It just wouldn’t be the Conference on Retroviruses and Opportunistic Infections (CROI) without at least one element of surprise and intrigue. At this year’s gathering, held earlier this month in Boston, an international team of researchers presented unexpected data from a large cohort study showing that two HIV drugs—abacavir (used in Ziagen, Epzicom and Trivizir) and didanosine (used in Videx tablets and Videx EC capsules)—significantly increased the risk of a heart attack, despite the fact that neither drug has ever been known to contribute in any way to cardiovascular disease.

When challenged to explain the most recent results from the large Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) study, even the most experienced HIV experts seemed perplexed. In a video interview with AIDSmeds, Donald Kotler, MD, chief of gastroenterology at St. Luke’s-Roosevelt Hospital in New York City, said, "I can’t deny the data, because there were really smart people who looked very carefully and found it, and yet I really can’t understand it, and I really can’t explain why it would have occurred."

Beyond potential explanations for the results is an even more pressing question-should HIV-positive people using either nucleoside reverse transcriptase inhibitor (NRTI) question their doctors about switching the medications? While it’s always a good idea to check in with a health care provider when confronted with information that sounds frightening-especially when it involves two commonly used HIV medications-making an informed decision requires understanding what the results really show and actually mean for HIV-positive people.

What D:A:D Says

The D:A:D study, a mega cohort that combines the medical records from 33,347 people living with HIV who are enrolled in a variety of smaller cohort studies in more than a dozen countries around the globe, was one of the first studies to establish an association between the use of protease inhibitors and the risk of heart attacks. Yet little is known about the potential association between NRTIs—which are almost always used in HIV drug combinations—and heart attacks. Zidovudine (used in Retrovir, Combivir and Trizivir) and stavudine (Zerit) are two NRTIs that have been linked to higher lipid levels and diabetes, two risk factors for cardiovascular disease. However, there hasn’t been any research showing that these two agents actually increase the risk of a heart attack.

Hypothesizing that stavudine and zidovudine do increase this risk, Jens Lundgren, MD, of the Copenhagen HIV Programme at the University of Copenhagen, in Denmark, and his colleagues examined follow-up results from the D:A:D study to determine what association may exist between heart attacks and the various NRTIs.

Of the 33,347 participants in the D:A:D study, 517, or 1.5 percent, had a heart attack over an average of seven years of follow-up. After controlling for known cardiovascular risk factors among those who suffered heart attacks—such as sex, age, smoking status and family history—the team analyzed the data to determine whether any particular NRTI was more strongly associated with a higher risk of having a heart attack than other NRTIs.

Contrary to expectation, people taking stavudine or zidovudine appeared to be at no greater risk of having a heart attack. The same was true for Epivir (lamivudine). To everyone’s surprise, however, taking abacavir was associated with a 90 percent increase in the risk of a heart attack, and the risk with didanosine was increased by 49 percent. The increased risk only persisted while a person was on either drug. If they stopped taking them, the risk decreased. Most unusual was the fact that the risk didn’t increase with increasing length of time on either drug. Typically, cumulative use leads to a cumulative risk for side effects.

It’s All Relative

Ninety percent may seem like a lot, especially for a side effect as serious as a heart attack, but it’s critical to put that number into proper perspective. Bob Huff, the antiretroviral director at the Treatment Action Group in New York City, worries about how people are hearing the results of the study, and says, "I think a lot of people read this as you have a 90 percent chance of having a heart attack if you take abacavir, which is ridiculous, and no one is saying that."

"Risk" can be used to measure both positive and negative treatment effects. Antiretroviral therapy can reduce the risk of HIV disease progression, while at the same time increase the risk of certain side effects. Most of the time, what researchers are looking for is a reduction, or increase, in "relative risk"—for example, how the risk of a heart attack in one group of patients (e.g., those taking abacavir) relates to another group of patients (e.g., those not taking abacavir).  What researchers don’t often report is "absolute risk"-to what extent the use of the medication changes the actual risk we face of a serious event, such as a heart attack, over a certain period of time. 

