Optimal glycemic control is fundamental to the management of diabetes. In epidemiological analyses, glycated hemoglobin (A1C) levels >7.0% are associated with a significantly increased risk of both microvascular and macrovascular complications, regardless of underlying treatment (1–3). Data from the Diabetes Control and Complications Trial (DCCT; type 1 diabetes) (2) and the United Kingdom Prospective Diabetes Study (UKPDS; type 2 diabetes) (3) demonstrated a continuous relationship between A1C and diabetes complications, with no apparent threshold of benefit. In the DCCT, a 10% reduction in A1C was associated with a 40% to 50% lower risk of retinopathy progression, although the absolute reduction in risk was substantially less at lower A1C levels (2). In the UKPDS, this relationship was directly linear, with each 1.0% (absolute) reduction in mean A1C associated with a 37% decline in the risk of microvascular complications, a 14% lower rate of myocardial infarction (MI) and a 21% reduction in deaths from diabetes (3).
Both fasting plasma glucose (FPG) and postprandial plasma glucose (PPG) are directly correlated to the risk of complications, with some evidence that postprandial might constitute a stronger risk factor for cardiovascular (CV) complications (4–9). In a meta-analysis of 102 prospective studies, FPG >5.6 mmol/L was associated with an increased risk of CV events (10). Postprandial hyperglycemia and the 2-hour post-challenge PG appears to be a better predictor of cardiovascular disease (CVD) and all-cause mortality than FPG (7). This association between CVD and 2-hour postprandial PG appears to be linear (6,7). Values >7.8 mmol/L are associated with an increase in all-cause mortality (8), and values >10.0 mmol/L are associated with both microvascular complications (11) and the highest risk of MI (12).
There is compelling evidence from randomized controlled studies that improved glycemic control reduces the risk of microvascular complications but has no significant effect on macrovascular outcomes in recently diagnosed type 1 (13) and type 2 diabetes (1,11,14), as well as more long-standing type 2 diabetes (15–19). The initial prospective randomized controlled trials were conducted in patients with recently diagnosed diabetes. These trials—the DCCT in type 1 diabetes (13) and the Kumamoto (11) and the UKPDS (1,14) in type 2 diabetes—confirmed that improved glycemic control significantly reduced the risk of microvascular complications but had no significant effect on macrovascular (particularly CV) outcomes. Subsequent observational data from long-term follow-up of both the DCCT and UKPDS cohorts showed a persistence of significant microvascular benefits in patients who had previously been in the intensively treated groups despite the fact that, during the subsequent follow-up period, their glycemic control became similar to that of patients who were previously in the standard arm (20–22). The follow-up data from these 2 studies also demonstrated a beneficial effect of improved glycemic control on CV outcomes. In the DCCT cohort, there was a significant reduction in CV outcomes (42%) as well as non-fatal MI, stroke and CV death (57%) in previously intensively treated patients compared with those who were previously in the standard arm (23). Similarly, there was a significant reduction in MI (15%–33%) and all-cause mortality (13%–27%) in the UKPDS cohort in patients who had been originally randomized to intensive treatment (22).
Three major trials—the Action to Control Cardiovascular Risk in Diabetes (ACCORD), Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation (ADVANCE), and Veterans Affairs Diabetes Trial (VADT)—looked at the effect of intensive glycemic control on patients with long-standing type 2 diabetes. The ACCORD trial randomly assigned 10 251 patients to intensive therapy targeting an A1C <6.0% or standard therapy targeting an AIC level of 7.0% to 7.9% (17,24). Patients included had either a previous history of CVD or multiple risk factors for CVD, and a baseline A1C level ≥7.5%. At inclusion, participants had a mean age of 62 years, diabetes duration of 10 years and a median baseline A1C level of 8.1%. A difference in A1C was rapidly obtained and maintained throughout the trial at 6.4% and 7.5% in the intensive and standard therapy groups, respectively. The primary outcome of this study was a composite of major CV events: non-fatal MI, nonfatal stroke or death from CV causes. The intensive glucose control arm was prematurely terminated after 3.5 years due to higher mortality associated with assignment to this treatment (17,24).
The ADVANCE trial randomly assigned 11 140 patients to standard (targeting A1C based on local guidelines) or intensive glucose control therapy aimed at reducing A1C levels to ≤6.5% (15). Patients were at least 55 years old with a history of major macrovascular or microvascular disease or at least 1 other risk factor for vascular disease. Median baseline A1C level and diabetes duration were lower than in the ACCORD trial at 7.2% and 8 years, respectively, whereas mean age was slightly higher at 66 years. The difference in A1C in both arms was obtained less rapidly, and, after a 5-year follow-up, mean A1C was 6.5% in the intensive group and 7.3% in the standard group. The primary outcome in the ADVANCE trial was a composite of microvascular events (nephropathy and retinopathy) and macrovascular disease defined by major adverse CV events.
VADT randomly assigned 1791 United States military veterans with poor glycemic control (≥7.5%) to either standard or intensive glucose therapy, which aimed for an overall reduction in A1C levels by 1.5%(18,19). Following a median follow-up of 5.6 years, A1C levels were 8.4% and 6.9% in the standard and intensive therapy groups, respectively. The primary outcome of the study was the time from randomization to the first occurrence of a major CV event(18,19).
