Baseline refraction by age and onset of diabetes mellitus. T1D indicates younger-onset diabetes;T2D, older-onset diabetes.
Klein BEK, Lee KE, Klein R. Refraction in Adults With Diabetes. Arch Ophthalmol. 2011;129(1):56-62. doi:10.1001/archophthalmol.2010.322
To examine refraction, change in refraction, and risk factors for change in refraction in adults with type 1 and type 2 diabetes mellitus.
Population-based study. Modified Early Treatment of Diabetic Retinopathy Study refractions and a standard history were obtained for all participants. Baseline and 10-year follow-up data were available.
Age and education were significantly associated with refraction in persons with younger-onset diabetes (T1D) and in those with older-onset diabetes (T2D); refractions were similar for both groups. Persons of similar age with T1D were likely to be more myopic than were those with T2D (P < .01). In those with T1D, on average, there was a −0.28-diopter (D) change in refraction in 10 years. Those with longer duration of diabetes and proliferative retinopathy were more likely to have hyperopic shifts in refraction. In persons with T2D, there was, on average, a +0.48-D change in refraction during the 10 years, but there was little consistency in the amount of change by age at baseline.
In persons of similar age, those with T1D were likely to be slightly more myopic than were those with T2D. Overall, mean refraction and the important risk factors of age and education were similar to those reported in nondiabetic populations.
Diabetes mellitus affects the eye, with the most commonly reported long-term changes being cataract and diabetic retinopathy. Acute hyperglycemia is associated with myopic refraction, but refraction becomes less myopic (or even hyperopic) with lowering of the levels of glycemia.1- 3 The distribution of refraction in a study of free-living persons with diabetes has not been well described. Some data are available from studies of eye diseases in general populations when persons with diabetes are included in the sampling frame. In the Tanjong Pagar study,4 the mean refractive error was −0.47 diopters (D) for those with diabetes and −0.39 D for those without diabetes, a difference that was not statistically significant. Similarly, in the Beaver Dam Eye Study,5 and the Blue Mountains Eye Study,6 there were no differences in refraction between those with and without diabetes. Some studies, however, have reported more myopic refraction in individuals with diabetes.7 Almost all the persons with diabetes in these studies had type 2 diabetes mellitus. There are no community-wide data on refraction for persons with type 1 diabetes mellitus. We report on the distribution of and change in refraction (spherical equivalent) and on risk factor associations for refraction in a large community-based study of persons with types 1 and 2 diabetes mellitus who were first identified in 1979-1980.
The Wisconsin Epidemiologic Study of Diabetic Retinopathy, a population-based survey of diabetic persons residing and receiving their health care in 11 counties in southern Wisconsin, has been described in detail previously.8- 13 Between July 1, 1979, and June 30, 1980, 10 135 diabetic persons were identified. A sample of 2990 persons selected for the baseline examination (1980-1982) was composed of all persons with a diagnosis of diabetes before 30 years of age who took insulin (n = 1210, the “younger-onset” group [T1D]) and a probability sample of the 5431 persons diagnosed as having diabetes at 30 years or older with their diagnosis confirmed by a casual or postprandial serum glucose level of at least 200 mg/dL (to convert to millimoles per liter, multiply by 0.0555) or a fasting serum glucose level of at least 140 mg/dL on at least 2 occasions (n = 1780, the “older-onset” group [T2D]). Of the T1D group, 996 participated in the baseline examination (1980-1982),9,10 891 in the 4-year follow-up,11 and 784 in the 10-year follow-up.13 Of the 1780 eligible persons with T2D, 1370 participated in the baseline examination,10 987 in the 4-year follow-up,12 and 550 in the 10-year follow-up.13 The reasons for nonparticipation and comparisons between participants and nonparticipants have been presented elsewhere.9- 13
At the time of the examinations of the cohort, a group of persons without diabetes (mostly spouses of the study participants) was examined using the same protocols. This group is referred to as the nondiabetic comparison group; 381 of these individuals participated at baseline, 269 at the second examination, and 211 at the third examination. Data from these persons are used for selected analyses, but direct statistical comparisons between them and the study groups are not given.
