A temporal deficit in juvenile diabetics

Graefe's Arch Clin Exp Ophthalmol (1999)  Springer-Verlag 1999 237:636±641

Anne Kurtenbach Andreas Neu Eberhart Zrenner

Received: 1 October 1998 Revised version received: 12 January 1999 Accepted: 14 January 1999

)

A. Kurtenbach ( ) ´ E. Zrenner University Eye Hospital, Department of Neuro-ophthalmology and Pathophysiology of Vision, Schleichstrasse 12±16, D-72076 Tübingen, Germany e-mail [email protected], Fax +49-7071-29-5361 A. Neu University Children's Hospital, Rümelinstrasse 19, D-72076 Tübingen, Germany

CLINICAL INVESTIGATION

A temporal deficit in juvenile diabetics

Abstract l Background: In this study we examined the temporal domain of visual function in diabetics without retinopathy by examining wavelength discrimination ability at two exposure durations. The results were compared to those found by heterochromatic brightness matching and anomaloscope matches. l Methods: Wavelength discrimination was performed between 440 and 540 nm at exposure times of 1 s and 0.04 s in eight juvenile diabetic patients without retinopathy. The monochromatic stimuli were presented in Maxwellian view and were set to be equally bright prior to the experiment using heterochromatic brightness matching. In addition, Rayleigh and Moreland anomaloscope matches were performed. The

Introduction Patients suffering from diabetes mellitus are known to show early visual deficits. Diabetics without visible retinal morphological changes have been shown to exhibit abnormalities in colour vision [2, 3, 8, 9, 11, 15, 22, 29, 31], spectral sensitivity [17, 35, 36] and contrast sensitivity [4±7]. The aim of this study was to investigate the temporal factor in the processing of colour information in juvenile diabetics. Our previous studies have shown a continuous decrease in the match range of the Moreland (blue-green) equation in anomaloscope matches in juvenile type 1 diabetic patients without retinopathy [14]. The same patients show a continuous change in heterochromatic

results of the diabetic group were compared to those of an age-matched control group of eight subjects with normal colour vision. l Results: Wavelength discrimination showed no difference between the groups for an exposure time of 1 s. With an exposure duration of 0.04 s, however, the diabetics show raised thresholds for the shortest wavelengths tested. In addition, brightness matches were increased at the short wavelengths, and anomaloscope matches showed a decrease in the match range for the Moreland (blue-yellow) equation. l Conclusion: The results indicate post-receptoral alterations in diabetic patients with no visible changes in their retinae.

brightness matching (HBM) but not in wavelength discrimination or in the Farnsworth-Munsell 100-Hue Test [12, 13]. However, wavelength discrimination judgements are known to be affected by the exposure time of the stimulus, becoming considerably worse with exposure times less than about 1 s [21, 23, 24, 26, 27, 30]. With a reduction in stimulus duration, a greater deficit in diabetics without retinopathy has been reported for the task of purity discrimination compared with control subjects or diabetics with retinopathy [25]: An increase in threshold to blue and yellow stimuli presented on a computer monitor was found, but not to red and green stimuli. The question addressed in this study was whether the juvenile diabetics previously examined would show a deficit in the wavelength discrimination task if the stimulus presentation time was reduced.

637

We report here the results from eight of the juvenile patients previously examined, retested 7 years after the initial measurements. We concentrated on the short- and middle-wavelength spectral region where the deficits in diabetic colour vision are most likely to occur. For comparison, we also show the results of brightness matching and anomaloscope matches performed in the yearly examination of these patients. The results for the diabetic group are compared to those of an age-matched control group.

