Chapter 45 The Epidemiology of Diabetic Retinopathy
Prevalence of diabetic retinopathy
Population-based studies such as the Wisconsin Epidemiologic Study of Diabetic Retinopathy (WESDR)1–3 that use stereoscopic fundus photographs of seven standard photographic fields and objective grading by standard protocols have provided precise estimates of the prevalence and severity of diabetic retinopathy. In 1980–82, the WESDR showed that 71%, 23%, and 11% of those with type 1 diabetes (insulin-dependent diabetes mellitus, IDDM) and 47%, 6%, and 8% of those with type 2 diabetes (noninsulin-dependent diabetes mellitus, NIDDM) had retinopathy, proliferative retinopathy, and macular edema, respectively.3,4 These prevalence estimates, derived from data collected approximately 30 years ago in an 11-county area of southern Wisconsin (99% white), are higher than more recent prevalence data reported in other population-based studies (Table 45.1, Figs 45.1, 45.2).
A recent effort to provide more up-to-date estimates of prevalence using pooled data from eight studies including the WESDR5 included 615 individuals who were black and 1415 who were Hispanic. The prevalence estimates were limited to persons 40 years of age and older. The estimates of retinopathy were higher in the WESDR group compared to the seven other studies, all of which were performed at least 10 years after the WESDR (Figs 45.1, 45.2). Based on pooled analyses from these studies, it was estimated that among persons with diabetes, the crude prevalence of diabetic retinopathy was 40% and the crude prevalence of severe vision-threatening retinopathy (pre-proliferative and proliferative retinopathy or macular edema) was 8%. Projection of these rates to the diabetic population 40 years of age or older in the United States resulted in an estimate of 4 million persons with retinopathy, of whom 900 000 have signs of vision-threatening retinopathy. Based on grading of fundus images in the 2005–8 National Health and Nutrition Examination Survey (NHANES), 4.2 million people 40 years of age or older with diabetes were estimated to have diabetic retinopathy, of whom 650 000 had signs of vision-threatening retinopathy. With expectations that diabetes will continue to become more prevalent, without a significant decline in the incidence of diabetic retinopathy, the actual number of persons with vision-threatening retinopathy is likely to continue to increase.
The lower prevalence of diabetic retinopathy in more recent studies is thought to be due, in part, to changes in the management of diabetes.6–46 In persons with type 1 diabetes in the WESDR, there have been dramatic changes in management that include an increase in the use of blood glucose self-monitoring (from 72% in 1984–6 to 91% in 2005–7) and a greater frequency of using three or more injections of insulin per day (from 4% in 1980–2 to 85% in 2005–7). In the WESDR, this was associated with a 25% drop in the mean glycosylated hemoglobin A1c (A1c) from 10.1% to 7.6% and a 29% increase in those achieving American Diabetes Association (ADA) guidelines of A1c of <7% from 4% to 33% over the same period.47
There have also been changes in the management of glycemia in people with type 2 diabetes. In 1988–94, the use of only one oral hypoglycemic agent was the primary treatment to manage hyperglycemia in people with type 2 diabetes. After the findings from the United Kingdom Prospective Diabetes Study (UKPDS), there was an increase in the use of more than one oral hypoglycemic agent over a 5-year period (1999–2004).48,49 This was associated with a decrease in the mean A1c levels from 7.8% to 7.2%, with a 41% increase (from 41% to 58%) in persons achieving A1c levels of <7.0% in the periods 1999–2000 and 2005–6.
Incidence and progression of diabetic retinopathy and incidence of clinically significant macular edema
There are fewer reports of incidence of retinopathy in population-based studies.8,21,50–63 The incidence of retinopathy in a 4-year interval in the entire WESDR population was 40.3%.50,51 The 4-year incidence and rates of progression of diabetic retinopathy and macular edema in the WESDR are presented in Table 45.2. Those with type 1 diabetes had a higher incidence of any retinopathy, progression, and progression to proliferative retinopathy than those with type 2 diabetes (Table 45.2).52 The highest 4-year incidence of clinically significant macular edema was in those with type 2 diabetes taking insulin, while the lowest was in those with type 2 diabetes not taking insulin. While the incidence of proliferative retinopathy was higher in those with type 1 diabetes, the estimates of the number of incident cases in the 4-year period were higher in the group with type 2 diabetes than in the group with type 1 diabetes (120 vs 83 persons) due to the higher frequency of people with type 2 diabetes.
