Risk of myocardial infaction and exposure to specific antiretrovirals from the PI, NNRTI and NRTI drug

classes: The D:A:D Study

Jens D. Lundgren, Peter Reiss, Signe W Worm, Wafaa El-Sadr, Stephane De Wit, Rainer Weber, Antonella D‘Arminio Monforte, Ole Kirk, Eric Fontas, Francois Dabis, Matthew G. Law, Nina Friis-Møller, Andrew N. Phillips, Caroline A. Sabin

On behalf of the D:A:D study group

CROI, 2009, late breaker presentation

Background

• Prior analyses* from the D:A:D study showed an increased risk of MI associated with:

- cumulative exposure to PIs but not NNRTIs

- current/recent exposure to abacavir (ABC) or didanosine(ddI)

• Drugs within PI and NNRTI classes associated with different metabolic profiles

• Previous NRTI analyses had insufficient follow-up to assess role of tenofovir

*: D:A:D Study Group, NEJM 2007; D:A:D Study Group, Lancet 2008

Background & aim

• Prior analyses* from the D:A:D study showed an increased risk of MI associated with:

- cumulative exposure to PIs but not NNRTIs

- current/recent exposure to abacavir (ABC) or didanosine(ddI)

• Drugs within PI and NNRTI classes associated with different metabolic profiles

• Previous NRTI analyses had insufficient follow-up to assess role of tenofovir

*: D:A:D Study Group, NEJM 2007; D:A:D Study Group, Lancet 2008

Aim: Sufficient follow-up has now accrued to exploreassociations between a total of 13 drugs from these

classes and MI risk

Drugs analyses and results presented

• The D:A:D Steering committee has decided to present data only when >30,000 person-years follow-up is available from individuals ever exposed to each drug

• 4 PIs – indinavir (IDV), nelfinavir (NFV), lopinavir/r (LPR/r) and saquinavir (SQV)

• 2 NNRTIs - efavirenz (EFV) and nevirapine (NVP)

• 7 NRTIs – zidovudine (ZDV), stavudine (d4T), didanosine (ddI), zalcitabine (ddC), lamivudine(3TC), abacavir (ABC) and tenofovir (TDF)

• Other drugs with less follow-up time are included in analyses but not presented

Methods

• 33,308 patients from 11 cohorts were included

• Follow-up counted from D:A:D enrolment until the first MI event, 1st February 2008 or 6 months after the patient’s last clinic visit (whichever occurred first)

• Poisson regression models assessed the association between MI risk and exposure to each drug

• Based on previous analyses, present analyses included

• NRTIs: cumulative* and recent (current use or use within last 6 months) exposure

• NNRTIs and PIs: cumulative* exposure only

*:*:*:*: ifififif drug isdrug isdrug isdrug is discontinueddiscontinueddiscontinueddiscontinued,,,, subsequent followsubsequent followsubsequent followsubsequent follow----up continuesup continuesup continuesup continues totototo be be be be attributedattributedattributedattributed totototo level of exposurelevel of exposurelevel of exposurelevel of exposure atatatat discontinuationdiscontinuationdiscontinuationdiscontinuation

Methods

• All models include adjustment for

• demographics, cardiovascular risk factors not thought to be affected by ART and use of other ARV drugs (primary model)

• Further analyses included adjustment for

• latest measures of lipids, metabolic parameters, CD4 and HIV-RNA to identify possible mechanisms for any association

• Exposure to IDV and SQV was calculated regardless of concomitant use of ritonavir

• sensitivity analyses explored separate associations for use with and without ritonavir

With MI Without MI

Number of patients 580 32728

Male sex; % 90.7 73.8

Age (years); median 49 44 BMI>26 kg/m2; % 18.8 17.3

Current smoker; % 44.8 28.7

Ex-smoker; % 29.8 30.1 Prior CVD event; % 20.0 2.5

Family history of CHD; % 13.6 8.3

Diabetes; % 16.6 5.3 Any hypertension; % 43.5 19.2

Use of lipid-lowering medication; % 36.0 12.5 Any dyslipidaemia; % 74.8 44.3

Predicted 10-year CHD risk; %

Moderate (10-20%) 30.3 14.5 High (>20%) 18.1 4.2

Characteristics of patients at time of MI/last follow-up

Characteristics of patients ever exposed to each of the drugs under study:

