icon-folder.gif   Conference Reports for NATAP  
  6th International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV
Washington, D.C., USA
25-28 October 200
Back grey_arrow_rt.gif
Lipodystrophy Meeting Report: lipoatrophy; facial fillers
  Can we prevent or improve lipoatrophy?
Written for NATAP by Graeme Moyle MD
Chelsea and Westminster Hospital, London UK
When asked to define pornography Supreme Court Justice Potter Stewart said "I can't define it but I know it when I see it". The sentiment of this response can equally be applied to lipodystrophy. Definitions of the syndrome have proliferated more rapidly than successful treatments since the widespread recognition of the metabolic and morphologic changes observed during antiretroviral therapy. The morphological changes observed are highly stigmatizing to individuals and the metabolic manifestations may contribute to a range of morbidities, most notably premature heart attacks and stroke. Anxiety about developing the lipodystrophy may lead to risking HIV disease progression through delaying the commencement of therapy or deciding to stop therapy to prevent or attempt to manage the problems, and risking long-term management through modification of established therapy to alternative regimens.
The issues typically considered under the lipodystrophy heading, but which may be observed individually or in combination in persons on antiretroviral therapy, include:
• Dyslipidaemia with raised total cholesterol, low HDL cholesterol and raised triglycerides
• Insulin resistance with hyperglycaemia, particularly in susceptible individuals
• Visceral, breast and/or local fat accumulation
• Generalized diminution of subcutaneous fat mass (lipoatrophy).
This NATAP report will focus on new considerations regarding the causes (etiology) and treatments for fat loss.
Only through understanding the etiology can optimal management of these problems be established. The etiology specifically of peripheral lipoatrophy arising antiretroviral therapy remains speculative. Evidence from cross sectional surveys point to an interaction between HIV disease, host genetics and antiretroviral medication. Opinions differ as to lipoatrophy's causation. Broadly, there are those who think the syndrome is multifactorial, others who believe that it is predominantly a drug-related side-effect with a particular contribution from NRTI-related mitochondrial DNA polymerase gamma inhibition (the mitochondrial toxicity theory), and those who believe it is an immune reconstitution or cytokine mediated phenomenon. Available data could be used to support all these points of view.
Lessons from Inherited Lipodystrophies
A plenary discussion at the conference looked at the genetic origins of familial lipoatrophy/lipodytstrophies. The most well studied are inherited total lipoatrophy Bernedelli-Siep syndrome and familial partial lipoatrophy, Dunnigan's lipoatrophy. In Bernedelli-Siep syndrome children are born with a paucity of fat and never develop subcutaneous fat. Children generally have severe dyslipidemia, constant hunger (due to a lack of the adipocyte produced 'satiety factor' leptin) and become diabetic. The children have visceral fat accumulation, hepatic steatosis and accelerated vascular disease. Two genetic errors have been identified, neither of which appear evident in persons with HIV lipodystrophy. Management is with lipid lowering, insulin sensitizing therapies and an experimental leptin analog. So far no good.
In Dunnigan's lipodystrophy, fat loss is predominately in the limbs with preservation or accumulation of head and neck fat. The syndrome usually expresses itself in adolescence although gene defects are, of course, present from conception. Similar metabolic derangements to Bernadelli-Siep syndrome are present although often to a less severe degree. Indeed, persons with the same gene defect may very substantially in the physical expression of the syndrome. Of note, individuals who remain active and have a normal body mass index (BMI) tend to have less overt manifestations and better metabolic markers. Treatment is as for Bernadelli-Siep syndrome although thiazolidione (TZD) drugs such as rosiglitazone may be particularly useful and a related TZD troglitazone has been associated with some fat recovery. Two gene defects have been reported. The more common defect is in a gene which encodes for the lamins that support the cell nucleous. Individuals with these defective lamins may have cells with bizarre nuclei. Other lamin-defect related conditions include some muscular dystrophies. Persons with HIV associated lipoatrophy do not appear to have lamin defects, although research in this regard are limited. In the other form of Dunnigan's lipodystrophy a defect in the cell signal ppar-gamma is present. This cell signal is associated with insulin sensitivity and has been reprted in vitro to be reduced in the presence of protease inhibitors. The TDZ drugs up-regulate ppar-gamma. This looks like a better line of enquiry...
TZD drugs have been tried in persons with HIV lipoatrophy. Over 12 weeks some studies have suggested benefits relative to placebo. However, over 24 and 48 weeks no benefits were seen, regardless of whether patients were taking or not taking protease inhibitors, taking or not taking d4T or AZT. Further data on the largest placebo controlled study of Rosigltazone 8mg/day indicated that even 96 weeks of therapy with this agent did not improve limb fat mass. Whilst insulin sensitivity improved with rosigltazone, lipids tended to get worse. This has also been seen with some studies of rosiglitazone in diabetic patients. Whilst the other approved TZD pioglitazone has not been tested (and is known in diabetics to have a slightly better lipid profile), the data with rosiglitazone and some drug interaction issues with pioglitazone, suggest large scale studies of pioglitazone are unlikely to be pursued. Drugs which act via several different ppars are currently under development so may warrant consideration after evaluation in diabetic subjects.
