Written for NATAP by Graeme Moyle, MD, MBBS, Chelsea & Westminster Hospital in London, UK

• In this Report from Dr Moyle, he discusses why or why not mitochondrial toxicity is related to lipoatrophy.
• Is d4T more associated with lipoatrophy than other NRTIs?
• Does the immune system dysfunction and restoration from HAART associated with HIV cause lipoatrophy?
• He discusses ongoing research and updates of treatment for lipodystrophy: metformin and glitazone drugs; Growth Hormone
• Anabolic steroids and fat loss, lipodystrophy, and LFT elevations
• Update on Switching Therapy to address lipodystrophy & metabolic abnormalities
• Bone Problems in HIV: update on research & treatment intervention

This conference involved around 500 participants from across Europe, Australia, the USA, Africa and Asia. The format of the meeting involved both plenary presentations and oral presentations of new research data. The vast majority of the focus of the meeting was about lipodystrophy and weight loss in persons with HIV infection, and included a debate regarding the mitochondrial hypothesis, which until recently had been considered the most important hypothesis on the developed of - particularly - peripheral lipoatrophy.

Factors associating with lipodystrophy
The Friday morning session began with a discussion by Simon Mallal from Perth, Australia, regarding the relative contributions of drugs, specifically the nucleoside analogues and protease inhibitors to the development of metabolic and morphological (body changes such as fat accumulation in stomach and fat loss in limbs) manifestations of the lipodystrophy syndrome. Dr Mallal split the different manifestations of the syndrome into fat loss, fat gain, lipid metabolism abnormalities and glucose metabolism abnormalities. He suggested that each should be regarded as separate end-points, albeit that they may co-exist within some or many individuals (in other words some individuals may have one or more of these specific syndromes). These manifestations may indeed have a range of contributing factors, including drug and genetic factors, as well as the possibility of contribution from HIV infection and immune reconstitution. He used a range of different levels of evidence to discuss the available data for different contributions of drugs to the lipodystrophy syndrome. Importantly, none of the available evidence is currently definitive - that is to say collected from randomised blinded placebo controlled studies with prospective evaluations of lipodystrophy based on validated definitions. Data are collected predominantly from cross-sectional studies or cohorts evaluating individuals established on therapy over multiple time points. Additional data are available from switch studies and shorter term data from prospective studies - where data on metabolic disturbances, glucose handling abnormalities and lactate elevation have been collected in a prospective manner. Further clinical data is available from small studies that have involved administering protease inhibitors to HIV uninfected individuals and then evaluated lipid or glucose handling outcomes. Additional, lower quality evidence, that was discussed included information gained from animal models, which are predominantly in rat or mouse, in-vitro data - particularly evaluating the potential of different nucleoside analogues to cause mitochondrial dysfunctions in cultured cells and corollaries derived from other clinical diseases, such as the development of lipid and glucose handling abnormalities and possibly morphological problems in individuals with immune problems related to other medical disease.

He provided evidence that there is a strong association between protease inhibitor therapy and abnormal lipid and glucose metabolism. In particular, data derived from studies in healthy volunteers indicates that small doses of ritonavir can lead to elevations in triglycerides and cholesterol values. As updated at this meeting by Dr Noor from San Francisco General, the use of indinavir monotherapy in healthy volunteers, whilst not initially leading to lipid abnormalities, does lead to a rise in insulin resistance over a short period of administration. Studies in healthy volunteers have not been of sufficient duration to establish the possibility that these metabolic disturbances could be in any way associated with body composition changes.

Dr Mallal discussed in detail how subcutaneous fat loss does not appear to be so closely associated with any individual drug or drug class. He drew attention to two studies which have illustrated the risk of lipoatrophy being particularly associated with the combination of protease inhibitors with nucleoside analogues. His own analysis from the Perth, Western Australia cohort of patients demonstrated that whilst individuals who received dual nucleoside combinations developed lipoatrophy over time, the cumulative prevalence of lipoatrophy was markedly greater over the same period of follow-up in individuals who received two nucleosides together with a protease inhibitor. More recently, the Prometheus study and a second Abbott study, both looking at dual protease inhibitor therapy, demonstrated, importantly, that lipoatrophy or fat wasting can occur in individuals who have never received, or are not currently receiving nucleoside analogue therapy, the chance of developing lipoatrophy during follow-up is markedly increased when both nucleoside analogues and protease inhibitors are combined together.

