1

Riesgo Cardiovascular

en

Paciente VIH

Riesgo Cardiovascular

en

Paciente VIHEsteban MartínezServicio de InfeccionesHospital ClínicUniversidad de BarcelonaEspaña

I Congreso GESIDA

Conferencia Especial GSK

Madrid, 22 Octubre 2009

CV disease

Patient

Antiretroviraldrugs

HIV(and other infections)

Drug consumption

TobaccoAlcoholCocaineOther?

There are reasons to consider that the risk of CV disease may be increased in HIV-infected patients

Metabolic abnormalities

DyslipidemiaInsulin resistance / DM

Body fat changes

LipoatrophyLipoaccumulation

Degree of immunedeficiency

PIsDyslipidemia

Insulin resistance ?Body fat changes?

Other?

NRTIsDyslipidemia?

Insulin resistance?Body fat changes?

Other?

HIV, HCV, HBV?, other?Dyslipidemia

Systemic inflammationInmune activationVascular infection

Aging

Other?

0

1

2

3

4

5

2003 2007 2008 2009

Inci

den

ce o

f M

I (p

er 1

000

PY

FU

)

No. MI 126 345 517 580

PYFU 36199 94969 157912 178835

NEJM 2003; NEJM 2007; Lancet 2008; CROI 2009

D:A:D Study: Low and stable incidence of myocardial infarction in HIV-infected patients

D:A:D Study

Contribution of the PATIENT

References Size Outcome HIV+ vs HIV-

Klein, JAIDS, ’02 4,159 / 39,877 CAD (6.5 vs 3.8/1000 PYFU)

Klein, CROI, ’07 5,000 / 43000 CAD (4.5 vs 2.9/1000 PYFU)

Currier, JAIDS, ’03 28,513 / 3 mill CAD (only in young)

Triant, JCEM, ’07 3,851 / 1 mill MI (75%)

Obel, CID, ’07 3,953 / 0.4mill CAD (39-112%)

Increased risk of CV disease in HIV+ persons relative to HIV- ones: impact of traditional factors

• Approximately 2-fold higher incidence of CAD / MI

• Higher prevalence of other types of CV disease in HIV+ vs HIV-:Peripheral arterial disease 1, 2

Increased cIMT 3-5

1. Palacios R et al. AIDS Res Human Retrovir 2008; 2. Periard D et al. Clin Infect Dis 2008; 3. Hsue PY et al. Circulation 2004; 4. Mercie P et al. HIV Medicine 2005; 5. Lekakis J et al. Clin Sci 2008; 6. Lorenz MW et

al. Atherosclerosis 2008

In these studies, higher prevalence of traditional CV risk factors (smoking, HT, DM, and dyslipidemia) in HIV+ relative to HIV-

No differences in CV disease between HIV+ and HIV- when prevalence of traditional CV factors is similar

References Size Traditional CV factors

Outcome HIV+ vs HIV-

Currier,

AIDS, ’07

•HIV+, PI+ (n=44)•HIV+, PI- (n=44)•HIV- (n=44)

Excluded if: DM, history of CV, uncontrolled HT, obesity

Matched for age, gender, race, smoking, and menopausal status

cIMT

Prospective (144w

change)

0.0096 vs 0.0058 vs 0.0085 mm/year

(P=0.71, for HIV+ vs HIV-)

Lebech,

Clin Physiol Funct Imag, ’07

•HIV+, ART+, normal chol (n=13)•HIV+, ART+, high chol (n=13)•HIV- (n=44)

Excluded if: DM, history of CV, HT, obesity, lipid-lowering therapy

Matched for age, gender, and smoking

cIMT

Cross sectional

0.683 vs 0.656 vs 0.657 mm

(P>0.05, for all comparisons)

DAD French Cohort

MI (n=580) No MI (n=32728)

MI (n=289) No MI (n=884)

Age (years) (median) 49 44 47 46 (matched)

Sex, male (%) 91 74 89 89 (matched)