% of follow% of follow% of follow% of follow----up time up time up time up time

6.56.56.56.56.06.06.06.06.96.96.96.96.96.96.96.96.46.46.46.47.87.87.87.86.26.26.26.27.07.07.07.06.06.06.06.06.86.86.86.87.67.67.67.66.76.76.76.76.26.26.26.2>20%>20%>20%>20%

14.114.114.114.111.911.911.911.914.914.914.914.915.115.115.115.113.213.213.213.215.315.315.315.315.115.115.115.114.814.814.814.812.712.712.712.714.214.214.214.215.515.515.515.514.614.614.614.613.013.013.013.010101010----20%20%20%20%

Predicted 10Predicted 10Predicted 10Predicted 10----year CHD riskyear CHD riskyear CHD riskyear CHD risk

51.551.551.551.549.449.449.449.458.658.658.658.658.058.058.058.051.851.851.851.855.855.855.855.850.150.150.150.153.453.453.453.449.249.249.249.255.055.055.055.059.059.059.059.054.754.754.754.749.849.849.849.8DyslipidaemiaDyslipidaemiaDyslipidaemiaDyslipidaemia

16.916.916.916.916.416.416.416.417.517.517.517.517.117.117.117.116.116.116.116.117.817.817.817.819.119.119.119.117.617.617.617.615.915.915.915.916.616.616.616.616.516.516.516.516.516.516.516.516.016.016.016.0HypertensionHypertensionHypertensionHypertension

6.16.16.16.15.35.35.35.36.16.16.16.15.55.55.55.55.45.45.45.46.36.36.36.36.06.06.06.06.36.36.36.35.15.15.15.16.06.06.06.07.37.37.37.36.06.06.06.05.25.25.25.2DiabetesDiabetesDiabetesDiabetes

2.42.42.42.42.52.52.52.52.52.52.52.52.42.42.42.42.42.42.42.42.82.82.82.82.52.52.52.52.82.82.82.82.22.22.22.22.42.42.42.43.03.03.03.02.42.42.42.42.22.22.22.2Previous CVDPrevious CVDPrevious CVDPrevious CVD

9.09.09.09.08.48.48.48.49.09.09.09.08.58.58.58.58.88.88.88.88.88.88.88.89.09.09.09.09.39.39.39.38.48.48.48.48.68.68.68.67.87.87.87.88.78.78.78.78.58.58.58.5Family history of CHDFamily history of CHDFamily history of CHDFamily history of CHD

25.225.225.225.229.429.429.429.427.927.927.927.928.028.028.028.023.723.723.723.725.525.525.525.530.730.730.730.726.826.826.826.825.725.725.725.724.924.924.924.925.425.425.425.424.724.724.724.725.925.925.925.9ExExExEx----smokersmokersmokersmoker

33.433.433.433.430.430.430.430.435.235.235.235.233.133.133.133.136.836.836.836.837.237.237.237.231.431.431.431.433.333.333.333.334.634.634.634.637.037.037.037.038.938.938.938.938.238.238.238.235.435.435.435.4Current smokerCurrent smokerCurrent smokerCurrent smoker

18.618.618.618.618.818.818.818.816.116.116.116.117.717.717.717.717.917.917.917.916.716.716.716.718.418.418.418.417.917.917.917.918.918.918.918.916.316.316.316.314.114.114.114.116.016.016.016.018.618.618.618.6BMI>26 kg/mBMI>26 kg/mBMI>26 kg/mBMI>26 kg/m2222

35.735.735.735.735.735.735.735.739.239.239.239.237.737.737.737.732.732.732.732.738.238.238.238.239.039.039.039.038.238.238.238.234.134.134.134.135.535.535.535.539.039.039.039.035.635.635.635.634.634.634.634.6>45 years (men)/ >55 years >45 years (men)/ >55 years >45 years (men)/ >55 years >45 years (men)/ >55 years (women)(women)(women)(women)

