icon-folder.gif   Conference Reports for NATAP  
  11th Annual Retrocirus Conference
(CROI-Conference on Retroviruses and Opportunistic Infections)
San Francisco
Feb 8-11, 2004
Back grey_arrow_rt.gif
Body Shape at CROI:
rosiglitazone, Newfill, Fuzeon, PI-switch study, d4T-XR, FTC, face ultrasound, Atlantic & NEFA studies

  Written by David Alain Wohl, MD
The University of North Carolina AIDS Research and Treatment Unit
There is much that still needs to be learned about the cause, prevention and management of body shape changes accompanying HIV and its therapy. However, those looking for new answers at the 11th CROI were likely to be disappointed. Some presentations, however, were notable and are detailed below. These include negative results for a once promising intervention, a reality check on a popular cosmetic procedure, additional information on the determinants of morphologic changes and the effect of newer antiretrovirals on body shape.
--Rosiglitazone-no benefit on Lipoatrophy
--NewFill - new study results find risks of treatment
--Old Studies, New Substudies: Atlantic Study, NEFA Study (switch off PI to ABC, NVP, or EFV)
--Fuzeon (T-20) and Body Shape
--Face ultrasound for fat loss
--d4T-XR (new extended release formulation) vs old d4T-IR (immediate release)
--FTC vs d4T (metabolics, body changes)
A Not So Rosy Picture for ROSEY
A long time ago, in a galaxy far, far away it seemed like a really good idea to try and treat lipoatrophy with PPAR-gamma agonists such as rosiglitazone. Theoretically, it looks attractive. The drug class increases cellular glucose intake, reduces fat breakdown and increases fat accumulation. In HIV-uninfected patients with congenital lipoatrophy a similar drug increased peripheral fat and deceased visceral fat, albeit modestly. However, the few small studies performed with HIV-infected individuals have demonstrated conflicting results regarding the effectiveness of this agent on fat redistribution. A recently reported placebo controlled trial involving 28 patients with lipoatrophy and insulin resistance found an overall increase in total body fat as measured by BIA with 4 mg/day of rosiglitazone for 3 months (Hadigan C, et al. 2nd IAS 2003, #50). Further improvement in total fat was seen during an open label dose escalation (8 mg/day) phase. This stands in stark contrast to an equally small single-arm study of the same agent at the higher dose in patients without known insulin resistance in whom no change in body shape was observed (Sutinen J, et al. Antivir Ther. 2003 Jun;8(3):199-207). The disparate results may be due to a number of factors, not the least of which is the study populations and the measures used to evaluate fat change.
To help settle the issue we can look two large clinical trials: an AIDS Clinical Trials Group (ACTG) placebo-controlled study of rosiglitazone, metformin, both or neither in patients with fat accumulation and insulin resistance and a study of rosiglitazone versus placebo for fat wasting - the ROSEY Study. The first study is still enrolling. The second has been completed and was simultaneously published in the Lancet and presented at CROI this week (Abstract 79). In this Australian study, 108 patients with limb fat less than 20% of limb tissue or limb fat percentage at least 10% less than truncal fat percentage by DEXA were enrolled. Importantly, these DEXA criteria were chosen so that participants would be similar to those recruited in two earlier randomized HIV lipodystrophy studies. Participants had to have been receiving combination antiretroviral therapy unchanged for 12 weeks. There was no requirement for an insulin or glucose disorder. Almost all (98%) of the subjects were male and, like most HIV+ Aussies, very white. There were a disproportionate number of participants receiving d4T in the rosiglitazone arm compared to the control arm (53% versus 29%). All subjects were randomized to 8 mg/day of rosiglitazone or matching placebo for 48 weeks.
It didn't work. There was no difference between the study arms in the change of fat as measured by DEXA at week 48. Limb fat increased by 0.14 kg (5%) in the rosiglitazone group but also increased 0.18 kg (7%) in the placebo group (mean difference -0.04 kg [95%CI -0.29 to 0.21]; p=0.74). CT scanning of the abdomen and thigh demonstrated similar decreases in intrabdominal fat in both groups but neither had an increase in thigh subcutaneous fat. Subjects in both study groups reported similar subjective improvements in lipoatrophy. Mean insulin levels, which were in the normal range at baseline, declined equally in both study arms. Adiponectin, a hormone produced by fat cells and may play a protective role against diabetes and atherosclerois, increased significantly in the rosiglitazone arm but not among the controls. Strangely, leptin levels were not significantly different.