Current smoker (%) 45 29 73 44

Previous CV disease (%) 20 3 0 (defined) 0

Family history CV disease (%) 14 8 19 7

Diabetes mellitus (%) 17 5 16 10

Hypertension (%) 44 19 21 12

Any dyslipidaemia (%) 75 44 - -

Hypercholesterolaemia (%) - - 52 33

10-year Framingham score

Moderate (10-20%) (%) 30 15 - -

High (≥20%) (%) 18 4 - -

Nr CV risk factors

0 (%) - - 1 18

≥3 (%) - - 39 19

Lundgren, J & DAD Study Group et al CROI 2009 abstract 44LB; Lang et al, CROI 2009, abstract 43LB

HIV+ patients with myocardial infarction have higher prevalence of traditional CV risk factors

• SMART

• D:A:D

• CASCADE*

• FIRST

• VACS†

• D:A:D

• CASCADE

• FIRST

• Aquitaine

• SMART

• D:A:D

• CASCADE

• FIRST

• SMART

• FIRST

• VACS

Cardiovascular disease

Non-AIDS cancer Liver Disease Renal disease

Goulet JL, et al. Clin Infect Dis 2007;Weber R, et al. 12th CROI;

CASCADE Collaboration, AIDS 2006;Baker JV, et al. AIDS 2008;

Weber R, et al. Arch Intern Med 2006;Bruyand M, et al. 15th CROI;

El Sadr WM, et al. N Engl J Med 2006

* Cardiovascular disease or type 2 diabetes; † Vascular disease

Causes of morbidity/mortality:

Studies suggesting association with lower CD4+ count:

Several studies suggest lower CD4 cell counts associated with higher rates of non-AIDS events

CD4<350/µL YES YES YES

CD4 ≥350/µL

Offer between 350-500 /µL if:

•>5 log c/mLHIV RNA•>50-100 CD4/year

decrease•>55 years•Hepatitis C

YES if:•Pregnancy

•HIV-associated nephropathy•Hepatitis B

Consider depending on patient scenarios and comorbidities

Consider if:•>5 log c/mLHIV RNA

•>100 CD4/year decrease•High cardiovascular risk

•HIV-associated nephropathy

•Hepatitis B or C

www.eacs.eu http://AIDSinfo.nih.govHammer S. et al JAMA

2008

From a purely CV perspective, it is better to administer ART than not to administer it

Contribution of the VIRUS

Riddler SA, et al. JAMA 2003; 289: 2978-82.

201

122

51

171

100

39

220

129

40

0

50

100

150

200

250

Total cholesterol LDL-chol HDL-chol

Pre-seroconversion

Pre-HAART

Post-HAART

Pla

sma

con

cen

trat

ion

(m

g/d

L)

MACS study

Total cholesterol / HDLc ratio increased

A relative HDL-cholesterol deficit is characteristic of HIV infection

Gordon T, et al. Am J Med 1977;62:707–14; *3rd Report of the NCEP Expert PanelLDL-chol, mg/dL

HDL-chol, mg/dL

Co

ron

ary

dis

ease

ris

k

0.0

1.0

2.0

3.0

100 160 220 8565

4525

Kannel WB. Am Heart J 1985;110:1100-1107.

Framingham study

HDL-cholesterol modulates CV risk associated with LDL-cholesterol

0

% w

ith

a M

ajo

r C

VD

Eve

nt

Years from Randomization

2752

0.5

1306

1.5

713

2.5

379

3.5

10

5

10

15

20

DC

VS

2720 1292 696 377 10DCVS

No. at Risk

3148All major CVD events

1422Coronary artery disease requiring surgery for invasive procedure

38Non-fatal stroke

511Non-fatal silent MI

1212Non-fatal clinical MI

47Death from CVD

VSDC

0 1 2 3 4

SMART Study

Phillips A, et al. 14th CROI, Los Angeles, CA, February 25-28, 2007

ART discontinuation increases the risk of CV disease relative to standard, continuous ART

SMART Study

-20

-15

-10

-5

0

HIV RNA at 1 Year

≤ 400

401-10,000

10,001-50,000

>50,000

HD

L D

ecl

ine

(m

g/d

l)

N = 239 148 158 164

P < 0.001 for trend

DC Patients on ART at Baseline with HIV RNA ≤ 400 copies /mL: HDL Decline at 1 Year According to HIV RNA Level at 1 Year

SMART Study: the higher the viral load rebound the higher the HDL-c decrease

Neaton J. Personal communication

Kuller L, et al. CROI 2008. Abstract 139.

