March 28, 2002
Several studies looked at the risk of heart attacks and strokes in people with HIV on antiretroviral therapy (ART) and found conflicting results. A large study conducted by the Centers for Diseases Control (CDC) found a higher rate of heart attacks (13 in 3,013) in people who had taken PIs, compared to those who did not take PIs (2 in 2,663). However, a very large review of veterans with HIV in the U.S from 1993 to 2001 found a large decrease in overall death rates due to the use of ART, and a slight decrease in both heart attacks and strokes over the last 8 years.
The focus on cardiovascular disease stems from the metabolic changes seen in people with HIV, especially those on antiretrovirals. Many studies tried to identify the causes of the elevated cholesterol and triglyceride levels seen in many people on ART, without any definite conclusions. What is clear is that people with HIV infection on ART should attempt to minimize their risks for heart disease through smoking cessation, increased exercise, healthier diets, and lowering of blood fats with drug therapy, if necessary.
The assessment of the cases and controls included their antiretroviral and metabolic drug history, HIV disease, fasting blood test results, DEXA scan and abdominal CT (two different types of x-ray imaging that can help quantify amounts of body fat). One model was proposed that included the following as indicating LD: a low trunk-to-peripheral-fat ratio and low percentage of leg fat by DEXA, report of abdominal bloating, lower alcohol consumption, higher waist-to-hip ratio, higher total-to-HDL cholesterol ratio, higher anion gap, report of increased bleeding tendency, higher intra-abdominal to extra-abdominal fat ratio by CT, and greater age. This model had a sensitivity of 84.4 percent (i.e., using this model, about 84 percent of the cases would be detected) and specificity of 81.4 percent (of the cases detected, only 81 percent of them would truly have LD). Unfortunately, definitions using only DEXA and CT or only labs were even less sensitive and specific.
Carr concluded that this model can diagnose HIV LD "simply and objectively" and the research team plans to make a scoring system available on the Internet to help healthcare providers diagnose LD in practice. It remains to be seen if this new definition, which is not all that simple, is clinically useful. The assessments used in the model, such as DEXA and CT scan, are often not used in clinical practice, because they are costly and their interpretation is not well standardized. Thus, although a step in the right direction, this definition for LD might be more useful for future research studies than for practice. Validation for the definition is still necessary as well.
The etiology of LD also remains in debate, which contributes to the difficulty of defining the syndrome. Various antiretroviral drugs have been implicated in the body changes and metabolic derangements associated with LD. "Switch" studies were presented at the Ninth Retrovirus Conference to determine if swapping one drug, used as part of a highly active anti-retroviral therapy (HAART) regimen, for another, would improve LD.
The NRTIs have been implicated as a cause of LD, mainly fat loss (lipoatrophy). Of the NRTIs, Zerit is thought most frequently to be the culprit. Studies adding NRTIs to regimens have noted an increase in the frequency of LD over time (Abstract 683). An earlier study conducted in France by T. Saint-Marc evaluated the switch from Zerit to either AZT or Ziagen in persons with lipoatrophy and found statistically significant improvements in subcutaneous fat in the abdomen and legs.
Three switch studies presented at this year's Ninth Retrovirus Conference confirm the improvement in body shape, but found that the amount of improvement is likely not clinically significant.
The first of these switch studies randomized 111 participants with moderate to severe lipoatrophy on Zerit or AZT with an undetectable viral load to either continue their therapy or switch to Ziagen (Abstract 32). After the switch to Ziagen, virologic control was maintained. After 24 weeks there were very small statistically significant increases in limb and abdominal subcutaneous fat. However, these increases were not clinically significant (they could not be seen with the naked eye). Of note, the people who continued their original regimen did not have progression of their lipoatrophy and switching to Ziagen for the minimal improvements seen was associated with a 10 percent risk of hypersensitivity.
