|Year : 2013 | Volume
| Issue : 2 | Page : 43-48
Color vision deficit in diabetes mellitus in presence of no or minimal diabetic retinopathy
Galal Mohamed Ismail
Faculty of Optometry and Visual Science, Al Neelain University, Khartoum, Sudan
|Date of Web Publication||10-Jan-2014|
Galal Mohamed Ismail
Faculty of Optometry and Visual Science, Al Neelain University, Khartoum,
Introduction: Color vision examination is of great clinical interest in diabetes mellitus, since changes in color vision can be accepted as an indication of pathological condition. Color vision as a psychophysical measurement was performed with a view to providing an assessment of early neural functional integrity in the presence of no or minimal diabetic retinopathy. Materials and Methods: The investigation was carried out on a healthy visual system control group and a number of age and sex-matched non-insulin dependent diabetic groups with different levels of diabetic retinopathy. The psychophysical test used was the Farnsworth Munsell 100-Hue test. Results: The results were considered in terms of the presence of functional changes relative to the severity of retinopathy and the duration of diabetes. The color vision test differentiated between normal and those with diabetes but without retinopathy. Clearly the color vision score error mean is increased in the diabetic groups compared with the normal subjects and the level of the score error mean increases with the severity of retinopathy, but not to duration. Conclusion : C0 hanges in color vision are not strongly related to the duration of diabetes, but more to the severity of the retinopathy.
Keywords: Color vision, diabetes mellitus, diabetic retinopathy
|How to cite this article:|
Ismail GM. Color vision deficit in diabetes mellitus in presence of no or minimal diabetic retinopathy. Sudanese J Ophthalmol 2013;5:43-8
|How to cite this URL:|
Ismail GM. Color vision deficit in diabetes mellitus in presence of no or minimal diabetic retinopathy. Sudanese J Ophthalmol [serial online] 2013 [cited 2019 May 21];5:43-8. Available from: http://www.sjopthal.net/text.asp?2013/5/2/43/124819
| Introduction|| |
Color vision examination is of great clinical interest, since changes in color vision can be accepted as an indication of pathological conditions. Tritan-like acquired color vision defects have been reported to occur in diabetes. ,,,,,, Therefore, color vision assessment may be a useful method of screening and monitoring diabetic retinopathy.
A related issue for diabetics is the subject of reliable self-monitoring of blood and urine glucose levels, for which good color vision is essential. Color vision defects could result in incorrect reading of urine and blood test strips and therefore mismanagement of diabetes mellitus. 
| Materials and Methods|| |
The non-insulin dependent diabetes patients were recruited from the University of Bradford Diabetic Retinopathy Screening Program. The patients were not in a tight diabetic control; therefore they can be accepted as reflecting a real image of the diabetic population. The non-diabetic control group was recruited from patients, partners and members of the University in departments other than optometry. Diabetics and non-diabetics tested within the same period of the study.
Subjects were excluded from the study if they had any sign of cataract within the undilated pupillary area using direct ophthalmoscopy, if they reported any congenital color deficiency on questioning or if they had any major systemic pathology other than diabetes.
The age and duration of diabetes for all subject groups were normally distributed. [Table 1] shows the values for the mean, standard deviation and ranges of age for the subject groups who participated in the study. [Table 2] shows the values for the mean, standard deviation and ranges of diabetic duration for the subject groups who participated in the study. None of the age means were significantly different from each other (F3,100 = 2.406, P > 0.05).
|Table 1: Means table of the age of the subject groups, showing the standard deviation, range and sex distribution |
Click here to view
|Table 2: Means table of the duration of diabetes for each diabetic subject group, showing the standard deviation and range |
Click here to view
Diabetic Retinopathy Grading Systems
Quantitative systems of grading retinal changes are more acceptable compared to qualitative systems in which statistical analysis is difficult to apply. The Airlie House Classification, established in 1968, graded each of 14 lesions on a three step scale : a0 bsence, mild-moderate and severe. A modification for this system appeared in 1981  and contained five steps. Further minor adjustment to the classification was applied by Klein et al. which classified in stages 1-6 and then by Davis et al.  containing stages 10-70 (i.e., seven fields). This latter classification was chosen in order to grade retinopathy in the present study.
