icon-folder.gif   Conference Reports for NATAP  
  9th Conference on Retroviruses and Opportunistic Infections
Seattle, Washington, February, 2002
Back grey_arrow_rt.gif
HIV and Cardiovascular Disease: True True or Unrelated? The Debate Continues
Reported by Judith Aberg, MD, Medical Director, HIV Services, washington University School of Medicine, St. Louis, and ACTG researcher
  Although there is increasing concern that the metabolic abnormalities associated with HIV and its therapy may lead to premature coronary heart disease (CHD), there has not been any conclusive data to prove a direct cause and effect. Having said that, I still believe it would be foolish not to have concerns that these metabolic aberrations may in fact lead to premature coronary heart disease. There was a lot of concern expressed after Dr. Sam Bozzette presented late breaker 9 at the Retrovirus Conference, "Cardio- and Cerebrovascular Outcomes with Changing Process of Anti-HIV Therapy in 36,766 US Veterans" stating that there was no evidence for premature CHD. One must take into a couple of considerations before we take Dr. Bozette's statements at face value. One, it is way to early to see a significant alteration in the natural history of CHD. CHD does not develop overnight. It takes 20-30 years. Dr. Bozette's downplaying of the significance of dyslipidemia and glucose intolerance over the past 6 years is like telling a teenager that it is okay to smoke. There will be no evidence of lung cancer or obstructive lung disease in a 20 year-old man who has been smoking for 6-8 years. But there is a strong association of both lung cancer and lung disease in individuals who have been smoking for more than 20 years. When investigators state that there are concerns that patients may develop premature CHD, we are talking not only about total amount of years it takes to develop atherosclerosis but also the age at which the person may develop disease. It is uncommon for individuals to have a myocardial infarction (MI or heart attack) before the age of 50. What will happen to the children or young adults with HIV on therapy that have developed glucose intolerance, hypertension, lipodystrophy and/or glucose intolerance in 20-30 years?
Another problem is the collection of data. As we all know, retrospective data is only as good as the data that has been entered. The VA database is a wonderfully engineered system and probably one of the best health care data systems. I have worked in several VA hospitals. Not all the data is entered correctly. I have found that medical history information is sometimes missing. I currently work on the medicine wards at a university hospital and am caring for 2 VA patients who intermittently get their care at the VA. I sincerely doubt that the admission information from my hospital will make it to the VA hospital. That will be 2 cardio-cerebrovascular events that may not be captured. So, in all fairness to Dr. Bozette, he presented the available data and the results are as one would expect. I disagree with the implication that providers should not use the fear of metabolic complications in the decision of choice of therapy. I think there are always risks and benefits to everything we prescribe. I personally do believe that one should weigh the long-term metabolic complications as a risk but that controlling immediate (in this case, HIV) disease should come first.
Risk For Women
There were several abstracts exploring the development of CHD from a more basic science perspective. Cell adhesion molecules (CAMs) are thought to play an important role in the early development of atherosclerosis. Elevated levels of CAMs are associated with CHD (MI) and carotid atherosclerosis (stroke). Dr. Baussermann presented Abstract 693-T "Circulating Cell Adhesion Molecules Are Elevated in HIV+ Women" describing a cross-sectional analysis of serum samples from 74 women collected over a 2 year period attempting to correlate levels of CAMs with metabolic tests. There was no statistical difference in amount of CAMs by ethnicity nor by treatment group (PI vs NNRTI vs NRTI vs no therapy). However the levels of CAMs were significantly elevated in all HIV positive women compared with those previously reported in HIV seronegative women. The CAMs appeared to be inversely related to LDL and HDL-3 results. Further studies are needed in this area to explore the relationship of vascular disease and HIV in women.