Table 45.2 Four-year incidences of any retinopathy, improvement or progression of retinopathy, progression to proliferative diabetic retinopathy (PDR) and incidence of clinically significant macular edema (CSME) in younger-onset type 1 diabetes and older-onset type 2 diabetes in the Wisconsin Epidemiologic Study of Diabetic Retinopathy, 1980–1986
There is also evidence that the prevalence and incidence of diabetic retinopathy may be decreasing in subjects more recently diagnosed with type 1 diabetes. Hovind et al.64 first showed a declining incidence of proliferative diabetic retinopathy and macular edema in a study of 600 patients with type 1 diabetes diagnosed between 1965 and 1984 in Denmark. In that study, the cumulative incidence of proliferative diabetic retinopathy and macular edema after 20 years of diabetes declined from 31% and 19%, respectively, in those diagnosed from 1965 to 1969, to 13% and 7%, respectively, in those diagnosed from 1979 to 1984. There was also significant improvement in visual acuity and lower prevalence of severe visual impairment in those diagnosed with type 1 diabetes more recently than those diagnosed in earlier periods. These changes were attributed by the authors to improved glycemic control, more aggressive treatment of blood pressure sooner after diagnosis of diabetes, and reduced smoking rates in the more recently diagnosed type 1 diabetic group than in previous years. There was also a decline in the cumulative proportion with severe laser-treated diabetic retinopathy after 25 years of type 1 diabetes from 47% in subjects diagnosed in 1961–5 to 24% in subjects diagnosed in 1971–5 in the Swedish Linköping Diabetes Complications Study.65,66 However, the Pittsburgh Epidemiology of Diabetic Complications Study did not show a significant decrease in proliferative diabetic retinopathy in those diagnosed more recently.67 In the WESDR, the annualized estimates for the progression of diabetic retinopathy (4.5 vs 2.5%) and the incidence of proliferative diabetic retinopathy (3.4 vs 1.5%), clinically significant macular edema (1.0 vs 0.4%), and visual impairment (0.7 vs 0.3%) were higher in the first 12 years of the study (1980–92) than in the latest 13 years of the study (1994–2007).68–71 While controlling for duration of diabetes, there was also evidence in the WESDR of lower prevalence of proliferative diabetic retinopathy (4% lower per more recent time period) and visual impairment (9% lower per more recent time period) but not of macular edema in those diagnosed with type 1 diabetes more recently than those diagnosed longer ago. The relationships remained when adjusting for hypertension and A1c levels over time.
The relationship of race/ethnicity to diabetic retinopathy
In contrast to whites, there are fewer epidemiological studies regarding the prevalence and incidence of diabetic retinopathy in other racial/ethnic groups in the United States, especially in persons with type 1 diabetes. Data from the New Jersey 725 study cohort, which used similar methods to detect and classify retinopathy severity as in the WESDR cohort, showed a similar frequency and severity of retinopathy in African Americans with type 1 diabetes as found in whites with type 1 diabetes in the WESDR.45,46 At the 6-year follow-up of the same cohort, 56% showed progression of diabetic retinopathy, 15% showed progression to proliferative diabetic retinopathy, and 16% developed macular edema.62 These findings were similar to those in whites in the WESDR (52).
In four population-based studies, the NHANES 1988–94 and 2005–8,36 the Atherosclerosis Risk in Communities (ARIC) study,72 the Cardiovascular Health Study,73 and the Multi-Ethnic Study of Atherosclerosis (MESA),74 retinopathy was more prevalent in African Americans with type 2 diabetes than in whites. In the NHANES III in 1988–94, compared to whites, African Americans had a higher frequency of people with poor glycemic control (A1c greater than 8.3%, 37% vs 30%), high systolic blood pressure (>142 mmHg, 42% vs 32%), longer duration of diabetes (>14 years, 29% vs 23%), and on insulin therapy (43% vs 24%). There was no difference (odds ratio [OR] 0.94; 95% confidence interval [CI] 0.54–1.66) in the prevalence of retinopathy between African Americans and whites while controlling for these factors (36). In addition, there were no statistically significant interactions of race with diabetes severity variables or systolic blood pressure, suggesting that the effect of risk factors was similar in both racial/ethnic groups. Similarly, the higher prevalence of retinopathy in the ARIC study (28% vs 17%) and in the MESA (37% vs 25%) in blacks as compared to whites was no longer statistically significant after controlling for differences in glycemic and blood pressure control between the races. Higher prevalence of retinopathy in African Americans with type 2 diabetes appears to be partially due to poorer glycemic and blood pressure control. These data suggest that programs designed to better control blood sugar and blood pressure in diabetic African Americans might be beneficial.
