Comparison of selected parameters of SITA |
129 |
|
|
COMPARISON OF SELECTED PARAMETERS OF SITA FAST AND FULL THRESHOLD STRATEGIES IN EVALUATION OF GLAUCOMA SUSPECTS
KATARZYNA NOWOMIEJSKA, PIOTR KAWA, TOMASZ ZARNOWSKI, MAGDALENA BIALEK and ZBIGNIEW ZAGORSKI
Department of Ophthalmology, First Eye Hospital, Medical University of Lublin, Lublin, Poland
Abstract
Purpose: To determine the differences between selected parameters (test duration, global indices and reliability indices) of the Swedish Interactive Threshold Algorithm (SITA) Fast and Full Threshold algorithms in glaucoma suspects. Methods: Fifty-five eyes of 30 patients (17 males and 13 females; mean age 44 ± 14.9 years; range 17-71 years) with glaucoma risk factors were included in this study. The risk factors were as follows: ocular hypertension (OHT) – intraocular pressure (IOP) around 30 mmHg (nine patients), OHT and family history (three patients), OHT and pigment dispersion syndrome (14 patients), and OHT and pseudoexfoliation syndrome (four patients). Each patient was examined at least twice (random sequence of the tests) with two strategies: SITA Fast and Full Threshold using the Humphrey Field Analyzer and Program 30-2 with appropriate optical correction. Main outcome measures were: test duration, global indices such as mean deviation (MD) and pattern standard deviation (PSD) and reliability indices: fixation losses, false positive and false negative responses. Paired Student’s t test was used for statistical analysis. Results: The mean ± SD test duration was 3.85 ± 0.56 minutes for SITA Fast strategy and 12.76 ± 3.61 minutes for Full Threshold strategy ( p < 0.001). The average MD was –1.16 ± 1.55 dB for SITA Fast strategy and –2.22 ± 2.31 dB for Full Threshold strategy (p = 0.0057). The mean PSD was 1.90 ± 1.11 dB and 2.42 ± 1.28 dB, respectively ( p = 0.027). The mean reliability indices were as follows: fixation losses 14.0 ± 14.7% for SITA Fast strategy and 10.0 ± 11.6% for Full Threshold strategy ( p = 0.12); false positive responses were 4.6 ± 4.9% and 3.1 ± 6.8%, respectively ( p = 0.18); false negative responses were 1.8 ± 3.4% and 1.0 ± 2.7%, respectively ( p = 0.18). Conclusions: There was a significant reduction of test duration for the SITA Fast strategy compared to the Full Threshold strategy. MD values were significantly lower and PSD values were significantly higher for the Full Threshold strategy in relation to those obtained with SITA Fast. Reliability indices did not significantly differ between two strategies.
Introduction
The Swedish Interactive Threshold Algorithm (SITA) is a new development in automated perimetry which uses a new sophisticated statistical approach. It estimates
Address for correspondence: Katarzyna Nowomiejska, Tadeusz Krwawicz Chair of Ophthalmology and First Eye Hospital, Medical University of Lublin, Chmielna 1, 20-079 Lublin, Poland. Email: k.sokolowska@mailcity.com
Perimetry Update 2002/2003, pp. 129–133
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
130 K. Nowomiejska et al.
measurement errors of threshold values with interruption of the staircase procedures when mathematical evaluations reach a predetermined level. Adaptation from a predefined visual field model allows further time-saving by reducing the number of presentations. The elimination of false-positive catch trials and improvement in timepacing during the examination are other sources of reduced test duration.1-3
The SITA group of strategies is available as SITA Standard, which is comparable to the standard full threshold program, and SITA Fast, which is comparable to Fastpac.4,5 In both strategies, a model of the visual field is constructed before the actual test. This model is repeatedly updated and modified during testing on the basis of the patient’s responses. Threshold values and measurement errors are estimated in the model.1 The SITA Standard algorithm uses a 4-2 dB step size and the SITA Fast algorithm a 4 dB step size.6 The difference between SITA Standard and SITA Fast is the amount of error that is allowed for the threshold estimate.7
SITA has been designed to be as accurate as, but considerably less time consuming than, Humphrey Full Threshold algorithm, which is widely accepted as a standard algorithm.7-9 There are data suggesting that SITA Standard and SITA Fast match the precision of older thresholding methods and may replace the Full Threshold strategy as the standard clinical test used in the diagnosis and management of glaucoma.
