Impact of the Data Collection on Adverse Events of Anti-HIV Drugs Cohort Study on Abacavir Prescription among Treatment-Naive,HIV-Infected Patients in Canada

Abstract

Objective:

To evaluate the trends in abacavir (ABC) prescription among antiretroviral (ARV) medication-naive individuals following the presentation of the Data Collection on Adverse Events of Anti-HIV Drugs (DAD) cohort study.

Methods:

We conducted a retrospective cohort study of ARV medication-naive individuals in the Canadian Observational Cohort (CANOC).

Results:

Between January 1, 2000, and February 28, 2010, a total of 7280 ARV medication-naive patients were included in CANOC. We observed a significant change in the proportion of new ABC prescriptions immediately following the release of DAD (−11%; 95% confidence interval [CI]: −20% to −2.4%) and in the months following the presentation of these data (−0.66% per month; 95% CI: −1.2% to −0.073%). A post-DAD presentation decrease in the odds of being prescribed ABC versus tenofovir (TDF) was observed (adjusted odds ratio, 0.72 per year, 95% CI: 0.54-0.97).

Conclusions:

Presentation of the DAD was associated with a significant decrease in ABC use among ARV medication-naive, HIV-positive patients initiating therapy.

Keywords

drug utilization pharmacoepidemiology HIV abacavir

Introduction

Initially presented at the 2008 Conference on Retroviruses and Opportunistic Infections, the findings from the Data Collection on Adverse Events of Anti-HIV Drugs (DAD) cohort study implicating recent (ie, within past 6 months) exposure to abacavir (ABC) with a heightened risk of myocardial infarction among patients living with HIV contributed to a reappraisal of the position of this drug within the antiretroviral (ARV) armamentarium.¹ Specifically, ABC was eventually relegated to second-line status among the nucleoside reverse transcriptase inhibitor (NRTI) class in both the Department of Health and Human Services (DHHS) and the International AIDS Society-USA treatment guidelines and was assigned a caveat of cautious use in patients with a higher baseline risk of cardiovascular disease.^2,3 Although the results of the DAD were corroborated by those obtained from the Strategies for Management of Antiretroviral Therapy (SMART) study,⁴ subsequent analyses did not support a relationship between ABC exposure and cardiac disease, reinforcing concerns that the initially reported association may have been obfuscated by residual confounding and selection bias.^5–8

Although the impact of high-quality, well-publicized randomized trials has been demonstrated to lead to rapid and significant changes in prescribing practices, it is less clear whether observational studies can exert similar effects on clinical practice.^9–11 Given the widespread promulgation of the DAD results within media directed toward persons living with HIV and clinicians managing these patients, it is possible that these data influenced physician prescribing behavior and ABC utilization despite the aforementioned concerns about confounding and bias. We therefore sought to evaluate the trends in the use of ABC among ARV medication-naive, HIV-infected patients in the periods preceding and following the initial presentation of the DAD findings.

Methods

Cohort Description

The Canadian Observational Cohort (CANOC) collaboration is a multisite cohort study of ARV medication-naive HIV-positive patients initiating combination antiretroviral therapy (cART) after January 1, 2000.¹² The collaboration was established in March 2008 and currently includes 8 participating centers across Canada. Eligibility criteria for inclusion in CANOC are documented HIV infection, residence in Canada, age 18 years and older, naive initiation of combination antiretroviral therapy (cART; defined as at least 3 individual ARV agents) and at least 1 HIV-1 RNA viral load and CD4 count measurement within 6 months of initiating cART. Patient selection and data extraction are performed locally at the data centers of the participating sites and are subsequently pooled and analyzed at the Project Data Centre in Vancouver, British Columbia, Canada. All participating sites have received approval from their respective institutional research ethics boards to contribute nonnominal patient-specific data.

Analysis

We examined monthly trends in the proportion of new prescriptions for ABC and tenofovir (TDF) in the 97 months preceding the presentation of the DAD and the 24 months immediately following the release of these data using segmented ordinary least squares regression, a type of interrupted time series model.¹³ For each monthly interval, we calculated the proportion of new prescriptions for ABC using all new prescriptions for ABC and ABC fixed-dose combination products (ie, ABC/lamivudine [3TC]; [Kivexa], ABC/3TC/zidovudine [Trizivir]) as the numerator and total new NRTI prescriptions as the denominator. Similarly, we calculated the proportion of new prescriptions for TDF using all new prescriptions for TDF and TDF fixed-dose combination products (ie, TDF/emtricitabine (Truvada) or TDF/emtricitabine/efavirenz [EFV]; [Atripla]) as the numerator and total new NRTI prescriptions as the denominator. We chose to contrast trends in ABC with those of TDF, as the latter agent is the most clinically relevant NRTI alternative to ABC.