In most population-based studies, Mexican Americans have been shown to have higher frequencies and more severe diabetic retinopathy than non-Hispanic whites.5,22,31,36,40,74–76 Haffner et al.22 found that after controlling for all measured risk factors, the frequency of retinopathy in Mexican Americans in San Antonio was 2.4 times as high as the frequency of retinopathy in non-Hispanic whites studied in the WESDR. Similarly, in the NHANES 1988–94 and 2005–8, the MESA, Proyecto VER, and the Los Angeles Latino Eye Study (LALES), retinopathy was more frequent in Mexican Americans compared to non-Hispanic whites 40 years of age or older.31,36,74,76 In the NHANES 1988–94, retinopathy was more prevalent in Mexican Americans (OR 2.15; 95% CI 1.15–4.04) compared to non-Hispanic whites, even while controlling for duration of diabetes, A1c level, blood pressure, and type of antihyperglycemic medication used.5 In the NHANES 2005–8, vision-threatening retinopathy was approximately 3.5 times (95% CI 1.05–12.56) as frequent in Mexican Americans compared to non-Hispanic whites.75 These variations in prevalence among ethnic groups may be a result of differences in how long it takes to diagnose diabetes after its onset, how it was defined, and levels of glycemia and blood pressure. Differences among Hispanic whites may be due to the degree of gene-sharing with Native Americans, a group with a high prevalence of retinopathy (see below).
Among population-based studies, only the LALES has provided data on the incidence and progression of diabetic retinopathy in Mexican Americans with type 2 diabetes.63 The 4-year incidence of diabetic retinopathy and clinically significant macular edema was 34% and 7%, respectively, and progression of retinopathy and progression from nonproliferative diabetic retinopathy to proliferative diabetic retinopathy was 39% and 5%, respectively, over the 4-year period. While these rates are comparable to those found in the WESDR, they are higher than in most other contemporaneous studies of whites with type 2 diabetes.
The prevalence and severity of retinopathy appears to vary among different Native American groups.41,77–80 In studies done in the 1970s, Native Americans were reported to have higher rates of severe retinopathy for a given duration of type 2 diabetes compared to whites.9,10 However, data from more recent studies on the incidence and progression of diabetic retinopathy in Pima Indians show a lower 4-year cumulative incidence and progression of diabetic retinopathy (17% and 18%, respectively) than reported in whites with type 2 diabetes, reflecting possible improvements in glycemic and blood pressure control.81
There are few data on the prevalence of retinopathy in Asian Americans and other racial/ethnic groups.27,37,43,74 The prevalence of retinopathy in second generation (Nisei) Japanese American men, 12%, was significantly lower than that reported in the diabetes clinic at Tokyo University Hospital (49% among patients with onset of diabetes at 20–59 years of age and 47% among those with onset after 59 years of age) and in whites reported in the WESDR (36%).3,27 In the MESA, the prevalence of any retinopathy (26% vs 25%) in Chinese Americans was similar to whites.74 However, clinically significant macular edema and proliferative diabetic retinopathy was higher (13% vs 2%) in Chinese than in whites. More data on the prevalence and incidence of retinopathy in Chinese and other Asian American groups are needed.