Rosiglitazone was well tolerated; only six (11%) rosiglitazone assigned subjects interrupted therapy (four permanently), two for adverse events and four because of patient choice, and six (11%) placebo recipients stopped treatment, one for an adverse event and five out of their own choice. ALT levels actually fell in the rosiglitazone arm, as has been seen in previous studies. This is speculated to be due to a reversal of hepatic steatosis by rosiglitazone. Notably, triglycerides, total cholesterol, and LDL cholesterol levels increased with rosiglitazone. Increases in triglycerides over the first 12 weeks of therapy were significantly associated with use of rosiglitazone (odds ratio 7á3 [95%CI 2á0-27á0]; p=0á003) - a phenomenon not typical in the HIV-uninfected diabetic. A companion poster presentation (Abstract 729) describing a substudy of brachial reactivity, a predictor of atherosclerosis, did not observe an effect of rosiglitazone on this parameter.
So what to make of these results? This is an important and well conducted trial. Several of us in the ACTG proposed and came close to launching an almost identical study but did not since it seemed it would be duplicating this rigorous investigation. The curious finding in this trial, however, is the relative increase in limb fat in the control subjects. Previous data on the natural history of lipoatrophy, much of it from Australia, indicated that continued worsening of fat wasting could generally be expected in the absence of NRTI therapy modification - the best demonstrated intervention to reverse lipoatrophy. While the authors report on the rate of antiretroviral switching during the study, only three subjects from each study group switched from d4T, there are no data presented on similar therapy changes prior to study enrollment (i.e. before the 12 weeks of stable therapy required for entry). If in the months leading up to study entry a significant number of participants had switched from their d4T or ZDV to abacavir this might explain the small improvement in limb fat seen in the controls. One third of subjects in each arm of the study were on abacavir at study entry, a potential marker of NRTI switching given there was little ZDV use (and therefore little Trizivir use) at baseline. Whether this degree of abacavir use is typical of the antiretroviral receiving population in Australia or not would be interesting to know. One would expect that a detailed antiretroviral history was obtained at study entry so these data should be available. These data are relevant to the interpretation of the results since the improvement of peripheral lipoatrophy among the controls may have obfuscated the relative benefits of rosiglitazone, particularly if there was an unequal distribution of persons who had switched from d4T in the control (as may or may not be hinted at by the low rate of d4T use among the control subjects). However, it must be appreciated that even if there had been stable or slight worsening of lipoatrophy in the control arm, the finding that rosiglitazone was able to increase limb fat only 0.14 kg over a year - substantially less than seen in switch studies - would remain an underwhelming result.
The ACTG study, although targeting patients with fat accumulation, undoubtedly has enrolled a large number of subjects with lipoatrophy. In addition, there are some experts who believe PPAR-gamma agonists like rosiglitazone are most likely to be effective in patients with concomitant insulin resistance, as are being enrolled in the ACTG trial. Results from that study will not be available for some time. In the meanwhile, an attempt to reverse peripheral fat wasting with rosiglitazone or its sister drugs is ill advised.
Newfill - Longer Term Safety and Efficacy Data
The select Ôbefore' and Ôafter' pictures that accompany every presentation regarding Newfill (polylactic acid) are striking and have played no small part in the increasingly fervent drum beat demanding access to the procedure in the U.S. However, despite the desire (and need) for a treatment for disfiguring facial fat loss, there are few published investigations of the safety and efficacy of Newfill even though the compound is available in Europe. A recent article in the journal AIDS (Valantin MA, et al. AIDS. 2003 21;17(17):2471-7) described 50 patients who received Newfill and positive response and minimal adverse effects were reported.
