|Year : 1996 | Volume
| Issue : 4 | Page : 219-223
Graded corneal sensitivity for screening of diabetic retinopathy
Jagjit S Saini, Sangeet Mittal
Postgraduate Institute of Medical Education and Research, Chandigarh, India
Jagjit S Saini
Department of Ophthalmology, Postgraduate Institute of Medical Education and Research, Chandigarh-160 012
Source of Support: None, Conflict of Interest: None
Several alternative approaches to screen diabetics followed by referral of patients with retinopathy changes, are being investigated. The intent is to demonstrate usefulness of a cost effective, easy and valid screening test. We investigated in this report the efficacy of graded corneal sensitivity for screening presence of diabetic retinopathy. In 105 randomly chosen subjects with (70 subjects) and without (35 subjects) diabetes mellitus, corneal aesthesiometry and status of retinopathy was determined independently. Corneal sensitivity in subjects of diabetes mellitus without retinopathy (1.17±0.29 gm/mm) was significantly different when compared to healthy controls (0.99±0.04 gm/mm) (p<0.01). Also corneal sensitivity in eyes with any type of retinopathy was significantly worse (1.94±1.33 gm/mm) when compared to eyes without retinopathy. At a cut off value of 1.20 gm/mm the sensitivity and specificity of corneal hypoesthesia as a test to detect diabetic retinopathy was 86% and 74%, respectively. For detecting presence of proliferative diabetic retinopathy at a cut off value of 1.49 gm/mm the sensitivity and specificity were 89% and 80% respectively. These observations indicate that corneal hypoesthesia may be a reasonable indication of the presence of diabetic retinopathy and could be used to screen diabetic populations for retinopathy, after its validity is confirmed in larger studies.
Keywords: Corneal sensitivity, corneal aesthesiometry, screening, diabetic retinopathy
|How to cite this article:|
Saini JS, Mittal S. Graded corneal sensitivity for screening of diabetic retinopathy. Indian J Ophthalmol 1996;44:219-23
|How to cite this URL:|
Saini JS, Mittal S. Graded corneal sensitivity for screening of diabetic retinopathy. Indian J Ophthalmol [serial online] 1996 [cited 2019 Aug 23];44:219-23. Available from: http://www.ijo.in/text.asp?1996/44/4/219/24675
Blindness from complications of retinopathic changes in diabetes mellitus, is increasing in both developed and developing nations because of increasing longevity and improving living standards. There is need to screen huge population of diabetes mellitus to identify those who are on the threshold of blindness., Ophthalmoscopic screening by non-ophthalmologist is inadequate. Objective but less cumbersome screening methods for those at high risk of diabetic retinopathy have been explored. Some of these suggested screening methods can be undertaken by non-ophthalmologist but need expensive equipment (autofluorescence of cornea) or recurrent cost (fundus photography). In diabetes mellitus, clinical corneal hypoesthesia,, because of altered corneal nerves, is known. Corneal and retinopathic effects in diabetes mellitus are demonstrated to be due to the metabolic changes. Further, investigators have demonstrated an association of maculopathy to cardiovascular autonomic neuropathy. These observations suggested to us that graded corneal hypoesthesia may be linked to retinopathy in diabetes mellitus. The measurement of corneal sensitivity by performing corneal aesthesiometry is a non-invasive and simple technique that can be done in a short time peroid.
In this study corneal sensitivity was measured in Non Insulin Dependent Diabetes Mellitus (NIDDM) subjects with or without retinopathic changes and healthy controls with an intent to assess its usefulness as an indicator for the presence of diabetic retinopathy (proliferative and non-proliferative).