75.775.775.775.774.074.074.074.078.178.178.178.177.077.077.077.071.571.571.571.577.577.577.577.575.175.175.175.175.775.775.775.774.674.674.674.675.175.175.175.177.077.077.077.075.475.475.475.474.474.474.474.4MenMenMenMen

% of FU contributed by:% of FU contributed by:% of FU contributed by:% of FU contributed by:

EFVEFVEFVEFVNVPNVPNVPNVPSQVSQVSQVSQVLPVLPVLPVLPVNFVNFVNFVNFVIDVIDVIDVIDVTDFTDFTDFTDFABCABCABCABC3TC3TC3TC3TCd4Td4Td4Td4TddCddCddCddCddIddIddIddIZDVZDVZDVZDVEver exposed to:Ever exposed to:Ever exposed to:Ever exposed to:

NNRTIsNNRTIsNNRTIsNNRTIsPIPIPIPINRTIsNRTIsNRTIsNRTIsClass of drugClass of drugClass of drugClass of drug

6.56.56.56.56.06.06.06.06.96.96.96.96.96.96.96.96.46.46.46.47.87.87.87.86.26.26.26.27.07.07.07.06.06.06.06.06.86.86.86.87.67.67.67.66.76.76.76.76.26.26.26.2>20%>20%>20%>20%

14.114.114.114.111.911.911.911.914.914.914.914.915.115.115.115.113.213.213.213.215.315.315.315.315.115.115.115.114.814.814.814.812.712.712.712.714.214.214.214.215.515.515.515.514.614.614.614.613.013.013.013.010101010----20%20%20%20%

Predicted 10Predicted 10Predicted 10Predicted 10----year CHD riskyear CHD riskyear CHD riskyear CHD risk

51.551.551.551.549.449.449.449.458.658.658.658.658.058.058.058.051.851.851.851.855.855.855.855.850.150.150.150.153.453.453.453.449.249.249.249.255.055.055.055.059.059.059.059.054.754.754.754.749.849.849.849.8DyslipidaemiaDyslipidaemiaDyslipidaemiaDyslipidaemia

16.916.916.916.916.416.416.416.417.517.517.517.517.117.117.117.116.116.116.116.117.817.817.817.819.119.119.119.117.617.617.617.615.915.915.915.916.616.616.616.616.516.516.516.516.516.516.516.516.016.016.016.0HypertensionHypertensionHypertensionHypertension

6.16.16.16.15.35.35.35.36.16.16.16.15.55.55.55.55.45.45.45.46.36.36.36.36.06.06.06.06.36.36.36.35.15.15.15.16.06.06.06.07.37.37.37.36.06.06.06.05.25.25.25.2DiabetesDiabetesDiabetesDiabetes

2.42.42.42.42.52.52.52.52.52.52.52.52.42.42.42.42.42.42.42.42.82.82.82.82.52.52.52.52.82.82.82.82.22.22.22.22.42.42.42.43.03.03.03.02.42.42.42.42.22.22.22.2Previous CVDPrevious CVDPrevious CVDPrevious CVD

9.09.09.09.08.48.48.48.49.09.09.09.08.58.58.58.58.88.88.88.88.88.88.88.89.09.09.09.09.39.39.39.38.48.48.48.48.68.68.68.67.87.87.87.88.78.78.78.78.58.58.58.5Family history of CHDFamily history of CHDFamily history of CHDFamily history of CHD

25.225.225.225.229.429.429.429.427.927.927.927.928.028.028.028.023.723.723.723.725.525.525.525.530.730.730.730.726.826.826.826.825.725.725.725.724.924.924.924.925.425.425.425.424.724.724.724.725.925.925.925.9ExExExEx----smokersmokersmokersmoker

33.433.433.433.430.430.430.430.435.235.235.235.233.133.133.133.136.836.836.836.837.237.237.237.231.431.431.431.433.333.333.333.334.634.634.634.637.037.037.037.038.938.938.938.938.238.238.238.235.435.435.435.4Current smokerCurrent smokerCurrent smokerCurrent smoker