Informed consent was obtained from each study participant at each examination. The procedures conformed to the tenets of the Declaration of Helsinki, and approval was obtained for all phases of the examinations from the Health Sciences Institution Review Board of the University of Wisconsin, Madison.
The study procedures remained the same with few additions for each examination phase. All the procedures were conducted according to codified protocols. Only pertinent parts are further described. A medical history was obtained that included information on age at onset of diabetes, medication use, and education. Examinations included measurement of blood pressures by protocol, height, weight, refraction, dilated fundus examinations with examination of the lens for cataract or cataract surgery, fundus examination, and 7 standard field fundus photography of both eyes. Refractions for all participants at each examination were performed according to the Early Treatment of Diabetic Retinopathy Study protocol14 before pupil dilation. Blood was collected for glycated hemoglobin, which was measured using a microcolumn technique15 on a casual blood specimen.
Cataract presence was defined by reference to photographic standards of nuclear and retroilluminated cataracts at the time of the examination. Refraction was converted to spherical equivalents. Early Treatment of Diabetic Retinopathy Study retinopathy severity was categorized as follows: none, mild nonproliferative diabetic retinopathy, moderate or severe nonproliferative diabetic retinopathy, and proliferative diabetic retinopathy.16
Data for the analyses in this study are from the baseline examination and the 10-year follow-up. Data were excluded for aphakic or pseudophakic eyes and for eyes in which the best-corrected visual acuity was 20/200 or poorer. Analyses are limited to individuals 20 years or older at the baseline examination (1370 in the T2D group, 724 in the T1D group, and 269 in the nondiabetic comparison group). Means and standard deviations are presented for the right eyes because the data for the left eyes are nearly identical. Model results are presented from generalized estimating equation models that include both eyes using an independent correlation structure. Unless otherwise noted, all the models include age, education, and the presence of nuclear cataract because these factors are important determinants of refraction in nondiabetic populations.
Individuals excluded from the analyses were significantly older, had more hyperopic refractive error, had higher blood pressures, had a longer duration of diabetes, were more likely to have nuclear cataract, and had more severe retinopathy than did those included (Table 1). The mean (SD) spherical equivalent was −1.24 (2.02) D. There was an increase in spherical equivalent with age (P < .001 per 5-year age category) from a mean refraction of −1.68 D in those 20 to 24 years of age to +0.11 D in those 60 years and older (Figure). There was no significant difference in refraction by sex, glycated hemoglobin level, or duration of diabetes in the adjusted models (Table 2). More years of education was associated with a more myopic refraction, as was more severe retinopathy. Nuclear cataract had no independent association with refraction. There was no significant effect of clinically significant macular edema on refraction.
Those excluded from the analyses were significantly older, had more hyperopic refraction, had a longer duration of diabetes, were more likely to have nuclear cataract, and had more severe retinopathy than did those included (Table 1). The mean (SD) spherical equivalent was 0.69 (2.05) D. There was an increase in spherical equivalent with increasing age (P < .001 per 5-year age category) from a mean refraction of −1.15 D in those younger than 45 years to +1.04 D in those 85 years and older (Figure). More education was associated with a more myopic refraction, as was female sex (Table 2). There were no consistent associations of glycated hemoglobin level, duration of diabetes, or severity of diabetic retinopathy with refraction in the adjusted models (Table 2). Cataract had no independent association with refraction. There was no significant effect of clinically significant macular edema on refraction.
The mean (SD) spherical equivalent was −0.15 (2.06) D. There was an increase in spherical equivalent with increasing age (P < .001 per 5-year age category) from a mean refraction of −1.53 D in those 20 to 24 years of age to +1.52 D in those 75 years and older (Figure). After adjusting for age, more education was associated with a more myopic refraction (data not shown).
We examined whether there was a difference in refraction between T1D and T2D persons of similar age (40-59 years of age). We found that after controlling for age and education, there was a significant difference in refraction between the groups (T1D persons were significantly likely to be more myopic than were T2D persons, P < .01).
The mean (SD) change in spherical equivalent was −0.28 (1.08) D. The youngest members of the group had, on average, a −0.23-D change during the 10-year interval, but the amount of the myopic shift decreased with increasing age at baseline (Table 3). After adjusting for age, women were more likely to have a myopic shift in refraction, and those with a longer duration of diabetes and more severe retinopathy were likely to have hyperopic shifts in refraction (Table 3).
