The Action to Control Cardiovascular Risk (ACCORD) study set out to define the role of intensified glucose, lipid and and blood pressure control in the prevention of coronary heart disease in type 2 diabetes. The aims were to compare outcomes in patients who strove for normoglycaemia (HbA1c <6.0%) as against more standard control (HbA1c 7.0-7.9%). A total of 10,251 patients were randomized to one or other glucose strategy, and 50% of these were then randomized either to intensified vs standard lipid-lowering or to intensified vs standard BP-lowering therapy. The major outcome measure was a coronary event or stroke; secondary outcomes included other CV outcomes, total mortality, microvascular outcomes, quality of life and cost-effectiveness. The unexpected outcome of the trial was increased mortality in the intensified glucose control arm, and this was terminated prematurely in 2009; the main study ended in 2010. Further, although intensive therapy delayed the onset of microalbuminuria, there was no overall difference in microvascular outcomes. The BP study provided no conclusive evidence that targeting a systolic BP of 120 vs 140 mmHg reduced major CVD events, and the lipid substudy showed no additive benefit of combining fenofibrate with a statin in the prevention of major cardiovascular outcomes.
ACCORD represented the logical sequel to two other major trials in type 2 diabetes, The University Group Diabetes Program and The United Kingdom Prospective Diabetes Study [UKPDS]. UGDP had shown no effect of glucose-lowering therapy upon vascular outcomes in diabetes (other than increased cardiovascular mortality with tolbutamide and phenformin). UKPDS, a larger and longer trial, showed a clear effect of glucose control upon microvascular outcomes, and a borderline reduction in cardiovascular outcomes. The long-term follow-up studies of both UKPDS and DCCT did however show a delayed reduction in cardiovascular events.
The smaller Steno-2 Study had meanwhile showed a marked improvement in major outcomes in recently diagnosed individuals offered intensive multifactorial control of glucose, lipids and blood pressure. Between them, these studies showed a clear benefit of glucose control upon microvascular outcomes. Glucose control had a possible long-term effect upon cardiovascular outcomes, but blood pressure and lipid therapy were more effective in this regard.
What remained unclear was the practical limit to these benefits. Should one aim for risk factor control to within the healthy range of glucose, BP and lipids, or was there a threshold benefit? ACCORD set out to address this question.
It is important to note that ACCORD participants, who were aged 40-79, had established diabetes, with an average duration ~ 10 years, and in addition had either clinical evidence of heart disease or at least two cardiovascular risk factors in addition to diabetes. The average HbA1c at entry was 8.2%.
ACCORD Glycaemic Control
10,251 participants were randomized to intensive glucose control (target ~ 6.0% HbA1c) or conventional control (7-7.9%). The study design did not consider specific treatment strategies, and the investigators were free to use any available treatment. The actual separation was HbA1c 6.4% (intensive) and 7.5% (conventional).
There were 257 deaths in the intensive treatment group and 203 in the conventional treatment group, equivalent to 3 excess deaths per year on intensive therapy. In the light of this finding the intensive limb was discontinued. It was initially anticipated that the excess death rate would be due to hypoglycaemia, but this did not emerge clearly from the analyses. Alternative possibilities were that the excess mortality was in some way attributable to specific therapies, for example glitazones, but the study design did not permit any firm conclusions to be drawn.
Points to note in evaluating this finding are (1) that very tight glucose control was achieved - better than in DCCT or UKPDS, (2) that this was superimposed rather rapidly following a long period of suboptimal control, and (3) that participants had either established heart disease or multiple risk factors for this. The outcome might therefore have been different if the intensive treatment had been offered earlier in the course of diabetes, or to people without heart disease, but the study findings can neither support nor disprove such speculation.
ACCORD Blood Pressure Trial
Before ACCORD, clinical trial data had demonstrated the benefit of lowering SBP levels below 140 mmHg, but observational studies had suggested benefits extending as low as 120 mmHg. 4,733 participants were accordingly randomized to intensive or standard blood pressure lowering regimens, using a range of existing therapies; targets were SBP of 120 or 140 mmHg respectively.
In the event, there were 208 cardiovascular events in the intensive group and 237 in the standard treatment group. Intensive BP lowering did reduce the risk of stroke (36 vs 62 strokes), but adverse events such as hypotension or hyperkalaemia were increased (77 vs 30), and renal function was adversely affected in some but without increasing the rate of renal failure. Other cardiovascular events were unaffected.
ACCORD Lipid trial
The rationale for the lipid component of ACCORD arose from the observation that people with diabetes obtained the same relative benefit from statin therapy as non-diabetic, yet still experienced higher absolute rates. This suggested that more aggressive lipid-lowering therapy might be needed. Fenofibrate was selected for combination therapy in this trial because it was expected to lower triglyceride levels by around 25% and raise HDL cholesterol by ~5-10%, and the risk of myositis is low when taken with a statin.
In this limb of the study, 5,518 participants were treated with simvastatin and randomized to fenofibrate or placebo. The outcome measure was a non-fatal coronary or stroke or cardiovascular death.
The study was negative; there was no benefit from addition of fenofibrate. Unexpectedly, there was a gender-specific effect: men did ~16% better on fenofibrate and women did ~38% worse, although these differences did not reach statistical significance.
Although the study appears negative in retrospect, since it showed no benefit for any of the three treatment strategies, it is important to recognise first, that these were people with long duration diabetes selected for the presence of established heart disease, and second that all three risk factors were already controlled to conventional treatment levels in the control limbs of the trial. The study showed either that further intensification produced no additional benefit, or (as with stroke reduction by blood pressure lowering), that the benefit was balanced by disadvantages.
Otherwise said, ACCORD suggests that there may be a point of therapeutic futility in the treatment of established type 2 diabetes beyond which benefits diminish and dysbenefits accumulate. This is an important outcome, not least for patients who feel frustrated at the difficulty of bringing their diabetes into perfect control. The message for these is that standard levels of control are (for most purposes) often good enough.
^ Gaede P et al: Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med 2008;358:580 –589
^ Cefalu W, Watson K. Intensive control and cardiovascular disease observations from the ACCORD Study. Now what can a clinician possible think? Diabetes 2008;57(5):1163-5
^ ACCORD Study Group. Effects of intensive blood-pressure control in type 2 diabetes mellitus. New Engl J Med 2010;362:1575-85
^ Ginsberg HN. The ACCORD lipid trial. What we learn from subgroup analysis. Diabetes Care 2011;34: Suppl 2 S107-8