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