Effects of HIV Drugs on Lining of Blood Vessels and Blood Flow
In Abstract 692-T "Zidovudine (AZT) Increases Endothelial Superoxide Generation, which Results in Abnormal Endothelium-Dependent Relaxation", Sutliff and colleagues looked at the effects of AZT on the endothelium (lining of the blood vessels including the aorta) in mice. Their results demonstrated that prolonged AZT exposure increases endothelial superoxide production, which results in impaired endothelium-dependent relaxation that ultimately may lead to cardiovascular events. Jumping from mouse to man, Dr. Michael Dube presented LB10 "Effect of Indinavir (IDV) Monotherapy on Endothelial Function in Men without HIV Infection". Dube and colleagues evaluated 6 healthy, HIV seronegative men (5/6 smokers) for various metabolic and endothelial function at baseline and again after taking indinavir for 4 weeks. Each man served as his own control so even though 5 of them did smoke, this effect would have been constant and not have significantly altered the results. They measured leg blood flow during an infusion into the femoral artery of methacholine (a drug which acts on the endothelium to cause dilation of the blood vessels). There was a significant impairment in the response after the subjects had taken the 4 weeks of indinavir compared with their baseline (before indinavir). Although the glucose levels remained similar pre and post indinavir, there was a slight increase in insulin levels suggesting that insulin resistance may play a role in endothelial dysfunction. Of note, these subjects did not experience any significant change in their lipid levels. These results are suggestive that indinavir, a protease inhibitor, may have a direct effect on endothelial function which may pose as a risk for the development of CHD. (editorial note: in speaking with Dr. Dube, he would like to study other protease inhibitors in this fashion).
Those are just a few examples suggesting there may be underlying cellular changes leading to an association of cardiovascular complications and HIV, but we still cannot prove a direct cause and effect. Nor can we conclusively state that patients will develop cardiac complications.
Atazanavir and Lipids
The next area to review is the metabolic abnormalities reported in the clinical trials. Dr. Haas reported the 48-week data on BMS Trial 009, "Atazanavir (ATV) Plus Saquinavir (SQV) Once Daily Favorably Affects Total Cholesterol (TC), Fasting Triglyceride (TG), and Fasting LDL Cholesterol (LDL) Profiles in Patients Failing Prior Therapy". This is a safety and tolerability study comparing ATV/SQV at 2 dosages with SQV/RTV in a randomized, active controlled, blinded fashion in patients who had previously failed HAART. The total cholesterol, fasting LDL and fasting triglyceride (TG) essentially remained stable or decreased for the 34 patients on ATV 400 mg/SQV 1200 mg and the 28 patients on ATV/SQV 1200 mg regimen. The decreases were incredibly small especially when one calculated the absolute decrease. Plus none of the patients had any clinically significant lipid abnormalities to start with. The investigators reported an 11% increase in cholesterol, 23 % in LDL and 93% TG in the patients on SQV 400 mg/RTV 400 mg twice daily. The number of patients participating in the SQV/RTV arm was 23, yet only 11 patients completed 48-week data for cholesterol and out of those 11, only 6 had fasting LDL. Does that mean that the cholesterol level of the other 5 patients was not fasting? If so, what does the 11% increase of total cholesterol mean in someone not fasting? In fact, what does an absolute change of 202 mg/dl to 222 mg/dl mean? Without the other lipid markers, this is impossible to interpret. They also did not report other cardiac risks. Of the 6 who did get a fasting LDL, the median LDL increased by 27 mg/dl. This may or may not be clinically significant. The fasting TG did go up from an absolute of 191 to 369 mg/dl but again it is difficult to interpret a small number of patients and it is also unclear how much risk is associated with this rise in TG. It is important to remember that although differences observed may be statistically significant, they may not always be clinically significant. Nevertheless, although this study was disappointing to me, there is other data that does suggest ATV may be more metabolic friendly and is a welcome therapy for us in the cardiovascular world.