As an example, an HIV-positive person—depending on a number of factors—might have a 2 percent risk of a heart attack over the next 10 years. If he or she continues to use abacavir, the risk of a heart attack increases by 90 percent. The 90 percent increase relates to the original 2 percent risk, and equates to an additional 1.8 percent. So the absolute risk of a heart attack among this abacavir user increases from 2 percent to 3.8 percent.

Based on another large and extremely long-term cohort study, known as the Framingham Heart Study, we know a lot about the various factors that can increase a person’s risk of cardiovascular disease. There’s even a simple calculator that can help determine your risk for developing heart disease in the next 10 years. It takes into account the various factors that are known to increase the risk of a heart attack: sex, age, smoking status, a diagnosis of diabetes, blood pressure and cholesterol levels.

According to an online calculator maintained by the National Cholesterol Education Program of the National Heart, Lung and Blood Institute, using data from the Framingham Heart Study, a 40-year-old male nonsmoker with normal blood pressure, no diabetes, and normal cholesterol levels faces a 1 percent risk of a heart attack over the next 10 years. By taking abacavir, with its relative risk of 90 percent, his 10-year absolute risk of a heart attack is still under 2 percent.  But what about a 45-year-old male with elevated total cholesterol levels, low "good" HDL cholesterol levels and high blood pressure readings? His 10-year risk of a heart attack, according to the Framingham calculator, is 19 percent. Add ongoing abacavir use and his absolute risk of a heart attack jumps to more than 36 percent-a substantial increase. 
Cohort Conundrum

Further confusing matters is the fact that no other studies, notably those closely evaluating the safety and effectiveness of abacavir and didanosine, have hinted at the possibility of a higher heart attack risk associated with the use of either drug.  "[An increased risk of heart attacks] hasn’t appeared in the comparative studies that have been done in the past," says Huff. That’s why the D:A:D study took so many people by surprise and left some wondering if the increased risk seen in the study was real.

A potential reason for the discrepancy may be the distinction between cohort studies, such as D:A:D, and comparative studies. Cohort studies are designed to detect associations—connections between two variables—whereas more rigorous comparative trials are designed to show a cause and effect. In cohorts, patients usually have incredibly diverse health and treatment histories, potentially creating a large number of extraneous factors that can falsely color the study results. Randomized, double-blind clinical trials are a good example of comparative studies in that they control for variability, allowing researchers to narrow in on the specific effects of a medication, be it healthful or hurtful.

But despite the limitations of cohort studies in general, and the fact that the D:A:D study couldn’t explain why the drugs were associated with an increased risk, neither Huff nor Kotler is ready to shrug off the data as unimportant. "Heart attacks get people’s attention, because on an absolute scale it seems like the worst thing to have. That’s enough to make people feel cautious, even though they don’t understand the data," says Huff.

Kotler says, "It’s concerning; somebody better figure out why, because they couldn’t explain it—it wasn’t because they had more diabetes, it wasn’t because they had higher bad cholesterol or lower good cholesterol, or any of those other factors, and they couldn’t find any reason to explain it. Yet it was there."

It will obviously take additional research to determine how meaningful the results of the D:A:D study are. "I think the next step," recommends Huff, "[is to] start data-mining some of the larger cohorts. And go forward from there. If you start to get the same signal from multiple cohorts then you start to take it more seriously."

Additional research will take time, however, and won’t help a person currently on abacavir or didanosine decide whether to hang on or make a switch. Huff points out that the alternative NRTIs all have their own particular sets of side effects, some of them potentially as serious as a heart attack in some people, but also says, "How do you know if [the D:A:D results are] going to apply to you? How do you know what your real, or relative, risks are other than taking a Framingham test and rating yourself?"

Some patients with a low absolute risk of a heart attack may decide that continued abacavir or didanosine use won’t push them into a danger zone, or may choose to work with their health care providers to reduce any modifiable heart disease risk factors—such as smoking, elevated lipid levels and high blood pressure—before writing off an HIV medication that is otherwise working well for them. People with a high absolute risk for heart disease or who are unable to modify their existing risk factors may face a more difficult decision. Tracy Swan, an HIV and hepatitis C activist who also works with the Treatment Action Group in New York City, agrees. She says, "What [this] really means in real life for people is [you need to] assess your risk factors. If you’re not at high risk otherwise, it’s not relevant information. It’s really only a big problem for people who are at high risk."