| Materials and methods|| |
| Subjects|| |
One hundred and five subjects included in this study were categorized in four groups. Group I consisted of 35 subjects with NIDDM without retinopathy changes. Group II consisted of 21 NIDDM subjects with manifest non-proliferative diabetic retinopathy changes. Group III included 14 NIDDM subjects with proliferative diabetic retinopathy. Group IV included another 35 subjects who were recruited as normal healthy controls. NIDDM subjects were recruited randomly from those attending Diabetes clinic and Ophthalmology OPD of the Post Graduate Institute of Medical Education and Research, Chandigarh. Diabetes mellitus was diagnosed on the basis of Fasting Blood Sugar (FBS) level more than 140mg/dl and Post Prandial Blood Sugar (PPBS) level more than 200mg/dl. Healthy controls were recruited from the staff at the hospital and their relatives. We recruited the subjects to achieve a stratified sample that was balanced for age and sex as explained in an earlier study. Before participation, informed consent was obtained from each subject and tenents of Declaration of Helnsinki were followed. The study protocol was approved by Institutional Ethics and Review Committee. Subjects with manifest anterior segment pathology or contact lens wearers were not included. Also patients who had undergone any ocular surgery or retinal photocoagulation were excluded. None of the patients was on any topical drugs.
Information regarding corneal sensitivity was collected independently as the observer remained masked for diabetic and retinopathic status of the subjects.
| Evaluation of Corneal Sensitivity|| |
The corneal thickness was measured using a Cochet Bonnet aesthesiometer under 1X - 1.6X magnification. In good illumination aesthesiometer filament was fully extended to 60 mm. The tip of the fiber was steadily advanced towards the cornea. When the end plate of nylon filament was found to be in contact with cornea a mild pressure was exerted such that fiber had the slightest bend just visible. The response was assessed either by subjective response of patient or by objective blinking or withdrawal response. If there was no response the fiber length was shortened in steps of 5 mm each time and procedure was repeated till a response was elicited. The readings were taken at five different points on the cornea (at centre, 12o' clock, 3o'clock, 6o'clock and 9o' clock) and was repeated three times. The average of all the readings in mm was noted. At times 'blanks' were given to test patient's reliability and only reliable data was included.
| Evaluation of Retinopathy|| |
Employing a modified Airlie House classification, diabetic retinopathy was grouped in three categories only. Detailed ophthalmoscopic (direct and indirect) evaluation including fluorescien angiography was performed. Group I included eyes with no or negligible retinopathy, demonstrating less than two microaneurysms per field. The category is identified as without retinopathy in the text. Group II included eyes demonstrating three or more microaneurysms per field and /or retinal haemorrhages, hard and soft exudates, intraretinal microvascular abnormalities and venous beading. This category is referred to as non-proliferative retinopathy in the text. Group III included eyes with new vessels and fibrous proliferation and/or vitreous haemorrhage in addition to findings of earlier categories. This category is identified as proliferative retinopathy in the text.
| Analysis of Data|| |
The data was assessed for normal distribution and was found to be normally distributed. Therefore, the parametric tests were applied for comparisons among the diabetic groups. The comparisons were made by performing analysis of variance (F-test). Whenever the analysis of variance showed significant result, then 't'-test based on mean square analysis of variance was performed for comparison of two groups. The student 't' test was employed between healthy controls and diabetic subjects without retinopathy to evaluate any significant difference in corneal sensitivity.
Evaluation of specificity, sensitivity and predictive values of corneal sensitivity for retinopathy was made. Determination of optimal cut off value of corneal sensitivity as a test for detection of retinopathy was obtained with receiver operating characteristics curve analysis.
| Results|| |
There was no effect of groups on age (p>0.05) or sex (p>0.05) [Table:1]. Hence the groups were effectively age and sex matched.
The mean and standard deviation of corneal sensitivity at the five tpoints is depicted in [Table:2]. The mean corneal sensitivity in eyes with proliferative diabetic retinopathy (2.63±1.83gm/mm) and non proliferative diabetic retinopathy (1.48±0.47gm/mm) were significantly higher when compared to mean value in subjects of diabetes mellitus without retinopathy (1.17±0.29gm/mm) and healthy controls (0.99±0.04gm/ mm) (p<0.01). The mean value of corneal sensitivity was significantly higher in subjects of diabetes mellitus with proliferative diabetic retinopathy when compared to subjects with non-proliferative diabetic retinopathy (p<0.01). Also the mean value of corneal sensitivity in subjects of diabetes mellitus with any type of retinopathy (1.94±1.33 gm/mm) was significantly higher than subjects of diabetes mellitus without retinopathy (p<0.01).
A correlation matrix of all independent variables was computed in diabetic subjects. In diabetic subjects there was no significant correlation of corneal sensitivity with the age of the patients (r=-0.041; p>0.05).