Dr. Pilliero presented Abstract 706-T, "Atazanavir: A Once-Daily Protease Inhibitor with a Superior Lipid Profile: Results of Clinical Trials Beyond Week 48". This poster actually presented the lipid data from two different BMS trials, BMS 007 which was a dose comparison of ATV 200/400/500 vs NFV 750 mg three times a day in ART naive subjects and BMS 008 which was ATV 400 or 600 compared with NFV 1250 twice daily in ART naive subjects. The lipid comparisons were based on the lipid results obtained on ATV 400 mg daily as follows:
  This data is quite encouraging in suggesting that ATV is associated with less lipid abnormalities than what we have encountered with the other protease inhibitors. The next task will be whether we see significant reductions in lipids in individuals with true hyperlipidemia who switch from another PI to ATV.
Dr. Eron reported the results of Abstract 409W, the once vs twice daily dosing of Kaletra in ART naive individuals. Unfortunately, as with all the Kaletra presentations, only extremely high lipids are reported. Out of 19 patients given once daily Kaletra, one subject had a cholesterol of >300 mg/dl and 2 subjects had TG >750 mg/dl. For the twice daily regimen, there was one subject with high cholesterol and TG. Without further breakpoints and fasting collections as recommended by the National Cholesterol Education Program (NCEP), it is difficult to interpret the lipid data in the Kaletra trials.
Dr. Keith Henry from University of Minnesota truly deserves an award for his poster presentation (Abstract 694 T) as he sat there in a chair with an IV heplock still in his arm after being diagnosed with active gallbladder disease. He subsequently underwent surgery and is fine. Dr Henry presented additional results for ACTG 5056s, which is the metabolic study of ACTG 372. C-reactive protein (CRP) was measured in a subset of patients who have been virologically suppressed on an Indinavir containing regimen for a median of 42 months. CRP is a marker of inflammation and elevated levels of CRP have been associated with increase CHD risk. The median CRP was 2.29 mg/L (range 0.18-42.9).
The distribution of CRP results is shown in the table below.
  The ACTG 5056s team showed that the CRP values were associated with increased age, higher fibrinogen levels (another maker for CHD risk), lower HDL levels, higher TG levels, higher insulin levels and higher Framingham CHD risk scores. Notably, LDL, other lipid markers such as apo A1 and B, total cholesterol, homocysteine levels, glucose and weight did not correlate with CRP levels. Nevertheless, this is also intriguing when one considers the definition of the NCEP metabolic syndrome, which includes 3 or more of the following: glucose intolerance (insulin resistance), increased TG (>150 mg/dl), increased waist circumference, hypertension, low HDL and a procoaguable or pro- inflammatory state. Many of the 5056 subjects therefore meet such criteria and are at significant risk for CHD.
Dr. Kumar presented Abstract 33, a prospective study of hyperlipidemia in ART-naive subjects taking Combivir (CBV)/Abacavir (ABC) vs CBV/Nelfinavir (NFV) vs d4T/3TC/NFV. There were two interesting observations in this study. One was that the d4t containing arms had a significant increase in LDL, TC, TG compared with CBV/NFV and CBV/ABC. The CBV/ABC arm had minimal changes in the lipid profiles. The other major observation was that men had a more significant increase in their LDL levels in both the NFV containing arms which is in striking contrast to that reported by Pernerstorfer-Schoen (AIDS 2001;15:725-34) who reported that women developed abnormal lipid and insulin levels while the men did not. Kumar and colleagues evaluated a much larger sample size (n= 258 of which 50% are female) compared with only 27 men and 13 women in the Pernerstorfer-Schoen study.
Abstract 34, "Distinguishable lipid profiles between PI and NNTRI therapy may carry different risk of cardiovascular disease" was a great concept but unfortunately fell short of meeting expectations. For the most part, the lipid data presented was not fasting and there were no results for gender differences available. Fortunately, some of the patients did get fasting levels and they do plan to evaluate those results.