18.618.618.618.618.818.818.818.816.116.116.116.117.717.717.717.717.917.917.917.916.716.716.716.718.418.418.418.417.917.917.917.918.918.918.918.916.316.316.316.314.114.114.114.116.016.016.016.018.618.618.618.6BMI>26 kg/mBMI>26 kg/mBMI>26 kg/mBMI>26 kg/m2222

35.735.735.735.735.735.735.735.739.239.239.239.237.737.737.737.732.732.732.732.738.238.238.238.239.039.039.039.038.238.238.238.234.134.134.134.135.535.535.535.539.039.039.039.035.635.635.635.634.634.634.634.6>45 years (men)/ >55 years >45 years (men)/ >55 years >45 years (men)/ >55 years >45 years (men)/ >55 years (women)(women)(women)(women)

75.775.775.775.774.074.074.074.078.178.178.178.177.077.077.077.071.571.571.571.577.577.577.577.575.175.175.175.175.775.775.775.774.674.674.674.675.175.175.175.177.077.077.077.075.475.475.475.474.474.474.474.4MenMenMenMen

% of FU contributed by:% of FU contributed by:% of FU contributed by:% of FU contributed by:

EFVEFVEFVEFVNVPNVPNVPNVPSQVSQVSQVSQVLPVLPVLPVLPVNFVNFVNFVNFVIDVIDVIDVIDVTDFTDFTDFTDFABCABCABCABC3TC3TC3TC3TCd4Td4Td4Td4TddCddCddCddCddIddIddIddIZDVZDVZDVZDVEver exposed to:Ever exposed to:Ever exposed to:Ever exposed to:

NNRTIsNNRTIsNNRTIsNNRTIsPIPIPIPINRTIsNRTIsNRTIsNRTIsClass of drugClass of drugClass of drugClass of drug

Characteristics of patients ever exposed to each of the drugs under study:

% of follow% of follow% of follow% of follow----up time up time up time up time

% of follow-up time at moderate and high CV risk after ever exposed to each of the drugs

0000

5555

10101010

15151515

20202020

25252525

ZDVZDVZDVZDV ddIddIddIddI ddCddCddCddC d4Td4Td4Td4T 3TC3TC3TC3TC ABCABCABCABC TDFTDFTDFTDF IDVIDVIDVIDV NFVNFVNFVNFV LPVLPVLPVLPV SQVSQVSQVSQV NVPNVPNVPNVP EFVEFVEFVEFV

Moderate risk* (10-20%) Moderate risk* (10-20%) Moderate risk* (10-20%) Moderate risk* (10-20%) High risk* (>20%)High risk* (>20%)High risk* (>20%)High risk* (>20%)% of follow-up time

*10 year predicted coronary heart disease risk (Framingham)

NRTIs and risk of MI:recent* exposure to each drug

* recent use= current or within the last 6 months; **: not shown (low number of patient currently on ddC)

139221554405148331523#MI:

39,15753,300152,00995,32029,67674,407138,109#PYFU:

TDFABC3TCd4TddCddIZDV

1

1.9

1.5

1.2

0.8

0.6

RRyes/no95%CI **

NRTIs and risk of MI:recent* and cumulative exposure

* recent use= current or within the last 6 months **: not shown (low number of patient currently on ddC)

139221554405148331523#MI:

39,15753,300152,00995,32029,67674,407138,109#PYFU:

TDFABC3TCd4TddCddIZDV

RRyes/no95%CI 1

1.9

1.5

1.2

0.8

0.6

1

1.9

1.5

1.2

0.8

0.6

RR per year95%CI**

PIs/NNRTIs and risk of MI:cumulative exposure to each drug

*: Approximate test for heterogeneity: P=0.02

221

44,657

SAQ EFVNVPLPV/rNFVIDV

228

61,855

221150197298#MI:

58,94637,13656,52968,469#PYFU:

NNRTIPI*

0.9

1

1.13

1.2

RR per year95%CI

1.1

LPV/rLPV/rLPV/rLPV/r and IDV: Effect of adjustment for latest metabolic factors on estimates of risk

