|
|
||||
|
Initial results from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) and Veterans Affairs Diabetes Trial (VADT) studies aroused concerns as to whether intensive glucose control targeting low A1C levels confers increased cardiovascular risk in some patients. New analyses of these studies presented at the final day of the 2009 ADA sessions shed light on this issue. ACCORD The ACCORD analysis confirmed that low A1C levels did not cause the increased mortality rate seen among patients who received intensive glycemic control (targeting A1C <6%) compared with those in the standard-control group (A1C target, 7%–9%). The trial was halted about 1-1/2 years ahead of schedule in February 2008 due to a 20% increase in deaths in the intensive arm. Matthew Riddle, MD, presented results on A1C and mortality risk for the new analysis:
Denise Bonds, MD, presented data relative to hypoglycemia and mortality. Hypoglycemia was more common in the intensive group over all years; there was an annual increase of 2% to 3.5% of hypoglycemia with intensive treatment versus 1% with standard treatment. Subjects with an increased albumin:creatinine ratio (hazard ratio, [HR], 1.74), African-Americans (HR, 1.43), women (HR, 1.21), and those with BMI >30 kg/m2 (0.65) had higher rates of hypoglycemia. There was also increased risk for every 1-year difference in age (HR, 1.03). The most common antecedent to hypoglycemia was delayed or missed meal, or consumption of fewer carbohydrates (58%) There was a nearly threefold higher risk for death in the standard vs intensive therapy group. The HR (95% confidence interval) for association between hypoglycemia and mortality was 2.81 (1.73–4.76) for standard therapy versus 1.28 (0.88–1.85) with intensive therapy. When the data were adjudicated, there was a slightly higher rate of cardiovascular deaths among subjects with hypoglycemia. However, there was no meaningful difference between standard and intensive therapy. When an etiologic fraction was used to assess the data, the number of deaths potentially attributable to hypoglycemia was 9 in the intensive group and 11 in the standard group. A1C was associated with hypoglycemia; higher baseline A1C was associated with a higher risk for severe hypoglycemia in the standard group. Severe hypoglycemia was associated with increased risk for mortality. However, severe hypoglycemia reported in ACCORD does not account for the difference in mortality observed at the conclusion of the intensive intervention. About the Study ACCORD was a double 2x2 factorial design study in which subjects (n=10,251) were randomized to two antihyperglycemia treatment strategies, two antihypertensive treatment strategies, and two lipid-lowering treatment strategies.1 Subjects received either standard or intensive therapy. Those in the intensive group had an A1C goal <6%, while the goal for standard group was 7% to 7.9%. The primary endpoint was a composite of cardiovascular death, nonfatal myocardial infarction (MI), and nonfatal stroke; total mortality was included as a secondary endpoint. Subjects had a mean age of 62 years, 10-year duration of diabetes, moderate obesity (average BMI, 32.2 kg/m2), and average baseline A1C of 8.1%; a total of 35% had a previous cardiovascular event. VADT The new analysis of VADT showed intensive therapy (targeting A1C <7%) with rosiglitazone did not increase deaths versus standard therapy (A1C 8–9%) among subjects with diabetes. While the study failed to meet its primary endpoint—time to occurrence of a major cardiovascular event—there was a trend toward reduction of all cardiovascular events (but not cardiovascular death) among subjects who received intensive treatment. William Duckworth, MD, presented results relative to diabetes duration, and severe hypoglycemia and cardiovascular outcomes. The primary endpoint occurred in 264 subjects receiving intensive therapy compared with 235 who received standard therapy. The HR for intensive versus standard therapy was 0.88 (95% CI, 0.74–1.05; P=0.14). Baseline predictors of first primary event were previous cardiovascular event, P<0.0001; insulin use, P=0.0202; and age, P<0.0001. Intensive therapy initiated prior to 15-year duration of diabetes demonstrated a significant decrease in risk in the primary outcome. When the initial primary endpoint was assessed after 6, 12, and 72 months of treatment, respectively, results showed that intensive therapy conferred benefits at all time points. There were no significant between-group differences in death from cardiovascular causes (HR, 1.32; 95% CI, 0.81–2.14; P=0.26) or death from any cause (HR, 1.07; 95% CI, 0.81–1.42; P=0.62). Hypoglycemic events were more frequent in the intensive group; any previous severe hypoglycemia increased risk for all-cause mortality (univariate P=0.0196). Increased HDL-C decreased risk for cardiovascular events and mortality such that for every 10-mg HDL-C increase above baseline, there was an 80% decreased risk for fatal or nonfatal MI or stroke, a 50% decreased risk for first MI or stroke, and a 55% decrease in the risk for death from any cause. Stephen Davis, MD, presented data regarding severe hypoglycemia (impaired loss of consciousness requiring assistance by another person) on primary outcomes and death. Factors that predicted first severe hypoglycemia events included age (P=0.0003), duration of disease (P=0.0001), C-peptide (P<0.001), insulin use at baseline (P<0.0001), and intensive treatment (P<0.0001). Gender, race, history of hypertension, and regular exercise did not predict first severe hypoglycemia events. Lower BMI and duration of diabetes increased the risk for severe hypoglycemia; the rate of severe hypoglycemia increased in proportion with A1C. In the intensive treatment group, longer duration of disease was associated with higher rates of hypoglycemia, although the rate was not significant. Severe hypoglycemia increased risk for those who had either one or multiple hypoglycemic events (P=0.02), and increased the risk for cardiovascular or macrovascular events in both treatment groups. The relative risk for cardiovascular death after severe hypoglycemia was amplified in the standard group only. Repeated hypoglycemia increased risk for all-cause mortality. Rates of severe hypoglycemia significantly increased with standard versus intensive therapy, and significantly increased (P<0.001) prior to primary outcome. About the Study VADT randomized 1,791 military veterans who had a suboptimal response to type 2 diabetes therapy (maximum doses of oral agents or insulin) to intensive or standard glycemic treatment.2 Blood pressure, lipid, diet, and lifestyle treatments were the same in the intensive- and standard-therapy groups. Subjects had a mean age of 60.4 years, and duration of diabetes was 11.5 years; a total of 40% of subjects had a prior cardiovascular event. The primary endpoint was time from randomization to first occurrence of a major cardiovascular event, a composite that included cardiovascular death, MI, stroke, heart failure, surgery for vascular disease, inoperable coronary disease, and amputation for ischemic gangrene.