At the 11th Conference on Retroviruses and Opportunistic Infections the authors of previous reports on Newfill provided longer term and more detailed safety and efficacy data than prior presentations (Abstract 726). A cohort of 94 patients receiving intradermal injections of Newfill to the face was followed longitudinally. Subjects had self-described and clinician confirmed lipoatrophy of the face, a CD4+ cell count of >200/uL, a viral load of <20,000 and were on stable ART. Those with bleeding disorders, low platelets (<80,000), or receiving non-steroidal anti-inflammatory medications (ibuprofen, aspirin, etc) were excluded. Subjects were almost all men (94%) and white (93%). Facial lipoatrophy at baseline was judged as "light" in 36%, "moderate" in 43% and "severe" in 21%. Injections of Newfill were given every 15 days. Approximately 3 mL of polylactic acid was injected along with lidocaine. Following injection, the area was iced and then massaged in an attempt to reduce the risk of focal accumulation of the compound.
Efficacy was primarily assessed using an Analogic Visual Scale Satisfaction Index (AVSSI) in which the subject draws a line on a scale of 0-10 (where 0 represents the least and 10 the most) indicating their satisfaction with the effect of Newfill. A median of 5 series of injections was administered. Notably, 41% of subjects also had injections at the temples. The AVSSI at baseline was 3.4 (recall 0 is least satisfied); at the end of treatment the median score was 6.8 and at 7.5 months following the end of treatment it was 7.0. Using any increase from baseline as the criterion for success, 87% of subjects experienced success at the end of treatment. This fell to 73.3% 7.5 months out from end of treatment. The AVSSI did not improve after the first 3-4 injections and despite the improvement in this measure a survey did not reveal enhancement in quality of life. To obtain a more objective assessment in appearance change three observers blinded to the sequence of sophisticated facial surface photographs of subjects taken at baseline, during and after therapy with Newfill were asked to order the images chronologically. This was correctly accomplished in just under 60% of the cases. An example of three images from a subject who bore an incredible resemblance to U.S. Senate Minority Leader Tom Daschle demonstrated to me the difficulty of the observers' task. The significance of the lower rate of success in this objective assessment versus self-report is likely moot as the most critical determinant of success is how the patient feels he/she looks.
Unfortunately, the probability that a repeat injection was needed during the 15 months following the last Newfill treatment was 43% - indicating the benefits of Newfill diminish with time.
Adverse events were not rare. All patients had some degree of post-injection edema, 77% experienced pain during the procedure (of these 28% required pain medication), 13% had post injection non-inflammatory nodules, 7.5% had vagal hypertonia during the injection which I think means they felt faint, 4% had bruising or bleeding, 1% had inflammatory nodule development, 1% experienced facial palsy of unreported duration ostensibly from the hitting of the facial nerve during treatment and 1% had an anaphylactic reaction (i.e. a very severe allergic reaction). Although, nerve palsy or anaphylaxis occurred in only 2 patients, a 2% rate of very serious adverse events is cause for concern.
Overall, these data may temper early expectations and praise for this procedure and may present some potential hurdles to eventual approval in the U.S. That repeated series of injections of this expensive agent are likely to be required and serious adverse events are to be risked may tilt the risk to benefit ratio for some. For others, the promise of improved appearance supercedes the risk of adverse events. This is understandable and not unique to those living with HIV - Botox is widely popular despite a 3.2% risk of causing eyelid droop. At present a perfect therapy for facial fat wasting does not exist and calls for prevention are meaningless to those already suffering from disfiguring facial fat loss. For these people, especially those living in areas where Newfill is available and who can afford it, this study provides import safety and efficacy data. Data they will need when deciding to inject or not to inject.
Old Studies, New Substudies: Atlantic Study, NEFA Study (switch off PI to ABC, NVP, or EFV)
Two clinical trials from the past continue to sputter data including long term observations regarding metabolic complications. The Atlantic Study was a throw back to an era when protease inhibitors were young and we were all full of hope. The more recent NEFA Study reflected our subsequent PI disillusionment by switching from this class to non-PI based regimens. Both provide opportunities to examine drug effects on body shape.