The precise grading classification used in the present study was as follows:
- Non-diabetics: Patients without a history of diabetes or any ocular abnormality.
- Diabetic retinopathy level 10 (DRL10): Patients with diabetes but without any clinical evidence of retinopathy.
- DRL20: Patients with diabetes and demonstrating a retinopathy consisting of microaneurysms (one or more) only.
- DRL30: Patients with diabetes and demonstrating a retinopathy consisting of microaneurysms and one or more of the following: retinal hemorrhage (but total of hemorrhages and microaneurysms less than those of [Figure 1] of Davis's classification, hard exudates but less than those in [Figure 2] of the classification, soft exudates questionably present, intraretinal microvascular abnormalities questionably present, venous beading questionably present, venous loops definitely present.
|Figure 1: Relationship between color vision root total error score and duration of diabetes|
Click here to view
|Figure 2: Plot of root partial error score against subject category for both blue-yellow and red-green color axes. Standard errors are shown|
Click here to view
Patients with grades above DRL30 were not included in the present study.
Color vision was assessed using the Farnsworth Munsell 100-Hue test. The test was used under standard illuminant "C" lighting conditions.  Subjects were requested to arrange the four boxes in sequence wearing their optimal near refractive error correction. Total error score was calculated from the design sheet for the test for all boxes combined and by taking the square root of this value to convert the data to a normal distribution, as suggested by Lakowski et al.  Following this, error scores were calculated individually for the blue-yellow color axis (by combining caps 1-12, 34-54 and 76-84) and for the red-green color axis (by combining caps 13-33 and 55-75), again by taking the square root of the data. This method of analysis is consistent with that suggested by Kinnear  and Smith et al. 
| Results|| |
Total Error Score
[Table 3] shows the mean, standard deviation and standard error for the root total error score for all four groups.
|Table 3: Means table of 100-Hue root error scores for each subject group |
Click here to view
[Figure 3] shows the error score means of the Farnsworth Munsell 100-Hue test plotted for each of the four subject groups. Standard error bars are shown. Clearly the score error mean is increased in the diabetic groups compared to the normal subjects and the level of the score error mean increases with the severity of retinopathy. Note the substantial increase in variability amongst the most severe group of diabetic subjects, indicated by the increased size of the standard error bars.
|Figure 3: Plot of root total error score against subject category. Standard errors are shown|
Click here to view
A single factor (root total error score) analysis of variance confirms the color defects with diabetic retinopathy (F3,100 = 121.1, P < 0.0001) and is shown in [Table 4].
|Table 4: ANOVA table for the effect of subject category upon color vision total error score |
Click here to view
A Scheffe-S post-hoc multiple comparison test was used to reveal significant differences in score error means between subject groups [Table 5]. Color vision error scores in all the diabetic groups differed significantly from the normal group (P < 0.05, P < 0.01 and P < 0.0001 for DRL 10, 20 and 30 respectively). Error scores for the DRL30 group were significantly different from all the other groups.
|Table 5: Post-hoc comparison of color vision total error scores for each subject group |
Click here to view
The correlation between the duration of the diabetic retinopathy and the color test error score was analyzed. Correlation did not reach a significant value within any of the three diabetic groups. Neither did the correlation reach significance when all diabetics were considered together r = 0.14, P > 0.1 [Figure 3]. This finding indicates that color vision changes are not strongly related to the duration of diabetes, but more to the severity of the retinopathy.
Blue-Yellow/Red-Green Error Scores
[Table 6] shows the mean, standard deviation and standard error for the root error score divided into the two categories of blue-yellow/red-green color axis for all four groups.
|Table 6: Means table for blue-yellow and red-green root partial error scores for the 100-Hue test |
Click here to view
[Figure 1] shows the root error score means of the Farnsworth Munsell 100-Hue test plotted for each of the four subject groups for both blue-yellow and red-green color axes. Standard error bars are shown. Clearly the score error mean is increased in the diabetic groups compared to the normal subjects and that the level of the score error means increases with the severity of retinopathy. Note the substantial increase in variability amongst the most severe group of diabetic subjects, indicated by the increased size of the standard error bars.