The aim of this study was to compare the accuracy of SITA Fast strategy with the Full Threshold strategy in suspected glaucoma patients.
Subjects and methods
Fifty-five eyes of 30 patients experienced in automated perimetry (17 males and 13 females with a mean age of 44 ± 14 years, ranging from 17 to 70 years) with glaucoma risk factors, were enrolled in the study. The sample consisted of: nine subjects with ocular hypertension (OHT) – intraocular pressure (IOP) of around 30 mmHg; three subjects with OHT and family history of glaucoma; 14 subjects with OHT and pigment dispersion syndrome (PDS); and four subjects with pseudoexfoliation syndrome (PEX). Each patient was examined at least twice within a time-span of four weeks.
Examinations were performed using the Humphrey Field Analyzer (HFA-II, Humphrey Systems, Dublin, CA, Model 740) with Program 30-2 Full Threshold and SITA Fast strategies, in random order. The same instrument was used for all tests. Appropriate optical correction was made. Ocular therapy was not changed during the study period. All visual field evaluations were carried out by the same perimetrist.
The main outcome measures were as follows:
•test duration,
•global indices: mean deviation (MD) and pattern standard deviation (PSD),
•reliability indices: fixation losses, false positive and false negative responses. The paired t test was used to compare the test results obtained with the two differ-
ent strategies.
Comparison of selected parameters of SITA |
131 |
Results
Test duration
The average test duration for SITA Fast strategy was 3.85 ± 0.56 minutes and was significantly shorter (p < 0.001) than the Full Threshold strategy (12.76 ± 3.61 minutes; Fig. 1).
Global indices
The average MD for SITA Fast strategy (-1.16 ± 1.55 dB) was significantly higher (p = 0.0057) than that for Full Threshold strategy (-2.20 ± 2.31 dB; Fig. 2). The mean PSD was 1.90 ± 1.11 dB and 2.42 ±1.28 dB, respectively. The differences between these parameters for both strategies were statistically significant (p = 0.027; Fig. 3).
Test duration (min)
Fig. 1. Average test duration and SD (minutes) of SITA Fast and Full Threshold strategies. The differences between these parameters for both strategies were statistically significant (p < 0.001).
Fig. 2. Average MD and SD (dB) of the SITA Fast and Full Threshold strategies. The differences between these parameters were statistically significant for both strategies ( p = 0.0057).
132 |
K. Nowomiejska et al. |
Fig. 3. Average PSD and SD (dB) of SITA Fast and Full Threshold strategies. The differences between these parameters were statistically significant for both strategies ( p = 0.027).
Reliability indices
The mean percentage of fixation losses was 14.0 ± 14.7% for SITA Fast strategy and 10.0 ± 11.6% for Full Threshold strategy. The mean percentage of false positive responses was 4.6 ± 4.9 and 3.1 ± 6.8%, respectively. The mean percentage of false negative responses was 1.8 ± 3.4 and 1.0 ± 2.7%, respectively. Statistical analyses showed that these results were not statistically significant (data not shown).
Discussion
Early automated threshold algorithms (the first and second generation) were timeconsuming and resulted in considerable patient fatigue and reduced patient compliance.13 However, these strategies became the standard for the comparison and evaluation of new perimetric strategies. The third generation of perimetric algorithms (SITA strategies) was designed to reduce test time without loss of accuracy.1
The present study was aimed to determine the differences between selected parameters of the SITA Fast and Full Threshold algorithms of HFA in glaucoma suspects. Our results showed that the SITA Fast strategy produced a 72% reduction of test duration when compared to the Full Threshold strategy. These data seem to be in agreement with those previously reported.11,15 In addition, it has been reported that SITA Fast showed similar savings in time for both normal and glaucomatous sub-
jects.1,4,5,10
The average MD for SITA Fast strategy was significantly less negative than that for Full Threshold strategy. Our results confirmed the earlier findings of Heijl et al. showing less variability around the MD regression lines with SITA than with the Full Threshold strategy. It has also been reported that the mean sensitivity and mean pattern deviation differed significantly.15 Moreover, the present study showed that the mean PSD value for SITA Fast strategy was significantly lower than for Full Threshold strategy. However, comparisons of PSD findings should be made with caution, since PSD values peak for moderate field loss and are low in normal cases and in those with advanced visual field loss.15 Additionally, we believe that our results are indicative of an important relationship between these strategies. SITA Fast showed a
Comparison of selected parameters of SITA |
133 |
tendency to underestimate visual field defects. Similar findings were reported by Nordmann et al.16
Interestingly, we did not observed any significant differences between the reliability indices of either the SITA Fast or Full Threshold strategies.