Materials and methods Subjects Eight patients (five female, three male) aged between 14.8 and 24.5 years (mean 20.03.3, SD) formed the juvenile diabetic (type 1) group. They had suffered diabetes for between 6.4 and 18.7 years (mean 12.93.9). All patients were examined 3±4 times a year for metabolic changes and underwent complete ophthalmological examination on the day of testing. Seven showed no visible retinal microvascular changes, while one showed a few microaneurysms. The detection of microvascular changes was undertaken with direct and indirect microscopy, as well as fundus photography. Fluorescein angiography, although more sensitive, was not undertaken as such an invasive procedure could be not ethically justified. The results of the juvenile group were compared to those of an age-matched control group of eight healthy subjects (four female, four male) with a mean age of 21.02.7 years and with normal colour vision as tested by the Lanthony D-15 test. Stimulus The stimulus was a 2-deg circular bipartite field presented to the subject by means of a three-channel maxwellian view system. Two channels provided monochromatic light (Jobin Yvon monochromator; 4 nm bandwidth) for the two halves of the bipartite field and the third channel provided an achromatic interstimulus field of 473 trolands (td), presented for 20 s between test wavelengths. The retinal illuminance of the achromatic field was calculated using side-by-side brightness matching with a 540-nm test hemifield. The equivalent brightness of the 540-nm field was 8.78 log Q/s/ deg2, and the troland value was calculated using the formula [34]: 2

±13

trolands (l) = (Q/s/deg ) / l * Vl * 4.454 * 10 Procedure

Wavelength discrimination was performed for wavelengths between 440 and 540 nm in 20-nm steps. The subject's head was held steady with a chin and forehead rest and he/she had to indicate by pressing a button whether a difference in hue between the two half-fields could be perceived. An ascending method of limits (1-nm steps) was used to find the just-noticeable difference (delta lambda) in wavelength. Five measurements were taken for each main wavelength: the first was discarded and the mean calculated from the remaining four values. For the first run of the experiment stimuli were presented for 1 s, for the second run, stimulus presentation time was reduced to 0.04 s. The duration of the test field was controlled by solenoid (RS-MSM) shutters with a rise time of 3 ms. The first measurement of the main wavelength in each session was discarded. Calibration, experimental procedure and data processing were computer controlled. Equal brightness for the chromatic stimuli used in the wavelength discrimination task was obtained by performing heterochro-

Fig. 1 Wavelength discrimination in control subjects (open squares) and juvenile diabetics without retinopathy (closed circles). The mean (1SEM) just-noticeable difference in wavelength (Delta Lambda) is plotted against wavelength. The upper panel show the results obtained with a presentation time of 1 s and the lower panel those with a presentation time of 0.04 s matic brightness matching prior to the discrimination measurements. HBM functions were measured between 420 and 560 nm in 10-nm steps. The subject was instructed to match the brightness of the test wavelength presented in the left hemifield to that of an achromatic field of 473 td presented in the right hemifield. Five measurements were taken using the method of ascending and descending limits. The first was discarded and the remaining four averaged. Metameric matches were performed using a computer-driven anomaloscope (Interzeag, Colour Vision Meter) allowing the examination of both red-green and blue-green axes by determination of Rayleigh (545 nm+670 nm=589 nm) and Moreland (436 nm+490 nm=desaturated 480 nm) matches.

Results In Fig. 1 we show the mean results for wavelength discrimination between 440 and 540 nm for the control and diabetic groups. The upper panel shows the results for a presentation time of 1 s and the lower panel, those

638

for a presentation time of 0.04 s. The curves in the upper panel are similar to those classically reported in the literature for this spectral region [33], showing a better discrimination ability around 500 nm than at 460 and 540 nm. There are no significant differences between the two groups of subjects. In the lower panel of Fig. 1 we show the results of wavelength discrimination for a presentation time of 0.04 s. At the shorter wavelengths, the diabetic group shows a larger delta lambda than the control group; the difference is significant at 440 nm (P=0.0445, t-test).

In Fig. 2 we show the results of brightness matching to an achromatic field, performed prior to the wavelength discrimination to obtain equibright chromatic stimuli. The inverse of the matched luminance, the relative sensitivity, is plotted as a function of wavelength. The data points indicate the mean  1SEM. It will be seen that between 420 and 490 nm the relative sensitivity of the diabetic group is greater than that of the control group. At the data points between 420 and 490 nm the differences are significant (P=0.0002, repeated-measures ANOVA). The results from metameric matching using the anomaloscope can be seen in Fig. 3. The left panel shows the mean match midpoints (+1SD) for the Rayleigh (redgreen) match and the Moreland (blue-green) match. On the right, the results for the match ranges for both equations are shown. Because the range of matches is not normally distributed, we show here the geometric mean (+1SD). Neither group shows changes in the match midpoints. Only the Moreland match range is significantly different from the control group, i.e. the values exceed the mean +2SD of the control group.