The Role of Thymidine Nuclesoides Analogs
While none of the inherited lipodystrophies described were associated with mitochondrial disorders, belief that mitochondrial toxicity is involved in fat loss remains prevalent. Nucleoside/tide analogues such as lamivudine, emtricitabine, abacavir and tenofovir which appear to have low potential for mitochondrial toxicity in vitro do not generally appear associated with limb fat loss, even over prolonged follow-up. Other mitochondrial toxicity associated events such as lactic acidosis, myopathy and peripheral neuropathy are also very infrequently observed with these drugs. Preference for these drugs as initial therapy may not therefore be simply based on their established potency and availability in two-drugs-in-one-pill formulations suitable for once daily dosing but also on the reduced risk of long term side effects.
Data from the GS903 study, in which tenofovir was compared with d4T, each with 3TC and efavirenz as initial therapy, was subjected to scrutiny by the HIV Lipodystrophy Case Definition.
This published definition of lipodystrophy is based one 417 'cases' and 371 'controls' (based on physicians' physical examination and patient questionnaire agreement on the presence of 'moderate or severe' lipoatrophy) who were subjected to demographic, HIV disease status, antiretroviral and metabolic drug histories, anthropometry, fasting metabolic bloods, dual-energy x-ray absorptiometry (DEXA) and abdominal computerized tomography (CT) scan. A definition ('best fit') model was then derived comprising age, gender, duration of HIV infection, HIV disease stage, waist:hip ratio, anion gap, high-density lipoprotein cholesterol, trunk:peripheral fat ratio, leg fat percent, and intra:extra-abdominal fat ratio, which had sensitivity of 79 percent (95% CI 70 -- 85%) and specificity of 80% (95% CI, 71 --87%) for diagnosis of lipodystrophy. (In other words the model and the physician/patient disagreed about 20% of the time on whether an individual had or didn't have lipodystrophy.
The conclusion being that not only do physicians and patients commonly disagree about lipodystrophy, but blood tests and scanners cannot necessarily determine what the physician and patient are seeing. Other models involving less technical assessment were also developed. Grades of severity were also established.
Spontaneous reports of lipodystrophy in GS903 by investigators suggested that at week 96, 1% of tenofovir treated and 12% of d4T treated patients had lipodystrophy. By week 144 these figures were 3% and 19%, respectively. DEXA scan data suggested that at both time points tenofovir patients had on average 'normal' limb fat (~8kg) while d4T patients had low limb fat (~5kg). Applying the version of the case definition which used DEXA but not CT data to the GS903 data, the figures for lipodystrophy cases in the TDF and d4T groups were 6% and 31%, respectively at week 96. At week 144, no patients in the TDF group had a grade 2, 3 or 4 lipodystrophy by the case definition while 7.9%, 2.6% and 3.5% of d4T patients had grade 2, 3 and 4 changes, respectively. While a few cases of mild lipodystrophy were present in the tenofovir group, the presenter, Dr Andrew Carr, indicated that a similar rate of 'mild' lipodystrophy would be present in the general population. That is to say the tenofovir population in GS903, after 3 years treatment with TDF+3TC+EFV had no more lipodystophy than would be expected in the general population. This appears to fairly well exonerate these drugs from a role in causing lipoatrophy but suggests that investigators in their spontaneous reports may be substantially underestimating the extent of lipodystrophy in the d4T group.
Now that d4T use has declined dramatically, the spotlight has turned to the contribution of the other widely prescribed analog AZT. Epidemiology and limited prospective data indicate AZT is a contributor to fat loss but that fat is lost at a slower rate, and perhaps to a lesser extent than with d4T. It is clear that AZT, in a similar way to stavudine, affects mitochondrial DNA content in adipocytes. A clinical study following changes in mitochondrial DNA content in biopsies of subcutaneous fat from individuals starting AZT or d4T-based therapy demonstrated after the commencement of antiretroviral therapy there was a progressive decline in adipocyte mitochondrial DNA content. The decline in mitochondrial DNA was more rapid and more severe with d4T but also occurred with AZT. Individuals who received regimens based on abacavir or tenofovir did not see declines in mitochondrial DNA content overtime and individuals who switched away from d4T or AZT to a thymidine-sparing regimens saw recovery in mitochondrial DNA content to 'normal' levels. Declines in mitochondrial DNA were noted in fat samples drawn relatively soon after initiation of ART despite this being a time when fat mass is typically rising. This may be because health status is still improving over the initial months of therapy, leading to fat and weight gain, while mitochondrial DNA levels may still be above the threshold at which cellular dysfunction begins. Once the threshold is reached fat mass begins to decline. Thus, the presenter Dr David Nolan, suggested that early changes in mitochondrial DNA in fat may predict, or at least be a harbinger of future limb fat loss. Unfortunately, previous data have suggested that mitochondrial DNA from more accessible sources, such as circulating PBMCs, may not be predictive of adipocyte mitochondrial DNA injury, as drug have tissue specific effects.
