Normal Values for Fundus Perimetry Withthe Scanning Laser Ophthalmoscope

KLAUS ROHRSCHNEIDER, MD, MATTHIAS BECKER, MD, NICOLE SCHUMACHER, MD,THOMAS FENDRICH, PHD, AND HANS E. VOLCKER, MD

● PURPOSE: To evaluate the normal light sensitiv-ity values for fundus perimetry and their shorttime fluctuation (reliability) in normal volunteersof different ages. After the development of full-threshold static fundus perimetry, age-correctedsensitivity values for normal subjects are requiredto interpret results and to compare them withconventional computerized perimetry.● METHOD: Full-threshold fundus perimetry of thecentral field (33 3 21 degrees) by means of thescanning laser ophthalmoscope was performed on152 eyes of 99 healthy persons aged 16 to 77 yearswith normal vision and no eye disease. Fixationwas simultaneously documented. Light sensitivityvalues were evaluated according to each subject’sage and test point location.● RESULTS: Linear regression analysis disclosed asignificant (P < .0001) decrease of the meansensitivity of 0.275 dB per decade of increasingage, starting with 16.6 dB at age 10 years.Standard deviation around the center of fixationwas 0.287 degrees in the first decade, and itincreased by 2.82 minutes of arc per decade(P < .0001). Variation in triple examinations ofsubjects did not differ from short time fluctua-tion.● CONCLUSION: Visual fields examined with fun-dus perimetry show reliable measurements in arange comparable to conventional computerizedperimetry. There is a significant correlation be-

tween increase of age and decrease of light sensi-tivity in fundus perimetry. Visual fields obtainedwith fundus perimetry seem to correlate well withknown data from computerized static thresholdperimetry. It should be recognized that even innormal subjects, the stability of fixation decreaseswith increasing age. (Am J Ophthalmol 1998;126:52–58. © 1998 by Elsevier Science Inc. Allrights reserved.)

P ERIMETRY OF THE CENTRAL VISUAL FIELD DUR-

ing routine clinical examination is mostlyperformed by means of computerized static

threshold testing. Besides conventional perimetry(Goldmann, Humphrey, or Octopus perimeter),fundus perimetry with the scanning laser ophthal-moscope (Rodenstock, Ottobrunn, Germany) is apromising technique that allows for exact correla-tion between morphologic appearance of the fundusand its function.1–6 Automated threshold proce-dures for fundus perimetry have been published.7,8

Although sensitivity values for conventional perim-etry in normal subjects are well established,9–13 thecorresponding values for fundus perimetry have notyet been described.

The aim of this study was to evaluate the normallight sensitivity values for fundus perimetry andtheir short time fluctuation (reliability) in normalvolunteers of different ages. The influence of ageand location of the test point on the fundus shouldespecially be investigated because the method ofstimulus presentation with monochromatic laserlight, which is modulated by an acousto opticmodulator, differs completely from that of conven-tional perimetry.

Accepted for publication Nov 3, 1997.Department of Ophthalmology, University of Heidelberg, Germany.

Supported in part by grant DFG Vo 437/2-1 from the DeutscheForschungsgemeinschaft.

Reprint requests to Klaus Rohrschneider, MD, Univ.-Augenklinik,Im Neuenheimer Feld 400, D-69120 Heidelberg, Germany; fax: 149-6221-565422; e-mail: Klaus_Rohrschneider@ukl.uni-heidelberg.de

© 1998 BY ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED.52 0002-9394/98/$19.00PII S0002-9394(98)00065-8

SUBJECTS AND METHODS

STATIC THRESHOLD FUNDUS PERIMETRY WAS PER-

formed by means of the previously described methodwith the scanning laser ophthalmoscope7 in 152healthy eyes (74 right, 73 left) of 99 subjects aged 16to 77 years (mean 6 SD, 41 6 15 years; 51 men and48 women). Normal eyes were defined as eitheremmetropic or with ametropia of less than 63.0diopter (astigmatism , 61.0 diopter). All eyesincluded were examined by means of dilated oph-thalmoscopy, with an assessment of the optic diskand retina. Visual acuity was 0.8 or more in all eyesexamined. Screening for red-green defects by Ishi-hara plates was performed, and volunteers withdeficiencies were excluded. There were no individ-uals with known eye pathology or family history ofglaucoma in the study group. After informed con-sent was obtained, examination with the scanninglaser ophthalmoscope was performed with dilatedpupils. Because it is well known that visual fieldtesting in subjects without experience shows con-siderably more variation,11 all volunteers enrolled inthis study performed a minimum of one conven-tional computerized static visual field test (program38, Octopus 500; Interzeag, Schlieren, Switzerland)before the actual examination with fundus perime-try. This visual field had to be normal, as well.

