Metformine et mortalitéchez des diabétiques en prévention secondaire

Ronan Roussel, Florence Travert, Blandine Pasquet, Peter F. Wilson,Sidney C. Smith Jr, Shinya Goto, Philippe Ravaud, Michel Marre, Avi Porath,

Deepak L. Bhatt, Ph. Gabriel Steg, for the REACH investigators

Paris, France; Atlanta, Georgia, USA; Chapel Hill, North Carolina, USA;Isehara, Japan; Beer Sheva, Israel; and Boston, Massachusetts, USA

Roussel, Archives of Internal Medicine, 2010

Conflits d’intérêt potentiels :

J’ai reçu des soutiens financiers pour la recherche et des congrès et des contreparties financières pour des conférences de

sanofi-aventis, MSD Chibret, Servier, Roche et Novo Nordisk.

Holman R, New Engl J Med, 2008

Primary prevention and metformin:

Secondary CV Prevention in Diabetes:Unmet Needs

- Diabetes is associated with a doubling of CV risk, but also with an increased risk for cancer

- In the REACH Registry, fatality rates were higher in diabetic patients, despite as intensive treatment of risk factors as in non-diabetic patients

Krempf M, Am J Cardiol, 2010

The Emerging Risk Factors Collaboration, the Lancet, 2010

Established CV Disease

Metformin Use is Restricted in High Risk Patients

Dormandy JA, The Lancet. 2005TThe ADVANCE Collaborative Group, The New Engl J Med. 2008The Action to Control Cardiovascular Risk in Diabetes Study Group, The New Engl J Med. 2008

40%60%ACCORD

39%61%ADVANCE

NoYes

Metformin at baseline

43%57%PROactive

27,746

1,931

17,886

846

10,951

2,872

5,656North America

Latin America

Eastern Europe + North Asia

Middle East

Australia

Western Europe

S. Asia (incl. Japan)

*up to 15 patients / site (up to 20 in the US)

1. Bhatt DL et al. JAMA 2006;295:180-189.2. Ohman EM et al. Am Heart J 2006;151:786.e1-10.

Patient Recruitment: > 67,000 Patientsfrom 5,473 Sites in 44 Countries

www.reachregistry.org

Must include:

Signedwritten

informedconsent

Patients aged≥ 45 years

At least of four criteria1

• Documented cerebrovascular diseaseIschemic stroke or trans ischemic attack

• Documentedcoronary diseaseAngina, myocardial infarction, angioplasty /stent / bypass

3. Documented historicalor current intermittentclaudication associatedwith ABI < 0.9

4. At least atherothrombotic risk factors3

• Male aged ≥ 65 yearsor female aged ≥ 70 years

• Current smoking> 15 cigarettes/day

• Type 1 or 2diabetes

• Hypercholesterolemia• Diabetic nephropathy• Hypertension• ABI < 0.9 in either

leg at rest• Asymptomatic carotid

stenosis ≥ 70%• Presence of at least

one carotid plaque

ABI, ankle brachial index.Ohman EM et al. Am Heart J 2006;151:786.e1-10.

Inclusion Criteria of the REACH Registry

Total patients 67888

Aims of the Analysis

•Evaluation of the risk/benefit ratio associated with metformin in secondary cardiovascular prevention

We studied the baseline characteristics and 2-year outcomes of the subset of 28700 diabetic patients undergoing a secondary prevention strategy in

the international Reduction of Atherothrombosis for Continued Health (REACH) Registry.

•To assess whether metformin use was associated with a difference in mortality after adjustment for baseline differences and the propensity to receive metformin in patients with established coronary artery disease (CAD), cerebrovascular disease (CVD), or peripheral arterial disease (PAD).

Results: Baseline General Characteristics

<.00128.6 (5.9)29.6 (6.0)BMI, kg/m2, mean (SD)

<.00199.6 (16.6)Waist, cm, mean (SD)

.9067997 (65.4)4881 (65.5)Male sex, n (%)

<.00169.2 (9.5)67.1 (9.3)Age, y, mean (SD)

P Value

No(n = 12 234)

Yes(n = 7457)Characteristics

Metformin

eGFR (mL/min/1.73 m²)median (SD)

<.001Ethnic origin

<.001Region

102.0 (16.2)

76.0 (37.5)78.3 (61.1) .003

Results: Propensity Score

Confounding factors were taken into account using the propensity score method.