Last but not least, we come back to the opening question on whether HIV and its therapy are associated with CHD. Dr Mauss and colleagues (Abstract 689-T) "Differentiating Hypercholesterolemia Associated with Antiretroviral Therapy (ART)" examined 187 fasting serum samples from patients that were either ART naive (17%), NRTIs only (11%), NNRTI regimen (37%) or PI regimen (35%). Only 1/3 of the patients had high LDL (bad cholesterol), which is similar to the general population and the rest of the subjects had high VLDL which when analyzed revealed large TG-rich particles, which are not thought to be as atherogenic. These subjects also had higher HDLs (good cholesterol) than one normally sees in familial hypertriglyceridemia or the combined dyslipidemia. They concluded that patients on ART might have a lower cardiac risk than others have reported.
Dr Bozzette LB9 (refer to beginning of article) analyzed 8.5 years of data from multiple VA databases for ARV use, hospital admissions, all cause mortality and cardio and cerebro-vascular mortality. As has been noted in previous cohorts, the use of HAART is associated with a significant decrease in morbidity and mortality from HIV associated conditions. Of note, Dr. Bozzette and colleagues showed that the cardiovascular event rate has essentially remained stable to trending downward over the past 8 years. Their conclusions are that this does not support the hypothesis that HIV and its therapy may be associated with CHD. Clearly, I disagree for the reasons stated in the opening and I look forward to the continuation of Dr Bozette's project over the next 20 years.
In contrast, other groups have already noted an increase in CHD events. Reisler and colleagues (Abstract 36) retrospectively reviewed 5 large CPCRA trials (n=3050), which revealed a hazard ratio of 9.29 for a cardiac event compared with the risk of developing AIDS. Dr. Tedaldi and colleagues (Abstract 659 M) reported a higher incidence of cardiovascular disease in HIV positive patients co-infected with HCV compared with HIV infection alone. There were four other abstracts presented, all of which suggested an increase risk of CHD in patients with HIV. Wall and colleagues (Abstract 965-T) prospectively followed 111 patients along with 25 HIV seronegative controls. The median risk for development of CHD in 10 years was 4% in the HIV positive cohort compared with 1% in the HIV negative control group. They further calculated risk based upon ART regimen and reported a risk of 6% in the PI group compared with 3% in those who never received a PI. Studies by Dr. Leport and colleagues (Abstract 697-T) reviewing French cohorts and Dr. Holmberg and colleagues (Abstract 698-T) reviewing the HOPS database support the findings by Wall in that HIV infected patients on PIs had a greater risk of having CHD. In the HOPS cohort, the strong association of having an MI and being on a PI remained even when one adjusted for other CHD risk factors such as smoking, hypertension, diabetes, age, gender and dyslipidemia. On the other hand, Klein and Hurley retrospectively reviewed 4159 HIV positive and 40,000 HIV seronegative charts in the Kaiser Permanente group and were unable to note any differences of CHD events based upon ART treatment type. However they did note increased CHD hospitalizations among HIV positive men compared with HIV negative men (6.5 vs 3.8 events per 1000 person years).
So, the debate continues. There is an increasing body of evidence to suggest that HIV and its therapies are associated with numerous metabolic aberrations such as hyperlipidemia, glucose intolerance, central obesity, increase of pro-inflammatory markers and hypertension, all of which are associated with CHD. Then there are the non-modifiable risk factors such as gender, age and family history. More emphasis needs to be placed on not smoking, diet, exercise and control of diabetes and hypertension. Although one cannot conclusively say that HIV itself or its therapies cause CHD, there is more data suggesting that it may increase one's risk for CHD. I personally do not believe we can disregard this. I am taking a pro-active approach and putting CHD risk in the equation for treatment of HIV. Many times providers do prescribe medications that have deleterious consequences and one has to carefully weigh the risks and benefits to all that we prescribe. But we shouldn't ignore what appears to be warning signs and put our patients at unnecessary risks. Hopefully, as we understand the pathogenesis of these metabolic abnormalities, we will learn ways to avoid or treat them effectively.