Guidelines Panel Responds to Abacavir Data

The Department of Health and Human Services (DHHS) HIV treatment guidelines panel has issued a response to recent negative findings about abacavir. Abacavir is an HIV drug sold as Ziagen and in the fixed dose combination pills Epzicom (with lamivudine/3TC) and Trizivir (with lamivudine/3TC + zidovudine/AZT). While acknowledging growing concerns about this widely used HIV drug, the panel stopped short of revising its recommendation of Epzicom as a preferred option for people taking HIV drugs for the first time.

The DHHS communication refers to two recent studies that have raised important questions about abacavir. The first was the analysis of the D:A:D study, first reported at CROI 2008 in February. Researchers looking at the D:A:D cohort found that people taking abacavir were about 1.9 times more likely to have a heart attack than those not on abacavir. The increased risk was greater for people who had other known risk factors for heart attack, like family history, diabetes and smoking.

As reported here, this finding came as a surprise. In contrast to protease inhibitors, which can increase cholesterol and triglycerides, there is no known biological explanation for the increased risk of heart attack with abacavir. While significant, the increased risk is also small, in absolute numbers. In the entire D:A:D study there was a .03% chance of heart attack per year. For people taking abacavir the rate was around .06%. That means if you followed 10,000 people for a year, 6 people taking abacavir would be expected to have a heart attack, compared to about 3 people overall.

It is important to point out that this increased risk associated with abacavir is much smaller than the risk found with cigarette smoking, poor diet and diabetes. The increased risk appeared to reverse once people had stopped taking abacavir for 6 months. This suggests that the effect is real, rather than a coincidental finding. It is also reassuring for anyone who has taken abacavir.

The second set back for abacavir came at the end of February 2008, when the AIDS Clinical Trail Group (ACTG) announced the decision to unblind almost half of the participants in ACTG 5202: a head-to-head study comparing Epzicom to Truvada (tenofovir + FTC), with either Sustiva (efavirenz) or Norvir (ritonavir)-boosted Reyataz (atazanavir).

The decision followed an early look at the results by a Data Safety Monitoring Board (DSMB), an independent group of scientists who periodically review data from ongoing studies to ensure the participants' safety. The DSMB found higher rates of treatment failure among people taking Epzicom who had started the study with HIV levels above 100,000 copies, referred to as the high viral load group. They also reported higher rates of moderate side effects for people in this group taking Epzicom. The DSMB decided to unblind the study participants who were in the high viral load group, and inform those taking Epzicom of the results. Importantly, they did not decide to stop the entire study, or unblind the lower viral load group.

The DHHS panel looked at these two sets of results and concluded they did not warrant changing the current recommendations. This decision was being watched closely, as these findings came closely after Epzicom had been added in January to the 'preferred' category for first line treatment.

Discussion

While these results can not be ignored, they raise as many new questions as the answer. The analysis of the D:A:D study did not look at the components of Truvada, which is the most likely option for someone to switch to. GlaxoSmithKline (the maker of abacavir) released its own analysis of several studies, finding no increased risk of hear attack from taking abacavir. The lack of a solid biological explanation, confirmatory studies and information on Truvada make it challenging to know what exactly to do with the D:A:D results.

The same can be said of the decision to unblind the high viral load group in ACTG 5202. The DSMB looked at all of the results and saw enough of a difference to warrant taking action only in the high viral load group. The fact that they didn't decide to unblind the entire study is curious. While differences in treatment results in people with higher and lower viral levels have been seen in other studies, those differences have always gone away over time -- a fact that members of the DSMB should be well aware of. Their decision not to unblind the lower viral load group suggests that they saw something of interest in the lower viral load group, and that they felt justified allowing that part of the trial to go forward.

If you take abacavir, these results do not necessarily mean you should switch. Discuss the D:A:D findings with your doctor, especially if you have known risk factors for heart attack, like smoking, family history and diabetes. If you started aregimen with abacavir when you HIV levels were above 100,000 copies, extra vigilance might be warranted.

Some might question why the DHHS panel didn't change its recommendation to include Epzicom as a preferred option. It is fair to say that these two studies have raised important questions about abacavir. It is less clear however that Epzicom should be viewed less favorably than other options, specifically Truvada. Until more is understood about both of these studies, Project Inform supports the DHHS decision to detail the findings, and to counsel people living with HIV/AIDS and their doctors to consider all of the available research when making decisions on HIV treatment.

The full text of the DHHS communication is below.

