RSS Feeds
Relation of Glycemic Control to Diabetic Microvascular Complications in Diabetes Mellitus
- Ronald Klein, MPH, MD;
- Barbara E. K. Klein, MPH, MD; and
- Scot E. Moss, MA
- From the University of Wisconsin Medical School, Madison, Wisconsin. Note: This article is one of a series of articles comprising an Annals of Internal Medicine supplement entitled “Risks and Benefits of Intensive Management in Non-Insulin-dependent Diabetes Mellitus: The Fifth Regenstrief Conference.” To view a complete list of the articles included in this supplement, please view its Table of Contents. Acknowledgments: The authors thank their collaborators, Karen Cruickshanks, PhD; Matthew D. Davis, MD; and Polly Newcomb, PhD, who provided scientific advice; the 452 Wisconsin physicians and their staffs who participated in and supported this study; and Terry Spennetta and Earl Shrago, MD, who provided laboratory support (P30 AM AG 26659, Public Health Service, National Institutes of Health, Bethesda, Maryland). Grant Support: By National Institutes of Health grant EYO-3083 (R Klein, BEK Klein) and, in part, by Research to Prevent Blindness (Dr. Klein, Senior Scientific Investigator Award). Requests for Reprints: Ronald Klein, MD, MPH, Department of Ophthalmology and Visual Sciences, University of Wisconsin-Madison, 610 North Walnut Street, 460 WARF, Madison, WI 53705-2397. Current Author Addresses: Drs. Klein. Klein, and Moss: Department of Ophthalmology and Visual Sciences. University of Wisconsin-Madison, 610 North Walnut Street, 460 WARF, Madison, WI 53705-2397.
Abstract
Objective: To describe the relation between glycated hemoglobin and the incidence or progression, or both, of diabetic microvascular complications in persons with insulin-dependent (IDDM) and non–insulin-dependent diabetes mellitus (NIDDM).
Design: Population-based cohort study.
Setting: An 11-county area in southern Wisconsin.
Patients: All persons with IDDM diagnosed before age 30 and taking insulin (n = 996) and a probability sample (based on duration of disease) of persons diagnosed with diabetes at age 30 or older who were either taking insulin (n = 674) or not taking insulin (n = 696) and who participated in a baseline examination from 1980 to 1982. Survivors of the cohort were re-examined again in 1984 to 1986 and 1990 to 1992.
Measurements: The incidence and progression of diabetic retinopathy was determined by masked grading of stereoscopic color fundus photographs using the modified Early Treatment Diabetic Retinopathy Study severity scale. Gross proteinuria was determined using a dipstick. Ten-year incidence of renal dialysis or transplantation or loss of tactile sensation or of temperature sensitivity was based on self-reported history.
Results: The glycated hemoglobin level at baseline was strongly related to the incidence or progression, or both, of diabetic retinopathy, the incidence of gross proteinuria, and the incidence of loss of tactile sensation or temperature sensitivity in persons with either IDDM or NIDDM.
Conclusions: These prospective epidemiologic data suggest that glycemic control is similarly related to the incidence and progression of diabetic microvascular complications in both IDDM and NIDDM. However, further evidence from clinical trials in persons with NIDDM is necessary to assess the risks and benefits of such treatment in preventing these complications.
Little information is available comparing the incidence and progression of chronic diabetic complications among persons with insulin-dependent diabetes mellitus (IDDM) and those with non–insulin-dependent diabetes mellitus (NIDDM) [1-4]. Such comparisons have been limited by the difficulty in accurately classifying the type of diabetes and by the paucity of studies in which persons with IDDM or NIDDM were assessed concurrently using the same standardized objective methods (such as masked grading of stereoscopic fundus photographs).
Epidemiologic data show that the natural history of retinopathy is similar in both types of diabetes [2, 5-8]; however, the prevalence is higher and the severity is greater in persons with IDDM than in those with NIDDM [5, 6]. When loss of vision is present, it is more likely to be associated with proliferative retinopathy in those with IDDM and with macular edema in those with NIDDM [9, 10]. In both types of diabetes, early detection and treatment of proliferative retinopathy and clinically significant macular edema is estimated to prevent 95% of occurrences of substantial loss of vision [11]. However, data from epidemiologic studies suggest that vision-threatening retinopathy is often not detected in a timely fashion in persons with either type of diabetes [12-14].