The regression model included an intercept, a linear trend variable, a dummy variable representing the month in which the DAD findings were first presented (February 2008), and a trend variable for the months following the presentation of these data. Specifically, the regression model took the form:

\begin{aligned} Y_{t} = β_{0} + β_{1} \times t i m e + β_{2} \times i n t e r v e n t i o n + β_{3} \\ \times t i m e a f t e r i n t e r v e n t i o n + e_{t} \end{aligned}

where Y_t is the proportion of all NRTI prescriptions per month for ABC at time t. In this model, β₀ is an estimate of the proportion of ABC prescriptions per month in 2000, β₁ is an estimate of the change in the proportion of prescriptions per month before the release of the DAD results, β₂ is an estimate of the change in proportion of ABC prescriptions immediately following the presentation of the DAD, and β₃ is an estimate of the change per month in the proportion of ABC prescriptions after the presentation of the DAD, where time after intervention is a continuous variable indicating the number of months that have passed since February 2008. We fit a similar model to determine whether the release of the DAD findings had an impact on the proportion of new prescriptions for TDF. We had hypothesized that there would be a decrease in ABC utilization and a corresponding increase in TDF initiation following the release of the DAD. We evaluated the adequacy of the models using standard graphical and statistical methods. Because time-series data are often autocorrelated, we used the Durbin-Watson statistic to test for autocorrelation when conducting the segmented regression analysis. Since second-order autocorrelation was evident, the final model included 2 terms to accommodate the dependence of the model residuals.

We complemented the segmented regression analysis with a structural break analysis to detect changes in the relationship between new ABC prescriptions and time. The timing of all breakpoints and associated 95% confidence intervals (CIs) occurring over the time series were estimated using the methodology described by Bai and Perron.¹⁴ Identified breakpoints were compared to the timing of the release of the DAD (February 2008) and the approval of competing fixed-dose ARV formulations, such as Truvada (TDF/emtricitabine; approved October 2007) and Atripla (TDF/emtricitabine/EFV; approved April 2006). We hypothesized that a structural break in the ABC time series would be found in the months following the presentation of the DAD, and that the 95% CIs would not encompass the periods in which competing ARV agents were released, thereby lending support for an association between a change in prescribing practice and release of the DAD findings.

In order to fully examine the impact of DAD on current prescribing practices, we also used multivariable logistic regression analysis to examine temporal trends in the odds of being prescribed ABC versus its leading competitor, TDF, adjusting for patient-level characteristics that may influence the choice of NRTI. We examined new prescriptions of either ABC or TDF from April 2004 (TDF approval in Canada) to February 2010. The change in prescribing patterns was modeled with a linear term for month (representing the monthly change in the odds of being prescribed ABC versus TDF), a dichotomous term for the DAD, representing the change in ABC prescription immediately after presentation of these findings, and a linear term for months since DAD release representing the change in slope that occurred after the release of the DAD findings. We adjusted our models for pertinent sociodemographic and clinical variables that may influence the choice of prescribing either ABC or TDF, including patient sex, province, age, and baseline values of viral load and CD4 count. We hypothesized that the odds of being prescribed ABC relative to TDF would decrease in the year following the release of the DAD.

All analyses were performed using either SAS version 9.3 (SAS Institute, Cary, North Carolina) or R version 2.12.1.

Results

Between January 1, 2000, and February 28, 2010, a total of 7280 ARV medication-naive patients were included in CANOC, of whom 1294 and 2985 initiated therapy with ABC and TDF, respectively. The baseline characteristics of these patients are summarized in Table 1.

Table 1.

Overall Baseline Characteristics of Study Cohort.

