icon-folder.gif   Conference Reports for NATAP  
  Lipodystrophy Workshop
July 8-11, 2003, Paris
Back grey_arrow_rt.gif
Glucose metabolism/insulin resistance, lipids and cardiovascular: report from Paris, IAS & Lipodystrophy Conferences
  Written by Michael Dube, MD, University of Indiana and the ACTG Lipodystrophy Workshop; contribution by Jules Levin July 8-10, Paris, 2003
With each passing year we are gaining considerable insight into these issues, but of course progress regarding effective interventions has lagged behind understanding of the mechanisms involved. Still, there is the hope that loss of fat from beneath the skin, or lipoatrophy, may be to some extent avoidable and partly treatable. Interventions that favorably affect glucose metabolism/insulin resistance, lipids and cardiovascular risk factors in the general population appear to have similar beneficial effects in HIV-infected individuals.
Glucose metabolism/insulin resistance:
Several studies have documented that the insulin sensitizers metformin and rosiglitazone can improve insulin resistance. Metformin does not appear to improve lipoatrophy, but rosiglitazone might. In a recent study subjects with HIV showed improved insulin sensitivity from rosiglitazone but no improvement in lipoatrophy (Yki-Jarvinen, ATVB 2003) but there was improvement in another (Gelato M, JAIDS 2002). At this meeting, Hadigan from Harvard presented a placebo-controlled study of rosiglitazone in 28 subjects with lipoatrophy (abstract 12). In the initial 3 months of study, rosiglitazone was given at a relatively low-dose of 4 mg per day or placebo; then in the second 3 months rosi was given at a full 8 mg per day to all subjects (open-label phase). During the initial 3-month placebo-controlled phase, while total body fat increased as measured by BIA (generally considered to be less accurate than DEXA scanning) it did not increase by DEXA scanning. There was a non-significant trend for increased leg fat (arm fat was not reported due to technical difficulties) and subcutaneous fat (beneath the skin) in the abdomen increased as measured by CT scanning in rosiglitazone recipients, as compared to placebo recipients. Subjectively, there was improvement in fat loss with rosiglitzone. But there was no significant increase in leg fat or total body fat by DEXA scanning in the open-label phase either. Rosiglitazone also increased cholesterol levels and adiponectin levels (see below notes on adiponectin). This study gives some hope that rosiglitazone might reverse fat loss but is certainly not definitive. Since this study enrolled only subjects who were highly insulin-resistant (as did the study by Gelato) but the negative study of Yki-Jarvinen did not, it may be that only those individuals who are highly insulin-resistant to begin with will benefit. Again, larger studies are needed, but I think it would be fair to say that if there is benefit from rosiglitazone in lipoatrophy, its benefits are likely to be relatively small overall. More information is needed to determine who is likely to benefit from it most. Ongoing studies in Australia and the ACTG may eventually clarify some of these issues.
Conflicting data exist regarding the ability of protease inhibitors to affect insulin production by the B-cells of the pancreas. A study by Koster et al from Washington University using B-cells obtained from 2 humans suggested that the PI indinavir could reduce the ability of the B-cell to produce insulin. Other recent data are conflicting. In HIV-infected subjects (Woerle H, Diabetes 2003) there was a 25% decrease in B-cell function 12 weeks after starting a PI-based regimen (usually nefinavir) but in healthy volunteers given indinavir for 4 weeks (Shankar S, ADA meeting 2003) B-cell function remained intact. Clearly, more work is needed to determine if the function of the pancreas to produce insulin is affected by antiretroviral drugs.
It has been observed that in "switch" studies in which PI's are replaced by other antiretrovirals, that an improvement in insulin resistance often occurs, but not always. Early on, a direct PI effect may be what is most responsible for insulin resistance, but later on once lipodystrophy is established, the lipodystrophy itself may be the predominant factor contributing to insulin resistance. This could explain why some studies show improvement after PI switching early in a course of PI treatment, while others done in more heavily-treated subjects with lipodystrophy do not. Marc van der Valk (abstract 8) provided confirmatory evidence that in 8 subjects with advanced lipodystrophy who had PI's stopped, insulin resistance did not improve significantly 96 weeks later. But then neither did fat redistribution. One interpretation of these finding is that "once severe lipodystrophy is present, the damage is done and stopping PI's is not enough to improve insulin sensitivity". Several studies now indicate that it may be lipodystrophy-associated decreases in adiponectin levels that may be a major cause for lipodystrophy-associated insulin resistance (see below).
Switch Studies
Written by Jules Levin
Andrew Carr (abstract 16) reported an update on the 108-patient MITOX study which evaluated switching off d4T to abacavir. In the initial 6-month report, limb fat increased about 10% but this was not clinically significant. The update reports a 2-year follow-up in which patients regained fat loss by 35% at the end of two years after switching. Carr reports that limb fat is generally 7-8 kg in white male HIV-negatives. At the beginning of the study patients in the abacavir arm had 3.54 kg of limb fat and patients in the d4T/AZT arm had 3.75 kg of limb fat. So the patients had lost 50% of their limb fat over the course of 6 years on their NRTI regimen when they started this study. In this study, limb fat had increased 1.26 kg. Since limb fat was 3.54 kg, this was a 36% total increase in limb fat after switching to abacavir in 108 weeks (2 years). DEXA showed right leg fat continually increased over course of 108 weeks for patients who initially switched to abacavir. Right thigh fat increased by regional CT scan by 24% for patients who switched to abacavir and the difference from baseline was significant. Total subcutaneous fat increased (32%) significantly in the patients who initially switched to abacavir after 2 years (p<0.0001). Carr reported that VAT (visceral apidose tissue; fat deep in belly) did not change at 2 years after switch to abacavir. Patient self-report at week 72 did not fully appreciate improved limb fat. Post-hoc analysis (analysis not planned before the study) suggests that fat gain in the right arm by DEXA was not as successful as in right leg. Fat gain occurred in right leg during first year and did not increase in second year. The study did not evaluate facial fat loss. CarrŐs analysis found baseline body mass index was the strongest predictor of greater increase in limb fat. Carr estimated that resolution of fat loss in this study group could be 6 years based on these results, this is the same time these patients were on NRTI therapy before switching. These study results are encouraging. I expect researchers and thought leaders will be scrutinizing this study. The study is expected to be published in the near future in a peer review journal. Of note, you can see below in TARHEEL study that fat gain occurred mostly in arm rather than leg.