The mean (SD) change in spherical equivalent was 0.48 (0.89) D. The youngest members of the group had, on average, a +0.58-D change in refraction during the 10-year interval, but the largest change happened in the next-oldest age group, with smaller positive changes occurring for each age group thereafter (Table 3). After adjusting for age, there were no significant effects of the other variables on change in refraction (Table 3).
There were no significant effects of age or other variables we considered on change in refraction (Table 3).
Because change was affected by age in most analyses, we examined change in refraction for those of similar ages (40-59 years of age at baseline). We found that after adjusting for age and education, there was a borderline significant difference in change in refraction between the groups (persons with T2D had smaller changes than did those with T1D, P = .08).
We found that risk factors for refraction and risk factors for change in refraction in adults with diabetes are similarly distributed in those with T1D and T2D after adjusting for age and education. In addition, the mean change in refraction in the present population with T2D is similar to that found in diabetic persons during a 10-year interval in another Wisconsin study.17 Most important, we found that refraction and its correlates in adults with diabetes regardless of type are similar to those reported in adults without diabetes.4- 6
In particular, the present data indicate the importance of current age and education in refraction. Regarding change in refraction during adult years, age seems to be a consistent factor. The primary purpose of the present investigation was not to compare refractions in those with diabetes with those without this condition but to compare refractions among those with T1D and T2D. The same risk factors for refraction were found in these groups. However, we found that in persons of the same age, those with T1D were likely to have more myopic refractions. It is possible that relatively elevated glucose levels earlier in life in those with T1D affected refraction and that there was a residual effect. Because those analyses were adjusted for education, the difference is unlikely to reflect greater education in the T1D group.
We had anticipated that glycemia would be an important determinant of refraction; it has been reported that refractions tend to be more myopic in persons with diabetes who have relatively acute elevations in their blood glucose levels, inferred from the rapid hyperopic shift in refraction that has been reported for markedly hyperglycemic persons who are rapidly brought under control.2 We did not find that glycemia, as reflected in the baseline glycated hemoglobin level, had a significant effect on refraction in either diabetic group. We could not assess the potential effects of glycemia as reflected in fasting blood glucose levels because we did not measure this analyte.
Other researchers18,19 have reported that individuals with high myopia are less likely to have more severe retinopathy than are those with emmetropia, hyperopia, or small to moderate myopic refractions. The present cross-sectional data do not support a relationship between severity of retinopathy and refractive error in patients with T1D or T2D. However, there were few high myopes in this study, and those with severe retinopathy were often excluded from refraction analyses because of poor visual acuity.
There are some limitations to this study that may have affected the results. Persons excluded from the baseline analyses were significantly older, had more hyperopic refraction, had a longer duration of diabetes, were more likely to have nuclear cataract, and had more severe retinopathy in both groups than did those included. Thus, these exclusions might be expected to bias the baseline estimates of refraction toward myopia. Age adjustment in subsequent analyses likely reduces some of these effects.
In summary, we examined correlates of refraction and change in refraction in a large population-based study of persons with T1D and T2D. The most important correlates for both groups were age and education, which are similar to those for the general population.
Correspondence: Barbara E. K. Klein, MD, MPH, Department of Ophthalmology and Visual Sciences, University of Wisconsin School of Medicine and Public Health, 610 N Walnut St, Fourth Floor, Wisconsin Alumni Research Foundation, Madison, WI 53726-2336 (firstname.lastname@example.org).
Submitted for Publication: October 20, 2009; final revision received April 27, 2010; accepted May 1, 2010.
Author Contributions: The authors have had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: None reported.
Funding/Support: This research was supported by grants EY03083 and EY016379 from the National Institutes of Health (Drs B.E.K. Klein and R. Klein) and, in part, by Senior Scientific Investigator Awards from Research to Prevent Blindness (Drs B.E.K. Klein and R. Klein). The National Eye Institute provided funding for the entire study, including collection and analyses of data, and Research to Prevent Blindness provided further additional support for data analyses.
Disclaimer: The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Eye Institute or the National Institutes of Health.