This score represents the probability of receiving metformin given an individual’s characteristics.

The list of co-variables was built in a two-step process. First, analyses were conducted to determine the variables associated with metformin prescription among all the available data variables.

The selected variables were then introduced in order to construct a multivariable logistic regression model.

The propensity score reached the quality requirements:

the likelihood associated with the model was strong (Wald test, P<.001)

the area under the ROC curve (0.72) exceeded the 0.7 threshold.

Results: Survival according to Metformin Use

Number at risk

Met

form

in

Yes

No

7397

12156

7234

11805

6848

10979

6119

9769

3340

5808

Hazard ratio: 0.67

95% CI: 0.59-0.75

P<0.0001

1270 fatality cases occurred.

28700 diabétiques DT2

-25% mortalité toute causeAssociées à Metformine

Results: Survival according to Metformin Use

Significant factors in univariate analysis:region, ethnic origin, education, employment, hypercholesterolemia, carotid surgery, atrial

fibrillation/flutter, congestive heart failure, aortic valve stenosis, abdominal aortic aneurysm, antiplatelet agents, anticoagulants, lipid-lowering agents, other cardiovascular agents, BMI, and SBP

No. of deaths/no. of patients

2-year mortality rate (95% CI)

Adjusted for sex and age

HR (95% CI)

HR (95% CI)

Adjusted for sex, age, propensity score, and significant factors

Metformin Use

No Yes

341/7397929/12 156

9.83(8.40-11.23)

6.33(5.24-7.41)

1

1

0.67(0.59-0.75)

0.76(0.65-0.89)

<.001

<.001

Results: Hazard Associated with Metformin UseAccording to Clinical Characteristics

Ageadjusted HR

Metformin Use

n/NYes

n/NNo

Favors metformin

P Value

P Value for interaction: p=0.07

78/2,987 176/3,859 0.63 p=0.008

191/3,791 532/6,768 0.77 p=0.016

71/ 598 220/1,492 0.92 p=0.605

Results: Hazard Associated with Metformin UseAccording to Clinical Characteristics

History ofCongestive Heart Failure

Favors metformin

P Value for interaction: p=0.39

221/6,002 488/9,120 0.80 p=0.034

116/1,220 419/2,790 0.69 p=0.006

adjusted HR

Metformin Use

n/NYes

n/NNo

P Value

Results: Hazard Associated with Metformin UseAccording to Clinical Characteristics

Renal FunctionAt Baseline

Favors metformin

P Value for interaction: p=0.12

14/118 90/455 1.06 p=0.890

86/1,572 336/3,388 0.64 p=0.003

188/4,442 379/6,326 0.89 p=0.302

adjusted HR

Metformin Use

n/NYes

n/NNo

P Value

Results: Hazard Associated with Metformin UseAccording to Clinical Characteristics

Insulin Use

Favors metformin

P Value for interaction: p=0.01

281/6,050 521/7,891 0.80 p=0.018

59/1,276 408/4,258 0.64 p=0.009

adjusted HR

Metformin Use

n/NYes

n/NNo

P Value

Merci de votre attention

Results: Hazard Associated with Metformin UseAccording to Clinical Characteristics

Gender

341/7,397 929/12,156 0.76 p<0.001

243/4,845 617/7,954 0.82 p=0.037

98/2,548 312/4,195 0.66 p=0.005

Favors metformin

P Value for interaction: p=0.07

adjusted HR

Metformin Use

n/NYes

n/NNo

P Value

Results: Baseline General Characteristics

EthnicOrigin

Region

North AmericaLatinAmerica

AustraliaJapan

Metformin Use: No

W Europe

W Europe

JapanAustralia

Asia

Asia

E Europe

E Europe

MiddleEast

MiddleEast

Latin America

Metformin Use: Yes Metformin Use: Yes

Metformin Use: No

Caucasian

Hispanic

East Asian

South Asian

Other Asian

BlackOther

Hispanic

East Asian

South AsianOther Asian

BlackOther

metformin could be associatedwith a better prognosis in older patients with diabetes

discharged after hospitalization for heart failure

Metformin: Unexpected Benefits?

Masoudi FA, Circulation. 2005;111(5):583-590

Adjusted mortality curves for patients hospitalized with heart failure and diabetes receiving prescription for metformin at hospital discharge and patients not treated with insulin-sensitizing drug.

HR=0.86, 95% CI 0.78 to 0.97