Genetic factors
Data from a number of studies that examined familial clustering suggested that genetic factors may be involved more strongly in the susceptibility to diabetic retinopathy than previously thought.82,83 In addition, data showing that the time of appearance of retinopathy and its severity are more likely to be similar among diabetic identical twins than dizygotic twins suggested that the tendency to develop diabetic retinopathy, and possibly its progression, are influenced by genetic factors. However, unlike the strong associations of complement factor H and other single nucleotide polymorphisms (SNPs) that have been found to be related to age-related macular degeneration, the putative genes and genetic variants have not been found to be as strongly or consistently associated with diabetic retinopathy (see Chapter 46, Diabetic retinopathy: Genetics and etiologic mechanisms). This may be a result of the stronger environmental influence of glycemic and blood pressure control than found for age-related macular degeneration. The fact that retinopathy is not specific to diabetes in its earliest stages may also contribute to inability to find and replicate genes associated with diabetic retinopathy.
Study of specific genetic factors associated with the hypothesized pathogenetic factors for retinopathy, such as aldose reductase activity, collagen formation, inflammatory processes, protein kinase activity, glycation, oxidative stress and platelet adhesiveness and aggregation may yield a better understanding of the possible causal relationships between genetic factors and diabetic retinopathy. There are already a number of studies that have reported associations between retinopathy and mitochondrial DNA mutations84 and polymorphisms of the aldose reductase gene,85,86 TNF-beta NcoI gene,87 epsilon4 allele of apolipoprotein E gene,88 paraoxonase (an enzyme that prevents oxidation of low-density lipoprotein cholesterol) gene,89 endothelial nitric oxide synthase gene,90 intercellular adhesion molecule-1 (ICAM-1),91 alpha2beta1 integrin gene (involved with platelet function),92 cytokine vascular endothelial growth factor (VEGF) gene, and many others.93,94 The reader is referred to a more comprehensive, in-depth discussion of the rapidly evolving field of genetic epidemiology of diabetic retinopathy in Chapter 46.
Sex
In the WESDR, higher frequencies of proliferative retinopathy were present in younger-onset men compared to women.2 However, there were no significant differences in the 4-, 10-, or 14-year incidence or progression of diabetic retinopathy between the sexes.50,53,58 There were no significant differences in the prevalence or 10-year incidence of retinopathy or rates of progression to proliferative retinopathy between the sexes in people with type 2 diabetes in the WESDR.3,51,53
Age and puberty
The prevalence and severity of diabetic retinopathy increased with increasing age in persons with type 1 diabetes in the WESDR.2 In persons under 13 years of age, diabetic retinopathy was infrequent, irrespective of the duration of diabetes. The 4-year incidence of retinopathy increased with increasing age, with the sharpest increase occurring in persons who were 10–12 years of age at baseline.50 Four-year rates of progression of retinopathy in younger-onset persons rose steadily with increasing age until 15–19 years of age, after which there was a gradual decline. No child younger than 13 years of age at baseline in the WESDR was found to have proliferative retinopathy at the 4-year follow-up. These findings have formed the rationale for guidelines for not screening for retinopathy in children with type 1 diabetes.95
In the WESDR, menarchal status, a crude marker of puberty, at the time of the baseline examination was related to the prevalence and severity of retinopathy.96 While controlling for other risk factors, those who were postmenarchal were three times as likely to have retinopathy as those who were premenarchal. In a follow-up study of 60 children with type 1 diabetes, Frost-Larsen and Starup97 found the incidence of retinopathy to be higher after puberty than before, independent of duration or metabolic control of diabetes or type of treatment. These findings have been observed in other studies.98,99 Increases in growth hormone, insulin-like growth factor I, sex hormones, and blood pressure as well as poorer glycemic control (due to increased insulin resistance, poorer compliance, and/or inadequate insulin dosage) have been hypothesized to explain the higher risk of developing retinopathy after puberty.
In older-onset persons taking insulin in the WESDR, the 4-year incidence of retinopathy and progression of retinopathy had a tendency to decrease with age.51 The 4-year incidence of improvement tended to increase with age. For those not taking insulin, the 4-year rate of progression to proliferative retinopathy decreased with age. Few persons 75 years of age or older with type 2 diabetes developed proliferative retinopathy over the 10 years of follow-up. These findings are consistent with data from other population-based studies.8,21 In one such study of people with type 2 diabetes in Rochester, Minnesota, Ballard et al.14 reported a lower incidence of retinopathy with increasing age in persons with diabetes older than 60 years of age. These findings might reflect a less severe disease in those with older-onset or selective survival, that is, older persons who develop severe retinopathy are at higher risk of dying and not being seen at follow-up in these studies.