These 3 trials confirmed the benefit of intensive glycemic control on microvascular outcomes. In the VADT study, the progression to albuminuria was significantly reduced in the intensive-treatment patients, with 9.1% of patients having significantly reduced progression compared to 13.8% in the standard therapy group(19). Similarly, intensive therapy in ACCORD patients showed a favourable effect on microvascular outcomes, particularly albuminuria and diabetic retinopathy (16). In ADVANCE, patients in the intensive control group demonstrated a reduction in the incidence of major microvascular events, mainly through a 21% relative reduction in nephropathy (15). A recent meta-analysis confirmed the positive impact of intensive glycemic control on microalbuminuria (25).
None of the above studies independently confirmed a significant benefit of tight glycemic control on macrovascular outcomes. However, meta-analysis of clinical trials designed to assess differences in CV outcomes in patients who had achieved lower versus higher levels of glycemia demonstrated that those treated with more intensive therapy, compared to less intensive glycemic control, were found to have a 10% to 15% reduction in the risk of major CV events, primarily because of a 15% reduced risk of MI, but with no effect on stroke, CV death or all cause mortality (26). Intensive glycemic control, however, was associated with more than a 2-fold increase in the risk of severe hypoglycemia (25).
The unexpected higher mortality rates seen in the intensive arm of the ACCORD study and the lack of clear macrovascular benefit in the ADVANCE and VADT trials have been further reviewed. Several potential reasons for these findings have been suggested, including patient age, duration of diabetes, presence of CVD, history of severe hypoglycemic events, weight gain and the rapid decrease in A1C values. Increased mortality associated with intensive treatment could not be explained by the type of pharmacological treatment, rapidity to implement the intensive strategy or weight gain (24). Hypothesis-generating secondary analysis from the ACCORD trial reported a nonsignificant trend toward lower all-cause mortality in individuals assigned to the standard arm who were younger than 65 years at baseline (27). Similarly, the ADVANCE trial also reported a nonsignificant trend toward fewer events among younger patients in the intensive therapy group (15). Duration of diabetes also may have played a role. Compared with the UKPDS and the DCCT, which were conducted in younger individuals with recent-onset diabetes, the duration of diabetes in the ACCORD, ADVANCE and VADT trials ranged from 8 to 11.5 years. Further emphasis of the importance of duration of diabetes was identified in a substudy of the VADT patients when measurement of the coronary calcium score, utilizing computed tomography, revealed fewer CVD events in these younger patients enrolled in the intensive treatment arm (28). The frequency of severe hypoglycemia in these trials was 2 to 3 times higher in the intensive therapy groups, and a higher mortality was reported in participants with 1 or more episodes of severe hypoglycemia in both the ACCORD (29) and the ADVANCE (30) trials, irrespective of the different treatment arms in which individual patients were allocated. However, these subanalyses confirmed that hypoglycemic events could not account for the difference in mortality between the intensive and standard therapy groups. Finally, in the ACCORD trial, mortality was increased in patients randomized in the intensive arm but who failed to reduce their A1C despite treatment intensification (31).
These findings suggest that microvascular and macrovascular events may be reduced by intensifying therapy targeting an A1C <7.0% in younger patients with recently diagnosed diabetes and a lower initial A1C value but with an increased risk of hypoglycemic risk. Individualized and higher A1C targets may be indicated in older type 2 patients with longer duration of diabetes, established CV risk factors, severe hypoglycemia episodes and/or without A1C reduction despite treatment intensification. Similarly, individualization of A1C targets may be needed in some patients with type 1 diabetes who are unable to achieve an A1C <7.0% without being at increased risk of severe hypoglycemia.
It also must be recognized that A1C measurement is a component of both the FPG and PPG. When A1C values are higher, the major contribution is the FPG levels, but as the A1C value approaches the target value of ≤7.0%, there is a greater contribution from PPG values (32,33). Another study using continuous glucose monitoring demonstrated that a 2-hour postprandial PG <8.0 mmol/L correlates best with A1C <7.0% (34). In view of this, if A1C targets cannot be achieved with a postprandial target of 5.0 to 10.0 mmol/L, further postprandial BG lowering to 5.0 to 8.0 mmol/L can be considered. The role of pre- vs. postprandial glucose control on reducing CV outcomes has been controversial (35,36).
A major difficulty in attempting to use evidence-based observations to determine the value of tighter postprandial glucose control has been the lack of well-designed, long-term outcome studies where assessing postprandial glucose values is the major objective of the study. Most of the large outcome trials conducted so far have been mostly based on preprandial glucose and A1C targets. Although there is evidence in type 2 diabetes that targeting postprandial hyperglycemia to <8.0 mmol/L reduces progression of carotid atherosclerosis (35) , a randomized controlled trial of type 2 diabetes patients treated with insulin therapy after acute MI showed no benefit of insulin regimen targeting postprandial hyperglycemia compared with the regimen targeting preprandial glucose (36).
Contrasting results from recent studies should not discourage physicians from controlling blood glucose levels. Intensive glucose control, lowering A1C values to ≤7% in both type 1 and type 2 diabetes, provides strong benefits for microvascular complications and, if achieved early in the disease, might also provide a significant macrovascular benefit, especially as part of a multifactorial treatment approach. More intensive glucose control, A1C ≤6.5%, may be sought in patients with a shorter duration of diabetes, no evidence of significant CVD and longer life expectancy, provided this does not result in a significant increase in hypoglycemia. An A1C target ≤8.5% may be more appropriate in type 1 and type 2 patients with limited life expectancy, higher level of functional dependency, a history of severe hypoglycemia, advanced comorbidities, and a failure to attain established glucose targets despite treatment intensification ( Figure 1).