The pathologic changes characterizing diabetic nephropathy—glomerular basement membrane thickening and expansion of the glomerular mesangium—are similar in IDDM and NIDDM [3]. Data from some studies suggest that the risk for developing diabetic nephropathy is higher in IDDM, but other studies have found the risk to be similar in IDDM and NIDDM [15-22]. Although renal dialysis and transplantation have been shown to prolong life and improve its quality in both types of persons, the treatment is costly and not without serious short- and long-term complications.
Less is known about the epidemiology of diabetic neuropathy in persons with IDDM and NIDDM. Data from one study suggest similar prevalences of symptomatic polyneuropathies in both types of diabetic persons, although those with IDDM had a higher frequency of a more severe stage of polyneuropathy than those with NIDDM [23]. However, a cross-sectional multicenter study of English diabetic patients showed that the prevalence of neuropathy in persons with NIDDM was higher (32%) than in those with IDDM (23%) [24].
The Diabetes Control and Complications Trial (DCCT) has shown that intensive management and control of hyperglycemia is of benefit in preventing the incidence and progression of diabetic retinopathy; loss of vision; the need for photocoagulation treatment; and the incidence of microalbuminuria, gross proteinuria, and diabetic neuropathy in persons with IDDM [25]. To date, no definitive clinical trial data have shown that glycemic control is of similar benefit in preventing these complications in persons with NIDDM. The purpose of this article is to examine the relation of hyperglycemia to the long-term incidence and progression of diabetic retinopathy, gross proteinuria, end-stage renal disease, and history of neuropathy in persons with either IDDM or NIDDM using data from a large, population-based study, the Wisconsin Epidemiologic Study of Diabetic Retinopathy (WESDR). Such information may help in determining whether information derived from persons with IDDM in the DCCT can be used to make recommendations about intensive glycemic control in persons with NIDDM.
Methods
WESDR Population
The WESDR population has been described in detail in other publications [2, 5-8]. In brief, the population included a probability sample selected from 10 135 diabetic patients who received primary care in an 11-county area in southern Wisconsin from 1979 to 1980. This sample included a “younger-onset” group (all patients were diagnosed before 30 years of age and were taking insulin, and almost all had IDDM [n = 1210]) and an “older-onset” group (a sample of persons [n = 1780]) stratified by duration of diabetes, of whom 824 were taking insulin (a mixture of persons with either IDDM or NIDDM) and 956 were not (all had NIDDM).
Among the younger-onset group, 996 participated in the baseline examination (1980 to 1982), 891 in the 4-year follow-up, and 765 in the 10-year follow-up. Of the 1780 eligible older-onset persons, 1370 participated in the baseline examination, 987 in the 4-year follow-up, and 533 in the 10-year follow-up. The population was 98.6% white. The reasons for nonparticipation and comparisons among participants and nonparticipants at baseline and the 4-year and 10-year follow-up examinations have been presented elsewhere [2, 5-8]. The main reason for nonparticipation in both groups over the 10-year course of the study was death.
Procedures
The baseline and follow-up examinations were done in a mobile examination van in or near the city where the participants lived. The ocular and physical examinations included measuring visual acuity using a modification of the Early Treatment Diabetic Retinopathy Study protocol [26], measuring blood pressure using the Hypertension Detection and Follow-up Program protocol [27], testing of urine for gross proteinuria (0.3 g/L or more) using a dipstick (Labstix, Ames, Iowa), administering a structured interview by examiners, taking stereoscopic color fundus photographs of the Diabetic Retinopathy Study seven standard fields [28], and measuring hemoglobin A1 ([HbA1] A1a, A1b, and A1c) using a microcolumn technique [29]. The normal range for glycated hemoglobin was 4.6% to 7.9%. Its coefficient of variation was 2.4%. The WESDR HbA1 microcolumn results compare with the DCCT HbA1c results as follows: DCCT = 0.003 + 0.935 (WESDR). Grading protocols have been described in detail elsewhere and are modifications of the Early Treatment Diabetic Retinopathy Study adaptation of the modified Airlie House Classification scheme of diabetic retinopathy [12].
Definitions
The severity scale for diabetic retinopathy is defined as follows. Level 10 represents no retinopathy, levels 21-53 represent nonproliferative retinopathy of increasing severity, and levels 60-80 represent proliferative retinopathy of increasing severity. The retinopathy level of a patient was derived by giving the eye with the higher level greater weight. Patients in a given level were divided into two groups: those with the same level in each eye and those with a lesser level in one eye. For example, the level for a participant with level 21 retinopathy in each eye was specified by the notation level 21/21. This scheme provides a 15-step scale (10/10, 21/< 21, 21/21, 31/< 31, 31/31, 37/< 37, 37/37, 43/< 43, 43/43, 47/< 47, 47/47, 53/< 53, 53/53, 60+/< 60+, and 60+/60+).