Variables	Total (n = 7280)
Demographics
Age	40.0 (34.0-46.0)
<35	2024 (28%)
35-45	3213 (44%)
>45	2043 (28%)
Sex
Male	5895 (81%)
Female	1379 (19%)
HIV risk factors
Men who have sex with men	2532 (49%)
Injection drug use	1646 (32%)
From endemic country	643 (12%)
Province of origin
British Columbia	3272 (45%)
Ontario	2439 (34%)
Quebec	1569 (22%)
Ethnicity
Caucasian	1937 (60%)
Black	472 (15%)
Aboriginal	393 (12%)
Mixed	151 (5%)
Other	302 (9%)
Year of first ARV	2005 (2003-2008)
Clinical
Baseline CD4 count, cells/mm³	200 (100-290)
<200	3607 (50%)
200-350	2535 (35%)
350-500	799 (11%)
>500	339 (5%)
Baseline viral load, log₁₀copies/mL	4.9 (4.3-5.1)
<25 000 copies/mL	1992 (27%)
25 000-100 000 copies/mL	2190 (30%)
>100 000 copies/mL	3098 (43%)
Viral load < 50 copies/mL	202 (3%)
Presence of ADI at baseline	991 (14%)
Hepatitis B+ at baseline	435 (6%)
Hepatitis C+ at baseline	1802 (25%)
NRTI
Abacavir	1294 (18%)
Tenofovir	2985 (41%)
Other	3001 (41%)

Abbreviations: ADI, AIDS-defining illness; ARV, antiretroviral; NRTI, nucleoside reverse transcriptase inhibitor. N (%) or median, interquartile range.

Trends in the proportions of new prescriptions for ABC and TDF are illustrated in Figure 1. In the segmented regression analysis, a significant segment change was observed in the proportion of new ABC prescriptions immediately following the release of the DAD (−11%; 95% CI: −20% to −2.4%). A significant negative trend was observed in the proportion of new prescriptions for ABC in the months following the release of the DAD (−0.66% per month; 95% CI: −1.2% to −0.073%). Respective results for the subset of patients with a viral load of less than 100 000 copies/mL were −7.7% (95% CI: −17.6% to 2.12%) and −0.69% (95% CI: −1.36% to −0.03%) per month. This result is consistent with the results of the structural break analysis, in that a statistically significant downward break in the ABC prescription time series was detected on May 2008 (95% CI, April 2008 to January 2009). An earlier upward break was also detected on July 2004 (95% CI: August 2003 to October 2004). The results of the segmented regression analysis for TDF indicated that there was no statistically significant change in the proportion of new prescriptions for this agent either immediately following the release of the DAD (6.3%; 95% CI: −4.6% to 17%) or in the months following the release of these findings (−0.52% per month; 95% CI: −1.3% to 0.24%).

Figure 1.

Temporal trend in prescriptions for abacavir and tenofovir.

Between April 2004 and February 2010, a total of 3927 patients were prescribed either ABC (n = 979) or TDF (n = 2948). The baseline characteristics of these patients are summarized in Table 2. A total of 2693 patients were included in the multivariable logistic regression analysis examining predictors of being prescribed ABC versus TDF (Table 3). Following multivariable adjustment, we observed a significant decrease in the odds of being prescribed ABC versus TDF before the release of the DAD findings (adjusted odds ratio [aOR] 0.83 per year, 95% CI: 0.74-0.93). The presentation of the DAD was associated with an immediate change in new ABC use (aOR 0.51, 95% CI: 0.36-0.72), while a significant postpresentation decrease in ABC use continued (aOR 0.72 per year, 95% CI: 0.54-0.97).

Table 2.

Baseline Characteristics of Individuals Prescribed Abacavir Versus Tenofovir from April 2004 to February 2010.

