Lipodystrophy and metabolic complications

By Andrew Carr, MD for


David Cooper reviewed the pre-conference data on pathogenesis at a closing plenary. Although the data are clearly incomplete, his single theme was that all of the in vitro data and most of the clinical data implicate nucleoside analogues (NAs) and protease inhibitors (PIs), especially when used in combination. Although there is no case definition and early studies showed wildly varying prevalence rates, a case definition is pending and prevalence rates and risk factors identified in recent large studies strongly suggest that it is time to set aside the notion that we should do nothing because it may not be real, we don’t know the cause, and “your drug does it but mine doesn’t”. There were some new data that did indeed improve our understanding.

A study by scientists at Bristol-Myers Squibb extended their previous data that stavudine (d4T), zidovudine (AZT) and ritonavir can indeed damage adipocytes in a dose-dependent and synergistic manner, although there was evidence (such as relatively unaffected ATP production and mitochondrial number) that this was not a result of damage to adipocyte mitochondria [Abstract 521].


A thoughtful study from Belloso and colleagues looked at lipodystrophy diagnosis [Abstract 507]. They compared a more “sensitive” diagnostic model, which was based on the presence of only 1 patient reported change in body fat of any severity, with a less sensitive diagnostic model, which was based on at least 2 physical parameters identified by both clinician and patient. The latter diagnostic model was, as one would expect, less sensitive but far more specific. In particular, with the specific model but not the sensitive model cases and controls could be distinguished on DEXA scanning and laboratories. This is further objective evidence that a patient report of lipodystrophy on its own is fairly unreliable diagnostically, and that a patient report is reliable only if there are at least 2 affected sites and their physician agrees.

Prevalence and incidence

The Vancouver group [Abstract 487] reported incidence rates of lipodystrophy of 35% in one year in patients on any therapy (the rate was somewhat higher in those receiving stavudine or a protease inhibitor), a higher rate than reported by the Spanish earlier this year in The Lancet. The regional rates for lipoatrophy, abdominal obesity and buffalo hump were 30%, 20% and 7%, respectively. Development of lipodystrophy was associated with a 50% greater risk of ceasing therapy (29%) than in those without, although it wasn’t clear if lipodystrophy or other adverse events was the reason.


Two studies reported on NA switching and surgery respectively. A pilot, randomised study from our group in Australia [Abstract 96] found that ceasing thymidine nucleosides (stavudine in 16 of 18 patients, all with undetectable plasma HIV RNA), with continuation of the remainder of the HAART regimen, resulted in an increase in limb fat from about 9% to 11% over a 6 month period. There are two disappointments from this study. Firstly, this increase, although statistically significant, still leaves the patients with very substantial peripheral lipoatrophy (a normal level in an adult man would be about 20-25%). Second, 5 of the 9 patients that ceased therapy had virological breakthrough, although 3 could be controlled with other agents.

On a brighter note, a French team [Abstract 500] found that autologous transplantation of fat resulted in objective and sustained (6 months) improvement in facial fat thickness and in appearance, as measured by patients (good or very good by 11 of the 15 patients), doctors, and independent observers.


A chart review from an HIV endocrinology clinic in Houston found that statins (mainly atorvastatin and pravastatin) reduced total cholesterol by a statistically non-significant 13%, much less than would be expected in HIV-uninfected adults [Abstract 489]. This was not due to adverse events. In contrast, gemfibrozil, but not other fibrates, significantly reduced triglycerides by 52%. No agent significantly increased HDL cholesterol, the most important lipid predictor of cardiac disease. A small switch study found that ritonavir-boosted indinavir (100/800 mg BID) had fairly equivalent effects on total and HDL cholesterol as indinavir 800 mg TID, but with a tendency to increase triglycerides. [Abstract 482]

Glucose metabolism

A study from Peter Reiss’ group in Amsterdam showed that a diabetic tendency in 6 men with lipodystrophy is a function of both peripheral and hepatic insulin resistance as well as increased hepatic glucose output. How much this was due to the drugs or the lipodystrophy or both is not clear [Abstract 495 – also in press at AIDS], although follow-up studies after protease inhibitor cessation are underway.