To determine the usefulness of corneal sensitivity in term of its ability to identify eyes with diabetic retinopathy, the criteria of corneal sensitivity values was varied to different cut off values. At each cut off value, studied eyes were resorted into three predefined categories of diabetic retinopathy. The sensitivity and specificity of corneal hypoesthesia as a predictor of diabetic retinopathy and proliferative diabetic retinopathy was then determined. By varying the criteria of cut off value of corneal sensitivity, and determining sensitivity and specificity values at each cut off value a receiver operating characteristics curve for corneal sensitivity was generated [Figure:1].
On employing a cut off value of corneal sensitivity of 1.20 gm/mm [Table:3], sensitivity and specificity of the test for detection of retinopathy was determined to be 86% and 74%, respectively. These observations indicate that corneal hypoesthesia was a reasonable indication of the presence of retinopathy. The positive and negative predictive values were calculated to be 77% and 84%, respectively.
On employing a cut off value of corneal sensitivity of 1.49 gm/mm [Table:4], the calculated sensitivity of the test for detection of proliferative retinopathy in our study was 89% and specificity of test was 80%. This indicates that corneal sensitivity was also a reasonable indication of the presence of proliferative diabetic retinopathy. The positive and negative predictive values were calculated to be 52% and 97%, respectively.
| Discussion|| |
In diabetes mellitus several abnormalities of cornea have been demonstrated. [8, 10, 12] Although the pathogenesis of these changes remain speculative, there is evidence now that like retinopathic changes, corneal abnormalities are effects of metabolic alterations in diabetes mellitus. In order to develop a similar but valid screening procedure to indicate presence of diabetic retinopathy, investigators have observed that study of corneal alterations may be useful in predicting presence of diabetic retinopathy. Studies on grading corneal autofluorescence and epithelial fragility have been demonstrated to be reasonable alternatives to screen patients of diabetes mellitus. We have explored the value of employing graded corneal sensitivity for predicting presence of diabetic retinopathy in patients of diabetes mellitus. We observed that mean corneal sensitivity in proliferative diabetic retinopathy eyes (2.63±1.83 gm/mm) and non proliferative diabetic retinopathy eyes (1.48±0.47gm/mm) was significantly higher when compared to values in eyes without retinopathy (1.17±0.29 gm/mm) and healthy controls (0.99±0.04 gm/mm). The corneal sensitivity in subjects with any type of diabetic retinopathy (1.94±1.33 gm/mm) was also significantly different from the subjects without retinopathy. Also, significant difference was observed between the healthy controls and subjects of diabetes mellitus without retinopathy. These observations indicate that corneal sensitivity was decreased in subjects of diabetes mellitus and this impairment increased with manifest retinopathy, which corroborate an earlier observation.
Graded corneal hypoesthesia demonstrated on employing receiver operating characteristics curve analysis, high sensitivity (86%) and specificity (74%) for predicting presence of diabetic retinopathy when values were higher than 1.20 gm/mm. On employing a cut off value of 1.49 gm/mm of corneal sensitivity, it was possible to predict presence of proliferative diabetic retinopathy with high sensitivity (89%) and specificity (80%). The prevalence of diabetic retinopathy in our study was 50% (35/70) resulting in positive and negative predictive value of 77% and 84% respectively. On assuming a 52% prevalence rate of diabetic retinopathy in diabetic patients the positive and negative predictive values would be 78% and 83%, respectively. The prevalence of proliferative diabetic retinopathy in our study was 20% (14/70) resulting in positive and negative predictive value of 52% and 97%, respectively. On assuming a prevalence rate of 10% of proliferative diabetic retinopathy in diabetic patientsthe positive and negative predictive values would be 33% and 99%, respectively. Low positive predictive value for proliferative diabetic retinopathy is due to the low prevalence rate.
