When it comes to lipodystrophy research, few clinicians have their hand in so many different studies—or are consistently consulted for expert advice and assistance—as Dr. Donald Kotler. Knowing that Dr. Kotler would be attending the 7th International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV, held in Dublin in November, Physicians’ Research Network asked him to check in with its membership at the November PRN meeting. What follows are highlights from the 7th International Workshop, with new data indicating that we are perhaps one step closer to better understanding what HIV-associated lipodystrophy really is (and what it’s not), factors that contribute to its development, and potential treatments to consider.
| I. Defining Lipodystrophy: Lumpers vs. Splitters | Top of page |
Lumpers include Dr. Andrew Carr and his colleagues, who published results from their lipodystrophy case definition (LDCD) study in The Lancet in 2003 (Carr, 2003). Now we have the splitters, most notably Dr. Carl Grunfeld and his colleagues, who recently published much anticipated data from the Fat Redistribution and Metabolic Change in HIV Infection (FRAM) study (Bacchetti, 2005).
| Lipodystrophy Case Definition Study | Top of page |
Characteristics included in the "best" model for the case definition were: age (> 40 years), sex (higher score for females), duration of HIV infection (> 4 years), Centers for Disease Control category C, higher waist-to-hip ratio, wider anion gap, lower HDL cholesterol, higher trunk-to-limb fat ratio, higher intra-abdominal-to-extra-abdominal fat ratio, and lower leg-fat percent. This model had a sensitivity of 78% and a specificity of 80% for diagnosis of lipodystrophy. “With the higher trunk-to-limb fat and intra-abdominal-to-extra-abdominal fat ratio, and lower leg-fat percentage, it appears that both lipohypertrophy and lipoatrophy are linked,” Dr. Kotler commented. “The FRAM study, however, came to some very different conclusions.”
| Fat Redistribution and Metabolic Change in HIV Infection Study | Top of page |
The HIV-positive men reported more fat loss than controls in all peripheral and most central depots. Peripheral lipoatrophy was more frequent in HIV-positive men than in controls (38.3% vs. 4.6%), whereas central lipohypertrophy was less frequent (40.2% vs. 55.9%). Among HIV-positive men, the presence of central lipohypertrophy was not positively associated with peripheral lipoatrophy (odds ratio = 0.71). On MRI, HIV-positive men with clinical peripheral lipoatrophy had less subcutaneous adipose tissue (SAT) in peripheral and central sites and less visceral adipose tissue (VAT) than HIV-positive men without peripheral lipoatrophy (see Figure 1). HIV-positive men, both with and without subjective lipoatrophy, had less SAT than controls, with legs and lower trunk more affected than upper trunk. Use of stavudine (Zerit) or indinavir (Crixivan) was associated with less leg SAT but did not appear to be associated with more VAT. Nevirapine (Viramune) use was associated with less VAT.
FRAM demonstrated that peripheral and central subcutaneous lipoatrophy is more common in HIV-positive individuals than in HIV-negative individuals. Of interest was the finding that increases in VAT—lipohypertrophy—were actually more common in HIV-negative men than in HIV-positive men. This conclusion has succeeded in turning conventional wisdom on its head. Is HIV-associated lipodystrophy a combination of lipoatrophy and lipohypertrophy, or is it just lipoatrophy? While the results of FRAM seem to suggest the former, there are a few potential study limitations that clinicians should consider when reviewing the results. For example, the HIV-negative control patients in FRAM were significantly heavier that the HIV-positive patients. This may have contributed to the observation that increases in VAT were more common in the control patients. Furthermore, FRAM is a cross-sectional study. Longitudinal studies, such as ACTG 5005 (a substudy of ACTG 384), have demonstrated an increase in truncal adiposity over time with institution antiretroviral therapy (Mulligan, 2005).
“FRAM also demonstrated that lipoatrophy occurs, whether or not it is visible to the patient or clinician,” Dr. Kotler added. “Some HIV-positive individuals who did not think they had lipoatrophy upon entering the study were found to have lipoatrophy on MRI. What we see doesn’t necessarily tell the whole story. In Harlem I study a lot of obese women, some of whom have elements of lipoatrophy. But the fact is, the skinniest of the lipoatrophic but obese women have more subcutaneous body fat than I do, and nobody has ever claimed that I have lipoatrophy. So it’s really complicated.” It is important to note that FRAM data involving HIV-infected women are forthcoming.