The reasons why this may occur are not clear, however these data underlines that the syndrome is not specifically caused by nucleoside analogues and protease inhibitors as the syndrome can develop in individuals who have neither received protease inhibitors nor received nucleoside analogues. However, it does appear that the optimal conditions for the development of wasting include the use of nucleoside analogues and protease inhibitors together. It is important to underline that this does not necessarily mean that the drugs are causative of the problem, but simply that events that occur when these drugs are combined are in some way contributing to an increased chance of having lipoatrophy at any one time point.

(Editorial note from Jules Levin: a number of studies, which I agree are not well done & definitive, have suggested an association between d4T and lipoatrophy. While studies also show people taking AZT can develop lipoatrophy, the likelihood of developing lipoatrophy with d4T seemed increased when compared to AZT. At the recent 8th Annual Retrovirus Conference Feb 2001, there was such a study showing d4T and AZT were associated with lipoatrophy but the incidence was greate in individuals taking d4T. Granted, this study also had flaws-researchers testing skinfolds were not blinded to which regimen the patients were taking, among other study limitations. In this study from Joly 30% of individuals taking d4T did not have symptoms of lipodystrophy and 57% receiving AZT did not. Here is the link to the NATAP report of this study: An Assessment of Lipodystrophy in Patients Receiving AZT or D4T regimen After Receiving AZT, DDI and/or DDC)

Dr Mallal then went on to discuss how data from predominantly in vitro and animal studies provide evidence how these drugs may contribute to the syndrome by separate mechanisms. He underlined, however, that the data in this regard, from a mechanistic standpoint are largely derived from short term in vitro studies looking at specific cell lines or looking at genetically modified animal species and the extent to which these data are therefore applicable to the human situation remains unclear. In particular, the level at which a defect exists in fat and sugar handling and the mechanisms by which fat cell or fat cell content are lost remain unclear. Dr Mallal proposed that the level of the problem lies in the adipocyte and proposed a 'tap and sink model'. - that is to say that in the normal circumstance an adult fat cell is a dynamic cell, taking in fats and sugar storing them as triglycerides after meals, then this fat is released when fat is required elsewhere. In the situation of antiretroviral therapy, there are contributing factors which are diminishing the flow of fat into fat cells which may involve both protease inhibitor and nucleoside analogue factors and that there may additionally be an increased outflow of fat in the process of lipolysis (release of fat). In vitro studies have demonstrated that when nucleoside analogues and protease inhibitors are combined in cell culture with adipocytes, there appears to be an increased lipolysis (cell decomposition) from those fat cells. The impact of nucleoside analogues on mitochondria could further contribute to diminished cellular energy production - hence diminished transporter activity and may also contribute to release of mitochondria factors that lead to cellular apoptosis (cell death). Biopsy studies performed on individuals at his clinic presented during this meeting indicated a diminution of mitochondrial DNA content and an increase in mitochondrial protein mass in individuals receiving antiretroviral therapy with fat wasting. He proposed that the decrease in mitochondria DNA represents the inhibition in mitochondria polymerase gamma by nucleoside analogues and the compensatory mechanisms within the cell which normally lead to an increase in mitochondrial numbers were driving increased mitochondrial division that was leading to an increased production in mitochondrial protein. In the absence of mitochondrial DNA, these new and enlarged abnormal mitochondria were unlikely to be the efficient energy producing powerhouses that normally exist within human cells. Data is now required to evaluate whether this is the case or whether the increase in mitochondrial protein production reflects an increased cellular energy need as the adipocye (fat cell) struggles with increased fat cycling - an abnormality previously reported by researches at Baylor College, Texas. In this case the diminution of DNA may simply reflect an artefact of sampling, i.e.a relative reduction not an actual reduction.