Analysis of variance with one between-subjects factor (subject category) and one within-subjects factor (blue-yellow or red-green color axis) confirms the color defects with diabetic retinopathy (F3,100 = 16.1, P < 0.0001) and is shown in [Table 7]. i0 n addition, the effect of color axis is highly significant (F1,100 = 45.54, P < 0.0001) indicating that the blue-yellow error scores are higher than those for the red-green axis. Interestingly, there is a significant interaction effect (F3,100 = 3.914, P < 0.025) indicating that the selective defect in blue-yellow performance becomes more pronounced with higher levels of diabetic retinopathy.
|Table 7: ANOVA table for the effect of color axis and subject category upon partial error score for the 100-Hue test |
Click here to view
A Scheffe-S post-hoc multiple comparison test was used to reveal significant differences in error scores between subject groups [Table 8]. Color vision error scores in the DRL20 and DRL30 groups differed significantly from the normal group (P < 0.05 and P < 0.0001 respectively), but the DRL 10 group difference just failed to reach significance (P > 0.05). Error scores for the DRL30 group were significantly different from all the other groups.
|Table 8: Post-hoc comparison of partial error scores for each subject group |
Click here to view
The correlation between the duration of the diabetic retinopathy and the color test error scores for both blue-yellow and red-green axes were analyzed. The correlation failed to reach significance for either the blue-yellow (r = 0.01, P > 0.1) or the red-green color axes (r = 0.10, P > 0.1). Again, this finding indicates that blue-yellow and red-green color vision changes are not strongly related to the duration of diabetes, but more to the severity of the retinopathy [Figure 4].
|Figure 4: Relationship between root partial error scores and duration of diabetes for both blue-yellow and red-green color axes|
Click here to view
| Discussion|| |
The impact of color in the world of work and other aspects has increased steadily in recent years. Measurement and clinical examination of those with abnormal color perception are particularly valuable in medical and general health assessment. Since changes in color vision can be an indication of pathological condition, it is important for clinicians to detect, link and deal with the cause to help in restoring the normal color perception.
Many of the studies have reported color vision defects in diabetes. These findings are supported by the results of the present study showing that the error score mean in the Farnsworth-Munsell 100-Hue test increases with the increase of the DRL [Figure 1].
Verriest noted that the alteration of color vision in diabetic patients was similar to the alteration in normal people of a much greater age and in younger normal people under greatly reduced condition of lighting. Both showed that 100-Hue error scores were significantly higher in diabetics than normal with no correlation between error score and age, sex, age of onset, duration of diabetes or its metabolic control. In a similar study, Moloney and Drury investigated 132 diabetic eyes and found a mean for the Farnsworth Munsell 100-Hue test score of 138.7 for the diabetics. This compares to a mean score of 60 for non-diabetics. They found that 66% of diabetics had scores greater than 100 and the score showed no correlation with age, duration of diabetes or surprisingly, retinopathy presence. 
Another study using the 100-Hue test by Monique who found that insulin dependent patients with minimal retinopathy had more severe color vision losses than controls of similar age and had higher mean scores. 
The general findings of the present study therefore confirm the findings of other authors. However, the results do not agree with reports which conclude that there is no correlation between color defects and the DRL. The present study found the error score mean for all diabetic groups to be significantly different from the normal group. In addition, the error seems to increase steadily with the increase of the retinopathy progression [Table 5] so that the DRL30 group produces scores significantly different from either the DRL10 or DRL20 groups. The present findings do, however, support the views of others that the 100-Hue score is not well correlated with the duration of diabetes [Figure 4].