P-values are for adjusted odds ratios for the 4th vs 1st quartile of each biomarker at baseline, estimated using logistic regression

IL-6

2

0

1

2

3

4

5

6

Od

d r

atio

(95

% C

I) C

V d

isea

se/d

eath

D-dimer

2.8

P=0.003

P=0.06

2.8

P=0.002

Amyloid P

N=499 stored baseline plasma samples

SMART Study: Inflammation and coagulation markers asscociated with risk of CVD/death

Contribution of the THERAPY

Holmberg SD, et al. Lancet. 2002;360:1747-1748

• Hypertension

• Smoking

• Male gender

• Age > 50 years

• Dyslipidaemia

• Diabetes

• Use of PI

HOPS cohort: Low, but increasing incidence of MI associated with

Inci

den

ce (

Per

100

0 P

YF

U) 3.5

3.0

2.5

2.0

1.5

1.0

0.5

01993 1994 1995 1996 1997 1998 1999 2000 2001

Potential confounding and biases:1.Competing risks of non-HIV diseases due to increased survival2.“Use of PI” was equivalent to “optimal ART” (vs suboptiomal ART)

PI use linked to Mi in some initial cohort studies, but there were some potential biases

Adjusted relative rate/year of PI: 1.15 (1.06, 1.25)

Adjusted relative rate/year of NNRTI: 0.94 (0.74, 1.19)

Nu

mb

er o

f M

Is p

er 1

000

PY

FU

(IC

95%

)

Years of exposure to PI or NNRTI

0

2

4

6

8

10

>65–62–3 3–4 4–5 0 1–2<1

Friis-Møller, et al. N Engl J Med. 2007;326:1723-1735

D:A:D study

Higher risk of MI with PI exposure(but not with NNRTI exposure)

NRTI: d4T or AZT en >75%PI: NFV, IDV (ritonavir boosting included)NNRTI: EFV, NVP

FIRST study

Shlay JC et al. J Acquir Immune Defic Syndr 2007; 44: 506-517

NNRTI-containing ART increases HDL-cholesterol more than PI-containing ART

1. Periard et al. Circulation. 1999; 2. Purnell et al. AIDS. 2000; 3. Noor et al. AIDS. 2001; 4. Lee et al. AIDS. 2004

0

1

2

3

4

5

RTV IDV NFV RTV IDV LPV/r

Tri

gly

ce

rid

es

mm

ol/

L

P <.01

P <.007

P <.01

LD

L C

ho

les

tero

l m

mo

l/L

0

1

2

3

45 P <.001

P <.05 P =.06

HIV-Infected 1 HIV-Negative 2, 3, 4

LDL is a “Return-to-Normality” effect not a drug effect

TG is a drug-specific effect

In each bar pair, the light color indicates baseline values and the darker color indicated values after receiving ART.

Not all lipid changes with ART are necessarily ART-related

1. Nadler JP et al. 44th ICAAC 2004. Poster H-156 (NEAT Study)2. Smith K et al. 46th ICAAC 2006; San Francisco. Abstract H-1670a (ALERT Study)

3. Eron J et al. XVI International AIDS Conference 2006; Abstract THLB0205 (KLEAN Study)

-10

0

10

20

30

40

50

60

70

Triglycerides

FPV1 FPV/r1002 FPV/r2003

Me

dia

n li

pid

ch

an

ge

at

we

ek 4

8 (

mg

/dL

)

(ABC+3TC) (TDF+FTC) (ABC+3TC)Better than NFV = ATV/r = LPV/r

Naive patients

PIs increase triglycerides and their effect depends on RTV boosting dose

*Adjusted for conventional risk factors (sex, cohort, HIV transmission group, ethnicity, age, BMI, family history of CVD, smoking, previous CVD events, lipids, diabetes, and hypertension).

†Unadjusted model.