Discussion

Fig. 2 Mean results (1SEM) of heterochromatic brightness matching to an achromatic (473 td) hemifield. Relative sensitivity (inverse of matched luminance) plotted as a function of wavelength. Symbols as for Fig. 1 Fig. 3 Results of metameric matching for the Rayleigh (redgreen) and Moreland (bluegreen) equations. On the left the mean results for the match midpoint are shown, on the right those for the match range. Grey bars indicate the results found for the control group and hatched bars the results for the diabetic group

The results for wavelength discrimination in juvenile diabetics without retinopathy show significant differences from the control group at the short-wavelength end of the spectrum for a stimulus presentation time of 0.04 s, but not for a presentation time of 1 s (Fig. 1). Our findings therefore confirm that there is a temporal deficit in diabetics without retinopathy, as found by Scase et al. [25] using the task of purity discrimination. Visual colour processing is believed to be completed in the retina [28]. Wavelength discrimination is thought to be governed by post-receptoral interactions between long (L-) and middle (M-) wavelength-sensitive cones throughout the majority of the spectrum [18, 20, 21].

639

Fig. 4 Data from Fig. 1 replotted to compare the functions obtained at both stimulus exposure durations for the control group (upper panel) and the diabetic group (lower panel). Continuous lines show schematically the deuteronopic function for an exposure duration of 1 s, discontinuous lines that for a stimulus duration of 0.04 s

Our ability to differentiate between hues depends on the ratio of quantum catches in the different classes of receptor, and there is normally a linear relationship between the intensity and the duration of a stimulus and its discriminability (Bloch's law) until saturation is reached. The opponent channel between the short (S-) wavelength-sensitive cones and the M- or L- cones, however, which is responsible for discrimination between about 460 and 520 nm, is already saturated at the luminance level of 473 td used in this study [19]. Reducing the exposure time of a stimulus, i.e. reducing stimulus energy, brings this ªdeuteranopicº opponent channel to its optimal working range, causing its displacement towards short wavelengths. This can be seen in the control subjects. In Fig. 4, we have replotted the data from Fig. 1. The upper panel shows the data at both exposure durations for the control group and the lower panel, the data from the diabetic group. The continuous lines in both panels show

schematically the deuteranopic function responsible for discrimination in this spectral range, based on that found by Walraven and Boumann [32] in deuteranopes. For the control group, the data obtained at the lower stimulus energy (shorter stimulus duration) are fit by a lateral movement of this function towards shorter wavelengths. In addition it is raised by about 1 nm and is slightly narrower. For the diabetic group, however the function which fits the data at 0.04 s is raised by about 2±3 nm and is considerably narrower than that required for the data found with a 1 s presentation time. The diabetics tested therefore appear more sensitive than the controls to the reduction in stimulus energy at the opponent site, although a deficit of the S-cones cannot be ruled out. The results of both HBM and anomaloscope settings (Figs. 2, 3) are in line with our previous reports including these subjects [12, 14], where they have been discussed in detail. Here we show the results at the shortwavelength end of the spectrum, where the HBM task differs significantly between the two groups. The short-wavelength increase in sensitivity is perhaps surprising in view of the decrease in sensitivity often found in diabetics in this spectral region using spectral threshold measurements. The reason for this is that in comparing HBM curves between groups, unlike spectral sensitivities determined by threshold measurements, there is no information about absolute differences in sensitivity between individual subjects, i.e. how dark or bright the comparison field appears. All curves come together at the comparison wavelength. Here matching was made against an achromatic field, and the curves appear to come together in the central spectral region. Adams [1], using spectral threshold measurements, has shown that diabetics can have an absolute sensitivity loss of about 0.2 log units in the middle of the spectrum, increasing to about 0.85 log units at 420 nm. Thus, it is most probable that the diabetics we studied also show a reduction in mechanisms responsible for sensitivity in the centre of the spectrum, which we cannot detect with this method. The shape of the HBM function is thought to be determined by post-receptoral mechanisms, with the central spectral region most probably being dominated by an achromatic channel formed by the additive interaction of the L- and M-receptors [10]. These results can therefore be explained by a reduction in sensitivity at the post-receptoral stage of processing in the diabetic eye. The anomaloscope matches show, as a single deficit, an increase in the Moreland match range: these patients suffer from a colour discrimination deficit involving the blue-to-green spectral region, which points to an alteration of the opponent interactions between the S-receptors and the M- and L-receptors. Diabetics are known to exhibit lens changes earlier than normals [16], but the fact that the Moreland match midpoints remain the same for control and diabetic groups indicates that lens opacity al-