It is of interest to consider color vision defects in terms of the color axes along which they occur. [Figure 4] shows a selective increase in error score along the blue-yellow color axis compared to the red-green. Further, the selective blue-yellow loss becomes more accentuated with increasing retinopathy level [Figure 4] and [Table 6]. The partial error scores, separated according to color axis, just fail to differentiate the DRL10 group from the non-diabetics [Table 6], indicating that color defects, especially along the blue-yellow axis, are primarily a feature of the presence of retinopathy rather than diabetes mellitus alone. Neither color error score was found to correlate with diabetic duration [Figure 4].
Other studies which have considered color vision errors along separate axes include Bronte-Stewart who looked at Farnsworth Munsell 100-Hue test results from 54 juvenile diabetics over a 4 year period from diabetes onset. They did not find any deterioration in color vision, although when the study was repeated after a 10 year period the results showed a gradual appearance of blue-yellow defects.  Kinnear studied color vision in a mixed group of 500 insulin and non-insulin dependent diabetic patients using several color vision tests. The results showed that diabetics were poorer in discriminating along the blue-yellow and the green-blue axes than the general population. Furthermore, diabetics who already had signs of retinopathy had poorer discrimination than diabetics with apparently normal fundi. Similar findings arose from the study of Lakowski et al. who found a specific association of blue-yellow and blue-green color vision losses in patients with diabetes mellitus. These losses were found significantly more often among diabetics with retinopathy than in diabetics without retinopathy. 
Primarily blue-yellow color vision deficits were noted in the 100-Hue study of Utku Atmaca who were also one of the few studies to find a relationship between diabetic duration and error score. 
Color vision is essential in the reliable self-monitoring of blood and urine glucose levels. Tests such as the Boehringer Mannheim (BM) test (blood glucose) and the Clinistix tests (urine glucose) rely on the patient having good color vision in order to interpret the results correctly. Poor color vision could lead to an incorrect reading of a test strip and therefore to a mismanagement of the diabetes mellitus. Lombrail et al. assessed the color vision of 103 insulin dependent diabetics using the Farnsworth Munsell 100-Hue test. They then asked the same diabetics to read a series of 30 pre-calibrated BM test glycemic strips. They found that the reading accuracy of the patients was strongly correlated to their 100-Hue score error and they suggested that self-monitoring of blood glucose should be avoided in patients suffering from axial color defects or having high 100-Hue score errors.  Sawicki et al. examined the importance of color discrimination ability regarding the accuracy in self-monitoring of blood glucose. They reported that the color discrimination ability decreased with age, diabetes duration and presence of retinopathy. It was independent of the degree of glycemia and accuracy of self-monitoring of blood glucose.  Mäntyjärvi studied color vision with various tests in a group of diabetic patients, specifically in relation to the blood-glucose strip-test matching. It was noted that the D-15 and 100-Hue tests would be useful in screening for those who cannot correctly interpret the color-dependent glucose test-strips and who would need a blood sugar meter for their blood glucose level testing. 
In summary, the majority of studies indicate a deficit of color vision in diabetes and specifically, diabetic retinopathy. Again, most agree that the blue-yellow color axis is predominantly affected. However, there is disagreement over whether color vision defects can be reliably observed in diabetics without any visible fundus signs of retinopathy when compared to a normal group. The present results suggest that, whilst approaching significance, the 100-Hue test scores for normals are not significantly different from diabetics, but scores rapidly increase as a function of retinopathy level.
| Conclusion|| |
This finding indicates that color vision changes are not strongly related to the duration of diabetes, but more to the severity of the retinopathy. The findings suggest the color vision test in early diagnosed diabetic cases of less significance. However, it would be worthy tested with progression of diabetic retinopathy.
| Acknowledgment|| |
The author would like to thank Professor David Whitaker for his help and support to study and analysis. My sincere thanks go to the participants, secretarial and technical personnel who organized for this study. Part of this study been published with a group of psychophysical and electrophysiological battery of tests in early date.