Ken Lichtenstein from Denver, USA, had earlier at this meeting discussed the interaction between nucleoside analogues and protease inhibitors but suggested that further factors, beyond drugs alone, were required for fat redistribution and lipoatrophy. Additionally, he provided data that suggested that in some individuals lipoatrophy may resolve or partially resolve over time. Dr Lichtenstein presented data from the CDC co-ordinated HOPS database which has used 8 large US HIV treatment practices to evaluate cross-sectionally individuals receiving or not receiving antiretroviral therapy at the time and evaluating them for the presence of moderate or severe lipoatrophy, according to strict definition criteria. Within this database analysis, a number of different statistical associates with lipoatrophy have arisen. These include the use of nucleoside analogues and protease inhibitors and the duration of antiretroviral therapy with both D4T and indinavir being specifically associated with the syndrome. The reason for this statistical association may represent that these drugs are the most commonly used nucleoside analogue and protease inhibitor combination with the population, and therefore have the greatest potential to be associated with a clinical outcome. Other statistical associates include host-factors such as age over 40, disease stage, individuals with AIDS being at greater risk than those with asymptomatic disease. A CD4 nadir of less than 100 and a rise in CD4 cell count of greater than 200 with therapy were also associated with increased risk suggesting an immune role..

One analysis that Dr Lichtenstein presented indicated looking at the two drugs which had become specifically associated with the syndrome in the data base, d4T and IDV. Individuals who received these agents - including for periods of greater than two years each, but who had none of the other additional host or immune response risk factors, did not develop the syndrome. However, the more of the non-drug factors, the greater the chance of the syndrome occurring in the individual. Dr Licthenstein suggested that the data suggested an hypothesis regarding the role of immune response in contributing to the syndrome in genetically or age-associated at risk individuals. This would suggest that it is not really the drugs that are driving the development of the syndrome but rather that the drugs that work best in leading to immune restoration and the drugs that are most well tolerated over time, are the drugs most likely to be associated with the syndrome. In their prospective database, Dr Lichtenstein reported that there were some individuals who had been assessed as having moderate to severe lipoatrophy in the first HOPS analysis but subsequently had had resolution of their lipoatrophy to either mild or absent. Individuals in this group were often those who had experienced either viral rebound during therapy and/or loss of CD4 count number over time. These physical improvements associated with loss of virological control and decline of CD4 count were independent of whether the individuals had discontinued therapy during follow-up. This provides support for the view that the immune system improvements may be involved in the development of lipoatrophy with subsequent loss of virological control and deterioration of CD4 cell number being associated with a degree of resolution.

Dr Caroline Pond, Milton Keynes UK, provided interesting evidence from animals, both laboratory and wild, that adipose (fat) tissue around lymph nodes is preserved even in lean animals (including such as after hibernation) and provided evidence that beyond simply providing fuel to the immune cells, the adipose tissue is acting as a paracrine (hormone or cytokine producing) tissue. Whist not specifically focussing on HIV lipodystrophy, these data provide insights into the intimate relationship between immune system and fat deposits which may be disrupted by HIV infection and immune reconstitution.

Mitochondrial toxicity hypothesis
A further session on Friday was devoted to mitochondrial toxicity. This session began with an elegant outline of how nucleosides analogues may interfere with mitochondrial polymerase gamma and possibly other human DNA polymerases, leading to a deterioration in both mitochondrial DNA quantity and, additionally, diminished mitochondrial function. The consequences, both in vitro and in vivo of diminished mitochondrial function are usually observed when mitochondrial capacity falls below 80% of normal values. In familial disease, this can lead to tissue specific events, most commonly in the most rapidly metabolising tissues, such as the central nervous system, cardiac muscle, skeletal muscle, endocrine and renal systems, and to metabolic derangements, specifically lactic acidosis.