The principle of threshold fundus perimetry withthe scanning laser ophthalmoscope has been de-scribed in detail.7 In summary, a modulated helium-neon laser beam (633 nm) of the scanning laserophthalmoscope is used for projection of the stimuliwhile the fundus is simultaneously observed with aninfrared diode laser source (780 nm). The area ofperimetric examination is 33 3 21 degrees. Theadaptation level of the image was set to 10 cd/m2

(100 Troland), the manufacturer’s setting accordingto subjective photometry. Stimulus intensity can bevaried in 0.1 log steps from 0 to 21 dB, where 0 dBrepresents the brightest luminance (71 cd/m2 or1.42 mW). The scaling was set following exactcalibration by use of a laser power monitor (PD2-Awith PD 300-SH; Ophir, Jerusalem, Israel), whichenabled a calibration curve between gray scalerating of the frame grabber card and the luminanceof the helium-neon laser. This dB scale is compara-ble to the scaling of the original Rodenstock soft-

ware in Germany for the same adaptation level. Inaddition, stimulus intensities of 23.6 and 26 dB canbe projected with our software.

The stimulus size was equal to the Goldmann IIIstimulus applied in Octopus perimetry (0.41-degreediameter), and presentation time was 120 msec. Weused a rectangular 3-degree grid (68 to 80 test pointlocations). Correction for changes of fixation wereperformed with the help of a landmark that was seton a reliable position before each and in the middleof each stimulus presentation. This enabled accu-rate correction for eye movements and rejection ofpresentations that were outside the correct funduslocation or were presented within a saccadic eyemovement.

We used a 4-2-1–dB double staircase procedure,which is comparable to procedures performed inconventional computerized static threshold perime-try. The last seen value was taken as thresholdinstead of the difference between seen and not seen,which is used in the Octopus perimeter software.About 5% of false-positive and false-negative testquestions were presented. In addition, only testswith less than 10% false-positive or false-negativetest questions were included. From the data, wecalculated the mean 6 SD sensitivity as well as thesensitivity for each test point location.

To test the reliability of the method in 10 eyes of10 volunteers, we additionally performed three fun-dus perimetric examinations with a minimum timeinterval of 2 days. None of these examinationsshowed more than 10% false-positive or false-nega-tive answers. We calculated the differences of thethreshold sensitivity for each test point location andcompared the results with the short time fluctua-tion, which was measured during each examination.

Stability of fixation during the examination wasmeasured as the standard deviation around themean fixation point14 and correlated with age.

Correlations were tested by linear regression anal-ysis and analysis of variation.

RESULTS

FOR STATIC THRESHOLD FUNDUS PERIMETRY, WE USED

4.8 6 0.3 stimulus presentations for each locationthat needed a mean time interval of 2.07 6 0.63

NORMAL VALUES FOR FUNDUS PERIMETRYVOL. 126, NO. 1 53

seconds between each stimulus presentation. Totalexamination time per eye varied between 6 and 22minutes (13 6 3 minutes) and was not correlated tothe subject’s age.

There was a linear decline in light sensitivitywith increasing age, as calculated by linear regres-sion analysis, of 0.275 dB per decade (P , .0001,r 5 0.337; Figure 1). There was no difference in thedecrease with increasing age toward the periphery(Table 1). Because fundus evaluation allows for anaccurate determination of the optic nerve headlocation, the blind spot was exactly delineated in alleyes. Mean sensitivity for all test point locationsoutside the blind spot was 16.9 dB in the firstdecade. Threshold sensitivity was significantly (P ,.01) lower in the peripheral locations compared tothe central part (Table 2). We can calculate thenormal hill of vision for each age, as demonstratedin a 40-year-old subject in Figure 2. Ninety-fivepercent confidence intervals were 62 dB around themean sensitivity and 64 dB around each single testpoint.