Variables	Total (n = 3927)	Tenofovir (n = 2948)	Abacavir (n = 979)	P value
Demographics
Age	41.0 (34.0-47.0)	41.0 (35.0-47.0)	40.0 (34.0-46.0)	.04
<35	1314 (27%)	728 (25%)	272 (28%)	.04
35-45	2087 (43%)	1288 (44%)	436 (45%)
>45	1437 (30%)	932 (32%)	271 (28%)
Sex
Male	3948 (82%)	2520 (86%)	798 (82%)	<.01
Female	884 (18%)	422 (14%)	181 (18%)
HIV risk factors
Men who have sex with men	1689 (50%)	1029 (52%)	386 (55%)	.12
Injection drug use	1042 (31%)	706 (35%)	148 (21%)	<.0001
From endemic country	401 (12%)	159 (8%)	109 (16%)	<.0001
Province of origin
British Columbia	2165 (45%)	1628 (55%)	264 (27%)	<.0001
Ontario	1640 (34%)	798 (27%)	460 (47%)
Quebec	1033 (21%)	522 (18%)	255 (26%)
Ethnicity
Caucasian	1139 (58%)	698 (59%)	205 (55%)	<.0001
Black	300 (15%)	131 (11%)	81 (22%)
Aboriginal	251 (13%)	169 (14%)	26 (7%)
Mixed	83 (4%)	48 (4%)	17 (5%)
Other	207 (10%)	128 (11%)	45 (12%)
Year of first ARV	2007 (2006-2008)	2008 (2006-2009)	2007 (2005-2008)	<.0001
Clinical
Baseline CD4 count, cells/mm³	215 (120-293)	220 (130-293)	220 (132-300)	.40
<200	2146 (44%)	1268 (43%)	409 (42%)	.89
200-350	1904 (39%)	1208 (41%)	406 (41%)
350-500	562 (12%)	344 (12%)	121 (12%)
>500	226 (5%)	128 (4%)	43 (4%)
Baseline viral load, log₁₀ copies/mL	4.8 (4.3-5.0)	4.8 (4.3-5.0)	4.8 (4.3-5.1)	.19
<25 000 copies/mL	1445 (30%)	859 (29%)	280 (29%)	.40
25 000-100 000 copies/mL	1451 (30%)	908 (31%)	284 (29%)
>100 000 copies/mL	1942 (40%)	1181 (40%)	415 (42%)
Viral load < 50 copies/mL	146 (3%)	47 (2%)	37 (4%)	<.0001
Presence of ADI at baseline	580 (12%)	339 (11%)	119 (12%)	.58
Hepatitis B+ at baseline	271 (6%)	168 (6%)	49 (5%)	.41
Hepatitis C+ at baseline	1115 (23%)	746 (25%)	166 (17%)	<.0001

Abbreviations: ARV, antiretroviral; ADI, AIDS-defining illness. N (%) or median, interquartile range.

Table 3.

Multivariable Logistic Regression Examining Odds of Being Prescribed Abacavir Versus Tenofovir (n = 2693).

Variable	Odds ratio	95% confidence interval	P value
Province
British Columbia	1	–	–
Ontario and Quebec	4.99	(3.87, 6.44)	<.0001
Injection drug use as risk factor	1.04	(0.80, 1.35)	.78
Age
<35 years old	1	–	–
35-45	0.89	(0.71,1.12)	.34
>45	0.91	(0.71,1.18)	.48
Sex
Male	1	–	–
Female	1.80	(1.41,2.31)	<.0001
Presence of ADI at baseline	1.14	(0.86,1.53)	.36
Time
Per year prior to DAD	0.83	(0.74,0.93)	.001
DAD	0.51	(0.36,0.72)	.03
Per year post DAD	0.72	(0.54,0.97)	.0001
Baseline CD4 count, cells/mm³
<200	1	–	–
200-350	1.30	(1.05,1.62)	.02
>350	1.32	(0.98,1.78)	.07
Baseline viral load, copies/mL
<25 000	1	–	–
25 000-1 00 000	0.99	(0.77,1.28)	.96
>1 00 000	0.96	(0.76,1.23)	.76

Abbreviations: ARV, antiretroviral; DAD, Data Collection on Adverse Events of Anti-HIV Drugs.

Discussion

In our study, we observed a steady decrease in the proportion of new ABC prescriptions being written for ARV medication-naive, HIV-infected patients initiating therapy in the months following the release of the DAD. The results of the segmented regression analysis were corroborated by the structural break analysis, which provided additional evidence of a decrease in new ABC prescriptions in May 2008, shortly following the presentation of the DAD findings. We did not observe changes in the ABC time series that could correspond to the approval of competitor combination products such as Truvada (approved April 2006) and Atripla (approved October 2007). The results of the logistic regression analysis lend further credence to our interpretation. Although the logistic regression model suggests that new prescription of ABC was decreasing before the release of the DAD, this trend became more pronounced following the release of these data. Although changes in prescribing practice have been documented to occur following the release of randomized controlled trials and meta-analyses, we are unaware of a contemporary example where the presentation of findings from a single cohort study has fueled a sustained change in physician prescribing behavior.