The other two similar NRTI switch studies also compared people on a Zerit-containing regimen to people who switched to other NRTIs such as AZT, Ziagen or Trizivir (Abstracts 700, 701). Using DEXA scans, both studies demonstrated small improvements in subcutaneous fat that was greater in the arms than in the legs. Of note, fat changes in the face were not addressed in any of these studies.
Taken together, these switch studies do provide confirmation that Zerit is likely responsible for at least some of the lipoatrophy observed in people. A switch to Ziagen appears to stop further fat loss and indeed resulted in slow and modest improvement in subcutaneous fat levels. Yet none of these studies had dramatic, or even clinically significant increases in fat. Participants and healthcare providers were unable to actually appreciate the changes detected by the imaging techniques used. The lack of significant improvement of fat levels could indicate long-term irreversible toxicity of the medications used, it may just mean either longer follow-up is needed, or that there are other yet unidentified factors contributing to the fat loss. In practice, Zerit has been generally well tolerated and easy to take. The minimal improvements with switching to another drug may not be worth the risk of toxicity and losing control of the HIV, although in the people studied there was continued viral suppression after the switch to Ziagen.
PIs have also been implicated in the development of LD, but mostly with fat accumulation rather than fat loss. They are thought to play a role in causing the metabolic changes (increased blood sugar and cholesterol) seen with HAART, as well as the changes in body shape. A large number of case reports and several studies have looked at the effect of switching from a PI-containing regimen to a PI-sparing regimen on cholesterol, triglycerides and body shape changes. Two studies presented at the Ninth Retrovirus Conference this year randomized people on a PI-containing regimen with stable viral suppression to either Ziagen, Sustiva or Viramune (Abstracts 17, 699).
Switching from a PI to an NNRTI is associated with metabolic improvements (decreases in bad cholesterol and improvements in good cholesterol). It is assumed that improved lipid profiles mean a decreased risk of heart disease and other vascular complications, because this has been clearly shown in HIV- people who lower their cholesterol. Changing regimens to decrease cholesterol will be most important in people with known heart disease or other cardiac risk factors. For those with no risk factors and only modest elevations of cholesterol it is not clear that changing antiviral regimens is necessary. Switches to Ziagen are less safe from the virologic point of view, especially if someone has extensive prior nucleoside experience, but better tolerated in regard to side effects.
Reversal of abnormalities in glucose and lipid metabolism when the PIs are replaced with an NNRTI or Ziagen (when viremia remains undetectable) supports the conclusion that PIs are, at least in part, responsible for these metabolic changes. The lack of significant improvement in body fat redistribution after switching from Zerit or a PI-containing regimen suggests that the causes of changes in body habitus are more complex. Other studies have implicated host factors, such as race, as well as disease factors such as viral suppression, history of low CD4 count, and change in CD4 count in the development of LD.
Without knowing the actual cause, treatment of LD remains difficult. There has been considerable interest in using "glitazones" (a class of drug used for diabetes) in the treatment of LD after two small studies had some promising results. The glitazones in general work by increasing the body's sensitivity to insulin, and in some diabetics can lead to weight gain and reduction in the LDL (the bad cholesterol). In the lab, rosiglitazone, one of the drugs from this class, prevents some of the deleterious effects of PIs, providing the rationale for using it in the treatment of HIV-related LD (Abstract 690). Unfortunately, very disappointing results were presented from the first prospective, randomized trial using high-dose rosiglitazone in 30 HIV+ people with LD (Abstract 13). At 24 weeks, although there was improvement in insulin resistance, there was no difference in either weight or fat mass between those who took rosiglitazone and those who took placebo. Several people, surprisingly, developed severe hyperlipidemia on rosiglitazone therapy, which caused one person to stop taking the drug. Despite the small sample size, the future of rosiglitazone as a treatment for HIV-associated LD does not look promising. Treatment with gemfibrozil and metformin (two other diabetic medications) did not look promising either (Abstract 702). More studies are certainly needed to better determine the causes of LD in order to develop better treatments for the syndrome.