Kees Dr Brinkman discussed in detail available data suggested that, whilst lactic acidosis remains a relatively rare condition, occurring in perhaps 1% or more of individuals treated with nucleoside analogue therapy per annum, a second syndrome of asymptomatic hyperlactataemia may be present in 10% or more of individuals receiving nucleoside analogue therapy. It remains to be clarified whether these episodes of hyperlactataemia are persistent or transient in the majority of individuals. He underlined the importance of sampling technique, to have the individual at rest when taking a blood sample un-cuffed to assess blood lactate. Evidence from a cross-sectional study of his cohort in Amsterdam found an association with nucleoside analogue treatment, the strongest association being with d4T-based regimens, the most commonly used regimens in the cohort. A second prospective evaluation from the Perth cohort in Australia indicated that lactic levels rise over the first six months of nucleoside analogue therapy and then appear to stabilise generally within the normal range, after approximately six months. This suggests a shift in either the production or clearance or both of lactate following nucleoside analogue therapy. Rises in lactate in the Perth cohort were observed to a similar degree in both AZT and D4T treated individuals, although the steady state lactate levels were marginally (0.2mmoll/litre) higher in the d4T group, relative to the AZT group. The clinical significance of such a small difference is not clear. Both these sets of data have not found any predictive value in performing lactate levels with regards to assessing an individual's risk of developing lactic acidosis in the future and therefore the routine use of lactate measurement in clinical practice remains to be evaluated. However, Dr Brinkman underlined that, in individuals that present with symptoms that are suggestive of lactic acidosis, such as fatigue, weight loss, abdominal pain, new onset of tender hepatomegaly, nausea, or vomiting then lactic acidosis should be considered and appropriate tests performed.

Ulrich Walker from Freiburg, Germany and Graeme Moyle from London, UK, then debated the evidence for the existence of mitochondrial toxicity in the pathogenesis of lipoatrophy in individuals receiving antiretroviral therapy. Dr Walker provided evidence, predominantly from in vitro studies and studies with mice where, in the in vitro setting, mitochondrial toxicity can be demonstrated with most nucleoside analogues, given an appropriate choice of cell line and cell activation state and high doses of nucleoside analogues. Similarly, when enough drug is administered to mice, particularly those with susceptible genetic backgrounds, declines in mitochondrial DNA content in a range of tissues, including white adipose tissue may be observed. Biopsy data from a cohort of patients in Freiburg, Germany and additional biopsy studies performed in Hawaii and Perth, Australia, have all demonstrated modest reductions on mitochondrial DNA content and abnormal mitochondrial forms in individuals receiving nucleoside analogue therapy. It remains curious that similar morphological appearances have been observed in areas of lipohypertrophy, such as buffalo humps, as well as areas of fat loss. It remains to be explained if the observed pathology is causative of the clinical appearance, how the same pathology could be observed in areas of fat increase as well as areas of fat decline.

Graeme Moyle provided a refutation for the mitochondrial toxicity hypothesis as the ætiology of lipoatrophy. He started with clinical data as discussed by Simon Mallal, indicating that individuals who have never received nucleoside analogues are still able to develop the syndrome, and that the syndrome would be unusual, relative to other drug-related mitochondrial toxicities, as the central nervous system was not involved. Drawing upon data presented by Dr Walker, he outlined that the levels of mitochondrial DNA depletion observed in fat biopsies are much more modest relative to those observed in familial mitochondrial diseases and that some individuals in lipoatrophy have normal mitochondrial DNA, whilst some control patients without lipoatrophy have depleted quantities of mitochondrial DNA. This suggests that mitochondrial DNA reduction in lipoatrophy is neither necessary, characteristic or diagnostic of the phenomenon. The finding of reduced mitochondrial DNA may be unrelated to the ætiology of the syndrome. This is underlined by recent biopsy work in muscles from people HIV infection receiving or naïve to treatment, which indicated that mitochondrial or respiratory chain abnormalities and histological features consistent with mitochondrial abnormalities may be observed in both groups of individuals. This implies that HIV infection per se may be leading to mitochondrial abnormalities and when individuals are biopsied to evaluate their fat, the findings that are observed may indicate changes that are present long before the development of the morophological changes and are unrelated to those changes. Data on fat oxidation previously presented at other meetings has suggested that individuals with lipoatrophy receiving nucleoside analogues, who are challenged with a fatty meal, oxidise fat normally or, possibly even to a greater degree, relative to healthy HIV positive individuals. Additional data presented at this meeting by Andersen and colleagues indicated that patients with HIV associated lipodystrophy has basal and insulin stimulated levels of glucose oxidation similar to a matched control group of HIV positive individuals without lipodystrophy. This indicates that mitochondrial function was similar within the two groups, although the study did note that individuals with lipodystrophy showed an impaired non-oxidative glucose metabolism relative to controls, suggesting that a defect in glycogen synthesis may have existed in those individuals. Another key issue regarding the mitochondrial hypothesis lies in the observation that, whilst in familial mitochondrial disorders the central nervous system is commonly effected, leading to encephalopathic or dementing processes (similar to that observed in some drug-induced mitochondrial toxicity states, such as lead poisoning), treatment with antiretroviral nucleoside analogues leads to a reduction in the chances of dementia and improvement in the neuro-cognitive function..