Duration of diabetes
Perhaps the most consistent relationship found in persons with diabetes is the increase in the frequency and severity of diabetic retinopathy and macular edema with increasing duration of diabetes.2 The prevalence of retinopathy 3–4 years after diagnosis of diabetes in the WESDR younger-onset group with type 1 diabetes was 14% in men and 24% in women. However, in persons who had had diabetes for 19–20 years, 50% of men and 33% of women had proliferative retinopathy. Shortly after diagnosis of diabetes, retinopathy was more frequent in persons with type 2 diabetes compared with those with type 1 diabetes (Figs 45.3, 45.4).3 In the first 3 years after diagnosis of diabetes, 23% of the type 2 diabetic group not taking insulin had retinopathy, and 2% had proliferative retinopathy (PDR).
Based on recent follow-up of the WESDR cohort, the prevalence estimates for a given duration likely overestimate the actual prevalence now found in the population.68 For a specific duration of type 1 diabetes, people diagnosed between 1975 and 1980 had a statistically significantly lower prevalence than persons diagnosed in earlier periods (P < 0.001). This difference remained while controlling for A1c, systolic and diastolic blood pressure, and presence of proteinuria. Similarly, for specific duration of type 2 diabetes, those diagnosed more recently had a lower prevalence of diabetic retinopathy than those diagnosed in earlier periods.
Harris et al.,100 using retinopathy prevalence data at different durations of diabetes from persons with type 2 diabetes in the WESDR and from a study in Australia, extrapolated to the time when retinopathy prevalence was estimated to be zero. They estimated that the onset of detectable retinopathy occurred approximately 4–7 years after diagnosis of type 2 diabetes in these populations.
In the WESDR, the 4- and 10-year incidence of diabetic retinopathy increased with increasing duration of diabetes at baseline.50,51,53 The risk of developing retinopathy in the younger-onset group was high (74%) after 10 years of diabetes. The 4-year incidence of proliferative retinopathy varied from 0% during the first 3 years after diagnosis of diabetes to 28% in those with 13–14 years of diabetes. Thereafter, the incidence remained stable.50 A similar trend was found in a cohort of patients with type 1 diabetes followed at the Joslin Clinic.101 In the older-onset WESDR group, 2% of those with less than 5 years and 5% of those with 15 or more years of diabetes who were not taking insulin at baseline developed signs of proliferative retinopathy at the 4-year follow-up.51
Age at diagnosis
Age at diagnosis was not related to incidence or progression of diabetic retinopathy in any of the diabetes groups followed in the WESDR.50,51 In contrast, while controlling for other risk factors, in a cohort with type 2 diabetes in Rochester, Minnesota, the development of retinopathy was significantly associated with younger age at diagnosis.14
Glycemia
In 1978, in his textbook on the epidemiology of diabetes and its complications, Kelly West wrote: “The extent to which hyperglycemia determines the risk of retinopathy is not at all clear. This is the most important issue at hand and deserves high priority in epidemiologic research.”102 Thirty years later, this issue has largely been resolved by epidemiologic studies and clinical trials.
The Diabetes Control and Complications Trial (DCCT) was “designed to compare intensive with conventional diabetes therapy with regard to their effects on the development and progression of the early vascular and neurologic complications of IDDM.”103 Two of the main questions asked in the study were: “Will intensive therapy prevent the development of diabetic retinopathy in patients with no retinopathy (primary prevention)?” and “Will intensive therapy affect the progression of early retinopathy (secondary intervention)?” In addition, the DCCT examined the magnitude of the effect of intensive insulin treatment on progression of retinopathy, the degree to which this effect changes over time, and the relation of the effect to the level of severity of the retinopathy at baseline.104–106
Randomization was used to assign conventional or intensive insulin therapy.103 Conventional therapy consisted of one or two daily injections of insulin per day, daily self-monitoring of urine or blood glucose, and education about exercise and diet. No attempts were made to adjust the insulin dosage on a daily basis. Intensive therapy consisted of administration of insulin three or more times daily by injections or an external pump. In addition, there was adjustment of the insulin dosage under the direction of an expert team, taking into account self-monitoring of blood glucose performed four times per day, dietary intake, and anticipated exercise.104
An important finding of the trial was the statistically significant reduction in risk of sustained progression of retinopathy by three or more steps by 76% (Table 45.3, Fig. 45.5). In the secondary-intervention cohort, the intensive therapy group had a reduction of average risk of progression by 54% during the entire study period compared to the patients assigned to the conventional-therapy group. In addition, when both cohorts were combined, the intensive therapy group also had a reduction in risk for development of severe nonproliferative retinopathy or proliferative retinopathy by 47% and of treatment with photocoagulation by 51% (Table 45.3). There was a decrease in the incidence of clinically significant macular edema in the group assigned to intensive therapy compared to those assigned to conventional therapy. However, this difference did not reach statistical significance.
