Abacavir, ddI and two protease inhibitors remain associated with heart attacks in largest cohort study

This article is more than 15 years old. Click here for more recent articles on this topic

The latest report on the D:A:D (Data Collection on Adverse Events of Anti-HIV Drugs) study, the largest cohort study monitoring the effects of HIV drugs, published in the Journal of Infectious Diseases, reports a continued association between current or recent use of the two nucleoside (NRTI) drugs abacavir and ddI (didanosine) and heart attacks.

Patients who were currently on the drugs, or who had stopped them no more than six months previously, had, respectively, a 70% and 30% increased risk of having a heart attack during this time, even after adjusting for other heart attack risks such as blood lipid level, diabetes and high blood pressure.

For the first time D:A:D also reports a smaller increasing risk of heart attacks related to the total time on abacavir as well. In other words the risk of heart attacks also increased per year of exposure to abacavir and persisted after the drug was stopped.



Fat or fat-like substances found in the blood and body tissues. Lipids serve as building blocks for cells and as a source of energy for the body. Cholesterol and triglycerides are types of lipids.

high blood pressure

When blood pressure (the force of blood pushing against the arteries) is consistently too high. Raises the risk of heart disease, stroke, kidney failure, cognitive impairment, sight problems and erectile dysfunction.


A group of diseases characterized by high levels of blood sugar (glucose). Type 1 diabetes occurs when the body fails to produce insulin, which is a hormone that regulates blood sugar. Type 2 diabetes occurs when the body either does not produce enough insulin or does not use insulin normally (insulin resistance). Common symptoms of diabetes include frequent urination, unusual thirst and extreme hunger. Some antiretroviral drugs may increase the risk of type 2 diabetes.

statistical significance

Statistical tests are used to judge whether the results of a study could be due to chance and would not be confirmed if the study was repeated. If result is probably not due to chance, the results are ‘statistically significant’. 


In HIV, an individual who is ‘treatment naive’ has never taken anti-HIV treatment before.

It also reports an increased risk of heart attacks per year of drug exposure associated with the protease inhibitors lopinavir/ritonavir (Kaletra) and indinavir (whether or not taken with ritonavir as a booster), but not for other protease inhibitors.

No increased risk was found for current exposure to these drugs, meaning that it was the total number of months spent taking the drug, even if that occurred in the past, that was associated with heart attack risk – not whether the patient was currently taking the drug.

The reason for distinguishing between current and cumulative exposure is that it distinguishes between a risk of acute side-effects that increases as soon as a patient starts taking a drug, but stops when they stop, and a risk of chronic and lasting side-effects, which may only become significant after several years, but which may also persist or even worsen after the drug is stopped. The former pattern appears to be associated with the two NRTI drugs and the latter with certain protease inhibitors.

Interestingly, when the researchers re-analysed the data only to include patients who had stopped protease inhibitors less than six months before a heart attack, they found virtually no change in heart attack risk, indicating that the increased risk really does persist well after these drugs are stopped.

No increased heart attack risk was found with any other HIV drug studied, including tenofovir, though in the case of tenofovir only three years’ data have been collected so far, resulting in a wide margin of uncertainty as to risk.

The D:A:D: researchers remind readers that, as a cohort study, its findings “cannot be assumed to reflect causal associations and must be interpreted cautiously because of the potential for unmeasured confounding”.

The findings published in JID expand on those already presented at the CROI conference in February 2009 – see this report .

To summarise the findings, the researchers found a 12% increased risk of heart attack per year of exposure to indinavir (with or without ritonavir boosting) and a 13% increased risk per year of exposure to lopinavir/ritonavir.

Previously, D:A:D has not had enough data on the levels of patients’ lipids (cholesterol and triglycerides) to take these into account. But it now has enough data to adjust for these factors. It found that after adjusting for lipid levels the increased risk per exposure year, independent of lipid levels, was now 8% for indinavir and 9% for lopinavir/r. These were lower, but still statistically significant.

Adjusting for lipid levels also explains why the increased risk of heart attack for patients currently or recently on abacavir fell from 90% in the last report to 70% in this one, and why it fell in the current report from 40% to 30% after lipid adjustment for ddI. However the finding that the risk of heart attack increased by 7% per year of exposure to abacavir is new, and did not change with adjustment for other risks.