Analysis of the data obtained in this study supports results of Roggen et al.9 who suggested that the SITA Fast strategy is ideal for screening purposes and for patients who cannot perform a reliable SITA Standard or Full Threshold strategy.
Acknowledgments
Katarzyna Nowomiejska was kindly supported by IPS travel grant.
References
1.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
2.Olsson J, Rootzen H, Heijl A: Maximum likelihood estimation of the frequency of false positive and false negative answers from the up-and-down staircases of computerized threshold perimetry. In: Heijl A (ed) Perimetry Update 1988/1989, pp 245-251, p 365. Amsterdam/Berkeley/Milano: Kugler & Ghedini Publ 1989
3.Olsson J, Bengtsson B, Heijl A, Rootzen H: Improving estimation of false-positive and false-nega- tive responses in computerized perimetry. In: Mills RP, Wall M (eds) Perimetry Update 1994/1995, p 219. Amsterdam/New York: Kugler Publ 1995
4.Sekhar GC, Naduvilath TJ, Lakkai M, Jayakumar AJ, Pandi GT, Mandal AK, Honavar SG: Sensitivity of Swedish interactive threshold algorithm compared with standard full threshold algorithm in Humphrey visual field testing. Ophthalmology 107:1303-1308, 2000
5.Wild JM, Pacey IE, O’Neill EC, Cunliffe IA: The SITA perimetric threshold algorithms in glaucoma. Invest Ophthalmol Vis Sci 40:1998-2009, 1999
6.Olsson J, Rootzen H: An image model for quantal response analysis in perimetry. Scand J Stat 21:375387, 1994
7.Hejl A: The Humphrey Field Analyzer, construction and 1999 concepts. In: Heijl, Greve (eds) Documenta Ophthalmol. Proceedings Series 42:77-84. Proceedings of the Sixth International Perimetric Society Meeting 1984. Dordrecht: Dr W Junk Publ 1985
8.Haley MJ: The Field Analyzer Primer. San Leonardo, CA: Allergan Humphrey 1986
9.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 Belge Ophtalmol 279:23-33, 2001
10.Bengtsson B, Heijl A, Olsson J: Evaluation of a new threshold visual field strategy, SITA, in normal subjects: Swedish Interactive Thresholding Algorithm. Acta Ophthalmol Scand 76:165-169, 1998
11.Bengtsson B, Heijl A: Evaluation of a new perimetric threshold strategy, SITA, in patients with manifest and suspect glaucoma. Acta Ophthalmol Scand 76:268-272, 1998
12.Budenz DL, Rhee P, Feuer WJ, McSoley J, Johnson CA, Anderson DR: Sensitivity and specificity of the Swedish interactive threshold algorithm for glaucomatous visual field defects. Ophthalmology 109:1052-1058, 2002
13.Johnson C A: Recent developments in automated perimetry in glaucoma diagnosis and management. Curr Opin Ophthalmol 13:77-84, 2002
14.Vivell PM, Lachenmayr BJ, Zimmermann P: Comparative study of various perimetric strategies. Fortschr Ophthalmol 88:819-823, 1999
15.Heijl A, Bengtsson B, Patella VM: Glaucoma follow-up when converting from long to short perimetric threshold tests. Arch Ophthalmol 118:489-493, 2000
16.Nordmann JP, Brion F, Hamard P, Mouton-Chopin D: Evaluation of the Humphrey perimetry programs SITA Standard and SITA Fast in normal probands and patients with glaucoma. J Fr Ophtalmol 21:549-554, 1998