640

terations are not a major cause of the perceptual changes in our subjects at this stage of the disease. The similarity of the Moreland match midpoints also indicates that changes in the shape of the S-cone sensitivity function are unlikely to be the cause of the deficit, although a reduction in the sensitivity or number of S-cones could also lead to this result. In conclusion, the group of diabetic patients tested in this study shows a diminished discrimination ability at the short-wavelength end of the spectrum, compared to control subjects, when the exposure time of the stimulus is reduced. The results can be explained by a reduc-

tion in stimulus energy at the opponent site between Scones and L-or M-cones in these patients, although this finding alone does not rule out a receptor deficit. An increased short-wavelength sensitivity in brightness matching and an increase in the match range of the Moreland (blue-green) anomaloscope matches are, however, compatible with post-receptoral changes in the diabetic eye. Acknowledgements We thank Dr. H. Knau for his help in gathering some of the data. This study was supported by the WilhelmSander-Stiftung, Munich

References 1. Adams AJ (1982) Chromatic and luminosity processing in retinal disease. Am J Opt Physiol Opt 59:954±960 2. Banford D, North RV, Dolben J, Butler G Owens DR (1994) Longitudinal study of visual functions in young insulin dependent diabetics. Ophthalmic Physiol Opt 14:339±346 3. Dean FM, Arden GB, Dornhurst A (1997) Partial reversal of protan and tritan colour defects with inhaled oxygen in insulin dependent diabetic subjects. Br J Ophthalmol 81:27±30 4. Della Sala S, Bertoni G, Somazzi L, Stubbe F, Wilkins AJ (1985) Impaired contrast sensitivity in diabetic patients with and without retinopathy: a new technique for rapid assessment. Br J Ophthalmol 69:136±142 5. Di Leo MA, Caputo S, Falsini B. Porciatti V, Minnella A, Greco AV Ghirlanda G (1992) Nonselective loss of contrast sensitivity in visual system testing in early type 1 diabetes. Diabetes Care 15:620±625 6. Dosso AA, Bonvin ER, Morel Y, Golay A, Assal JP, Leuenberger PM (1996) Risk factors associated with contrast sensitivity loss in diabetic patients. Graefe's Arch Clin Exp Ophthalmol 234:300±305 7. Ghafour IM, Foulds WS, Allan D, McClure E (1982) Contrast sensitivity in diabetic subjects with and without retinopathy. Br J Ophthalmol 66:492±495 8. Greene FD, Ghafour IM, Allan D (1985) Colour vision of diabetics. Br J Ophthalmol 69:533±536 9. Hardy KJ, Fisher C, Heath P, Foster DH, Scarpello JH (1995) Comparison of colour discrimination and electroretinography in evaluation of visual pathway dysfunction in aretinopathic IDDM patients. Br J Ophthalmol 79:35±37

