← Back

ACCORD

Action to Control Cardiovascular Risk in Diabetes - Effects of Intensive Glucose Lowering in Type 2 Diabetes

Year of Publication: 2008

Authors: Hertzel C. Gerstein, Michael E. Miller, Robert P. Byington, ..., William T. Friedewald; The Action to Control Cardiovascular Risk in Diabetes Study Group

Journal: New England Journal of Medicine

Citation: N Engl J Med 2008;358:2545-59

Link: https://doi.org/10.1056/NEJMoa0802743

Clinical Question

Does intensive glucose-lowering therapy targeting HbA1c below 6.0% reduce cardiovascular events compared with standard therapy targeting HbA1c 7.0-7.9% in middle-aged and older patients with type 2 diabetes and established cardiovascular disease or additional cardiovascular risk factors?

Bottom Line

Intensive glucose lowering to target HbA1c <6.0% did NOT significantly reduce major cardiovascular events but INCREASED all-cause mortality by 22% (HR 1.22, p=0.04) after mean 3.5 years. The trial was stopped early due to this excess mortality. While nonfatal MI was reduced, this was offset by increased cardiovascular death. This landmark trial identified a previously unrecognized harm of aggressive glucose lowering in high-risk diabetics.

        Major Points
        Primary composite outcome (nonfatal MI, nonfatal stroke, CV death) not significantly different: 6.9% intensive vs 7.2% standard (HR 0.90, p=0.16)
ALL-CAUSE MORTALITY INCREASED with intensive therapy: 5.0% vs 4.0% (HR 1.22, 95% CI 1.01-1.46, p=0.04)
CV death also increased with intensive therapy: 2.6% vs 1.8% (HR 1.35, 95% CI 1.04-1.76, p=0.02)
Trial stopped early (February 2008) by DSMB due to safety concerns, 17 months before scheduled completion
Nonfatal MI was REDUCED with intensive therapy: 3.6% vs 4.6% (HR 0.76, p=0.004)
No difference in nonfatal stroke: 1.3% vs 1.2% (HR 1.06, p=0.74)
Achieved HbA1c: 6.4% intensive vs 7.5% standard (difference maintained throughout follow-up)
Hypoglycemia requiring assistance much higher with intensive therapy: 10.5% vs 3.5% (p<0.001)
Weight gain >10kg more common with intensive therapy: 27.8% vs 14.1% (p<0.001)
91.2% of intensive group received rosiglitazone vs 57.5% standard - drug safety questions raised
Exploratory analyses did not identify clear explanation for mortality finding
Subgroups without prior CVD or with baseline HbA1c ≤8.0% showed possible primary outcome benefit (hypothesis-generating)

      

Design

Study Type: Multicenter randomised controlled trial with double 2×2 factorial design (blood pressure and lipid interventions tested concurrently)

Randomization: 1

Blinding: Open-label glucose intervention; outcomes adjudicated by central committee blinded to treatment assignment; central laboratory blinded for HbA1c analysis

Enrollment Period: January-June 2001 (phase 1: 1174 patients); February 2003-October 2005 (phase 2: 9077 patients)

Follow-up Duration: Mean 3.5 years (median 3.4 years); stopped early February 2008

Centers: 77

Countries: United States, Canada

Sample Size: 10251

Analysis: Intention-to-treat; Cox proportional-hazards regression with stratification for BP trial/lipid trial assignment, clinical network, and prior CVD; likelihood-ratio tests

Inclusion Criteria

Type 2 diabetes mellitus
HbA1c ≥7.5%
Age 40-79 years with established cardiovascular disease
OR age 55-79 years with anatomical evidence of significant atherosclerosis, albuminuria, left ventricular hypertrophy, or ≥2 additional CV risk factors (dyslipidemia, hypertension, current smoking, obesity)

Exclusion Criteria

Frequent or recent serious hypoglycemic events
Unwillingness to do home glucose monitoring or inject insulin
Body mass index >45 kg/m²
Serum creatinine >1.5 mg/dL (133 µmol/L)
Other serious illness

Arms

Field	Intensive Therapy	Control
Intervention	Comprehensive individualized glucose-lowering strategy targeting HbA1c <6.0% using combinations of metformin, secretagogues (glimepiride, repaglinide), TZDs (rosiglitazone), alpha-glucosidase inhibitors, incretins (exenatide), and insulin. Monthly visits for first 4 months, then every 2 months. Mean 4.4 medication adjustments per year.	Glucose-lowering strategy targeting HbA1c 7.0-7.9% using similar drug classes with fewer medications and combinations. Visits every 4 months. Mean 2.0 medication adjustments per year.
Duration	Mean 3.5 years (stopped early)	Mean 3.5 years