The incidence of any retinopathy was estimated from all persons who had no retinopathy at the baseline examination (severity level 10/10) and who participated in the follow-up examination(s) (n = 261 for younger-onset persons, 146 for older-onset persons taking insulin, and 301 for older-onset persons not taking insulin). Progression to proliferative retinopathy was estimated from all persons who were free of this complication at the baseline examination (n = 712 for younger-onset persons, 417 for older-onset persons taking insulin, and 487 for older-onset persons not taking insulin). For persons with nonproliferative or no retinopathy, progression was defined as an increase in the severity of retinopathy by two steps or more at either of the follow-up examinations (n = 712 for younger-onset persons, 417 for older-onset persons taking insulin, and 487 for older-onset persons not taking insulin).
Presence of macular edema was defined as thickening of the retina with or without partial loss of transparency within one disc diameter from the center of the macula. The incidence was estimated from all persons who had no macular edema and had not been previously treated with photocoagulation at baseline (n = 688 for younger-onset persons, 329 for older-onset persons taking insulin, and 444 for older-onset persons not taking insulin).
For each eye, the best-corrected visual acuity was recorded as the number of letters read correctly from 0 (≤ 20/250) to 70 (20/10). Loss of vision was defined as a doubling of the visual angle (a loss of 15 letters or more, for example, a change from 55 to 40 letters, which corresponds to a visual acuity change from 20/20 to 20/40).
The incidence of proteinuria was estimated from all persons who had less than 0.30 g/L urine protein concentration at the baseline examination and who participated in the follow-up examination(s) (n = 666 for younger-onset persons, 376 for older-onset persons taking insulin, and 418 for older-onset persons not taking insulin). Incidence of proteinuria was defined as a urine protein concentration of 0.30 g/L or more at either of the follow-up examinations. The incidence of renal failure was defined as being on renal dialysis or having had a renal transplantation at one of the follow-up examinations in all persons with a urine protein concentration less than 0.30 g/L at the baseline examination (n = 666 for younger-onset persons, 376 for older-onset persons taking insulin, and 418 for older-onset persons not taking insulin). The incidence of loss of tactile sensation or loss of temperature sensitivity was defined as reporting a history of these complications at one of the follow-up examinations in persons who did not have them at the baseline examination (n = 444 for younger-onset persons, 148 for older-onset persons taking insulin, and 258 for older-onset persons not taking insulin).
Data Analysis
To determine the relative risks and 95% confidence intervals for the incidence and progression of retinopathy, the baseline glycated hemoglobin level was divided into quartiles for each group. Tests for trends in rates were done by the Mantel-Haenszel procedure stratified on time period [30]. Multivariate analyses for predicting incidence of any retinopathy, progression of retinopathy, and progression to proliferative retinopathy were done by discrete linear logistic regression [31]. All models were evaluated internally for fit by the Hosmer-Lemeshow test [31]. Except for an occasional statistically significant result, the tests indicated reasonable fit for most of the models (see Table 1).
Results
Glycated Hemoglobin Levels in the WESDR
For the younger-onset group, the older-onset group taking insulin, and the older-onset group not taking insulin, the mean age at baseline was 29.3 years, 65.2 years, and 68.0 years and the mean duration of diabetes was 14.7 years, 15.0 years, and 8.8 years, respectively. There were statistically significant differences (P < 0.0001) in mean glycated hemoglobin among the groups at baseline (younger-onset, 10.8%; older-onset taking insulin, 10.2%; and older-onset not taking insulin, 8.9%) and at the 4-year follow-up (younger-onset, 10.1%; older-onset taking insulin, 9.7%; and older-onset not taking insulin, 8.9%) in the WESDR. Over the first 4 years of the study, the mean glycated hemoglobin decreased in both WESDR groups taking insulin (P < 0.001) but not in the group taking no insulin (P = 0.84).
To examine whether the risk for incidence and progression of complications varied with the baseline level of glycated hemoglobin among the diabetic groups, quartiles were calculated for glycated hemoglobin at baseline for the total population in the WESDR. The first quartile varied from 5.4% to 8.5%; the second quartile, from 8.6% to 10.0%; the third quartile, from 10.1% to 11.5%; and the fourth quartile, from 11.6% to 20.8%. Equivalents for glycated HbA1c, as determined in the DCCT, were 5.1% to 7.9% for the first quartile; 8.0% to 9.3% for the second quartile; 9.4% to 10.8% for the third quartile; and 10.9% to 19.5% for the fourth quartile.