Finally, Dr Moyle discussed a Finnish study, published in the Lancet in 2000, indicating that individuals recovering from bone marrow transplants were frequently observed to have dyslipdaemia, hyperinsulinaemia and, in some cases, morphological changes, predominantly obsesity. These phenomenon were more commonly observed in individuals who had chronic inflammatory problems with immune reconstitution. This provides a corollary to HIV associated metabolic disturbances, occurring predominantly in individuals undergoing immune restoration.

Possible Treatments For Metabolic & Body Abnormalities
The conference also discussed possible treatments for the metabolic and physical problems, although, because of the issues in definition of lipodystrophy there was frequent overlap in the discussion of treatments for wasting as well as treatments for predominantly lipoatrophic lipodystrophy. Surprisingly, given that this was a nutrition conference, there was very little discussion of the use of the dietary interventions to manage lipodystrophy.

Possible treatments discussed included the use of acetyl-L-Carnitine which was discussed by Mike Youle from the Royal Free Hospital in London. The focus of his talk was largely about the potential to manage peripheral neuropathy, a nucleoside related problem better documented to be related to mitochondrial toxicity, although in some cases known to be associated with Carnitine deficiency. Dr Youle has observed benefits in individuals with a nucleoside related peripheral neuropathy with the administration of acetyl- L-Carnitine either orally or, initially, intra-muscularly followed by oral administration. Evidence for use of L-Carnitine in lipodystrophy is currently lacking, although one small study previous presented by Stefan Mauss from Germany has indicated that L-Carnitine may lead to some improvements in triglyceride levels, although over six month's follow-up benefits were not observed in terms of body shape abnormalities.

Drugs which may be used to manage lipodystrophy include growth hormone, insulin sensitising agents such as metformin or glitazone drugs and, at least for wasting associated with lean body mass loss rather than fat loss, anabolic steroid use may be considered. Other alternative managements described include switching between antiretrovirals and the use of plastic surgery.

(Editorial note: the use of metformin and glitazone drugs in lipodystrophy are being studied. There may be safety concerns associated with the use of these drugs for certain individuals. Fat loss and elevated liver enzymes are of concern. The ACTG is planning a study combining both metformin and a glitazone. Until more information is available from these studies, experimentation may be harmful.)

Insulin resistance and sensitisers
The use of insulin sensitising agents was reviewed by Dr Stephen Grinspoon, Boston, USA, who discussed the available data with metformin and the potential to use glitazone drugs such as rosiglitazone or pioglitazone to manage some of the clinical manifestations of fat redistribution syndrome. In particular, insulin resistance is thought to be associated with central fat accumulation. Dr Grinspoon described a randomised placebo-controlled study of metformin 500mg bd, which led to improvements in insulin sensitivity, triglycerides, and modest falls in weight, with weight loss predominantly occurring in the intra-abdominal fat, leading to an improvement in the clinical appearance of these individuals. However, as metformin causes weight loss in many individuals and carries a risk of lactic acidosis, something which has not been observed in clinical trials in HIV to date, it may not be an ideal medication for everyone. (editorial note: In other words, there is a risk for aggravating fat loss in limbs.)