The mean short time fluctuation for all eyesincluded in this study was 2.0 6 0.8 dB.

The mean reliability for the three independentexaminations in 10 eyes was 1.5 6 0.7 dB (range,1.1 to 3.9 dB). As expected, the largest differenceswere observed at the border of the optic disk.15,16

However, there were increasing deviations in thesurrounding of angioscotomas, as well. There was nosignificant dependency of the variation from the testpoint location, that is, variation did not increase

toward periphery (Table 3). The mean short timefluctuation, which was measured at six locations ineach visual field, was 1.8 6 0.7 dB in these 10 eyes.

The standard deviation around the mean fixationpoint as a parameter inversely correlated to thestability of fixation increased significantly with sub-jects’ age, from 0.287 degrees by 2.82 minutes of arcper decade (P , .0001, r 5 0.298; Figure 3). Astandard deviation of less than 0.6 degrees aroundthe mean fixation point indicates a sufficient stablefixation. We could not detect an increase of thevariation of fixation during the course of the peri-metric examination.

DISCUSSION

AGE-CORRECTED NORMAL VALUES OF THE VISUAL

field are needed for an accurate interpretation ofperimetric examinations. Because of this necessity,normal isopters for kinetic perimetry based on theGoldmann perimeter have been reported.17 Never-theless, reliability of manual kinetic visual fieldtesting is low.

Because computerized static perimetry providesthreshold values for light sensitivity, it seems to beeasier to obtain quantitative normal values. Variousresearch groups have observed a decrease of thresh-old sensitivity from the center to the peripheralfield. In addition, static perimetry also shows adecreasing sensitivity with age. There is a lineardecrease of 0.58 to 0.88 dB per decade in 60-degree–diameter visual field testing.10,18,19 According to thetest point location, there is a more pronouncedthreshold reduction in the peripheral visual field ofmore than 10 degrees of eccentricity.12

Our findings show comparable results when staticthreshold fundus perimetry is used. One majordifference from conventional perimetry is the re-duced field size tested in the scanning laser ophthal-moscope. Whereas the former usually tests a circlewith a diameter of 60 degrees, the maximal projec-tion area in the scanning laser ophthalmoscopemeasures 33 by 21 degrees. The decrease of themean threshold with age of 0.275 dB can partiallybe explained by this smaller test area. Nevertheless,we could not observe a significant difference in theloss of sensitivity with age depending on the test

FIGURE 1. Linear regression between mean light sen-sitivity threshold values and age shows a significantdecrease with increasing age (y 5 –0.0275x 1 16.89;r 5 0.337; P < .0001).

AMERICAN JOURNAL OF OPHTHALMOLOGY54 JULY 1998

point locations (Table 1). This might be caused bythe fact that the whole measured area belongs to thevery central part of the visual field.

The completely different kind of stimulus presen-tation should influence the differences betweenconventional perimetry and fundus perimetry, aswell. The difference in color of the stimulus, whichis monochromatic red light in the scanning laserophthalmoscope (633 nm) but white light in theother perimeters, may be responsible for variationsof the sensitivity threshold.7 Furthermore, there is achromatic aberration of white stimuli that blurs thetarget, and the contrast at the retina is always lessfor white light. Because blurring increases with age,this difference might lead to an increased loss ofsensitivity with age compared to scanning laserophthalmoscope fundus perimetry thresholds. Thedifferent properties of the stimulus generation are

another factor that may influence our results. Thesummation of several light stimuli, each of which ispresented for a very short period of time as per-formed by the scanning laser ophthalmoscope (pixelduration, 15.26 msec) requires much higher lightlevels during the presentation time. Our results mayindicate that the scanning laser ophthalmoscopesetting reaches Talbot brightness (correspondingluminance level of continuous light), which must beconsidered when interpreting data obtained byscanning laser perimetry.20

The quantitative data obtained with staticthreshold perimetry allows for an easier calculationof the reliability than does data obtained fromkinetic perimetry. Nevertheless, with static perime-try, there must be a variation of sensitivity thresh-olds resulting from the staircase strategy, as well.21

In addition, the subjective method of measuring the

TABLE 1. Decrease of Sensitivity Threshold Values With Increasing Age for Each Stimulus Location*

y-Axis

(Degrees)

x-Axis (Degrees)

