While this is the first large study to suggest an association between longer time on ARV and an increased risk of serious heart problems, there are a number of limitations. First, there was no link shown between heart disease and any particular drug or even class of drugs. Second, a large number of participants in this study had additional risk factors for developing heart disease, such as smoking and being older. Finally, while time on therapy might contribute, the association between the length of time in the study and heart attack could just as well be due to advancing age or the long-term effects of HIV itself.
One counter-intuitive finding was that either fat accumulation or fat wasting seemed to have a protective effect against heart attack. Some have speculated that this might be because people with physically obvious lipodystrophy may be motivated to begin measures to forestall heart disease, such as taking lipid-lowering statins, stopping smoking or getting more exercise. With the risk of heart attack in the D:A:D study running a modest 1.26 times above averages in non-HIV populations, adopting such protective measures is likely to more than counteract any increased risk due to antiretroviral drugs. And one fact is undisputed: the risk of dying from untreated HIV is dramatically reduced by antiretrovirals.
One limitation of the VA study is the relatively short period of time that most patients had been receiving drugs. Only about 1,000 people had been on protease inhibitor-containing ARV combinations for periods greater than four years.
So while these large observational cohorts give contradictory signals about whether there is a link between heart disease and HIV meds, they both dispel the worry that there is a huge, tip-of-the-iceberg epidemic of treatment-related illness lurking just ahead. However, the authors of both studies agree that more study is warranted and that focused research on the role of individual drugs needs to be conducted.
At the 10th Retrovirus Conference, researchers from San Francisco General Hospital reported on an investigation into the incidence of abnormal intima-media thickness (IMT) of the carotid artery within a cohort of 106 HIV patients. The carotid artery conducts the main supply of blood to the brain. IMT is measured by ultrasound assessment of the thickness of part of the arterial wall and is recognized as a sensitive predictor of myocardial infarction and stroke. In this observational cohort, in an attempt to find independent predictors of increased IMT, abnormal IMT values were correlated with laboratory and patient history parameters such as cholesterol levels and cigarette smoking, as well as with ARV use and duration of HIV infection. In addition to the snapshot taken of IMT in the main cohort, a subset of 22 patients was evaluated for progression of IMT over a period of one year.
The results revealed increased IMT in this cohort as compared to averages obtained from studies in non-HIV populations. Statistically significant predictors of increased IMT included age, LDL cholesterol, high blood pressure, and having had a fall in CD4 cell count below 200 at any time in the past. A patient's duration of HIV infection, HIV viral load, or current CD4 count were not found to be predictive of IMT. Non-significant associations were reported for smoking and for the length of time that patients had been taking protease inhibitors.
In the subset of 22 patients assessed for changes in IMT over one year, IMT progressed at a rate four to five times greater than changes reported in previous progression studies in non-HIV populations. The investigators concluded that carotid intima-media thickness is increased in this cohort of HIV-infected individuals as compared to non-HIV historical controls, and that increases were independently associated with the traditional risk factors for cardio-vascular disease, such as cholesterol and hypertension, as well as with HIV-specific factors such as history of low CD4 count.
Another ongoing study of IMT by the federal AIDS Clinical Trials Group (ACTG) is looking more closely at the role of ARV drug classes. Preliminary results reported from three matched cohorts found no significant differences in IMT between those on PI-containing or non-PI-containing regimens, and no difference between HIV-positive and negative participants. The PI patients had a mean time on treatment of over four years. These observations were baseline snapshots only; the study will continue to see if progression rates differ between the groups.
One question remaining unanswered by these studies is the role -- if any -- played by ARV therapy in increased IMT and the risk of vascular disease. Some researchers feel that chronic inflammation from HIV infection itself may be responsible for the observed IMT changes. For patients with additional risk factors such as smoking or family history of heart disease, the answer may never be clear.
Sam Bozzette on the VA Study
The study we just published about the risk of cardiovascular or cerebrovascular disease in people on HAART took advantage of the fact that the very large VA medical system has a common database across all centers. We had 36,000 patients, with 1.5 million antiretroviral prescriptions, and there were 2000 cardiovascular or cerebrovascular events. Retrospective, administrative database studies like this have limits on what kinds of answers they can give, but ours does describe the experience of this very large cohort. It's kind of like asking 36,000 of your friends how they did. There's a definite message in the answer to that question so we thought it would be a contribution if we could pull that data together.
When we started this VA study, we weren't asking whether or not effective antiretroviral therapy with protease inhibitors was going to marginally affect the rates of cardiovascular disease in some small way. People were worried that we had a disaster brewing. People were worried that cardiovascular and cerebrovascular events were going to be so common as to be canceling out the benefits of therapy in people with advanced disease who clearly needed it. But if you look at the admission and mortality graphs in our paper, you can comfortably say, "Okay, there's no emergency here." Is there an effect? There could be. We still can't forget about lipids or heart attacks -- they need additional careful investigation that cannot be done in the administrative databases.
Back to the GMHC Treatment Issues April 2003 contents page.