SMART study and 54-trial pooled analysis produce conflicting data about abacavir heart attack link

Data from the SMART trial provide further evidence that abacavir (Ziagen) may be associated with a higher rate of cardiovascular events, but a pooled analysis of more than 50 clinical trials conducted by abacavir’s manufacturer GlaxoSmithKline did not find any increased risk, attendees heard on Thursday at the XVII International AIDS Conference in Mexico City.

Researchers first reported a link between recent abacavir use and myocardial infarction (heart attack) among more than 33,000 participants in the observational D:A:D cohort at the 2007 Conference on Retroviruses and Opportunistic Infections in February.

Following up on these findings, investigators analysed data from the large international SMART trial to look for a similar association. Briefly, SMART included 5472 participants who were randomly assigned either to interrupt antiretroviral treatment when their CD4 count was above 350 cells/mm³ and resume when it fell below 250 cells/mm³, or else to remain on continuous therapy throughout the study.

In the current analysis, presented in a late-breaker session by Jens Lundgren of the University of Copenhagen, the researchers gathered data on heart attacks, strokes, and other cardiovascular events, as well as six biomarkers associated with inflammation or coagulation (clotting).

They compared these events and biomarkers in three groups of patients classified according to which nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs) they were taking when they entered the trial: abacavir but not ddI (Videx, which was also associated with a higher myocardial infarction (MI) risk in D:A:D), ddI (with or without abacavir) and neither abacavir nor ddI.

Participants were also categorised as having less than five vs five or more, cardiovascular risk factors. As a group, SMART participants were at moderate risk for cardiovascular disease. About three-quarters were men, the average age was 44 years, about 40% were smokers, about 20% were on blood pressure or lipid-lowering medications and 7% had diabetes; about 15% had five or more cardiovascular risk factors.

Current use of abacavir was associated with an increased risk of cardiovascular events assessed in several ways. Abacavir recipients had an elevated rate of MIs (adjusted hazard ratio [AHR] 4.3); major cardiovascular disease including MI, stroke, coronary artery disease (CAD) surgery and cardiovascular death (AHR 1.8); an expanded definition of major cardiovascular disease that also included congestive heart failure (CHF), peripheral vascular disease, CAD treatment and unwitnessed death (AHR 1.9); and minor cardiovascular events including CHF and use of CAD medications (AHR 2.7).

Among the subset of SMART participants with available biomarker data, levels of the inflammatory marker high sensitivity C-reactive protein (hsCRP) were 27% higher in the abacavir-treated patients (p = 0.02), while interleukin-6 (IL-6) levels were 16% higher (p = 0.02).

These associations remained in effect when the researchers compared patients taking abacavir with those taking tenofovir (Viread). Tenofovir, a newer NRTI, was not included in the D:A:D MI analysis. In contrast with D:A:D, use of ddI was not associated with an increased risk of cardiovascular events or biomarker changes.

“Consistent with the D:A:D study, in SMART, abacavir was associated with an increased risk of cardiovascular disease,” the investigators concluded. They suggested that this elevated risk may be due to vascular inflammation caused by the drug.

After the D:A:D results were released last February, researchers with abacavir manufacturer GlaxoSmithKline (GSK) conducted a pooled analysis of the company’s past clinical studies to see if they could find a similar association between the drug and cardiovascular events. The investigators said the D:A:D findings were “unexpected” and they were unable to identify a potential biological mechanism to explain this outcome.

GSK’s John Pottage presented findings from the analysis on Wednesday, the day before the SMART data were released.

The GSK HIV Data Repository includes results from 54 company-sponsored Phase 2-4 clinical trials in which participants took abacavir for at least 24 weeks as part of combination antiretroviral therapy. The database included 13 randomised clinical trials in which participants were assigned to receive abacavir or a comparator drug, 33 trials in which all participants received abacavir as part of a background regimen, and eight in which participants did not receive abacavir. Study durations ranged from 24 to 96 weeks, which was shorter than the D:A:D follow-up period.

Overall, data from 14,683 HIV-positive study participants (14,174 adults and 509 children) were included in the pooled analysis, representing 9639 people who took abacavir (7845 person-years) and 5044 who took regimens without abacavir (4653 person-years).

Most study participants (about 80%) were men, the median age was about 38 years, and the median CD4 count was about 300 cells/mm³. About two-thirds were treatment-naïve at study entry. The abacavir and non-abacavir groups had similar baseline demographic and HIV disease characteristics, including lipid levels and glucose values.

Importantly, these studies were not designed to look at cardiovascular outcomes, and therefore did not collect complete data on risk factors such as smoking or high blood pressure.

As in the SMART analysis, the GSK investigators looked at both MIs and an expanded definition of cardiovascular disease. This analysis was based on terms related to ischaemic coronary artery disorders found in the MeDRA glossary used for regulatory affairs, including CAD, atherosclerosis, angina pectoris and myocardial ischaemia.

The overall rate of cardiovascular events in this analysis was lower than the rate in SMART, and was similar in the abacavir and non-abacavir groups.

In the abacavir group, the MI rate was 0.114% or 2.04 per 1000 person-years, compared with 0.139%, or 2.36 per 1000 person-years in the non-abacavir group, giving the abacavir recipients about a 14% lower MI risk (p = 0.706). Looking at the broader measure of CAD, the respective rates were 0.249% or 3.45 in the abacavir group vs 0.416%, or 5.82 per 1000 person-years in the non-abacavir group, for about a 40% lower risk in patients taking abacavir (p = 0.055). Patterns were similar when looking only at the randomised clinical trials.

The studies included in this analysis did not collect data on biomarkers related to inflammation and cardiovascular disease, which have only recently been recognised as relevant for HIV-positive people receiving antiretroviral therapy.

During his conference presentation, however, Pottage showed late-breaking data from the more recent HEAT study comparing the abacavir/3TC fixed-dose combination pill (Kivexa or Epzicom) against the tenofovir/emtricitabine coformulation (Truvada).

In this trial, which did look at several inflammatory biomarkers, there were no significant differences in levels of hsCRP or IL-6 in the (Kivexa and Epzicom) arms, in contrast with the SMART findings.

The investigators concluded that there was “no difference in incidence of ischaemic coronary artery events or myocardial infarction” in patients who received abacavir-containing vs non-abacavir-containing antiretroviral therapy.

As with all medications, they continued, “physicians and patients must weigh the risks of HIV disease against the risks and benefits of the antiretroviral agents available.”

At this time, there is no clear explanation for the conflicting findings in the D:A:D and SMART cohort studies compared with the clinical trials in the GSK analysis.

The study populations were different in some important respects. Overall, participants in the GSK studies were slightly younger. D:A:D participants were mostly European, and commentators have noted the high rate of smoking in both D:A:D and SMART (this was not systematically assessed in the GSK trials).

In discussing the findings at the conference, several questioners brought up the possibility that cardiovascular events might be related to abacavir hypersensitivity reactions, perhaps due to unrecognised subclinical inflammation.

Looking to the future, Pottage recommended that all trials of antiretroviral therapy going forward should include comprehensive data collection on cardiovascular risk factors and biomarker assessment. Lundgren said a definitive answer to the puzzle would require a randomised clinical trial consisting of some 15,000 participants followed for two to three years.

With regard to current clinical practice, the SMART authors stated, “This adverse effect appears to be only clinically relevant to consider among patients with elevated underlying cardiovascular risk.”