221 218 215 212 29 26 23 0 3 6

29 20.27 20.34 20.30 20.07 20.25 20.43 20.33 20.12 20.14 20.14

26 20.30 20.55 20.22 20.24 20.23 20.13 20.19 20.22 20.26 20.21

23 20.54 20.28 20.38 20.34 20.41 20.20 20.19 20.02

0 20.50 20.32 20.13 20.36 20.47 20.28 20.29 20.16

3 20.39 20.18 0.18 20.24 20.02 20.30 20.28 0.10 20.15 20.06

6 20.39 20.26 20.22 20.17 20.24 20.15 20.11 20.22 20.11 20.16

9 20.28 20.61 20.22 20.40 20.31 20.10 20.42 20.41 20.41 20.32

*Given in dB per decade. The x- and y-axes show the distance from the fovea at the fundus in horizontal and vertical direction in degrees

(0 degrees/0 degrees represents the fovea with 20.28 dB sensitivity loss per decade). Four locations in the area of the blind spot are

excluded.

TABLE 2. Threshold Sensitivity (dB) in Normal Subjects as Calculated for Age Zero for Each Stimulus Location*

y-Axis

(Degrees)

x-Axis (Degrees)

221 218 215 212 29 26 23 0 3 6

29 15.5 15.6 15.5 15.1 16.2 17.6 17.3 16.6 16.7 16.4

26 15.3 14.3 11.7 15.8 16.7 16.6 17.8 18.1 18.1 17.7

23 17.2 13.4 17.3 17.5 18.9 19.0 18.7 17.6

0 17.2 14.0 16.8 18.1 20.2 20.1 20.1 18.8

3 17.0 14.9 12.4 15.7 16.6 18.1 18.6 17.9 18.9 17.8

6 15.9 15.4 15.1 15.4 16.4 16.7 16.7 18.0 17.2 17.1

9 14.2 15.6 14.7 15.5 15.2 15.3 16.3 17.2 17.2 16.8

*The x- and y-axes show the distance from the fovea at the fundus in horizontal and vertical direction in degrees (0 degrees/0 degrees

represents the fovea with 20.1 dB threshold). Four locations in the area of the blind spot are excluded.

NORMAL VALUES FOR FUNDUS PERIMETRYVOL. 126, NO. 1 55

sensitivity renders a minimum of variation betweendifferent examinations.22,23 The reliability of thetest method may be estimated either by comparisonof multiple examinations of the same eye or byobservation of the number of false-positive andfalse-negative questions.21,24–26 Because we ex-cluded all eyes with more than 10% false-positive orfalse-negative answers, the accuracy of the thresh-old values should allow for a reliable database. Inaddition, the stability of fixation gives an estimate ofthe compliance during the perimetry.27 We are ableto give a very accurate value for the stability offixation since the actual point of fixation is docu-mented during each stimulus presentation. There-fore, it is possible to give age-corrected data fornormal deviation around the mean point of fixation,

as well (Figure 3). As observed for the thresholdsensitivity, the stability of fixation decreases withage slightly but significantly (P , .0001). Neverthe-less, the poor correlation means that age onlyaccounts for about 9% of the variance in fixationstability, which might be important in patientssuffering from age-related macular degeneration. Inpatients suffering from macular degeneration ormacular dystrophy, for example, this deviation issignificantly increased.14 Therefore, fundus perime-try will be a very helpful tool, especially for theexamination of patients with macular pathology,because it allows for exact correlation of morpho-logic changes with functional impairment.28 Be-cause we could not detect an increase in thevariation of fixation during the course of the peri-

FIGURE 2. Age-corrected sensitivity values for a 40-year-old subject with standard optic disk size. The hillon the right represents the macular area, whereas thedeep hole on the left is the location of the blind spotwith an absolute scotoma.

FIGURE 3. Stability of fixation was measured as thestandard variation around the mean fixation point indegrees and correlated with age. Linear regressionanalysis revealed significant increase with increasingage (y 5 0.0047x 1 0.2874; r 5 0.298; P < .0001).