Lopinavir/rIndinavir

1 1.3

Primary model

1.1 1.4

RR of MI 95%CI

LPV/rLPV/rLPV/rLPV/r and IDV: Effect of adjustment for latest metabolic factors on estimates of risk

1111 1.31.31.31.3

Primary model

Dyslipidaemia

Elevated blood pressure

Diabetes mellitus

Lipodystrophy

Glucose

All above+lipid-lowering and antihypertensive medication

1.11.11.11.1 1.41.41.41.4

RR of MI 95%CI

Lopinavir/rIndinavir

Impact of concomitant use of ritonavir onassociation between SQV and IDV on risk of MI

# PYFU:

44,657 24,72726,145

68,46922,18657,961 Indinavir

Indinavir+rtvIndinavir (overall)

Saquinavir

Saquinavir+rtvSaquinavir (overall)

Adjusted RR* of MI 95% CI

1 1.1 1.2 1.3 1.4

*: Primary model

Conclusion

• Abacavir, didanosine, indinavir and lopinavir/r associated with an increased risk of MI

• No significant associations between exposure to other NRTIs including tenofovir, either of the NNRTIs nor 2 other PIs (SQV/NFV) and the risk of MI

• For indinavir and lopinavir/r,

• the risk increased with cumulative exposure,

• partly explained by dyslipidaemia

• Concomitant use of ritonavir did not appear to modify the effects of saquinavir or indinavir

Discussion

• Findings stress the need to continue to examine the associations with individual drugs, as drugs from the same class may confer different risks of MI

• Given the observational design and the possibility of unmeasured confounding, cautious interpretation warranted

• Additional follow-up required to examine the risk of MI for more recently introduced PIs

Acknowledgements

• Cohort Cohort Cohort Cohort PIPIPIPI’’’’ssss:::: W E-Sadr * (CPCRA), G Calvo * (BASS), F Dabis * (Aquitaine), O Kirk *(EuroSida), M Law * (AHOD), A d’Arminio Monforte * (ICONA), L Morfeldt *(HivBIVUS), C Pradier * (Nice), P Reiss * (ATHENA), R Weber * (SHCS), S De Wit *(Brussels)

• Cohort coordinators and Cohort coordinators and Cohort coordinators and Cohort coordinators and datamanagersdatamanagersdatamanagersdatamanagers:::: S Zaheri, L Gras (ATHENA), M Bruyand, S Geffard (Aquitaine), K Petoumenos (AHOD), S Mateu, F Torres (BASS), M Delforge(Brussels), G Bartsch, G Thompsen (CPCRA), J Kjær (EuroSIDA), P Pezzotti (ICONA), E Fontas, C Caissotti (Nice), A Sundström, G Thulin (HivBIVUS), M Rickenbach, O Keiser (SHCS)

• Statisticians:Statisticians:Statisticians:Statisticians: CA Sabin *, AN Phillips *

• Community representative:Community representative:Community representative:Community representative: S Collins *

• DAD coordinating office:DAD coordinating office:DAD coordinating office:DAD coordinating office: N Friis-Møller, S Worm, A Sawitz, JD Lundgren *¢

• Steering Committee:Steering Committee:Steering Committee:Steering Committee: Members indicated w/*; ¢ chair;

Additional members: S Storfer *, D Pizzuti *, I Weller *

• Funding: Funding: Funding: Funding: ‘Oversight Committee for The Evaluation of Metabolic Complications of HAART’ with representatives from academia, patient community, FDA, EMEA and a consortium of ”Abbott, Boehringer Ingelheim, Bristol-Myers Squibb, Gilead Sciences, GlaxoSmithKline, Merck, Pfizer, Roche, and Tibotec”

Back-up slides

Unadjusted rates of MI according to duration of exposure to 4 PIs and 2 NNRTIs

Results – PIs and NNRTIstest of heterogeneity

• Purpose: to assess whether some drugs have a greater impact on the risk of MI than others

• Approach 1: Comparison of the model containing the separate PI effects to one that included a single PI ‘class’ effect only

• Finding: p=0.01 suggested that there is some heterogeneity between the effects

• Approach 2: Akaike’s Information Criteria (AIC) also suggested that a model including separate PI effects was more useful than a model including the class effect only