Humphrey 30-2 visual fields (separated by seven years) of 76 patients with open-angle glaucoma were entered into a computer simulation program, which generated 14 interim semi-annual fields under conditions of high, moderate and no variability. Progression was analyzed using the methods of the Advanced Glaucoma Intervention Study (AGIS), the Collaborative Initial Glaucoma Treatment Study (CIGTS), three criteria based on a glaucoma change probability-like (GCP-like) analysis, and two criteria based on point-wise linear regression analysis (PLRA). Specificities were calculated by using the same visual field of each patient as both the initial and final field (no progression), under conditions of moderate and high variability.

Results

For moderate variability, all criteria that were studied had high specificity (92% or higher). The AGIS, CIGTS, PLRA, and one of the GCP-like determinations were relatively robust in their specificity under high variability conditions. With no variability, progression rates were 18% for AGIS, 36% for CIGTS, 47% to 62% for the three GCP-like criteria, and 67-72% for the two PLRA criteria. Progression rates increased with greater variability for the three GCP-like criteria, and decreased for all other criteria. The time to detect confirmed progression was longest for the two PLRA criteria and shortest for the CIGTS and GCP-like criteria.

Conclusions

AGIS and CIGTS criteria had high specificity, but classified fewer cases of progression than the other criteria. GCP-like criteria had the shortest follow-up times to confirmed progression, but were not as specific. Criteria based on PLRA were specific but follow-up times to confirmed progression were the longest.

Address for correspondence: Eija Vesti, MD, Helsinki University Eye Hospital, 00029 Hus, Finland. Email: eija.vesti@hus.fi

Perimetry Update 2002/2003, pp. 39–40

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

Acknowledgment

Supported in part by grants from the Glaucoma Research Foundation (BCC, CAJ); EY03424 (CAJ) from the National Eye Institute, Bethesda, MD; MT-11357 (BCC) Medical Research Council of Canada; the Eye Foundation (Silmäsäätiö), Finland (EV); the Eye and Tissue Bank Foundation (Silmäja Kudospankkisäätiö), Finland (EV); and the Finnish Cultural Foundation, Finland (EV).

COMPARISON OF CAPRIOLI’S DECIBEL CRITERIA AND ANDERSON’S PROBABILITY CRITERIA FOR THE DETECTION OF GLAUCOMATOUS DEFECTS WITH SITA

AIKO IWASE,1,3 DAISUKE TAKAHASHI,1 YOSHIKI KONO,2 TETSUYA YAMAMOTO3 and YOSHIAKI KITAZAWA3

Departments of Ophthalmology, 1Tajimi Municipal Hospital,2Gifu Municipal Hospital and 3Gifu University School of Medicine, Gifu, Japan

Introduction

Swedish interactive threshold algorithms (SITA) tests realized the same high test accuracy as the standard full-threshold algorithm, and the reduction in the test duration, but on the other hand, have totally different characteristics compared to the full threshold program. We have to recognize the difference between these two algorithm and adopt the most useful strategy clinically.

For interpretation of the visual field, we often apply the decibel criteria to analyze the visual field test for detecting glaucomatous visual field defects with the full threshold strategy. In this prospective study, in order to investigate the efficacy of the traditional threshold-comparison method for interpreting a single SITA field, we evaluated the two common criteria for detecting glaucomatous visual field defects.

Subjects and methods

One eye of 47 experienced and reliable patients with glaucoma or ocular hypertension were examined. The clinical background is shown in Table 1.

Three strategies: the traditional full threshold strategy (FT), SITA standard (SS), and SITA fast (SF) were tested with the central 30-2 program of the Humphrey Field Analyzer, in random order on the same day. The visual fields were judged according to the two common criteria, i.e., Caprioli’s decibel criteria (1991)1 and Anderson’s criteria (1992)2,3 (Tables 2 and 3).

SITA results were compared with FT results assuming the latter as the ‘gold standard’ in each category.

Address for correspondence: Aiko Iwase, MD, Department of Ophthalmology, Tajimi Municipal Hospital, 3-43 Maebata-cho, Tajimi, Gifu, 507-8511, Gifu, Japan. Email: aiko-gif@umin.ac.jp

Perimetry Update 2002/2003, pp. 41–44

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

Age 57 ± 14 yr (Mean ± SD)

MD -3.84 ± 6.04 (dB)

Table 2. Caprioli’s criteria

Strict

≥4 adjacent points of ≥ 5 dB loss each

≥3 adjacent points of ≥ 10 dB loss each

Difference of ≥ 10 dB across nasal horizontal meridian at 3 adjacent points

Exclusions: physiologic blind spot; superior and inferior rows

Moderate

≥3 adjacent points of ≥ 5 dB loss each

≥2 adjacent points of ≥ 10 dB loss each

Difference of ≥ 10 dB across nasal horizontal meridian at ≥ 2 adjacent points

Exclusions: physiologic blind spot; superior and inferior rows

Liberal

≥2 adjacent points of ≥ 5 dB loss each

≥1 adjacent points of ≥ 10 dB loss each

Difference of ≥ 5 dB across nasal horizontal meridian at ≥ 2 adjacent points

Table 3. Anderson’s criteria

1Pattern deviation probability plot: a cluster of three or more non-edge points having the sensitivity with p < 5%, and one of the points had the sensitivity with p < 1%

2PSD or CPSD with p < 5%

3“Outside normal limits” by Glaucoma Hemifield Test (GHT)

Results

The results are shown in Table 4. The specificity value with Caprioli’s decibel criteria tended to be high, but the sensitivity was not high enough. This disagreement in the results was significant in cases with early glaucomatous changes (p < 0.01; Table 5). The sensitivity values with Anderson’s probability criteria for Statpac or Statpac for SITA were higher than those of Caprioli’s criteria. This disagreement was not significant according to stage of the visual field.

Agree: all judge of three categories showed agree.

Disagree: If one of the three categories were disagree, the eye were counted to the Disagree group.

Discussion

Computerized threshold perimetry is a clinical standard, but the test time is usually long. Fatigue and lack of concentration influence the results of the test and the compliance of long-term glaucoma patients. The new perimetric threshold strategy has been developed for these reasons. SITA with the Humphrey Field Analyzer realized the same high test accuracy as the standard full-threshold algorithm as well as a reduction in test time, and SITA has become the standard visual field test in many glaucoma clinics due to the time saved. We know that the light sensitivity of SITA is slightly higher than that of the traditional standard full-threshold strategy. We are concerned whether the limit of normality (confidence interval ) is strict in SITA.3 The raw numeric dB data obtained with the two strategies have different characteristics and meanings. We have to use judgment criteria to interpret the visual field properly.

The results according to Anderson’s criteria are satisfactory. The single field in SITA should be evaluated based on probability criteria using Statpac for SITA, but not on the threshold value, and certainly not on the Gray scale.

References

1.Caprioli J: Automated perimetry in glaucoma. Am J Ophthalmol 111:235-239,1991

2.Anderson D, Patella M: In: Automated Perimetry, pp 152-153. St Louis, MO: CV Mosby Co 1999

3.Bengtsson B, Olsson J, Heijl A, Rootzen H: A new generation of algorithms for computerized threshold perimetry, SITA. Acta Ophthalmol Scand 75:368-375, 1997