TABLE 3. Variation (dB) Between Three Independent Examinations in 10 Eyes Given for Each Test Point Location*

y-Axis

(Degrees)

x-Axis (Degrees)

221 218 215 212 29 26 23 0 3 6

29 1.0 1.3 0.6 2.2 2.8 1.2 1.0 2.0 1.5 1.7

26 3.0 2.0 1.0 1.0 1.2 1.4 1.0 1.3 .75 1.3

23 2.5 1.8 1.2 1.0 0.5 2.5 1.3 1.4

0 3.5 2.4 1.4 0.8 1.4 1.5 1.2 0.7

3 1.0 2.3 2.2 2.2 0.8 1.4 1.2 1.0 2.3 0.3

6 0.5 1.3 1.2 1.4 0.6 1.8 1.2 1.5 1.8 0.3

9 2.5 1.5 1.6 1.4 1.2 2.0 3.0 2.5 2.8 1.0

*The x- and y-axes show the distance from the fovea at the fundus in horizontal and vertical direction in degrees (0 degrees/0 degrees

represents the fovea with 1.5 dB variation). Four locations in the area of the blind spot are excluded.

AMERICAN JOURNAL OF OPHTHALMOLOGY56 JULY 1998

metric examination, there seems to be no need toestablish rest phases during the examination for animprovement of the stability of fixation.

Short time fluctuation, which is the standarddeviation of identical test point locations during thesame examination, is another parameter for estimat-ing the reliability of the perimetry. This variation isknown to increase in patients with pathologic visualfields.15 In our normal population, we found resultscomparable to those of conventional perimetry(2.0 6 0.8 dB). The variation between three inde-pendent examinations (long time fluctuation) wasin the same range, as well (1.5 6 0.7 dB).25,29

In addition to the age-corrected normal valuespresented in this study, the comparison with con-ventional perimetry needs to be performed in alarger cohort of subjects. In an earlier study inves-tigating young subjects, we found a good correlationbetween fundus perimetry and conventional perim-etry that allowed us to compare visual fields bysimple linear recalculation of the threshold values.7

Further studies are needed to show whether thiscorrelation is valid for patients with visual fielddefects, as well. A comparison between both meth-ods with an increasing number of normal subjectsmay confirm the data given in the present study.

In conclusion, our results indicate that staticthreshold fundus perimetry provides reliable visualfields that enable a more precise follow-up of pa-tients with macular pathology. The actual locus offixation can be observed during the perimetricexamination, and the influence of unstable fixationcan be determined. The knowledge of age-correctedsensitivity thresholds enables an easier detection ofdiscrete defects. Such values should be knownbecause a comparison made with locations outsidethe macula or with areas that might not be influ-enced by the disease is incorrect.

REFERENCES

1. Sturmer J, Schrodel C, Rappl W. Scanning laser ophthal-moscope for static fundus controlled perimetry. In: Nas-emann JE, Burk ROW, editors. Scanning laser ophthalmos-copy and tomography. Munich: Quintessenz, 1990:133–146.

2. Acosta F, Lashkari K, Reynaud X, Jalkh AE, Van de VeldeF, Chedid N. Characterization of functional changes in

macular holes and cysts. Ophthalmology 1991;98:1820–1823.

3. Van de Velde FJ, Jalkh AE, Elsner AE. Microperimetrywith the scanning laser ophthalmoscope. In: Mills RP,Heijl A, editors. Perimetry update, 1990/91. Amsterdam:Kugler & Ghedini, 1991:93–101.

4. Sjaarda RN, Frank DA, Glaser BM, Thompson JT, MurphyRP. Assessment of vision in idiopathic macular holes withmacular microperimetry using the scanning laser ophthal-moscope. Ophthalmology 1993;100:1513–1518.

5. Fletcher DC, Schuchard RA, Livingstone CL, Crane WG,Hu SY. Scanning laser ophthalmoscope macular perimetryand applications for low vision rehabilitation clinicians.Ophthalmology Clinics of North America 1994;7:257–265.

6. Rohrschneider K, Becker M, Fendrich T, Volcker HE.Kinetische funduskontrollierte Perimetrie mit dem Scan-ning-Laser-Ophthalmoskop. Klin Monatsbl Augenheilkd1995;207:102–110.

7. Rohrschneider K, Fendrich T, Becker M, Krastel H, KruseFE, Volcker HE. Static fundus perimetry using the scanninglaser ophthalmoscope with an automated threshold strat-egy. Graefes Arch Clin Exp Ophthalmol 1995;233:743–749.