10. Kalloniatis M, Pianta MJ (1997) L and M cone input into spectral senstivity functions: a reanalysis.Vision Res 37:799±811 11. Kinnear PR, Aspinall PA, Lakowski R (1972) The diabetic eye and colour vision. Trans Ophthalmol Soc UK 92:69± 78 12. Kurtenbach A, Schiefer U, Neu A, Zrenner E (1999) Development of brightness matching and colour vision deficits in juvenile diabetics. Vision Res 39:1221±1229 13. Kurtenbach A, Wagner U, Neu A, Schiefer U, Ranke MB, Zrenner E (1994) Brightness matching and colour discrimination in young diabetics without retinopathy. Vision Res 34:115±122 14. Kurtenbach A, Schiefer U, Neu A, Zrenner E (1999) Pre-retinopic changes in the colour vision of juvenile diabetics. Br J Ophthalmol 83:43±46 15. Lakowski R, Aspinall PA, Kinnear PR (1972/1973) Association between colour vision losses and diabetes mellitus. Ophthalmic Res 4:145±159 16. Lutze M, Bresnick GH (1991) Lenses of diabetic patients ªyellowº at an accelerated rate similar to older normals. Invest Ophthalmol Vis Sci 32:194±199 17. MarrØ E, MarrØ M (1972) The influence of the three color vision mechanisms on the spectral sensitivity of the fovea. Mod Prob Ophthalmol 11:219±223 18. Mollon J, EstØvez O (1988) Tyndall's paradox of hue discrimination. J Opt Soc Am A 5:151±159 19. Mollon JD, Polden PG (1977) Saturation of a retinal cone mechanism. Nature 265:243±246 20. Mollon J, EstØvez O, Cavonius CR (1991) The two subsystems of colour vision and their roles in wavelength discrimination. In: Blakemore C (ed) Vision: coding and efficiency. Cambridge University Press, Cambridge, UK, pp 119±131

21. Mollon J, Astell S, Cavonius CR (1992) A reduction in stimulus duration can improve wavelength discriminations mediated by short-wave cones. Vision Res 32:745±755 22. Peduzzi M, Longanesi L, Ascari A, Cascione S, Galletti M, Roncaia R, Pacchioni C, Maione M (1989) Screening of early colour vision loss in diabetic patients. J Fr Ophtalmol 12:839±840 23. Regan D, Tyler CW (1971) Temporal summation and its limit for wavelength changes: an analog of Bloch's law for color vision. J Opt Soc Am 61:1414± 1421 24. Reitner A, Sharpe LT, Zrenner E (1991) Is colour vision possible with only rods and blue sensitive cones? Nature 352:798±800 25. Scase MO, Foster DH, Honan WP, Heron JR, Gulliford MC, Scarpello JHB (1990) Abnormalities in hue discrimination revealed with very brief stimuli in diabetes mellitus and in optic neuritis. Clin Vision Sci 6:49±57 26. Siegel MH (1965) Color discrimination as a function of exposure time. J Opt Soc Am 55:566±569 27. Smith VC, Bowen RW, Pokorny J (1984) Threshold temporal integration of chromatic stimuli. Vision Res 24:653±660 28. Sperling HG (1997) Visual colour processing is completed in the retina. In: Cavonius CR (ed) Colour vision deficiencies XII. Doc Ophthalmol Proc Ser 59. Kluwer, Dordrecht, pp 135±139 29. Trick GL, Burde RM, Gordon MO, Santiago JV, Kilo C (1988) The relationship between hue discrimination and contrast sensitivity in patients with diabetes mellitus. Ophthalmol 95:693±698

641

30. Uchikawa K, Ikeda M (1987) Color discrimination and appearance of shortduration equal luminance monochromatic lights. J Opt Soc Am 4:1097±1103 31. Verrotti A, Lobefalo L, Chiarelli F, Mastropasqua L, Ciancaglini M, Morgese G (1995) Colour vision and persisent microalbuminuria in children with type 1 (insulin -dependent) diabetes mellitus: a longitudinal study. Diabetes Res Clin Pract 30:125±130

32. Walraven PL, Bouman MA (1966) Fluctuation theory of colour discrimination of normal trichromats. Vision Res 6:567±586 33. Wright JA, Pitt FH (1935) Hue discrimination in normal color vision. Proc Phys Soc 46:459±473 34. Wyszecki G, Stiles WS (1982) Colour science. Concepts and methods: qualitative data and formulae. Wiley, New York

35. Zisman F, Adams A (1982) Spectral sensitivity of cone mechanisms in juvenile diabetics. Doc Ophthalmol 33:127± 131 36. Zwas F, Weiss H, McKinnon P (1980) Spectral sensitivity measurements in early diabetic retinopathy. Ophthalmic Res 12:87±96