Limited data is available on the safety and efficacy of glitazone agents in HIV infection. The first of these drugs, troglitazone, has demonstrated benefit in a cohort of individuals with a familial lipodystrophy. These drugs, which act as agonists of Ppar-gamma, lead to improvements in peripheral insulin sensitivity, reductions in hepatic VLDL and LDL production and may stimulate accelerated differentiation of pre-adipocytes into mature adipocytes. A small cohort of 6 individuals treated with troglitazone has previously been reported from a cohort in Germany where improvements in insulin sensitivity, triglycerides and total cholesterol were demonstrated over a three month period of follow-up. However, troglitazone has now been withdrawn due to concerns over liver function test abnormalities. Liver function test abnormalities with the approved glitazone agents does not appear common, but clearly concerns exist with regards to whether people with HIV infection are at increased risk, and in particular those individuals who are co-infected with hepatitis B or C will require particular monitoring if these agents are considered. In general glitazone therapy lead to increases in weight together with a potential to increase fat mass, hence they would be most suitable for individuals who have lost weight and have predominantly a lipoatrophic presentation.

A subsequent presentation by Mustapha Noor, San Francisco, USA, demonstrated that the protease inhibitor indinavir causes insulin resistance when given to healthy volunteers as a monotherapy. One possible mechanism for this effect has been reported from an in vitro study that suggested that protease inhibitors including indinavir may interfere with the glucose transporter GLUT-4. These data suggests that in individuals with significant insulin resistance, switching away from protease inhibitors may be a beneficial approach.

Anabolic steroids
Available data with anabolic steroids suggest that this may be useful for pesons with lean body mass loss related to HIV wasting. Two studies were presented at this meeting, one looking at an oral antibody steroid - oxymetholone, and the second looking at injectable nandrolone decanoate. Additionally, data on anabolic steroid use was reviewed by Julian Gold from Syndey, Australia. In a study of 92 individuals receiving oxymetholone, Dr Hengge, from Essen in Germany, demonstrated weight gain of 3-4 kilograms at 50mg BID or TID dosing, compared with a gain of only one kilogram in the placebo group. These statistically significant differences in weight were observed in body cell mass and lean body mass but with no change in body fat. 40% of patients receiving oxymetholone developed 'significant' elevations of AST, ALT or GGT, with two patients in the BID arm experiencing grade 3 or 4 liver toxicity and six (21.4%) of those in the TID arm experiencing shifts in liver function tests to at least five times the upper limit of normal. In a study of nandrolone presented by Kathleen Mulligan from San Francisco, USA, 25 eugonadal men with involuntary weight loss of greater than 5% or BMI bellow 20kg/m2 were given 200mg of nandrolone or a placebo for twelve weeks. The nandrolone therapy was associated with increases in lean body mass but no change in fat. Of note, HDL cholesterol decreased with nandrolone therapy, whereas haemoglobin and hematicat increased with nandrolone. Nandrolone is only licensed in Europe and the USA for the treatment of anaemia. Liver function test abnormalities were not observed in the nandrolone study. Neither study evaluated the time to recovery in gonadal function after the discontinuation of anabolic steroid therapy. These data also supports the view than anabolic steroids, whilst being potentially useful in individuals who have lost lean body mass, many not be suitable for individuals with lipodystrophy, where the predominant weight loss relates to loss of fat cell mass. Anabolic steroids in both these studies did not appear to replete fat cell mass and, indeed, in some previous studies, have been linked to a fall in fat mass. Additionally, as indicated in these studies, there are risks of liver function test abnormalities (and, possibly, long term hepatic malignancies) and worsening of HDL to total cholesterol ratios, which may already be low in individuals with HIV infection.