| References|| |
|1.||Sherman J, Bass JJ, Richardson V. The differential diagnosis of retinal disease from optic nerve disease. J Am Med Assoc 1981;52:933-9. |
|2.||Green FD, Ghafour IM, Allan D, Barrie T, McClure E, Foulds WS. Colour vision of diabetics. Br J Ophthalmol 1985;69:533-6. |
|3.||Bresnick GH, Palta M. Temporal aspects of the electroretinogram in diabetic retinopathy. Arch Ophthalmol 1987;105:660-4. |
|4.||Bresnick GH, Palta M. Oscillatory potential amplitudes. Relation to severity of diabetic retinopathy. Arch Ophthalmol 1987;105:929-33. |
|5.||Roy MS, McCulloch C, Hanna AK, Mortimer C. Colour vision in long-standing diabetes mellitus. Br J Ophthalmol 1984;68:215-7. |
|6.||Roy MS, Rick ME, Higgins KE, McCulloch JC. Retinal cotton-wool spots: An early finding in diabetic retinopathy? Br J Ophthalmol 1986;70:772-8. |
|7.||Trick GL, Burde RM, Gordon MO, Santiago JV, Kilo C. The relationship between hue discrimination and contrast sensitivity deficits in patients with diabetes mellitus. Ophthalmology 1988;95:693-8. |
|8.||Lombrail P, Cathelineau G, Gervais P, Thibult N. Abnormal color vision and reliable self-monitoring of blood glucose. Diabetes Care 1984;7:318-21. |
|9.||Diabetic retinopathy study. Report Number 6. Design, methods, and baseline results. Report number 7. A modification of the Airlie house classification of diabetic retinopathy. Prepared by the diabetic retinopathy. Invest Ophthalmol Vis Sci 1981;21:1-226. |
|10.||Klein R, Klein BE, Moss SE, Davis MD, DeMets DL. The Wisconsin epidemiologic study of diabetic retinopathy. III. Prevalence and risk of diabetic retinopathy when age at diagnosis is 30 or more years. Arch Ophthalmol 1984;102:527-32. |
|11.||Davis MD, Hubbard LD, Trautman J, Klein R. Conference on insulin pump therapy in diabetes. Multicenter study effect on microvascular disease. Studies of retinopathy. Methodology for assessment and classification with fundus photographs. Diabetes 1985;34 Suppl 3:42-9. |
|12.||Verriest G, Van Laethem J, Uvijls A. A new assessment of the normal ranges of the Farnsworth-Munsell 100-hue test scores. Am J Ophthalmol 1982;93:635-42. |
|13.||Lakowski R, Aspinall PA, Kinnear PE. Association between colour vision losses and diabetes mellitus. Ophthalmic Res 1972;3:145-59. |
|14.||Kinnear PR. Proposals for scoring and assessing the 100-Hue test. Vision Res 1970;10:423-33. |
|15.||Smith VC, Pokorny J, Pass AS. Color-axis determination on the Farnsworth-Munsell 100-hue test. Am J Ophthalmol 1985;100:176-82. |
|16.||Moloney J, Drury MI. Retinopathy and retinal function in insulin-dependent diabetes mellitus. Br J Ophthalmol 1982;66:759-61. |
|17.||Bronte-Stewart JM, Cant JS, Craig JO. The detection of early visual loss in young diabetics. Proc R Soc Med 1970;63:786-8. |
|18.||Kinnear PR, Aspinall PA, Lakowski R. The diabetic eye and colour vision. Trans Ophthalmol Soc U K 1972;92:69-78. |
|19.||Utku D, Atmaca LS. Farnsworth-Munsell 100-hue test for patients with diabetes mellitus. Ann Ophthalmol 1992;24:205-8. |
|20.||Sawicki PT, Karschny L, Stolpe V, Wolf E, Berger M. Color discrimination and accuracy of blood glucose self-monitoring in type I diabetic patients. Diabetes Care 1991;14:135-7. |
|21.||Mäntyjärvi M. Screening of diabetics who read incorrectly colour-dependent glucose test-strips. Doc Ophthalmol 1992;80:323-8. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7], [Table 8]