July 14, 2003
The impact of antiretroviral therapy on the rate of cardiovascular complications is one of the major controversies in HIV management. In particular, there has been concern about the use of protease inhibitors (PIs) in view of their association with metabolic side effects including elevated cholesterol, elevated triglycerides, insulin resistance and diabetes mellitus. On the other hand, non-nucleoside reverse transcriptase inhibitor (NNRTI)-based therapy has been associated with fewer metabolic disturbances.
Studies in treatment-naive patients have also shown better metabolic measures and lipid levels in patients on NNRTIs compared with PIs. Additionally, a number of studies have been done in which patients were switched from PI to NNRTI-based therapy and had improved lipid profiles. Moreover, many clinicians make initial therapy choices and modify antiretroviral therapy (ART) based on metabolic concerns, so this topic is timely.
Because of the widely differing opinions on the level of cardiovascular risk posed by ART among HIV treaters, it was fitting that this issue was addressed in the format of a debate session devoted to this topic. The question debated was "should we modify antiretroviral treatment based on cardiovascular risk?"
The Pro side of the debate was undertaken by Dr. Marc van der Valk, from the Academic Medical Center in Amsterdam. Dr. van der Valk has previously published data that showed improved lipid profiles in patients on nevirapine (NVP, Viramune)-containing regimens. The Con argument was put forth by Dr. Sam Bozzette, from the University of California in San Diego. Dr. Bozzette was lead author of a study recently published in the New England Journal of Medicine that showed an overall reduction in cardiovascular-related hospitalizations in the Veterans Administration (VA) medical system, despite the increasing use of ART and PIs.
Dr. van der Valk was selected to speak first and presented the evidence and arguments for the Pro side of the argument -- that, based on cardiovascular risk, clinicians should modify ART.
Dr. van der Valk began his presentation with a cartoon of an ostrich with its head in the sand. The analogy was that some HIV doctors would say that until vascular complications of ART are proven no action is warranted. The next slide was a picture of a shark which represented the overly aggressive physician who perhaps modified therapy even in situations where it might not be necessary. The final introductory slide was a cartoon of Bugs Bunny with the caption "Be wise, doc" -- reflecting the middle ground approach to this issue.
The first study Dr. van der Valk discussed was a study from Dr. Sharon Riddler in the Journal of the American Medical Association. Dr. Riddler tracked lipid levels following HIV seroconversion and found that total cholesterol, LDL cholesterol and HDL cholesterol all declined after acquisition of HIV infection. Her findings were consistent with prior studies that showed reduced cholesterol levels associated with chronic HIV infection.
Dr. van der Valk then discussed the metabolic effects attributed to PIs which included increased plasma cholesterol, LDL cholesterol and triglycerides. Other metabolic changes linked to PIs include insulin resistance, fat redistribution and reductions in HDL cholesterol.
Dr. van der Valk referenced four surrogate marker studies that linked PIs with higher risk for vascular complications. He selected studies that showed PIs were associated with increased carotid intima-media thickness and increased coronary artery calcium scores. These tests (i.e., ultrasound to measure carotid artery intima-media thickness as well as ultrafast CT of heart to measure coronary artery calcium scores. Some cardiologists also consider endothelial dysfunction a surrogate marker [lots of ways to measure -- most common is non-invasive flow-mediated vasodilation]) both determine presence of subclinical atherosclerosis and are widely used in the cardiovascular literature. He also discussed one study that showed endothelial dysfunction with PIs based on measurement of brachial artery flow mediated vasodilation. The last study showed that PIs were associated with blood markers of diminished fibrinolysis, suggesting a higher risk of thrombosis.
Switching species from humans to mice, Dr. van der Valk presented some interesting data from a model of genetically modified mice with knockout of LDL receptor genes. These mice are prone to the development of atherosclerosis. The mice were treated with either low-dose PIs or high-dose PIs. The high-dose group had elevated lipid levels but the low-dose group had no significant change in lipids compared to control. Notably, despite the lack of lipid elevations in the mice treated with low-dose PIs, they were found to have significant increases in atherosclerotic lesions in the aorta. The implication of this study is that atherosclerosis promotion by PIs may occur by mechanisms independent of lipid levels.
Dr. van der Valk then presented several studies comparing the lipid effects of PI-based therapy to NNRTI and triple nucleoside-based therapy for both initial therapy and in switch studies. Not surprisingly, these studies showed that non-PI regimens had more favorable lipid profiles compared to PI-based therapy.
Dr. van der Valk reminded us of the Framingham study in which low HDL had been associated with increased risk of vascular events. In the VA HIT study, patients with known coronary artery disease (CAD) and isolated low HDL were treated with gemfibrozil or placebo. Patients on gemfibrozil had a 6 percent increase in HDL and a 22 percent reduction in cardiac events. Dr. van der Valk also referenced a recent study in the journal Circulation that showed HDL infusion leading to direct improvements in endothelial dysfunction.
