Reports

This report is based on medical evidence presented at sanctioned medical congress, from peer reviewed literature or opinion provided by a qualified healthcare practitioner. The consumption of the information contained within this report is intended for qualified Canadian healthcare practitioners only.

15th Conference on Retroviruses and Opportunistic Infections

Boston, Massachusetts / February 3-6, 2008

In an unexpected outcome from an analysis conducted to explore the hypothesis that thymidine analogue nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs) increase cardiovascular (CV) risk, recent use of didanosine (ddI) or abacavir (ABC) was associated with a significantly increased rate of myocardial infarctions (MIs). No increased risk was associated with recent use of thymidine analogues or lamivudine (3TC). Several other NRTIs, including emtricitabine (FTC) and tenofovir (TDF), have not yet been subjected to the same analysis. As an observational study capable of generating associations without causality, the absolute risk and significance of these findings cannot be determined, but the authors of the analysis believed the data are sufficiently strong to warrant consideration in clinical decisions.

“Only a randomized clinical trial can demonstrate a definitive link between use of a specific drug and a specific side effect, such as MI,” observed Dr. Jens Lundgren, Head, Copenhagen HIV Program, and Professor, University of Copenhagen, Denmark. As one of the senior investigators for the D:A:D (Data Collection of Adverse Effects of Anti-HIV Drugs) study, he stated, “We do feel the data suggest that there is a link between these two NRTIs and CV risk, but it must be weighed in the context of other issues, particularly the efficacy of the antiretroviral regimen and the patient’s specific CV risk.”

D:A:D Revisited

The D:A:D study is a prospective collection of clinical data generated by 33,347 HIV-positive patients who are being managed at 212 clinics participating in 11 patient cohorts. These include cohort studies from Europe, Australia and the US. Previously published data from D:A:D were among the first to associate protease inhibitors (PIs) with an increase in CV risk, a finding that has been reinforced subsequently. However, in this new study, a completely different pattern of risk has been observed, confusing efforts to understand a mechanism of action.

“With PIs, there was a progressive increase in risk with the period of time on therapy, and we did not see an immediate reduction in risk when PI therapy was discontinued. With these NRTIs, the peak in risk appears to be reached soon after treatment is started. When treatment is discontinued, the risk appears to return to normal,” Dr. Lundgren explained. He told listeners that the unexpected finding is not consistent with any known effect of these drugs on mechanisms that might increase the risk of MI.

New Analysis

In a study with 157,912 person-years of follow-up, there have been 517 MIs. In this new analysis of the data, the hypothesis was that exposure to zidovudine (AZT) or stavudine (d4T) would increase the risk of heart disease. The hypothesis was based on the association between these thymidine analogues with dyslipidemias and insulin resistance, both of which would be expected to increase CV risk. The same analyses were conducted with ddI, 3TC and ABC on the presumption that these would provide controls with which to compare relative risk.

A Poisson regression analysis was undertaken to determine the impact of cumulative exposure, recent exposure (still using or stopped using within the last six months) and past exposure (previously used but not within the last six months). Exposures were evaluated from the time patients entered the D:A:D analysis until MI, death or February 2007. The MI rate was then evaluated in the context of that predicted by the Framingham equation for 10-year coronary heart disease (CHD) risk.

Contrary to the study hypothesis, neither of the two thymidine analogues, AZT or d4T, nor 3TC was associated with an excess rate of MI whether assessed by cumulative, current or past exposure. However, recent use of ddI raised the odds ratio (OR) of MI to 1.49 (95% CI, 1.14-1.95; P<0.01) while recent use of ABC raised the OR of MI to 1.90 (95% CI, 1.47-2.45; P<0.01). In the model that produced this result, neither cumulative use nor past use was associated with an increased risk. The risk of MI associated with recent ddI and ABC use was retained after adjusting for dyslipidemias, other metabolic factors, CD4+ cell count and HIV-RNA levels. Although there was adjustment for CV risks, the authors did point out that the populations treated with ABC differed from those who were not.

“Patients with recent exposure to ABC were more likely to be male, older, and to have diabetes, hypertension, dyslipidemia, or a family history of CHD than those with no recent exposure to ABC, although they were less likely to be smokers or to have a high body mass index]” reported Dr. Caroline A. Sabin, Royal Free Hospital, University College, London, UK.

Interpreting the Results

When these findings were evaluated in the context of the 10-year CHD risk as calculated by the Framingham equation, the association of the NRTIs with greater rates of MI was found to be concentrated in those who were already at risk of CV disease. This was particularly true of ABC for which the interaction between the predicted 10-year CHD risk and recent use reached statistical significance (P=0.04). This is an important point in analyzing the results in a clinical context.

“If you are already at risk of CHD, ddI and ABC appear to exacerbate this risk,” Dr. Lundgren observed. However, patients who developed CV risk factors did not appear to be at any increased risk of MI if they had previous exposure to ABC. “Therefore, the clinical implications of the results of this abstract must take into account an individual patient’s underlying CV risk.”

The importance of understanding these results in the context of relative risk was a point repeatedly emphasized by Dr. Lundgren. As antiretroviral drugs are life-saving medicines, the immediate and absolute benefit of an effective regimen may outweigh the relative long-term risks. This is not to suggest that these long-term risks should be ignored, but Dr. Lundgren indicated that control of HIV is the first priority.

“We are now in a position with effective antiretroviral therapies that patients are living long enough to face age-related risks such as heart disease. It is important that we are aware of these and choose therapies with these long-term risks in mind, but effective control of HIV is essential so we should switch therapies only when we have an alternative that is likely as effective and well tolerated,” Dr. Lundgren observed.

Moreover, patients taking ABC or ddI face a very low MI risk. “The absolute risk of having a heart attack in this group of patients as a whole was low. The risk for the study overall was 1.6% over five years,” Dr. Lundgren reported. He noted that as smoking is associated with a twofold greater risk of CHD than ddI or ABC, smoking cessation “would do a lot more [than stopping ddI or ABC] to reduce the risk of having a heart attack and other serious diseases.”

Although both Drs. Lundgren and Sabin believe that the increased risks posed by ddI and ABC are real, especially in individuals who already have CHD, one concern is the absence of a biological basis. Neither of these NRTIs are associated with metabolic abnormalities, and there are no prior reports of adverse effects on pump function, including electrical disturbances. This is the first report of an association between ddI, ABC and MIs, and there was no association with events other than MI, such as stroke. From this perspective, the findings are confusing.

“An observational study such as D:A:D that follows patients cannot definitely prove that one thing causes another,” Dr. Lundgren acknowledged. Although he feels that data are sufficiently strong to influence clinical decisions, he recognizes that whether patients decide to take these agents should depend on baseline CV risk and the other available treatment options.

Summary

New data from the D:A:D trial indicates that patients who have been recently exposed to ddI or ABC have an increased risk of CV disease than those who have not, but the absolute risk was low and concentrated in patients who already have high risk of CHD. While the authors of this study suggested that these NRTIs may not be the optimal first-line therapies when other effective agents are available to patients with CHD, the mechanism by which these agents increase risk is unclear. More data would be useful to understand how these findings should be incorporated into clinical care.