Growth Hormone
Steve Grinspoon from Boston, USA, presented data indicating that individuals with lipodystrophy have lower level of growth hormone relative to HIV infected treated individuals without lipodystrophy, or control subjects without HIV infection. Of note, no relationship was observed between particular drugs and the low levels of growth hormone in these individuals. However, an association was observed between visceral fat and low growth hormone levels, which has previously been reported in other individuals with the metabolic syndrome, sometimes called syndrome X. Don Kotler from New York, USA went on to discuss how low growth hormone in a range of medical settings is associated with visceral adipose tissue increase (fat accumulation in stomach), insulin resistance, elevation in triglycerides and diminishment of HDL cholesterol, increases in fibrinogen and an increase risk of atherosclerosis. Studies outside the HIV setting, in individuals with increased adipose tissue, using doses as low as only 1mg od of Growth Hormone over a period of nine months had previously been reported to lead to decreases in visceral adipose tissue, with relative maintenance of subcutaneous adipose tissue (fat loss in limbs), leading to an improvement in the visceral adipose to sub-cutaneous adipose tissue fat mass ratio. These benefits were also associated with an improvement of diastolic blood pressure and whilst, after the initiation of growth hormone in this setting lead to an initial increase in insulin resistance, over the course of nine month's therapy this abnormality had resolved. Studies in AIDS wasting, where there appears to be relative growth hormone resistance, have indicated that doses of 6mg od for twelve weeks leads to approximately a 3kg gain in fat free mass, although there may be loss of fat in these circumstances. Data from a study evaluating just 3 mg/day of growth hormone indicated that fat mass loss may predominantly be truncal rather then peripheral. Doses of growth hormone evaluated in people with HIV infection are associated with some risk of diabetes, predominantly in the phase when insulin resistance is increased. Individuals who may have an initial worsening of insulin resistance in these circumstances without the development of diabetes may be able to treat through with growth hormone and observe the disappearance of the insulin resistant increase over time.

Dr Kotler presented a recently completed study in New York where patients were treated for an initial 24-week period with 6mg od of growth hormone, followed by a 12 week wash-out period and then a further 24-week treatment phase of 4mg on alternate days. Whilst 30 patients were initially evaluable, only 24 completed 24 weeks of therapy and a further 10 patients did not attend for follow-up after the wash-out period, with only 9 patients completing the second phase of growth hormone therapy. In part this was related to the patients' feeling that the 4mg dose, despite evidence from DEXA and CT scan to the contrary, was less effective than the 6mg dose and they therefore discontinued the study to access growth hormone at higher doses elsewhere. Over 24 weeks of the initial study phase, visceral adipose tissue fell by 45% whereas subcutaneous adipose tissue diminished by just 8%. Most of the benefit that was observed within 12 weeks of initiation of growth hormone therapy, but most was then lost over the 12 week wash-out period. In both treatment phases, significant increases in skeletal muscle mass at approximately 1.5kg were observed. However, adverse events were common at the higher dose, including joint pain and stiffness, carpal tunnel syndrome, flu-like symptoms, headache and worsening of hypertriglyceridaemia. Of considerable concern in this study was that three individuals over the course of the one year of treatment developed malignancies, in one case a basal cell carcinoma (common skin cancer), and in the second case a squamous cell carcinoma of the anus, and the third a cholangiocarcinoma at the site of a pre-existent AIDS-related sclerosing cholangitis and stent (a known risk factor for this tumour). It is unclear whether these events were related to growth hormone therapy. Additionally, three patients in this study developed diabetes; the risk factor observed by Dr Kotler's group being that all individuals that developed diabetes had markedly elevated insulin at baseline, all patients having insulin values at over 35U/l. Several of the participants at the meeting discussed how treatment with growth hormone leads to improvements in appearance of individuals both with fat accumulation and fat atrophy. The mechanism by which the appearance of fat atrophy improves during growth hormone therapy is not know although it may in part relate to fluid retention during therapy. Studies looking at individuals specifically with lipodystrophy and lipoatrophy are now ongoing with a view to potentially combining growth hormone therapy with insulin sensitising agents such as rosiglitazone.

Switching
The potential for therapy switching was discussed by Graeme Moyle, London, England, who provided evidence from available studies that suggested that individuals switching from first line therapy with protease inhibitors to either non-nucleoside or abacavir-based regimens were likely to maintain virological control. Those individuals switching to abacavir appeared to do best, with significant improvements in cholesterol and triglycerides described in several studies, as well as improvements in insulin resistance being observed. Similar data is available from nevirapine from several studies, whereas with efavirenz, whilst the HDL cholesterol may improve, triglycerides and total cholesterol appear to remain stable. Insulin resistance also improves with switching to efavirenz. (Editorial note: my recollection is that study results are mixed on whether EFV improves insulin resistance.)