Relative Rate of MI* (95% CI)

0.72 (0.52–0.99); P=0.05*

1.58 (1.43–1.75); P<0.001†

1.26 (1.19–1.35); P<0.001*

1.10 (1.01–1.18); P=0.002*

1.00 (0.93–1.09); P=0.92*

Total cholesterol(per mmol/L)

Triglycerides(per log2 mmol/L higher)

HDL cholesterol(per mmol/L)

PI exposure (per additional year)

NNRTI exposure(per additional year)

1 100.1

Friis-Moller N et al. N Engl J Med. 2007

Contribution of dyslipidemia to MI risk

Adjusted for sex, age, cohort, year, prior CVD, family CVD , smoking, body mass index, and the other 3rd drug

Exposure to NNRTI(por year)

0.9 1.0 1.1 1.2 1.30.9 1.0 1.1 1.2 1.3

Relative rate of MI (95% CI)

RR 1.00

RR 1.10

RR 1.05

RR 1.16Exposure to PI(per year)

Part (but not all) of CV risk attributable to PI can be explained by lipids

Part (but not all) of CV risk attributable to PI can be explained by lipids

Friis-Møller N et al. N Engl J Med 2007

0.9 1 1.1 1.3 1.4

Primary model

Dislipidemia

Increased blood pressure

Diabetes mellitus

Lipodystrophy

Glucose

All above + lipid-lowering and antihypertensive medication

Lopinavir/r

Indinavir

RR of IAM (95%CI)

Traditional factors explain little on CV risk associated with IDV and LPV/r in DAD studyTraditional factors explain little on CV risk

associated with IDV and LPV/r in DAD study

Lundgren JD et al., CROI 2009. Abst 44LB

0.9 1 1.1 1.2 1.3 1.4

# PYFU:

44, 657 Saquinavir (global)

24,727 Saquinavir + rtv

26,145 Saquinavir - rtv

68,469 Indinavir (global)

22,186 Indinavir + rtv

57,961 Indinavir - rtv

RR of IAM (95%CI)

Lundgren JD et al., CROI 2009. Abst 44LB

Ritonavir boosting does not modify the risk associated with individual PIs in DAD studyRitonavir boosting does not modify the risk associated with individual PIs in DAD study

PI use different over time; patients treated with same PI also differ over time

PI use different over time; patients treated with same PI also differ over time

Datos del Hospital Clínic, Barcelona 2008

Sabin C et al., CROI 2008; #957c.Recent = still using or stopped within last 6 months

Zidovudine Stavudine LamivudineRecent

useNo

recent use

Recent use

No recent

use

Recent use

No recent

use

MI events 214 303 134 383 377 140

Event rate (95% CI)/1000 person-years

3.4(3.0, 3.9)

3.2(2.8, 3.5)

3.7 (3.1, 4.4)

3.1(2.8, 3.5)

3.7 (3.3, 4.1)

2.5(2.1, 2.9)

Didanosine AbacavirRecent

useNo

recent use

Recent use

No recent

use

124 393 192 3254.5

(3.7, 5.3)3.0

(2.7, 3.3)6.1

(5.3, 7.0)2.6

(2.3, 2.9)

35

30

25

20

15

10

5

0

Overall Low Moderate High Not known

Rat

e (p

er 1

000

PY

)

Predicted 10-year CHD risk

No recent didanosine

Recent didanosine

Stratified by recent* didanosine use

35

30

25

20

15

10

5

0

Overall Low Moderate High Not known

Rat

e (p

er 1

000

PY

)

Predicted 10-year CHD risk

No recent abacavir

Recent abacavir

Stratified by recent* abacavir use

Rates of MI for Recent Use of Didanosine and of Abacavirby Predicted 10-year Coronary Heart Disease (CHD) Risk

D:A:D studyRates of MI for Recent or No Recent Use of different NRTIs

Higher risk of myocardial infarction with recent ABC or ddI use?

DAD 2007 DAD 2008 DAD2009 FHDB2009

PI (as a family)

16% per year(relative to NNRTI)

(10% per year, after adjustment for lipids, BP, and DM)

- - 16% per year(relative to SQV)

LPV/r - - 13% per year 37% per year

IDV - - 12% per year Not significant

APV/fAPV - - Not significant 52% per year

ABC - 90%recent exposure *

14% per year

68% recent exposure *

7% per year

97% recent exposure *

but only if exposure <1 year

ddI - 49%recent exposure *

6% per year

41% recent exposure *

Not significant

1. DAD NEJM 2007; 2. DAD Lancet 2008; 3. DAD CROI 2009; 4. FHDB CROI 2009

* Recent exposure means current exposure or stopped in the previous 6 months

DAD and French Hospital Database Studies: discrepancies among results

Increase in the risk of MI by individual antiretroviral drugs and families

% adjusted difference

using ”ABC (no ddI)”versus

using ”other NRTIs”