Our results have important implications. Most notably, we have demonstrated that a well-publicized cohort study can precipitate changes in prescribing practice. Despite the gravity of the DAD findings, the observed association between ABC exposure and myocardial infarction remained difficult to interpret because of the inability to adjust for variables that could be important confounders of this relationship. Although subsequently corroborated by the results of the SMART study, other investigators were not able to replicate the DAD findings once potential confounders such as intravenous drug use or chronic kidney disease were incorporated into analyses. For example, in a nested case–control study analyzing the impact of individual ARV agents on the risk of myocardial infarction in hospitalized patients with HIV infection, an association with both recent and past ABC use and the outcome of interest was no longer observed after the analysis was restricted to nonusers of cocaine and intravenous drugs.⁶ Similarly, no association was found between the use of ABC and myocardial infarction following adjustments for chronic kidney disease and traditional cardiovascular risk factors in a study using the Veterans Administration Clinical Case Registry.⁷ However, studies documenting an association between ABC and cardiovascular disease following adjustment for kidney function have also been published, challenging the confounding role of this covariate and contributing to the uncertainity in the literature.^15,16 In addition to the manner in which potential confounders were considered, discrepant results across the studies could be attributable to the different patient populations examined. Specifically, studies finding an association between ABC use and myocardial infarction included predominantly ARV medication-experienced patients,^1,4 whereas those in which no association could be found assessed mainly ARV medication-naive patients.^6,7,17,18 Nonetheless, irrespective of the limitations of the DAD and the fact that the majority of patients in the study were ARV medication experienced, our study suggests that clinicians acted upon the release of the DAD findings and changed their prescribing practices when starting treatment for ARV medication-naive patients.

Several limitations of our study should be noted. First, because of the varied determinants of medical practice, it is difficult to infer conclusively that changes in ABC prescription could be attributable to the DAD. However, we used a combination of ecological and individual-level analyses to corroborate our findings. In addition, our study was retrospective in nature, and we had limited data available regarding comorbidities. We were therefore unable to examine whether the baseline presence of cardiovascular and/or renal disease modified the impact of the DAD on ABC initiation, and whether the observed changes in prescribing practice resulted in important clinical outcomes. Furthermore, we analyzed only the impact of the DAD on ABC initiation and did not consider switches from ABC to TDF in the months following the presentation of the DAD. However, this likely underestimates the impact of the DAD on ABC use, as we expect that some clinicians may have substituted this drug for an agent perceived to confer a smaller risk of myocardial infarction among their patients with risk factors for or a history of cardiac disease. Because of a lack of statistical power, we were unable to ascertain the impact of the presentation of the ACTG5202 findings and the change in status of ABC to an alternative agent in the DHHS guidelines in August 2008 and November 2008, respectively.^3,19 Finally, it is impossible for our analyses to separate the effects of the presentation of DAD to clinicians by professional publications and conference reports from the marketing efforts of pharmaceutical companies expected to benefit from the promotion of these findings. Our results should therefore be interpreted as reflecting the aftermath of all combined efforts at disseminating these findings. However, the role of the pharmaceutical industry in conflating research dissemination with product promotion warrants further examination in the field of HIV, particularly when the research in question is limited in its ability to conclusively establish cause–effect relationships.

In conclusion, we found that the presentation of the DAD study was associated with a significant decrease in the use of ABC among ARV medication-naive, HIV-positive patients initiating therapy. The change in ABC prescription trends were observed despite methodological concerns and the lack of external generalizability to the ARV medication-naive patients comprising the CANOC. Our study therefore suggests that shifts in physician prescribing practice are possible following the presentation and promotion of data from a single, well-publicized observational study. However, the subsequent publication of studies, which both contradict and corroborate the results of the DAD, serves as a reminder that conclusions drawn from observational studies must be interpreted with the potential for residual confounding and selection bias in mind and therefore incorporated into clinical practice cautiously.