Relation of Glycated Hemoglobin Level to the 10-year Incidence and Progression of Diabetic Retinopathy, Macular Edema, and Loss of Vision in the WESDR
The 10-year incidence of any retinopathy, progression, or progression to proliferative retinopathy was 89.3%, 75.8%, and 29.8%, respectively, in the younger-onset group; 79.2%, 68.7%, and 23.6%, respectively, in the older-onset group taking insulin; and 66.9%, 52.9%, and 9.7%, respectively, in the older-onset group not taking insulin. The 10-year incidence of macular edema (25.4%) and loss of vision (32.8%) was higher in the older-onset group taking insulin than in the younger-onset group (20.1% for macular edema and 9.2% for visual loss) and the older-onset group not taking insulin (13.9% for macular edema and 21.4% for visual loss).
Although the incidence of proliferative retinopathy was highest in the younger-onset group because of the higher frequency of older-onset diabetes in the population, more persons in the WESDR cohort who developed proliferative retinopathy over the 10-year period had older-onset diabetes (387 in the older-onset group as compared with 226 in the younger-onset group). Similarly, those with older-onset diabetes accounted for more persons who developed macular edema (426 in the older-onset group as compared with 157 in the younger-onset group).
Similar significant trends for 10-year incidence, progression, progression to proliferative retinopathy, or incidence of macular edema with increasing baseline glycated hemoglobin occurred for each of the three WESDR diabetic groups (Figure 1 and Figure 2) [32]. Few differences existed among the groups in incidence or rates of progression of retinopathy within specific quartiles of glycated hemoglobin at baseline. For example, the 10-year rate of progression for those in the WESDR younger-onset group whose glycated hemoglobin was in the third quartile (10.1% to 11.5%) at baseline was 82.2%; for the older-onset group taking insulin, it was 74.0%; and for the older-onset group not taking insulin, it was 84.2% (P = 0.11; Figure 1, center). Thus, these data suggest that the level of hyperglycemia is more important in determining risk for progression of retinopathy than the “type” of diabetes. These data also suggest that the lower prevalence and decreased incidence and progression of retinopathy in persons with NIDDM compared with those with IDDM may be caused, in part, by a “better” level of glycemic control in persons with NIDDM.
To evaluate the relative influence of several variables on the 10-year incidence of proliferative retinopathy or macular edema, we developed models based on logistic regression. Independent variables included in the models were age, duration of diabetes, baseline retinopathy level, and glycated hemoglobin at baseline and the change in glycated hemoglobin from the baseline to the 4-year follow-up examination. Glycated hemoglobin levels and the change in glycated hemoglobin levels were forced into all models. Using these models, we estimated that a 2 percentage point difference in glycated hemoglobin levels (for example, from 11.0% to 9.0%) from baseline to the 4-year follow-up would be expected to lead to decreases in the 10-year incidence of proliferative retinopathy and macular edema that were substantial and generally comparable in all groups (Table 1). We have previously reported finding no threshold effect within the range of hyperglycemia found in our population [32]. Thus, a similar decrease in glycated hemoglobin levels from 9% to 7% would be expected to lead to similar reductions in the risk for developing proliferative diabetic retinopathy or macular edema.
Relation of Glycated Hemoglobin Level to the 10-Year Incidence of Gross Proteinuria and Renal Failure in the WESDR
The 10-year incidences of gross proteinuria and renal failure were 28.3% and 2.2%, respectively, in the younger-onset group; 40.0% and 1.7%, respectively, in the older-onset group taking insulin; and 33.4% and 0.6%, respectively, in the older-onset group not taking insulin. These differences among the groups were statistically significant (P < 0.005) for incidence of gross proteinuria but not renal failure (P = 0.32). There was a trend for increasing incidence of gross proteinuria with increasing baseline glycated hemoglobin for the three groups (Figure 3). The 10-year incidence of gross proteinuria is similar among younger-onset group (36.3%), older-onset group taking insulin (42.4%), and older-onset group not taking insulin (43.5%) whose glycated hemoglobin was in the third and fourth quartile (P = 0.47). However, the incidence of renal failure is substantially lower in the older-onset groups with high glycated hemoglobin compared with the younger-onset group, suggesting that factors other than the level of glycemia are important in the progression to end-stage renal disease in diabetic persons. These differences may also be caused, in part, by the higher risk for death in both older-onset groups.