Vascular disease

Diagnosis of hypertension in HIV-infected women was strongly associated with use of protease inhibitor therapy of greater than 2 years duration, after adjustment for other confounders such as obesity, race and age. [Abstract 512] Hypertension was not associated with plasma lipids, which suggests, but certainly doesn’t prove, that protease inhibitors are not the culprit. However, it was not clear whether this increased diagnosis of hypertension was merely a result of more intensive blood pressure monitoring in such women, given the concerns of vascular disease in patients on therapy and because the rates of AIDS and AIDS-related mortality would have fallen substantially. Prospective regular measurement in patients on and not therapy are clearly needed.

Lactic acidemia

An in vitro study [Abstract 95] reported that the peripheral white blood cells of patients with moderately symptomatic lactic acidemia had significantly lower amounts of mitochondrial DNA (mtDNA), when corrected for nuclear DNA (nDNA) content, than did antiretroviral-naïve adults or HIV-uninfected adults (mtDNA:nDNA ratios of 0.3, 0.7 and 1.3, respectively). The mtDNA levels increased after NA withdrawal and resolution of the illness. Whether mtDNA levels can predict symptomatic lactic acidemia remains to be seen. Given that lactate is an easy and very cheap way to diagnose the illness, the value of this test will be if it can predict future illness in asymptomatic patients on therapy. The authors could not explain why they found mtDNA levels to be lower in adults naïve to antiretrovirals, although there is limited data that HIV can impair mitochondrial function in cell culture.

Two studies of asymptomatic lactic acidemia [Abstracts 519 and 520] found very similar prevalence rates (9 and 11%) and risk factors (stavudine and lipodystrophy positively associated; abacavir negatively associated) to what has been reported at previous meetings. There was no data on management of, or on risk factors for, symptomatic lactic acidemia.


Sub study data from the CHARM study (an ongoing randomised factorial study of zidovudine, lamivudine (3TC) and abacavir with either nevirapine, hydroxyurea or both) found that additional randomisation to prednisolone 40 mg daily for 2 weeks did not prevent hypersensitivity. [Abstract 92] If anything the risk was increased with prednisolone (hydroxyurea had no impact). The rash rate in the abacavir/nevirapine group was 20%, versus 6% in the abacavir/placebo group suggesting that most rashes were due to nevirapine not abacavir. However, the mean time to rash was 14 days in the prednisolone group and only 7 days in the placebo group, suggesting that some of the rashes may have been due to abacavir.

GlaxoSmithKline, in a review of 5332 patients who participated in abacavir studies of at least 24 weeks duration, found no factor that predicted the development of hypersensitivity to abacavir, except that black Africans appeared to have a lower risk. [Abstract 527] The rate remained relatively constant at 3.7%.


Two presentations, one from Boerhinger-Ingelheim, clarified the frequency, severity and risk factors for nevirapine hepatotoxicity. [Abstracts 44 and 45] Hepatotoxicity (increased ALT/AST greater than 3 times the upper limit of normal) occurred in 9% of HIV-infected adults with CD4+ T cell counts greater than 350 cells/mm3 within 6 weeks of commencing nevirapine, but only 3% of those with CD4+ T cell counts less than 200 cells/mm3. Hypersensitivity remains strongly associated with hepatitis B or C infection. Clinical hepatitis is rarer of course, about 3%. The authors suggested regular monitoring of liver enzymes.


Unless otherwise stated, all references in the text are to the 1st International IAS Conference on HIV Pathogenesis and Treatment. July 8-11, 2001. Buenos Aires, Argentina.


Copyright: United States and international copyright laws protect the entire design and contents of HIV and Information posted on the site may be used for personal, scientific or informational purposes, but NOT for commercial reproduction

Links to other websites are current at date of posting but not maintained.