Attention to AZT's possible role in fat loss was initially limited not only due to the evidently greater effects of d4T but also because in vitro and enzymatic studies with AZT suggested this drug had modest if any effects on mitochondrial DNA or DNA polymerase gamma. This was despite evidence from clinical practice of AZT's effects on mitochondrial function, notable with regards to myopathy and lactic acidosis. A number of studies in the literature have indicated that mitochondrial toxicity with AZT may arise through mechanisms other inhibition of DNA polymerase gamma, the mechanism of mitochondrial damage with 'd'-NRTIs. A study of AZT in non-mitotic cells where it is only activated by the mitochondria-expressed thymidine kinase-2 enzyme suggested that AZT converted to the monophosphate form but very little of the antiviral triphosphate forms. However, AZT or AZT-monophosphate inhibit of phosphorylation of the 'natural' nucleoside thymidine. By blocking of natural thymidine phosphorylation deficiencies in this key building block of mitochondrial DNA eventually may lead to depletion of mitochondrial DNA content or defects in mitochondrial. These effects of AZT on the pool of natural phosphorylated thymidine were observed in both myocardial cells and hepatocytes, cell lines where AZT toxicity may be observed in clinical practice. Repetition of these experiments in adipocytes and pre-adipocytes in certainly warranted.
However, not only do d4T and AZT act via mitochondrial DNA. When administered to healthy volunteers for two weeks both AZT/3TC and d4T/3TC lead to a change in a number of lipid metabolism genes and inhibited mitochondrial RNA transcription. These effects were noted to gradually reverse on withdrawal of the drugs. These data suggest that thymidine nucleosides may hit cellular activity at a number of points which are likely to influence adipocyte function.
This leaves us with no new data regards manipulations of therapy, be it drug switches or drug additions which are likely to benefit fat loss. Wherever possible the approach is to try and switch away from d4T or AZT to a 'benign' backbone, most well evidenced for abacavir, but increasingly likely to be true for tenofovir. The other management tool is facial filling.
Which Facial Filler?
The FDA, EMEA and some European funding bodies, have appreciated that there is a need to make cosmetic surgical treatment available for people with stigmatizing facial lipoatrophy. The agents used in these surgical approaches are known as implants or "facial fillers."
Facial fillers may be natural (such as transferred fat) or synthetic, and they may be biodegradable (temporary) or non-biodegradable (permanent). Of note, some "permanent" fillers may be wholly or partially removable if the underlying condition improves. Fillers are necessary because of the loss of tissue mass from the lipoatrophy process. Face lifts cannot make up for lost tissue but merely tighten some of the skin that has become looser following the loss of tissue mass. As such, the benefits of face lifts for improving facial lipoatrophy are modest and generally transient.
Polylactic acid (NewFill or Sculptura) is an injectable bio-absorbable material that has been approved in Europe for a number of years and is the only agent currently FDA approved for HIV-associated facial lipoatrophy. The substance stimulates dermal fibroblasts to produce collagen, leading to thicker but natural feeling skin. In many ways, its action is similar to that of injectable collagen, except that it has a more durable effect. The injected area requires frequent massage in the days following injection to avoid nodule formation. Results in the temporal area are generally less satisfactory than those in the cheek area. Polylactic acid gradually reabsorbs over a 2- to 3-year period and therefore can be expected to be active over most of this time. 'Top-ups' may be needed at regular intervals to maintain optimum benefits.
Polyacrylamide is not FDA-approved but has been used in clinics in Europe (where various related preparations are approved) and Latin America. Reported data consist mostly of large clinical series accompanied by impressive photographs. These products are relatively viscous and are injected subcutaneously. Following injection, the substances encapsulate as an implant. Long-term safety issues have not been reported.
Fat transfer is a more problematic approach. Many patients do not have fat elsewhere to transfer to facial areas and, if the pathologic process that led to the fat loss is ongoing, the transferred fat can be rapidly lost. Cosmetically, lumpiness may be a problem. The most commonly used method for fat transfer is known as Coleman's technique.
A partially randomized study evaluated the benefits of polylactic acid relative to non-resorbable polyacrylamide and fat transfer by Coleman's technique in 59 individuals with facial lipoatrophy. Patients were treated in a non-standaridzed way, so that treatment was until surgeon and patient were satisfied. Not surprisingly, therefore, visual analog scores of initial treatment satisfaction did not suggest differences and the number of injections required for a satisfactory response did not differ between polylactic acid and polyacrylamide. Fat transfer raised some problems. Three individuals who had fat transferred from a buffalo hump and one person with fat transferred from a standard subcutaneous site developed 'hamster syndrome', as the authors described, with fat accumulation developing and some shifting of the transferred fat. As a result, the authors suggested that fat transfer should not be done with fat from accumulated areas. The study is ongoing and whilst methodologically imperfect, is likely to provide some indication of the durability, pros and cons of these approaches.