The ADA meeting ended with a review of two widely discussed trials: ACCORD and VADT. These studies both examined the impact of intensive glycemic control on cardiovascular outcomes. The original data from ACCORD, a 2x2 design study in which subjects were randomized to two glycemic treatment strategies, two hypertensive treatment strategies, and two lipid treatment strategies, initially reported a slight increase in cardiovascular mortality in the intensively managed patient population. This finding was not substantiated either in the similarly designed ADVANCE trial or in the VADT trial, which reported no increase in cardiovascular mortality, but also demonstrated little macrovascular benefit to intensive control in their patient population. At the ADA in June 2009, four presentations reviewed some of the subset analyses from these two studies. Matthew Riddle, MD, reviewed some of the data surrounding the increased mortality risk in ACCORD. This patient population had a mean age of 62 years, a 10-year duration of diabetes, and a baseline A1C of 8.1%, indicating relatively advanced stage of diabetes and a likely advanced stage of underlying cardiovascular disease (CVD). There were four hypotheses suggested for the possible causes of increased mortality: hypoglycemia, weight gain, choice of therapy, or rapid reduction in A1C. There was no signal correlating increased mortality with any specific therapeutic intervention. There was no evidence that rapid reduction of blood glucose correlated with increased mortality. And, in fact, those patients with the greatest response to intensive glycemic management had a lower risk of CVD. Weight gain would not be an issue as a risk factor in that a large percentage of these patients were treated with thiazolidinediones. In these patients, weight gain would be associated with a favorable impact on inflammatory cytokines and a reduction in A1C, and would likely not be reflected in increased cardiovascular events. Hypoglycemia itself does not explain the increased mortality in these patients. The one persistent signal was that the greatest risk for mortality seen only in the ACCORD trial, and not in any other intensive-treatment studies, was that excess mortality occurred in those patients receiving intensive management who had residual A1C >7%. Comment: It is essential to remember that the ACCORD trial is the only study which demonstrated any increased risk with intensive glycemic management—this was not substantiated in either VADT or ADVANCE. It is not likely that a targeted A1C of 6% is either achievable in patients with such advanced disease, nor would it be likely to demonstrate favorable effects of such glycemia in long-duration diabetes, as the predominant determinant of events in this patient population is LDL-C and blood pressure. The excess mortality seen in non-responders to intensive management is not unexpected, as these patients are by definition the most insulin resistant and, therefore, most likely to be at greater cardiovascular risk. Indeed, the therapy may not have engendered the risk, but merely identified those patients who have intense insulin resistance and, therefore, may be at greater risk. The second subset data were presented by Denise Bonds, MD, reviewing the hypoglycemia seen in the ACCORD trial. There was a significant increase in hypoglycemia with intensive management, with an annual increase of 2% to 3.5% with intensive treatment. Patients with impaired renal function and African-Americans were at significantly increased risk for hypoglycemia, and caution should be exercised when using intensive management in these patient populations. Again, there was no correlation between hypoglycemia and cardiovascular death, although hypoglycemia was associated with increased mortality overall. The context of these data when combined with the finding of no improved cardiovascular benefit in the study population would suggest that the risk of hypoglycemia does not justify A1C targets substantially below 7% in this patient population. The third subset analysis was presented by William Duckworth, MD, and Stephen Davis, MD, reviewing data from the VADT trial. These patients were somewhat older and had a higher baseline A1C than subjects in the ACCORD trial. The primary endpoint in this study was a composite of cardiovascular death, myocardial infarction, stroke, heart failure, and peripheral vascular disease. The duration of diabetes was somewhat longer, at 11.5 years. This study showed no increased risk but also no benefit of intensive glycemic therapy overall. However, intensive therapy in patients with shorter duration of diabetes did demonstrate a significant decreased risk in the primary outcome. These data, along with the coronary calcium subset from the VADT trial, suggest that intensive management of glycemia early in disease may indeed accrue cardiovascular benefit. This would make sense in that hyperglycemia plays a role in atherogenesis, but much less of a role in event generation. Dr Duckworth’s conclusion, which we agree with, stated that duration of diabetes should be a factor when initiating and determining parameters for intensive glycemic control. Stephen Davis, MD, reviewed the impact of hypoglycemia in the VADT trial, and the findings were quite similar to that seen in ACCORD, with increased rates of severe hypoglycemia in both intensive- and standard-management groups, and an increased risk of all-cause mortality. Therefore, individualized target for A1C is essential and dependent upon duration of diabetes and underlying risk for cardiovascular events | ||||