*

*

**

*: p=0.02**: p=0.07

For all others:p>0.1

n=791

(nmol/L)(µg/mL)(µg/L)(mg/mL)(pg/mL)(µg/mL)

0.4 (0.3-0.5)0.3 (0.2-0.5)65 (51-86)3.6 (1.9-6.8)2.2 (1.4-3.7)2.3 (1.0-5.3)Median (IQR)

levels in ”Others NRTIs”

Lundgren J et al XVII IAC, Mexico, 2008 & AIDS 2008

* For known CVD risk factors, CD4 count, HIV-RNA, hepatitis status, and use of NNRTIs and PI

72% < 400 copies/mLART was not randomized

Higher biomarkers with ABC (no ddl) at entry in SMART study do not mean causality

0

2

4

6

8

10

AbacavirSparingRegimen

AbacavirContainingRegimen

P =0.01

End

othe

lium

-de

pend

ent

FM

D (

%)

After adjustment for age, gender, traditional risk factors, HIV-specific factors, and baseline brachial artery diameter, current abacavir use was independently associated with lower FMD (p=0.02)

Cross-sectional studyNo data on antiretrovirals other than ABC and other epidemiological characteristics

Hsue P et al, CROI 2009 abstract 723

ABC use associated with endothelial function among patients on ART with HIV

suppression

•13 [23]% .v. 1 [4]%

P = 0.009

30 [25]% .v. 18 [18]%

P = 0.032

•

•12 [19]% .v. 5 [2]%

P = 0.06

31 [20]% .v. 21 [17]% P = 0.043

Satchell C et al, CROI 2009 abstract 151LB

Cross-sectional studyNo data on antiretrovirals other than ABC and other epidemiological characteristics25% patients had detectable viral load

Significance lost when adjusted for HIV RNA for ADP and epinephrine

ABC use associated with platelet reactivity

N exposed

N exposed

cases

Univariate model

OR [ 95% CI ]

Model 1: cumulative exposure only

OR [ 95% CI ]

Abacavir, cumul expo 410 127 1.05 (0.96 - 1.15) 0.97 (0.86 - 1.10)

Didanosine, cumul expo 691 186 1.02 (0.95 – 1.09) 0.91 (0.82 – 1.01)

Lamivudine, cumul expo 1043 269 1.06 (1.00 – 1.13) 0.96 (0.86 – 1.08)

Stavudine, cumul expo 718 199 1.09 (1.02 – 1.16) 1.11 (0.99 – 1.24)

Tenofovir, cumul expo 238 65 1.19 (0.99 – 1.44) 1.01 (0.79 – 1.30)

Zalcitabine, cumul expo 314 92 1.08 (0.94 – 1.24) 0.99 (0.82 – 1.21)

Zidovudine, cumul expo 998 256 1.03 (0.98 – 1.08) 1.09 (1.00 – 1.19)

Adjusted for hypertension, smoking, family history of premature CAD, use of cocaine and/or IV drug use, plasma HIV-1 RNA level, CD4/CD8 cells ratio, exposure to emtricitabine, atazanavir, ritonavir and tipranavir

This is different from D:A:D

Without D:A:D, we would have found no association

Costagliola. IAS 2009. MOAB201

A re-analysis of the FHDH Study has not found a MI signal for NRTIs, including ABC

15

0.1

• Patients switching to ABC/3TC (n = 46) or TDF/FTC (n = 34)

• Markers of inflammation, endothelial dysfunction, insulin resistance, and hypercoagulability compared at BL and Wk 48

• None of these markers were significantly different between arms at Wk 48

Med

ian

Cha

nge

Fro

m

BL

at W

k 48

(%

)

-5

0

5

10

20

ABC/3TCTDF/FTC

-3.9hsCRP

5.94.0

MCP-1

5.1

-2.8OPG

6.65.2

ICAM-1

8.4

-2.0VCAM

-1

7.8

-0.4Selectin-E

-2.5

8.8

Insulin

0.050.1

D-dimer

12.6

Selectin-P

4.6

15.4

Adiponectin

-2.2

P = .26

P = .52P = .59

P = .12

P = .84P = .96 P = .13

P = .33

P = .69

P = .73

Martinez et al. IAS 2009. MOAB203

Relative to TDF/FTC, ABC/3TC does not promote mechanisms known to be involved in MI

BICOMBO Sub-study:patients with sustained viral suppression, clinical stability, and no prior CVD

1. HR: Unadjusted HR of AMI for each PY of exposure to each one of the categories2. Most recent estimated GFR (by MDRD method; carried forward) .3. Age, hypercholesterolemia, HTN, type 2 DM, and tobacco use.