Footnotes

Acknowledgments

We would like to thank all the participants for allowing their information to be a part of the CANOC collaboration. The CANOC Collaboration includes: Community Investigators: Sean Hosein and Shari Margolese. Investigators: Gloria Aykroyd (Ontario HIV Treatment Network, OHTN), Louise Balfour (University of Ottawa, OHTN Cohort Study, OCS Co-Investigator), Ahmed Bayoumi (University of Toronto, OCS Co-Investigator), Ann Burchell (Ontario HIV Treatment Network), John Cairney (University of Toronto, OCS Co-Investigator), Liviana Calzavara (University of Toronto, OCS Co-Investigator), Angela Cescon (British Columbia Centre for Excellence in HIV/AIDS), Curtis Cooper (University of Ottawa, OCS Co-Investigator), Kevin Gough (University of Toronto, OCS Co-Investigator), Silvia Guillemi (British Columbia Centre for Excellence in HIV/AIDS, University of British Columbia), P. Richard Harrigan (British Columbia Centre for Excellence in HIV/AIDS, University of British Columbia), Marianne Harris (British Columbia Centre for Excellence in HIV/AIDS), George Hatzakis (McGill University), Robert Hogg (British Columbia Centre for Excellence in HIV/AIDS, Simon Fraser University), Don Kilby (University of Ottawa, Ontario HIV Treatment Network), Marina Klein (Montreal Chest Institute Immunodeficiency Service Cohort, McGill University), Richard Lalonde (The Montreal Chest Institute Immunodeficiency Service Cohort and McGill University), Viviane Lima (British Columbia Centre for Excellence in HIV/AIDS, University of British Columbia), Mona Loutfy (University of Toronto, Maple Leaf Medical Clinic, OCS Co-Investigator), Nima Machouf (Clinique Medicale l’Actuel, Université de Montréal), Ed Mills (British Columbia Centre for Excellence in HIV/AIDS, University of Ottawa), Peggy Millson (University of Toronto, OCS Co-Investigator), Julio Montaner (British Columbia Centre for Excellence in HIV/AIDS, University of British Columbia), David Moore (British Columbia Centre for Excellence in HIV/AIDS, University of British Columbia), Alexis Palmer (British Columbia Centre for Excellence in HIV/AIDS), Janet Raboud (University of Toronto, University Health Network, OCS Co-investigator), Anita Rachlis (University of Toronto, OCS Co-Investigator), Stanley Read (University of Toronto, OCS Co-Investigator), Sean Rourke (Ontario HIV Treatment Network, University of Toronto), Marek Smieja (McMaster University, OCS Co-Investigator), Irving Salit (University of Toronto, OCS Co-Investigator), Darien Taylor (Canadian AIDS Treatment Information Exchange OCS Co-Investigator), Benoit Trottier (Clinique Medicale l’Actuel, Université de Montréal), Chris Tsoukas (McGill University), Sharon Walmsley (University of Toronto, OCS Co-Investigator), and Wendy Wobeser (Queens University, OCS Co-Investigator).

Analysts and Staff: Svetlana Draskovic (British Columbia Centre for Excellence in HIV/AIDS), Mark Fisher (OHTN), Sandra Gardner (University of Toronto), Nada Gataric (British Columbia Centre for Excellence in HIV/AIDS), David Milan (British Columbia Centre for Excellence in HIV/AIDS), Sergio Rueda (OHTN), and Benita Yip (British Columbia Centre for Excellence in HIV/AIDS).

Authors’ Note

All additional research team members of CANOC collaboration are listed in Acknowledgment.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr Antoniou has received unrestricted research grants from Merck Frosst for different studies. Dr Loutfy has received unrestricted research grants for other projects from, and has acted as a speaker and advisor for, Abbott Canada, Merck Frosst, Pfizer, Bristol-Myers Squibb, Tibotec, Boehringer Ingelheim, and Glaxo-Smith-Kline Inc. Dr Klein has acted as a consultant for, received research funding, and honoraria from ViiV. Dr Montaner has received grants from Abbott, Biolytical, Boehringer Ingelheim, Bristol-Myers Squibb, Gilead Sciences, Janssen, Merck, and ViiV Healthcare.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This work was supported by CANOC, which is funded through an Emerging Team Grant from the Canadian Institutes of Health Research (grant #53444) and is supported by the CIHR Canadian HIV Trials Network (CTN 242). Dr Tony Antoniou is supported by a post-doctoral fellowship from the Ontario HIV Treatment Network. Drs Janet Raboud and Curtis Cooper are recipients of Career Scientist Awards from the Ontario HIV Treatment Network. Dr Mona Loutfy receives salary support from the Canadian Institutes for Health Research. Dr Marina Klein is supported by a Chercheur-Boursier clinicien senior career award from the Fonds de recherche en santé du Québec (FRSQ). Dr Julio Montaner is supported by the British Columbia Ministry of Health Services and the Ministry of Healthy Living and Sport, and an Avant-Garde Award from the National Institute of Drug Abuse, National Institutes of Health. He has also received support from the International AIDS Society, United Nations AIDS Program, World Health Organization, National Institute on Drug Abuse, National Institutes of Health Research-Office of AIDS Research, National Institute of Allergy & Infectious Diseases, The United States President’s Emergency Plan for AIDS Relief (PEPfAR), Bill & Melinda Gates Foundation, French National Agency for Research on AIDS & Viral Hepatitis (ANRS), and Public Health Agency of Canada.