Since these findings were initially presented at CROI, a couple of studies, in particular a large US one of army veterans, have failed to find a link between abacavir and heart attack – see this report – though others have found links between abacavir and specific changes in indicators of vascular inflammation – see here and here.

Some researchers have suggested that the increased risk seen with abacavir is due to “channelling bias” – meaning that patients with a certain characteristic may be more likely to be prescribed a certain drug – and that characteristic could also make heart attacks more likely. Thus, though there is no direct chain of causation, characteristic Z would make both characteristic A (being on abacavir) and characteristic B (having a heart attack) more likely.

In the case of the US Veterans’ study, the researchers think they may have found their characteristic Z in the shape of kidney disease. Because this may be exacerbated by tenofovir use, the argument goes, patients with it are more likely to be prescribed abacavir instead.

Kidney disease certainly increases the long-term risk of heart attack by about 48%, according to one US study, and in the Veteran’s study it increased it enormously: HIV patients with kidney disease were four times (400%) more likely to have a heart attack. But taking abacavir only raised the risk by 17%, and this was not statistically significant.

However adjusting for kidney disease in the Veterans’ study made very little difference to the risk seen with abacavir.

The D:A:D study does not collect kidney disease data (or hasn’t done so far). But the researchers emphasised that there was very little difference in the heart attack risk data they did collect between patients on tenofovir and ones on abacavir.

Patients who took abacavir had the highest rate of diabetes and family history of heart attack, but on the other hand tenofovir patients had the highest rate of high blood pressure, itself strongly associated with kidney disease. However the differences between risk profile were only small and not statistically significant. The D:A:D researchers therefore continue to maintain that the increased risk seen in patients cannot be explained solely by differences in other heart attack indicators.

They also point out that the accumulated risk seen with protease inhibitors (and possibly with abacavir), though it may seem small, translates into considerable clinical risk over time. The 13%-a-year increase in exposure to lopinavir/r, for instance, became an 84% risk over five years. This is roughly the same as the increased risk seen with every ten year increase in age.

Having said that, the increased risks seen with HIV drugs were small compared with some other characteristics. In the D:A:D cohort having a high Framingham score (predicted risk of a heart attack over the next decade) quadrupled the risk, and a moderate Framingham score doubled it. (A low score halved it.) High blood pressure or diabetes more than tripled the risk, having high lipids raised the absolute risk of a heart attack by 69%, and smoking raised it by 56%.

And of course untreated HIV infection itself increases the risk of heart attack. In the SMART study, which compared continuous use of HIV drugs with intermittent, CD4-guided use, intermittent use of HIV drugs raised the heart attack risk by 57% - and taking abacavir raised it by 80%.

In SMART, the vast majority of patients who started taking intermittent therapy had previously been taking HIV therapy or had done so recently. Very few were drug naive.

One theory is that the studies that show an association between abacavir and heart attacks largely feature patients who were already on HIV drugs while ones that show no association feature drug-naive patients who started abacavir. The argument is that any increase in risk due to abacavir was masked by the decrease in risk due to uncontrolled HIV replication.

The debate about whether the apparent risk posed by abacavir is real or due to some unrelated factor remains unsettled. A couple of studies, for instance, have emphasised the importance, in an HIV-positive population, of also adding in recreational drug use as a risk factor to be taken into account, given that in one US study, using cocaine raised the risk of a heart attack by 770%.

Until the question is settled – and it may be eventually if studies like D:A:D accumulate enough data – all that guidelines like the BHIVA ones can say is that “Patients currently on abacavir-containing regimens should be carefully reviewed to see whether other options are available”.

But, they emphasise, “For any individual the excess risks of abacavir are small, but [though] it is important to address any avoidable risk, [this] should not be at the expense of addressing the other and more key modifiable risk factors.”


Worm SW et al. Risk of myocardial infarction in patients with HIV infection exposed to specific individual antiretroviral drugs from the 3 major drug classes: the Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) Study. Journal of Infectious Diseases 210: 3108-30, 2010.

Reiss P Abacavir and cardiovascular risk. Sixteenth Conference on Retroviruses and Opportunistic Infections, Montreal, abstract 152, 2009.

Supplement to BHIVA Guidelines for the Treatment of HIV-1 Infected Adults with Antiretroviral Therapy.