Dr. van der Valk's argument was completed with the mention of two very informative theoretical case studies. The first was of a 40-year-old male with no CAD risk factors. The 10-year risk for this person on no ART was about 1 percent based on Framingham calculations. With PI-based ART, this low-risk person would have an increase to about 3 percent for 10 years (assumption of cholesterol elevations derived from lopinavir/ritonavir [LPV/r, Kaletra] studies). With NNRTI-based ART, this low-risk person would have the same 1 percent 10-year risk as a non-ART 40-year-old male.
The second case study involved a 40-year-old male smoker with hypertension and a family history of CAD. The 10-year risk of CAD with no ART for this person would be about 5 percent. If on PI-based ART, the risk would increase to about 18 percent. If this person could be convinced to quit smoking, the risk would decrease to about 9 percent. If hypertension-controlled risk would be about 5 percent, and if lipids were normalized, the risk would decrease to about 3 percent. Alternatively, if NNRTI-based ART were used, the 10-year risk would be the same 5 percent as no ART.
These two case studies illustrated his point: that thoughtful selection or modification of ART, together with appropriate attention to other cardiovascular risk factors, both have a role in risk reduction.
Dr. Bozzette had the more difficult task of arguing the less popular Con side of the argument. Most HIV-treating physicians do not believe that HIV-infected patients are magically protected against the metabolic effects of PIs, so defending the Con argument (that we should not modify ART based on cardiovascular risk) posed a challenge.
Dr. Bozzette began his discussion by asking a series of questions about the question in debate: "Should we modify antiretroviral treatment based on cardiovascular risk?" He asked who is the "we" in the question? Are we talking just about the developed world, or are we including the less developed world? When we use the word "modify," does this mean treatment initiation decisions or the content of therapy?
He brought up the fact that here we are in Paris debating the nuances of therapy at an International AIDS Society meeting while the vast majority of people affected by this disease worldwide do not even have access to medications. His point was pertinent and underscores the vast chasm separating treatment of HIV infection in the less developed world from the developed world.
Dr. Bozzette then showed a slide which summarized the large cohort studies that have been reported on cardiovascular events in HIV-infected patients. He made the critical point that it is difficult to interpret cohort data and multivariate analysis. He showed how the results from the VA cohort differed when assessed by multivariate analysis compared to the first go-round with univariate analysis. He did concede that the majority of these studies showed higher risk of events with ART, although there were relatively few events.
He then showed data from the D:A:D cohort study that looked at attributable risk. In this study, there was a relative risk (RR) of 1.27 for myocardial infarction associated with year of ART. However, smokers had a RR of 2.63, and patients with a previous diagnosis of cardiovascular disease had a RR of 5.41. Given these data, shouldn't we be spending more energy getting patients to quit smoking?
(How about getting the European doctors and healthcare providers attending the conference in Paris to quit smoking as well?)
Dr. Bozzette then posed the question, "what is the downside to modifying therapy?" Changing to another medication is not without risk. What about considerations of drug-induced hepatitis, adherence, new toxicity and development of resistance? This is in effect a restatement of the adage, "if it's not broke, don't fix it." This point is well taken; we may not have enough proof yet to say that what we are doing is broken.
He also indicated that when ART became available in the VA system there was virtually no use of lipid-lowering therapy in HIV-infected patients. Statin use has increased dramatically, and this trend will mitigate negative metabolic effects. The point remains that cardiovascular events are lower overall in this very large VA patient population.
Questions from the audience were entertained and the question of patient preference was addressed. Both speakers agreed that the consideration of patient preference or choice in medications was a given in HIV management.
An overlooked point was raised by a statement from a member of the audience. Risk factors and metabolic changes may be very different between ART regimens at one or two years, but by year four there may be little difference. Our time frame for this analysis of the topic may be too brief. The point was also made that the initial choice of NNRTI may not matter so much because, if resistance develops, the patient will probably end up on a PI after a year or two anyway.
Another physician from the audience made the point that surrogate marker studies comparing PI and NNRTI-based therapy showed mixed results, but that it made sense to select lipid-neutral therapy in patients with multiple cardiovascular risk factors.
At this point, Dr. Bozzette conceded that it was reasonable to modify therapy in patients with multiple risk factors. It must be said that both speakers did an excellent job and presented thoughtful and interesting arguments in favor of their side of the controversy.
In summary, we know that opportunistic complications of HIV infection and mortality rates have been hugely reduced with ART and overall benefits clearly outweigh risks. However, based on the metabolic changes we are seeing now, we justifiably can predict higher rates of vascular complications in the future. This means that prevention efforts, as well as strategies for early diagnosis and treatment, should be put into place now.
|Please note: Knowledge about HIV changes rapidly. Note the date of this summary's publication, and before treating patients or employing any therapies described in these materials, verify all information independently. If you are a patient, please consult a doctor or other medical professional before acting on any of the information presented in this summary. For a complete listing of our most recent conference coverage, click here.|