Outcomes

Outcome	Type	Control	Intervention	HR / OR / RR	P-value
First occurrence of nonfatal myocardial infarction, nonfatal stroke, or death from cardiovascular causes	Primary	371/5123 (7.2%); 2.29% per year	352/5128 (6.9%); 2.11% per year	0.9	0.16
Death from any cause	Secondary	203/5123 (4.0%); 1.14% per year	257/5128 (5.0%); 1.41% per year	1.22	0.04
Death from cardiovascular causes	Secondary	94/5123 (1.8%); 0.56% per year	135/5128 (2.6%); 0.79% per year	1.35	0.02
Nonfatal myocardial infarction	Secondary	235/5123 (4.6%); 1.45% per year	186/5128 (3.6%); 1.11% per year	0.76	0.004
Nonfatal stroke	Secondary	61/5123 (1.2%); 0.37% per year	67/5128 (1.3%); 0.39% per year	1.06	0.74
Fatal or nonfatal congestive heart failure	Secondary	124/5123 (2.4%); 0.75% per year	152/5128 (3.0%); 0.90% per year	1.18	0.17
Hypoglycemia requiring medical assistance	Adverse	179 (3.5%)	538 (10.5%)		<0.001
Hypoglycemia requiring any assistance	Adverse	261 (5.1%)	830 (16.2%)		<0.001
Weight gain >10 kg	Adverse	713/5042 (14.1%)	1399/5036 (27.8%)		<0.001
Fluid retention	Adverse	3378/5054 (66.8%)	3541/5053 (70.1%)		<0.001
Any nonhypoglycemic serious adverse event	Adverse	82 (1.6%)	113 (2.2%)		0.03
ALT >3x ULN	Adverse	77/5061 (1.5%)	51/5065 (1.0%)		0.02

Subgroup Analysis

Mortality effect consistent across all prespecified subgroups (no heterogeneity). For primary outcome, possible heterogeneity suggested: patients without prior CVD (P interaction=0.04) and those with baseline HbA1c ≤8.0% (P interaction=0.03) may have had fewer events with intensive therapy. These findings are hypothesis-generating only.

Criticisms

Open-label design for glucose intervention may have influenced patient management and outcomes
Rapid reduction in HbA1c (1.4% drop in 4 months) may have contributed to adverse outcomes rather than achieved level
High use of rosiglitazone (91% in intensive group) raises questions about drug-specific vs glycemic target effects
Multiple drug combinations at high doses may have caused unmeasured drug interactions
Hypoglycemia much higher in intensive group (10.5% vs 3.5%) and may have contributed to CV mortality
Unable to identify clear mechanism for mortality excess despite extensive exploratory analyses
Study not designed to test individual components of intensive strategy
Population limited to high-risk patients; results may not apply to lower-risk diabetics or primary prevention
Concurrent BP and lipid trials within factorial design complicate interpretation

Funding

National Heart, Lung, and Blood Institute (N01-HC-95178 through N01-HC-95184, IAA-Y1-HC-9035, IAA-Y1-HC-1010); National Institute of Diabetes and Digestive and Kidney Diseases; National Institute on Aging; National Eye Institute; Centers for Disease Control and Prevention; General Clinical Research Centers. Study medications/supplies from Abbott, Amylin, AstraZeneca, Bayer, GlaxoSmithKline, King, Merck, Novartis, Novo Nordisk, Sanofi-Aventis, Schering-Plough

Based on: ACCORD (New England Journal of Medicine, 2008)

Authors: Hertzel C. Gerstein, Michael E. Miller, Robert P. Byington, ..., William T. Friedewald; The Action to Control Cardiovascular Risk in Diabetes Study Group

Citation: N Engl J Med 2008;358:2545-59

Content summarized and formatted by NeuroTrials.ai.