DHHS Adults and Adolescents Antiretroviral Treatment Guidelines Panel's Communication Regarding Abacavir -- April 4, 2008 Data from studies addressing the safety of abacavir (ABC) and efficacy of abacavir/lamivudine (ABC/3TC) as part of a combination antiretroviral regimen recently became available. At the 15th Conference on Retroviruses and Opportunistic Infections (CROI), the D:A:D study group reported their analysis of association of NRTI use and risk of myocardial infarction (MI) in a large multi-national observational cohort.1 Among the 33,347 patients enrolled, with 157,912 person-years follow-up, 517 subjects were diagnosed with an MI; 192 and 124 of these subjects reported use of ABC and didanosine (ddI), respectively, within the previous 6 months of the occurrence of the MI. In this analysis, recent (within 6 months) but not cumulative or past use (last use >6 months) of either ABC or ddI predicted risk of MI (relative risk of 1.94 [95% CI: 1.48-2.55] for ABC and 1.53 [95% CI: 1.10-2.13] for ddI). The heightened risk of MI with recent ABC exposure was accentuated in subjects with pre-existing cardiac risk factors (as defined by 10-year predicted coronary heart disease risk >20%). Use of tenofovir (TDF) or emtricitabine (FTC) was not analyzed in this study. A separate analysis conducted by GlaxoSmithKline using their internal database containing data from 54 clinical trials and post-marketing reports identified 9,369 patients who received ABC for 7,845 person-years and 5,044 patients who did not receive ABC with 4,653 person-years follow-up. Eleven ABC-treated and seven subjects not on ABC were reported to have an MI during the follow-up period.2 GlaxoSmithKline did not find any evidence of an increase in cardiovascular disease in their clinical trials among patients who received ABC. It should be noted that these studies were not designed to evaluate cardiovascular events as secondary endpoints, and the follow-up periods in these trials were 24-48 weeks. On February 28, 2008, the National Institute of Allergy and Infectious Diseases released an announcement regarding a modification of the AIDS Clinical Trial Group (ACTG) 5202 study, a randomized controlled trial of antiretroviral-naïve participants evaluating the efficacy and safety of ABC/3TC vs. TDF/FTC when used in combination with either efavirenz (EFV) or ritonavir (RTV)-boosted atazanavir (ATV).3 Treatment randomization was stratified based on screening HIV RNA of <100,000 copies/mL or >100,000 copies/mL. Over 1,800 participants enrolled in this study. At a planned interim meeting of the independent Data Safety Monitoring Board (DSMB), the board noted that the majority of patients in all treatment arms had good virologic responses. The DSMB noted that among the participants with screening HIV RNA >100,000 copies/mL, those who were randomized to ABC/3TC had a significantly shorter time to study-defined virologic failure than those randomized to the TDF/FTC arm, regardless of whether they were using EFV or RTV-boosted ATV in combination. The DSMB also noted that within the stratum of subjects with a screening HIV RNA level >100,000 copies/mL, the ABC/3TC group had a shorter time to development of certain grade 3 or 4 side effects, including body aches and laboratory abnormalities such as lipid elevations. Based primarily on the difference in virologic efficacy, the DSMB recommended unblinding of the subjects who had screening HIV RNA >100,000 copies/mL, and counseling subjects on ABC/3TC about the findings, giving them the option to switch to other regimens. Subjects with HIV RNA <100,000 copies/mL at study screening are to remain blinded to their NRTI assignment. The results of the HEAT study, a head-to-head trial of 688 subjects comparing ABC/3TC to TDF/FTC, (both in combination with once-daily lopinavir/ritonavir) were presented at the 15th CROI.4 A subgroup analysis according to baseline HIV RNA < or >100,000 copies/mL (43% and 57% of subjects, respectively) yielded similar percentages of subjects with HIV RNA <50 copies/mL at 48 weeks for the two regimens: 71% (ABC/3TC) vs. 69% (TDF/FTC) for those with baseline HIV RNA <100,000 copies/mL and 63% vs. 65% for those with HIV RNA >100,000 copies/mL. At this point, the Panel concludes that the preliminary information available from these studies does not warrant a change in its current recommendations regarding the use of antiretroviral drugs in adults and adolescents. The Panel will continue to review additional data as they become available and will make further recommendations if needed. Meanwhile, the Panel recommends clinicians consider all available information so that the optimal therapeutic choice for each patient is based on individual patient characteristics and the potential risks and benefits of each treatment component.