To evaluate the relative influence of several variables on the 10-year incidence of gross proteinuria, we developed models based on logistic regression. Independent variables included in the models were sex, systolic or diastolic blood pressure, antihypertensive medications, other medications, total pack-years smoked, retinopathy level, and the glycated hemoglobin level at baseline and the change in the glycated hemoglobin level from baseline to the 4-year follow-up examination. Using these models, we estimated that a 2 percentage point difference in the glycated hemoglobin level from baseline to the 4-year follow-up would be expected to lead to a 29% decrease in the 10-year incidence of gross proteinuria in younger-onset and a 19% and 0% decrease in older-onset persons taking and not taking insulin, respectively.
Relation of Glycated Hemoglobin Level to the 10-Year Incidence of History of Symptoms of Diabetic Neuropathy in the WESDR
The 10-year incidence of loss of tactile sensation or temperature sensitivity is shown in Table 2. There was a trend for increasing 10-year incidence of these symptoms with increasing glycated hemoglobin level at baseline in these groups (Figure 4). To evaluate the relative influence of several variables on the 10-year incidence of loss of tactile sensation or temperature sensitivity, we developed models based on logistic regression. Independent variables included in the models were age, duration of diabetes, sex, and the glycated hemoglobin level at baseline and the change in the glycated hemoglobin level from baseline to the 4-year follow-up examination. Using these models, we estimated that a 2 percentage point difference in the glycated hemoglobin level from baseline to the 4-year follow-up would be expected to lead to a 19% decrease in the 10-year incidence of loss of tactile sensation in younger-onset persons, a 23% decrease in older-onset persons taking insulin, and a 9% decrease in older-onset persons not taking insulin as well as a 16% decrease in the incidence of self-reported loss of temperature sensitivity in younger- and older-onset persons taking insulin and a 29% decrease in older-onset persons not taking insulin (Table 1). These findings in younger-onset persons are consistent with findings from the DCCT [25].
Discussion
The WESDR data suggest that poor glycemic control is associated with increased risk for incidence and progression of diabetic microvascular complications, independent of the type of diabetes. Other epidemiologic data in persons with IDDM or NIDDM are consistent with these findings [25, 32-43]. However, these data do not permit extrapolation of intensive treatment with insulin for the control of hyperglycemia to prevent microvascular complications in IDDM, as shown in the DCCT, to people with NIDDM. One must exercise caution in such extrapolation because epidemiologic data do not provide estimates of the risks for and benefits of reduction of hyperglycemia in persons with NIDDM. Only clinical trials, some of which are under way, can provide such estimates. Persons with IDDM studied in the DCCT were young, nonobese, and normotensive, with normal lipid profiles and no history of cardiovascular disease at baseline; they are very different from persons with NIDDM who are more likely to be older, obese, hypertensive, and dyslipidemic and have a history of cardiovascular disease. Thus, until results of such trials in persons with NIDDM are available, attempts at glycemic control should be made cautiously, especially in patients at high risk (see “United Kingdom Prospective Diabetes Study 17: A 9-Year Update of a Randomized, Controlled Trial on the Effect of Improved Metabolic Control on Complications in NIDDM” and “The Feasibility of Intensive Insulin Management in NIDDM: The Implications of the Veterans Affairs Cooperative Study on Glycemic Control and Complications in NIDDM”).
Few population-based incidence data exist using objective measurements of diabetic neuropathy. For persons with IDDM, the Epidemiology of Diabetes Complications Study [43] reported a 15% incidence of distal symmetrical polyneuropathy, as measured using the DCCT protocol, and a statistically significant relation with the baseline glycated hemoglobin level. In the DCCT, intensive insulin therapy reduced the occurrence of clinical neuropathy by 60% compared with conventional insulin therapy [25]. Data from ongoing population-based studies [23] and clinical trials such as the United Kingdom Prospective Diabetes Study [44] should provide insights into the relation of glycemic control and incidence of neuropathy in persons with NIDDM.
In the absence of achieving complete normoglycemia, it is important to control other potential risk factors that may be involved in the pathogenesis of diabetic microvascular complications in persons with IDDM or NIDDM, for example, high blood pressure. Other potential modifiable risk factors may also result in reducing the risk for chronic microvascular complications [43].
In summary, epidemiologic data suggest a similar course and prevalence of diabetic microvascular complications in persons with either IDDM or NIDDM and a strong relation to glycemic control. However, epidemiologic data do not provide information about the risk/benefit ratios associated with glycemic control, which are necessary in translating results from clinical trials involving persons with IDDM.
- Copyright ©2004 by the American College of Physicians