Results on employing graded corneal sensitivity in screening diabetic population for diabetic retinopathy are comparable to employing corneal autofluorescenceand corneal epithelial fragility [Table:5]. Grading of corneal sensitivity is a simple, non-invasive, portable and inexpensive procedure which can be performed by non- ophthalmologists without any discomfort to the patient. Corneal autofluorescence is measured using a commercial fluorophotometer and the instrument used is expensive, is not easily portable and is not available in most clinics. Corneal epithelial fragilitymeasurement is cheap but needs a slit lamp to observe the punctate staining as end point which requires trained manpower. Ophthalmoscopic screening when performed by diabetologist as reported in one study was considerably inferior (missed diagnosis 33%) to screening with corneal aesthesiometer (missed diagnosis 8.3%). Screening of diabetics by grading corneal sensitivity should be useful in identifying those patients who need more comprehensive retinal evaluation, which otherwise may not be possible given the inadequate facilities for a large diabetic population in several parts of the world. Also, in situations when retina can not be seen adequately due to media haze a corneal sensitivity evaluation may help in indicating presence of diabetic retinopathy and the need for early treatment.
| References|| |
Moss SE, Klein R, Klein B. The incidence of vision loss in a diabetic population. Ophthalmology. 95:1340-1348, 1988.
Merimee TJ. Diabetic retinopathy. A synthesis of perspectives. N Engl J Med. 322:978-982, 1990.
L' Esperance FA Jr, James WA Jr. The problem of diabetic retinopathy. In: Diabetic Retinopahy. Little HL, Jack RL, Patz A, Forsham PH, eds. New York, Thieme -Stratton Inc. 1983, pp 11-20.
Will JC, Geiss LS, Wetterhall SF. Diabetic retinopathy. N Engl J Med. 323:613-618, 1990.
Retinopathy working party: A protocol for screening diabetic retinopathy in Europe. Diabetic Med. 8:263-267, 1991.
Begg IS. Screening for diabetic retinopathy. Changes in direction? Can J Ophthalmol: 28:3-4, 1993.
Sussman EJ, Tsiaras WG and Soper KA : Diagnosis of diabetic eye disease. JAMA. 247:3231-3235, 1982.
Stolwizk TR, Van Best JA, Oosterhuis JA, Swart W: Corneal autofluorescence: An indicator of diabetic retinopathy. Invest Ophthalmol Vis Sci. 32:92-97,1992.
Moss SE, Klein R, Kessler SD. Comparison between ophthalmoscopy and fundus photography in determining severity of diabetic retinopathy. Ophthalmology. 92:62-67, 1985.
Saini JS, Khandalvala B. Corneal epithelial fragility in diabetes mellitus. Can J Ophthalmol. 30:142-146, 1995.
Schwartz DE. Corneal sensitivity in diabetics. Arch Ophthalmol. 91:174-178, 1974.
Rogell GD. Corneal hypoesthesia and retinopathy in diabetes mellitus. Ophthalmology. 87:229-233, 1980.
Ishida N, Rao GN, Del Cerro M and Aquavella JV. Corneal nerve alterations in diabetes mellitus. Arch Ophthalmol. 102:1380-1384, 1984.
Hosotani H, Ohashi Y, Yamada Y, Tsubota K. Reversal of abnormal corneal epithelial cell morphologic characteristics and reduced corneal sensitivity in diabetic patients by aldose reductase inhibitor CT, 112. Am J Ophthalmol. 119:288-294, 1995.
Datiles MB, Kador PF, Fukui H, et al. Corneal re-epithelialisation in galactosemic rats. Invest Ophthalmol Vis Sic. 24:563-569, 1983.
Kinoshita JH. Aldose reductase in the diabetic eye : (Edward Jackson memorial lecture). Am J Ophthalmol. 102:685-692, 1986.
Awata T, Sogo S, Yamamoto Y : Effects of aldose reductase inhibitor, CT-112, on sugar alcohol accumulation in corneal epithelium of galactose-fed rats. Jpn J Ophthalmol. 30:245-250,1986.
Steinke J, Soeldner JS. Diabetes mellitus. In Principles of Internal Medicine, Thorn GW, Adams RD, Braumwald E, et al, eds. Tokyo, Mcgraw Hill, Koga Kush. 1977, pp 563-583.
Dosso A, Golay A, Morel Y, et al. Cardiovascular autonomic neuropathy in diabetic patients with macular edema. Diabete Metab. 21:41-45, 1995.
The Diabetic Retinopathy Study Research Group: Report 7: A modification of Airlie House Classification of diabetic retinopathy. Invest Ophthalmol Vis Sci. 21:210-226, 1981.