References

D:A:D Study Group, Sabin

Worm

Use of nucleoside reverse transcriptase inhibitors and risk of myocardial infarction in HIV-infected patients enrolled in the D:A:D study: a multi-cohort collaboration. Lancet. 2008;371(9622):1417–1426.

Thompson

Aberg

Cahn

Antiretroviral treatment of adult HIV infection: 2010 recommendations of the International AIDS Society-USA panel. JAMA. 2010;304(3):321–333.

Panel on antiretroviral guidelines for adults and adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. Department of Health and Human Services. January 10, 2011:1-166.

http://www.aidsinfo.nih.gov/ContentFiles/AdultandAdolescentGL.pdf. Accessed September 5, 2011.

Strategies for Management of Anti-Retroviral Therapy/Insight; DAD Study Groups. Use of nucleoside reverse transcriptase inhibitors and risk of myocardial infarction in HIV-infected patients. AIDS. 2008;22(14):F17–F24.

Brothers

Hernandez

Cutrell

Risk of myocardial infarction and abacavir therapy: no increased risk across 52 GlaxoSmithKline-sponsored clinical trials in adult subjects. J Acquir Immune Defic Syndr. 2009;51(1):20–28.

Lang

Mary-Krause

Cotte

Impact of individual antiretroviral drugs on the risk of myocardial infarction in human immunodeficiency virus-infected patients: a case-control study nested within the French hospital database on HIV ANRS Cohort C04. Arch Intern Med. 2010;170(14):1228–1238.

Bedimo

Westfall

Dreschler

Vidiella

Tebas

. Abacavir use and risk of acute myocardial infarction and cerebrovascular events in the highly active antiretroviral therapy era. Clin Infect Dis. 2011;53(1):84–91.

Ding

Andraca-Carrera

Cooper

No association of myocardial infarction with ABC use: an FDA meta-analysis. In: Program and Abstracts of the 18th Conference on Retroviruses and Opportunistic Infections, Boston, MA, February 27 to March 2, 2011. Abstract 808; 2011.

Mamdani

Jacka

Forde

Rothwell

. The striking effect of the heart outcomes prevention evaluation (HOPE) on ramipril prescribing in Ontario. CMAJ. 2003;168(5):553–557.

10.

Col

McLaughlin

Soumerai

The impact of clinical trials on the use of medications for acute myocardial infarction: results of a community-based study. Arch Intern Med. 1996;156(1):54–60.

11.

Shah

Juurlink

Austin

Mamdani

. New use of rosiglitazone decreased following publication of a meta-analysis suggesting harm. Diabet Med. 2008;25(7):871–874.

12.

Palmer

Klein

Raboud

Cohort profile: the Canadian observational cohort collaboration. Int J Epidemiol. 2011;40(1):25–32.

13.

Wagner

Soumerai

Ross-Degnan

Zhang F

. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther. 2002;27(4):299–309.

14.

Bai

Perron

. Computation and analysis of multiple structural change models. J Appl Econom. 2003;18(1):1–22.

15.

Sabin

Worm

Phillips

Lundgren

. Abacavir and increased risk of myocardial infarction, reply. Lancet. 2008;372(9641):804–805.

16.

Choi

Vittinghoff

Deeks

Weekley

Shlipak

. Cardiovascular risks associated with abacavir and tenofovir exposure in HIV-infected persons. AIDS. 2011;25(10):1289–1298.

17.

Sax

Tierney

Collier

Abacavir–lamivudine versus tenofovir–emtricitabine for initial HIV-1 therapy. N Engl J Med. 2009;361(23):2230–2240.

18.

Ribaudo

Benson

Zheng

No risk of myocardial infarction associated with initial antiretroviral treatment containing abacavir: short and long-term results from ACTG A5001/ALLRT. Clin Infect Dis. 2011;52(7):929–940.

19.

Sax

Tierney

Collier

. ACTG 5202: shorter time to virologic failure with ABC/3TC than tenofovir/FTC in treatment-naive subjects with HIV RNA > 100, 000. In: 17th International AIDS Conference, Mexico City, Mexico, August 3-8, 2008. Abstract THAB0303; 2008.