8. Sunness JS, Schuchard RA, Shan N, Rubin GS, DagnelieG, Haselwood DM. Landmark-driven fundus perimetryusing the scanning laser ophthalmoscope. Invest Ophthal-mol Vis Sci 1995;36:1863–1874.

9. Drance SM, Berry V, Hughes A. Studies on the effects ofage on the central and peripheral isopters of the visual fieldin normal subjects. Am J Ophthalmol 1967;63:1667–1672.

10. Haas A, Flammer J, Schneider U. Influence of age on thevisual fields of normal subjects. Am J Ophthalmol 1986;101:199–203.

11. Heijl A, Lindgren G, Olsson J. Normal variability of staticperimetric threshold values across the central visual field.Arch Ophthalmol 1987;105:1544–1549.

12. Johnson CA, Adams AJ, Lewis RA. Evidence for a neuralbasis of age-related visual field loss in normal observers.Invest Ophthalmol Vis Sci 1989;30:2056–2064.

13. Haas AL. Normalwerte fur den Octopus und ihre Bedeu-tung bei der Befundinterpretation. Klin Monatsbl Augen-heilkd 1989;194:320–321.

14. Rohrschneider K, Becker M, Kruse FE, Fendrich T, VolckerHE. Stability of fixation-results of fundus-controlled exam-ination using the scanning laser ophthalmoscope. Ger JOphthalmol 1995;4:197–202.

15. Haefliger IO, Flammer J. Increase of the short-term fluctu-ation of the differential light threshold around a physio-logic scotoma. Am J Ophthalmol 1989;107:417–420.

16. Haefliger IO, Flammer J. Fluctuation of the differentiallight threshold at the border of absolute scotomas: compar-ison between glaucomatous visual field defects and blindspots. Ophthalmology 1991;98:1529–1532.

17. Egge K. The visual field in normal subjects. Acta Ophthal-mol 1984;169(suppl):1–64.

18. Brenton RS, Phelps CD. The normal visual field on theHumphrey field analyzer. Ophthalmologica 1986;193:56–74.

19. Jaffe GJ, Alvarado JA, Juster RP. Age-related changes ofthe normal visual field. Arch Ophthalmol 1986;104:1021–1025.

20. Hart WM Jr. The temporal responsiveness of vision. In:Hart WM Jr, editor. Adler’s physiology of the eye: clinicalapplication. Baltimore: Mosby 1992:548–578.

21. Flammer J, Niesel P. Die Reproduzierbarkeit perimetrischer

NORMAL VALUES FOR FUNDUS PERIMETRYVOL. 126, NO. 1 57

Untersuchungsergebnisse. Klin Monatsbl Augenheilkd1984;184:374–376.

22. Barlow HB. Retinal noise and absolute threshold. J Opt SocAm 1956;46:634–639.

23. Koch P, Roulier A, Fankhauser F. Perimetry: the informa-tion theoretical basis for its automation. Vision Res 1972;12:1619–30.

24. Katz J, Sommer A. Reliability indexes of automated visualfields. Arch Ophthalmol 1988;106:1252–1254.

25. Flammer J, Drance SM, Schulzer M. Covariates of thelong–term fluctuation of the differential light threshold.Arch Ophthalmol 1984;102:880–882.

26. Nelson–Quigg JM, Twelker JD, Johnson CA. Response

properties of normal observers and patients during auto-mated perimetry. Arch Ophthalmol 1989;107:1612–1615.

27. Katz J, Sommer A. Screening for glaucomatous visual fieldloss: the effect of patient reliability. Ophthalmology 1990;97:1032–1037.

28. Rohrschneider K, Gluck R, Becker M, et al. Scanning laserfundus perimetry before laser photocoagulation of well-defined choroidal neovascularisation. Br J Ophthalmol1997;81:568–573.

29. Young WO, Stewart WC, Hunt H, Crosswell H. Staticthreshold variability in the peripheral visual field in normalsubjects. Graefes Arch Clin Exp Ophthalmol1990;228:454–457.

Authors InteractivetWe encourage questions and comments regarding this article via theInternet on Authors Interactivet at http://www.ajo.com/ Questions, com-ments, and author responses are posted.

AMERICAN JOURNAL OF OPHTHALMOLOGY58 JULY 1998