ACCORD

(2008)

Objective

To determine whether intensive glucose lowering (HbA1c <6.0%) reduces cardiovascular events compared to standard therapy (HbA1c 7.0-7.9%) in patients with type 2 diabetes and established cardiovascular disease or risk factors

Study Summary

• Intensive glucose lowering increased mortality by 22% compared to standard therapy (5.0% vs 4.0%, HR 1.22, p=0.04)
• No significant reduction in primary cardiovascular outcome (6.9% vs 7.2%, HR 0.90, p=0.16)
• Trial was terminated early due to increased mortality in the intensive therapy group

Intervention

Intensive glucose lowering targeting HbA1c <6.0% using multiple medications vs standard therapy targeting HbA1c 7.0-7.9%

Inclusion Criteria

Type 2 diabetes with HbA1c ≥7.5% and either age 40-79 with cardiovascular disease or age 55-79 with atherosclerosis, albuminuria, left ventricular hypertrophy, or ≥2 additional cardiovascular risk factors

Study Design

Arms: Intensive therapy (N=5,128) vs Standard therapy (N=5,123)

Patients per Arm: 5,128 intensive, 5,123 standard

Outcome

• Primary outcome: nonfatal MI, nonfatal stroke, or cardiovascular death occurred in 6.9% vs 7.2% (HR 0.90, p=0.16)
• Death from any cause: 5.0% vs 4.0% (HR 1.22, p=0.04)
• More hypoglycemia and weight gain in intensive group

Bottom Line

Major Points

Primary composite outcome (nonfatal MI, nonfatal stroke, CV death) not significantly different: 6.9% intensive vs 7.2% standard (HR 0.90, p=0.16)
ALL-CAUSE MORTALITY INCREASED with intensive therapy: 5.0% vs 4.0% (HR 1.22, 95% CI 1.01-1.46, p=0.04)
CV death also increased with intensive therapy: 2.6% vs 1.8% (HR 1.35, 95% CI 1.04-1.76, p=0.02)
Trial stopped early (February 2008) by DSMB due to safety concerns, 17 months before scheduled completion
Nonfatal MI was REDUCED with intensive therapy: 3.6% vs 4.6% (HR 0.76, p=0.004)
No difference in nonfatal stroke: 1.3% vs 1.2% (HR 1.06, p=0.74)
Achieved HbA1c: 6.4% intensive vs 7.5% standard (difference maintained throughout follow-up)
Hypoglycemia requiring assistance much higher with intensive therapy: 10.5% vs 3.5% (p<0.001)
Weight gain >10kg more common with intensive therapy: 27.8% vs 14.1% (p<0.001)
91.2% of intensive group received rosiglitazone vs 57.5% standard - drug safety questions raised
Exploratory analyses did not identify clear explanation for mortality finding
Subgroups without prior CVD or with baseline HbA1c ≤8.0% showed possible primary outcome benefit (hypothesis-generating)

Study Design

Study Type: Multicenter randomised controlled trial with double 2×2 factorial design (blood pressure and lipid interventions tested concurrently)
Randomization: Yes
Blinding: Open-label glucose intervention; outcomes adjudicated by central committee blinded to treatment assignment; central laboratory blinded for HbA1c analysis
Sample Size: 10251
Follow-up: Mean 3.5 years (median 3.4 years); stopped early February 2008
Centers: 77
Countries: United States, Canada

Primary Outcome

Definition: First occurrence of nonfatal myocardial infarction, nonfatal stroke, or death from cardiovascular causes

Control	Intervention	HR/OR	P-value
371/5123 (7.2%); 2.29% per year	352/5128 (6.9%); 2.11% per year	0.9 (0.78-1.04)	0.16

Limitations & Criticisms

Open-label design for glucose intervention may have influenced patient management and outcomes
Rapid reduction in HbA1c (1.4% drop in 4 months) may have contributed to adverse outcomes rather than achieved level
High use of rosiglitazone (91% in intensive group) raises questions about drug-specific vs glycemic target effects
Multiple drug combinations at high doses may have caused unmeasured drug interactions
Hypoglycemia much higher in intensive group (10.5% vs 3.5%) and may have contributed to CV mortality
Unable to identify clear mechanism for mortality excess despite extensive exploratory analyses
Study not designed to test individual components of intensive strategy
Population limited to high-risk patients; results may not apply to lower-risk diabetics or primary prevention
Concurrent BP and lipid trials within factorial design complicate interpretation

Citation

N Engl J Med 2008;358:2545-59

View Original Publication →

ACCORD

Action to Control Cardiovascular Risk in Diabetes - Effects of Intensive Glucose Lowering in Type 2 Diabetes

Clinical Question

Bottom Line

Major Points

Design

Inclusion Criteria

Exclusion Criteria

Arms

Outcomes

Subgroup Analysis

Criticisms

Funding

Ask about ACCORD