Bedimo R et al. IAS 2009. MOAB202

The association between ABC and MI may be biased by chronic kidney disease

Veterans Cohort Study

0

20

40

60

80

100

%

19

85

19

86

19

87

19

88

19

89

19

90

19

91

19

92

19

93

19

94

19

95

19

96

19

97

19

98

19

99

20

00

20

01

20

02

20

03

20

04

20

05

20

06

20

07

year

aba azt d4t ddi3tc ten inf ftc

Datos del Hospital Clínic, Barcelona 2008

NRTI use different over time; patients treated with same NRTI also differ over time

NRTI use different over time; patients treated with same NRTI also differ over time

Management

Law MG, et al. 11th CROI. 2004. Abstract 737.

Duration of HAART (years)

MI

per

100

0 P

YF

U

0

1

2

3

4

5

6

7

8

< 1 1-2 2-3 3-4 4+0

Observed

Predicted

Risk of myocardial infarction in HIV-infected patients can be estimated with the Framingham score

DAD Study

Law MG et al. HIV Med 2006; 7: 218-230

P = .81

Lichtenstein K et al. CROI 2006. Abstract 735.

0

5

10

15

20

25

30

35

% p

ts o

n a

nti

-hyp

erte

nsi

ve o

r an

ti-

lipem

ic a

gen

ts

MI incidence per 1000 pt-yrs, by year

Inci

den

ce o

f M

I per

100

0 P

YF

U

0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2004 20052003

0.5

1.0

Effect HRadj95% CI P-value

LLAs 0.34 0.14–0.85 0.021

Age >40 y 2.38 0.88–6.43 0.087

Diabetes 2.45 0.99–6.05 0.052

Smoking 2.22 0.98–5.05 0.057

HOPS Cohort

The incidence of myocardial infarction can be satisfactorily modified with intervention

Assess CVD risk in next 10 years (Framingham score)

Advise on diet and lifestyle in all patientsConsider ART modification if CVD risk ≥20%

Identify key modifiable risk factorsSmoking

Blood pressure Coagulation Glucose Lipids

Drug treatment if:SBP ≥140 orDBP ≥90 mmHg(especially if 10 yearCVD risk ≥20%)

Drug treatment if:Established CVD orage ≥50 and 10 yearCVD risk ≥20%

Confirm DM anddrug treatment if:HBA1c ≥6.5%

Drug treatment if:Established CVD orT2DM orTC:HDL ratio >6 or10 year CVD risk ≥20%

TargetDM or Non-DM;CVD or no CVD:CKD+prot:<130/<80 <140/<90

TargetHBA1c <6.5%

Target Optimal StandardTC 4 5 (155) (190)LDL 2 3 (80) (115)

Target – N/A

Treat withacetylsalicylic acid

75-150 mg/d

2009 EACS guidelines on non-infectious co-morbidities in HIV: Prevention of CVD

To be released at EACS, Köln 2009.

• Higher risk of CVD in HIV + than HIV – persons:

– Rationale for CVD prevention in clinical practice

• Traditional risk factors account for a substantial portion of CV risk in HIV + persons:

– Rationale for early and aggressive intervention

• Uncontrolled HIV and other concurrent co-infections further increase the risk for CV disease:

– Rationale for early and continuous ART and for HCV and other co-infections

• Individual ARVs and classes associated with increased risk of CV disease through mechanisms not yet completely understood:

– Contribution of ART lower than traditional factors & uncontrolled infection.

– Rationale for choosing ART in patients with high CV risk (≥20%):

– PI’s associated with MI at least through lipid impact

– Controversial role of ABC: it may be a marker of increased CV disease, but causative role not demonstrated and potential mechanism not clear

Conclusions