At the IAS Conference, Grace McComsey (abstract 51) and GSK reported 48-week results from the TARHEEL Study in which 118 patients with lipoatrophy switched from d4T to abacavir (86) or AZT (32). Baseline DEXA results showed median increases in actual (grams) and percent changes from baseline: arms (239, 35.3%), legs (269, 12.0%) and trunk (859, 16.4%). CT Scan findings confirmed DEXA results. Median increase in subcutaneous abdominal tissue (SAT) was 4% and median decrease in visceral abdominal tissue (VAT) was 2%. 79% of subjects increased SAT and 54% decreased VAT. ABC and ZDV subgroups independently demonstrated positive changes in regional body fat. Similar to CarrŐs findings, increase in trunk fat in DEXA (OR: 2.854, P=0.037). Baseline arm fat in DEXA approached significance in predicting lipoatrophy regression (P=0.059). The positive changes in fat seen by DEXA and CT Scans did not correlate with any changes in anthropometric measurements, which can be insensitive in picking up small changes; only circumferences and waist-hip ratio were done, and not skinfold measurements. The latter is believed by some to be more sensitive in picking up changes. In a biopsy substudy that included 16 subjects, mtDNA in muscle, fat and PBMC improved after switching off of d4T and that improvement was associated with improvement in fat apoptosis.
Lipids and Cardiovascular Diseases
Not a great deal of new information was presented at this meeting on this topic. Levy (abstract 66) compared lipid values in HIV-infected men on HAART with Canadian population norms. Most, but not all, were fasting values. They reported similar findings to those that were recently reported using longitudinal data from the MACS cohort (Riddler, JAMA 2003). Total cholesterol was increased after HAART, after being lower than controls before HAART; this increase appeared to be from triglyceride-rich lipoproteins rather than LDL-cholesterol. Regardless, it suggested some increase in cardiovascular risk profile among HAART-treated men compared to the general population. Hopefully the authors will ultimately report additional analyses from these data, including items like the "non-HDL" cholesterol values and the ratio of total/HDL-cholesterol, which may give better insight into elevated risk among HAART recipients.
A study from France (Mercie, abstract 2) used carotid intimal media thickness (IMT) by ultrasound imaging as a surrogate measure of atherosclerosis. Other groups have used similar methods and have reported conflicting data regarding whether HIV infection or protease inhibitor use is associated with increased IMT. These investigators showed an increased rate at which IMT increased among an HIV infected group, most of whom were on HAART. In a related study (Hsue, abstract 35), reported an increased rate of increase in IMT among HIV-infected subjects. This suggests, but does not prove, that atherosclerosis may progress more rapidly in these patients. HIV disease itself, the drugs used, and immune reconstitution (return of immune function which had been damaged by HIV, for example the increase in CD4 cell count seen with therapy) are all possible reasons. However, Mercie et al was limited by lack of an appropriate control group and more data is clearly needed from long-term studies with proper controls before firm conclusions can be made about what is responsible for more vascular problems among those with HIV.
Shankar (abstract 1), in an update of data presented at Retrovirus last year, reported that when healthy men received indinavir for 4 weeks their endothelial function deteriorates to a similar degree as in type 2 diabetics. The endothelium lines all the blood vessels in the body, and when it does not function properly blood vessels become more prone to atherosclerosis (hardening of the arteries/cholesterol buildup) and heart attacks. Her group will be studying the newer protease inhibitors Kaletra and Reyataz to see if similar effects occur. Certainly, we will want to use drugs that have minimal effects on endothelial function if at all possible. Endothelial dysfunction is one more thing we need to be concerned about in regards to cardiovascular disease, in addition to cholesterol increases and insulin resistance. Measuring endothelial function is currently only a research technique that it not available in routine clinical care.
Adiponectin: Association with insulin resistance, fat redistribution, and vascular disease
A surprising and interesting finding from Shankar et al was that adiponectin levels increased with indinavir treatment. Adiponectin is a hormone known as an "adipocytokine" that is produced by fat cells and improves insulin sensitivity. So, increases in adiponectin are generally thought of as being "good" and decreases in adiponectin (such as has been reported by several groups, including Khatami at this meeting [abstract 44]) that are seen with central obesity and loss of peripheral fat are "bad" because insulin resistance worsens and diabetes and cardiovascular risks increase. But in healthy subjects, adiponectin increased during indinavir treatment, even though indinavir also causes insulin resistance. A potential explanation is that the increased adiponectin might "protect" insulin resistance from being worse than it would have been otherwise, i.e. if the increase had not occurred. In any case, it will be of interest to see how different antiretroviral regimens affect adiponectin levels, in both short and long-term over time in treated individuals. There may very well be a good connection between nucleoside-induced lipoatrophy and fat tissue dysfunction leading to reduced adiponectin and ultimately to worsening of insulin resistance and cardiovascular risk.