The Atlantic Study randomized approximately 300 participants to ddI+d4T plus one of the following: indinavir (IDV) versus nevirapine (NVP) versus 3TC. At the time this study was launched, Ôbody shape change' was the money in the piggy bank my patients were saving for a sex change operation. When it became evident that persons taking HIV meds were experiencing redistribution of fat, a substudy of Atlantic was developed. Results of this substudy, which enrolled 81 subjects, were presented as a poster (Abstract 718). The median time following study drug initiation among substudy participants was 144 weeks (range 72-204). Assessment for body shape alterations by whole body DEXA and CT scan of abdomen, as well as clinical examination, lipid testing and other blood based metabolic parameters, occurred every 24 weeks.
Clinical diagnosis of fat accumulation decreased overall among the substudy participants and in all three substudy arms with the greatest decrease among the triple nucleoside arm. Lipoatrophy was reported in a quarter of subjects at week 96, increasing to 39% by week 192 with no differences in reported fat wasting between study arms. Abdominal CT scanning at entry into the substudy demonstrated significantly more visceral fat in the IDV group than the other study groups. However, visceral fat increased over time in the NVP and 3TC arms while it remained stable in the IDV assigned subjects. Looking at DEXA results, there were no differences at substudy entry between study assignments in extremity fat. Over time limb fat decreased in the legs but not the arms. The 3TC arm had a significantly greater decrease in fat percent of the legs. There were no differences in lipid profiles between arms, however, over time HDL cholesterol increased more in the NVP arm. Overall, no surprises here: IDV treated subjects had more visceral fat, everyone lost fat on d4T+ddI and NVP raised HDL.
The NEFA Study also provided a unique opportunity to detect the relationship between drug regimens and body shape. In this well known switch study (New England Journal of Medicine 2003, 349:1036), participants entering on a suppressive PI-based regimen, substituted either abacavir (ABC), NVP or efavirenz (EFV) for their PI. In an oral presentation (Abstract 78), a metabolic substudy of this trial was described. The substudy enrolled 96 subjects entering the parent study and evaluated body shape and lipids every three months for 24 months. Data from the 69 subjects on their originally assigned therapy were presented. Following study entry, insulin, insulin resistance, total cholesterol, HDL cholesterol, LDL cholesterol and total:HDL cholesterol improved. The NNRTIs performed about the same as far as metabolic improvement; neither led to changes in total cholesterol but HDL cholesterol rose and LDL cholesterol dropped. The switch to ABC did result in a decrease in total cholesterol as well as LDL cholesterol.
At baseline 40 subjects had lipodystrophy based on clinical examination. Patients with lipidystrophy tended to have higher insulin and triglyceride levels. Importantly, switching off of the PI did not impact body shape no matter what agent was substituted.
T-20 and Body Shape
Does it or doesn't it? That is the question many have regarding T-20 and metabolic complications. To answer, investigators from the two large an practically identical clinical studies of T-20, TORO 1 and TORO 2, involving almost 1,000 patients, examined anthropometric (body shape measurements) from all study participants and DEXA plus single slice abdominal CT scan in a metabolic substudy of 155 subjects (Abstract 715). Lipids, insulin and glucose parameters were also evaluated. Both clinical trials randomized (2:1) highly antiretroviral experienced participants to an optimized regimen plus T-20 versus an optimized regimen alone.
Rates of fat redistribution on treatment with T-20 were no different than optimized therapy alone (9.2 versus 11.7 events/100 patient-years) during the year long study. However, there was an overall greater increase in weight at week 48 in the T-20 arm (1.7 kg) compared to the control arm (1.0 kg) - likely reflecting the benefits T-20 patients accrued with better virologic and immunologic responses. By DEXA leg fat increased slightly from baseline in T-20 assigned subjects but decreased in the controls. CT scanning of the abdomen demonstrated increases in visceral and subcutaneous fat, again with decreases seen among the control subjects. The changes in fat visualized by DEXA and CT in T-20 receiving subjects likely reflects the overall greater weight gain experienced by these subjects. There were no differences between study arms in the change from baseline in fasting glucose, insulin, C-peptide, triglycerides or LDL, HDL, VLDL or total cholesterol.
Although these data were generated from a population of highly treatment experienced subjects starting a variety of new antiretrovirals, factors which complicate the assessment of the metabolic effects of T-20, these data are reassuring.