Less data are available for people who have experienced prior nucleoside analogue based regimens before being fully suppressed on PI-based therapy. However, data with abacavir and nevirapine suggest that there are higher rates of virological failure when these drugs are used for substitution in this circumstance, relative to individuals on first-ever therapy. Therefore, the risk versus benefit of switching in individuals who have had multiple prior treatment regimens, particularly documented nucleoside analogue resistance, may suggest that these patients are not suitable for a switch approaches or require greater intensification in those circumstances.

Data with nucleoside analogues which are not available from randomised comparative cohorts. Data from first line studies such as previously reported by Bogner and Rubio, as well as from Chelsea & Westminster in London suggest no differences in metabolic parameters or rates of lipodystrophy between stavudine and zidovudine. However, in an uncontrolled study in France of individuals switched from stavudine to either zidovudine or abacavir has suggested that some benefit in fat mass and triglyceride levels may be observed. Several studies are now going on evaluating thymidine analogue to abacavir switch protocols. This strategy cannot be recommended until evaluation of the risk versus benefit (if any) of this approach has been established.

In no studies, and anecdotally not in individuals who have discontinued therapy for more than six months, have benefits in terms of facial lipoatrophy been observed. Additionally, as lipoatrophy has been observed on both nucleoside analogue containing and nucleoside analogue sparing, protease inhibitor containing and protease inhibitor sparing regimens, it seems unlikely that simply modifying the treatment is likely to lead to clinical benefit. Furthermore the outcomes of switch studies are unable to inform us about the ætiology of lipodystrophy, but may potentially by used as starting points for hypothesis generation and subsequent testing in future clinical studies.

Bones
The meeting concluded with some comments about bone mineral density from Juilan Gold, Sydney, Australia. The bone strength relates to issues of bone geometry, architecture and bone mass, the bone being a dynamic structure with constant turnover between bone-building osteoclasts and bone modelling osteoblasts. Bone mineral density is known to decline with age and with falling body mass index (BMI, a measure of height and weight). It is most regularly evaluated using DEXA scanning. Risk of fracture related to low bone mineral density is greatest in individuals with the lowest bone mineral density, with the lowest body mass index, with higher levels of osteocalcin (a marker of bone turnover) and history of prior fracture, as well as in individuals who smoke and take excessive alcohol. Falls in bone mineral density have been associated with age, low testosterone or oestrogen levels, elevated parathyroid hormone, low levels of vitamin D, high levels of alcohol consumption, the use of corticosteroids, high levels of inflammatory markers such as tumour necrosis factor, Interleukins 1 and 6 and low levels of physical activity or exposure to light. Data available within the HIV setting broadly suggests that low bone mineral density is associated with HIV disease per se, rather than specific drugs, albeit that one initial evaluation by Tebas et al published in AIDS in 2000 suggested an association the protease inhibitor use. Dr Gold presented new data from an evaluation of 110 patients from his clinic in Sydney and found that osteoporosis (severe bone mineral density loss) was present in 10% of individuals and osteopenia (a more modest loss) in as many as 44.5%. Those with lower bone mineral density had lower weight, longer duration of HIV infection, longer duration of antiretroviral therapy and tended to have higher levels of osteocalcin and triglycerides. In a multi-variate model, age, nucleoside analogue use and levels of osteocalcin ware all associated with moderate elevations in risk of low bone mineral density. Dr Gold suggested that the association with nucleoside analogue therapy was likely to relate to the duration of HIV infection and the advancement of HIV disease in these individuals rather an association that was linked in any way to the ætiology of the problem.

He then went on to briefly discuss possible treatments. Firstly, eliminating possible secondary causes of low BMD by reducing alcohol and smoking, increasing exercise, replacing hormone deficiencies, improving metabolic disturbances and ultimately managing HIV disease. Limited available data suggest that neither growth hormone nor anabolic steroids are beneficial for managing bone mineral density. Therefore standard treatments such as disphosphonates or calcitriol are the likely therapies. Although the statins, used to manage cholesterol elevations, have also been reported recently to improve bone mineral density in the elderly. Studies evaluating these drugs in individuals, specifically with osteoperosis related to HIV are now required.