Table 45.3 Development and progression of long-term complications of diabetes in the study cohorts and reduction in risk with intensive as compared with conventional therapy*
Early worsening of retinopathy in the first year of treatment of the intensive therapy group in the secondary-intervention cohort was observed, as had been reported previously.107–109 On average, it took about 3 years to demonstrate the beneficial effect of intensive treatment. After 3 years, the beneficial effect of intensive insulin treatment increased over time.
The DCCT investigators also examined whether there was an association of A1c values <8% versus those of ≥8% for progression of retinopathy. When they combined the two groups (conventional and intensive treatment), they found no evidence to support the concept of a glycemic threshold regarding progression of retinopathy, as had been described by others.110
Intensive insulin treatment reduced but did not prevent the incidence and progression of retinopathy in persons without signs of retinopathy at the baseline examination. The 9-year cumulative incidence of one microaneurysm or more severe retinopathy in eyes with no retinopathy present at baseline was 70% in persons with <2.5 years of type 1 diabetes and 62% in persons with >2.5 years of type 1 diabetes at baseline. Approximately 40% of these individuals developed a three-step progression of their retinopathy.105
The DCCT examined whether intensive therapy was more beneficial when started earlier in the course of type 1 diabetes. They found that the 9-year cumulative incidence of sustained three-step progression in persons without retinopathy with <2.5 years of type 1 diabetes in the intensive therapy group was 7% compared to 20% in those with >2.5 years. The 9-year cumulative incidence of sustained three-step progression in the intensive therapy group was lower in eyes with minimal to early nonproliferative retinopathy at baseline compared to eyes with more severe nonproliferative retinopathy at baseline (11.5–18.2% vs 43.8%). These data suggested a benefit of beginning intensive treatment earlier in the course of diabetes, prior to the onset of diabetic retinopathy.105
From the trial, it was estimated that intensive therapy would result in a “gain of 920,000 years of sight, 691,000 years free from end-stage renal disease, 678,000 years free from lower extremity amputation, and 611,000 years of life at an additional cost of $4.0 billion over the lifetime” of the 120 000 persons with IDDM in the United States who meet DCCT eligibility criteria.111 The incremental cost per year of life gained was $28 661, and when adjusted for quality of life, intensive therapy costs $19 987 per quality-of-life year gained. These findings were similar to cost-effectiveness ratios for other medical interventions in the United States.