Face Ultrasound
The ever busy ROSEY investigators performed ultrasonographic evaluation of the maximal subcutaneous fat depth over the right malar bone (under the cheek) and correlated the result to DEXA derived limb fat, mid-thigh subcutaneous fat by CT scan, self-report and their lipodystrophy case definition score among ROSEY Study participants (Abstract 727). During the 48 week study, ultrasound did register a drop in facial fat but this did not correlate with the other measures. The investigators conclude that ultrasound of the malar region of the face is not a useful measure of facial fat wasting.
D4T-XR - Kinder Gentler?
Few antiretrovirals since ddC have taken the bashing heaped on to d4T the past few years. The accumulation of damning studies linking the drug with fat wasting, mitochondrial toxicities and lipid abnormalities have significantly decreased overall use of the drug. The extended release formulation of d4T (d4T-XR) offers once a day convenience but also carries the baggage of it's twice a day antecedent. To describe predictors for lipoatrophy and whether d4T-XR carries the same risk for fat wasting as standard d4T (also called immediate release, or d4T-IR) investigators from Bristol Myers Squibb performed an analysis using data from two completed clinical trials comparing d4T-XR and d4T-IR.
Over 800 treatment nave subjects participating in the trials were analyzed. All received either d4T-XR or d4T-IR plus 3TC+EFV. In a multiple logistic regression analysis risk patient-related factors for investigator-defined lipoatrophy were found to be baseline triglyceride levels > 200 mg/dL and age > 40 years. Further, report of lipoatrophy during the 48 week studies was almost two times more likely among d4T-IR assigned subjects compared to those assigned d4T-XR (OR=1.975 [95% CI 1.252-3.117] p=.0003). The combination of these three factors (high triglycerides, age 40+ and d4T-IR) seemed to have, at least, an additive effect. The report states that among subgroups with different risk factors such as < 40 years of age with triglycerides < 200 mg/dL , age > 40 years and triglycerides > 200 mg/dL, etc, the risk of lipoatrophy was greater with d4T-IR than d4T-XR in each stratum. Whether these differences were statistically significant is not stated.
There will need to be additional evidence to indicate the extended release formulation of d4T is safer vis--vis metabolic toxicity than standard d4T. A comparison with competitive nucleosides such as ZDV or tenofovir and truely objective evaluations of body shape will need to be produced before the pendulum of the HIV-treater opinion swings anywhere near this nucleoside, whatever its formulation.
FTC is not d4T
The above report notwithstanding, it will come as no surprise to most that in a head-to-head study comparing d4T (standard formulation) + ddI + EFV versus FTC + ddI +EFV fewer morphologic and other metabolic abnormalities would be observed among those receiving the latter regimen. And so it was in a clinical trial sponsored by Gilead, the makers of FTC. In a placebo controlled study of 571 subjects randomized to each regimen, body shape measurements but no DEXA or CT scanning were performed at regular intervals. Subjects in both study arms gained weight initially but, as has been seen in other studies, the d4T receiving subjects experienced a later decline in weight. By week 72 of the study there was a significant decline in weight from baseline in the d4T arm versus a net gain in the FTC arm. Waist, hip and chest circumference and abdominal girth were significantly lower in the d4T group but waist to hip ratio was no different between study arms. Investigator reported lipodystrophy was more common in those randomized to d4T group.
Fasting triglycerides increased in both arms but the increase was significantly greater in the d4T arm (+27 mg/dL for FTC, +97.8 mg/dL for d4T, p <.001). HDL cholesterol also increased in both groups but more so with FTC (+13.5 mg/dL) versus d4T (+8.8 mg/dL) (p=.001). There was no difference between study arms in the magnitude of the increase in LDL cholesterol (+15-17 mg/dL), total cholesterol (+37-44 mg/dL) or glucose (+2.5-4.5 mg/dL). As has been reported previously, FTC performed better virologically.
These data serve to illustrate the metabolic profile of FTC rather than provide any new data regarding d4T. Overall, this new nucleoside appears to have a relatively favorable impact on body shape, although more rigorous measurements are absent, and to be better than d4T insomuch as dyslipidemia.