Fourteen years of additional follow-up of the DCCT cohort after the study was stopped revealed that despite convergence of A1c levels in the intensive and conventional groups, the protective effect of glycemic control was maintained in the intensive group.112–114 This has been labeled “metabolic memory” and has been found also in persons with type 2 diabetes in the UKPDS (see below).115 The reason for this finding remains speculative. Recent data suggest that biochemical pathways involving advanced glycation endproducts and oxidative stress may affect genes and proteins involved in the pathogenesis of diabetic microvascular and macrovascular complications.114
The UK Prospective Diabetes Study (UKPDS) was a randomized controlled clinical trial involving 3867 patients newly diagnosed with type 2 diabetes.116–118 After 3 months of diet treatment, patients with a mean of two fasting plasma glucose concentrations of 6.1–15.0 mmol/L were randomly assigned to intensive glycemic control with either a sulfonylurea or insulin or conventional glycemic control. The latter group was further divided into those who were overweight or not. Metformin was included as one of the treatment arms for 1704 overweight patients, and analyses included comparison of the effect of metformin against conventional therapy in overweight patients. After 12 years of follow-up, there was a reduction in rate of progression of diabetic retinopathy of 21% and reduction in need for laser photocoagulation of 29% in the intensive versus the conventional treatment group. In addition, there were no differences in reduction in the incidence of the retinopathy endpoints among the three agents used in the intensive treatment group (chlorpropamide, glibenclamide, and insulin) but the chlorpropamide treatment group failed to show a reduced rate of retinopathy requiring photocoagulation. Furthermore, there was no difference in vision outcomes between conventional and intensive treatments. It was concluded that metformin was preferred as the first-line pharmacological therapy in newly diagnosed type 2 diabetic patients who were overweight, based on their finding of a significant (39%) reduction in myocardial infarction compared to the conventional treatment group. When metformin was added to sulfonlyureas (in both obese and nonobese patients), however, it was associated with increased diabetes-related (96%) and all-cause mortality (60%) when compared to conventional therapy. The intensive treatment group suffered significantly more major hypoglycemic episodes and weight gain than patients in the conventional group. Economic analyses of the clinical trial data suggested that intensive glucose control increased treatment costs but substantially reduced complication costs and increased the time free of such complications.118
The development of new treatment modalities for achieving glycemic control has resulted in two recently completed randomized clinical trials that permitted evaluation of near normalization of glycemic level on the incidence of cardiovascular disease and retinopathy. The first trial involved 1791 military veterans with an average age of 60 years and an average duration of 11 years of type 2 diabetes, who had a suboptimal response to therapy for their diabetes. They were randomly assigned to receive either intensive or standard glucose control, with an aim in the intensive therapy group of achieving an absolute reduction of 1.5 percentage points in the A1c level as compared with the standard therapy group. The primary outcome was the time to the first occurrence of a major cardiovascular disease event, and a secondary objective was to evaluate the effect of glycemic control on the incidence and progression of diabetic retinopathy and other microvascular complications.119,120 The subjects were followed for up to 7.5 years (median: 5.6 years). Despite reaching their glycemic goal (median A1c level at 6 months: 8.4% in the group receiving standard therapy and 6.9% in the intensive therapy group), there were no statistically significant differences in any of the retinopathy outcomes between groups receiving intensive and standard therapy (incidence of retinopathy 42% vs 49%, P = 0.27; progression of retinopathy by two or more steps on the Early Treatment Diabetic Retinopathy Study [ETDRS] severity scale 17% vs 22%, P = 0.07; progression to proliferative diabetic retinopathy 4% vs 5%, P = 0.27) or in progression to clinically significant macular edema (3% vs 5%, P = 0.31). While it is possible that a benefit might have been seen if the study was continued, these data lead to the conclusion that decreasing the A1c level from 8.4% to 6.9% in persons with relatively long-standing type 2 diabetes has little benefit in preventing the incidence and progression of retinopathy.
Another recently concluded large randomized controlled clinical trial is Action to Control Cardiovascular Risk in Diabetes (ACCORD), which examined whether intensive treatment with an even lower targeted A1c level (<6.0%) than in the military veteran study versus standard treatment (targeted A1c level 7.0–7.9%) would reduce the risk of morbidity and mortality from cardiovascular disease (primary endpoint) and microvascular events, such as the incidence of photocoagulation treatment for diabetic retinopathy and incidence of microalbuminuria and macroalbuminuria over a 5-year period (secondary endpoints) in persons with a mean age of 60 years with an average duration of 10 years of type 2 diabetes.121,122 They reported findings on the same composite microvascular endpoints measured in the UKPDS. They also reported the incidence and progression of diabetic retinopathy on the basis of the grading of fundus photographs in a sample of 4065 of the 10 251 participants. They did not find a statistically significant difference in their composite microvascular endpoints (one combining a history of advanced kidney and eye disease, and the other adding peripheral neuropathy to that outcome) or for some of the other specified ocular or renal outcomes. In the eye study, using the grading of fundus photographs to assess intensive glycemic control, they reported a 33% reduction in the relative risk of progression from 7.3% with intensive glycemia treatment, versus 10.4% with standard therapy (adjusted OR 0.67; 95% CI 0.51–0.87; P