Now that we have two conflicting answers, we have to go back to the question, “What is lipodystrophy?” Is it something that is visible, or is it something that exists, whether or not it is visible? This distinction is important, since Dr. Carr’s group found that subjects deemed to have lipodystrophy by the case definition also were more likely to fit a consensus definition of the metabolic syndrome than those without lipodystrophy.
| MITOX and ROSEY | Top of page |
There was a significant, although relatively weak, positive correlation between the changes in limb-fat mass and VAT. In other words, an increase in limb fat in men recovering from lipoatrophy was not associated with a reduction in VAT, but rather an increase. Additional data presented by the authors indicated that changes in limb fat and VAT have differing risk factors. These data are consistent with previously reported longitudinal results from ACTG 5005, in which changes in limb and trunk fat following initiation of antiretroviral therapy were found to be positively correlated (Mulligan, 2005).
“The hypothesis underlying the lipodystrophy case definition study is that visceral fat increases and subcutaneous fat decreases are related,” Dr. Kotler commented. “However, FRAM and data from the MITOX/ROSEY analysis indicate that this isn’t true. In fact, if they are related, it’s in a positive way, whereby both SAT and VAT increase or decrease together. So, in effect, the splitters win this round.”
| II. Identifying Risk Factors | Top of page |
| Thymidine Analogues and Lipoatrophy | Top of page |
In the original 2002 MITOX study published by Dr. Carr and his colleagues, some basic comparisons between patients with and without lipoatrophy were reported (Carr, 2002). Total fat measured 17.1 kg in a control group of patients, compared to 12.8 kg in the patients with subjective lipoatrophy. Trunk fat weighed in at 9.1 kg in both lipoatrophic patients and control patients. Limb fat, however, was found to be 7.3 kg in control patients and only 3.7 kg in the patients with lipoatrophy, a difference of 3.6 kg.
Two years later, long-term follow-up data from MITOX were reported (Martin, 2004). After 104 weeks, the median limb fat increase among patients who switched to abacavir (Ziagen) was only 1.3 kg, compared to a 0.24 kg increase among patients who remained on a stavudine- or lamivudine-containing regimen (see Figure 2). “The patients who switched got somewhat better—a 39% increase in their limb fat,” Dr. Kotler commented. “What we want to know is why these patients didn’t get completely better.”
Possible answers may come from two adipose tissue studies reported at the 7th International Workshop. In one study by Dr. Cecilia Shikuma and her colleagues at the John A. Burns School of Medicine in Honolulu, levels of proviral HIV-DNA in peripheral blood mononuclear cells (PBMCs) were significantly higher in patients with lipoatrophy (80 copies in 106 cells), compared with HIV-infected patients without lipoatrophy (22 copies in 106 cells) (Shikuma, 2005). While the jury is still out regarding the significance of this finding, it does suggest the PBMCs are a potential reservoir of HIV in patients with lipoatrophy. Of even greater interest was another finding of the study: increased numbers of macrophages in SAT from patients with lipoatrophy. Evaluating PBMCs in adipose tissue expressing CD68—a hallmark marker of macrophages—Dr. Shikuma’s group reported macrophage contents of 3.2% in HIV-negative study volunteers; 9.5% in HIV-positive, antiretroviral-naïve volunteers; 4.0% in non-lipoatrophic HIV-positive patients, and 9.4% in lipoatrophic HIV-positive patients. “It wasn’t just that there was less fat,” Dr. Kotler said; “there was also an inflammatory infiltrate in the adipose tissue from these patients.”
In an Australian study, conducted at Royal Perth Hospital, Dr. Emma Hammond and her colleagues investigated risk factors for lipoatrophy in 32 HIV-infected patients participating in a Western Australia cohort (Hammond, 2005). None of the patients had evidence of an effect of HIV disease on adipose tissue prior to initiating treatment. Histologic analysis of subcutaneous fat biopsies from patients on antiretroviral therapy revealed a distinct pathologic signature associated with lipoatrophy, including the loss of tissue architecture, adipocyte pleomorphism, mitochondrial toxicity, cell loss, increased macrophage infiltration, and elevated proinflammatory cytokines. “As they evaluated people over time, macrophage infiltration became more prominent,” Dr. Kotler said. “This infiltration clearly preceded lipoatrophy. This research team—and others—have shown that macrophages produce proinflammatory cytokines, which may be at least partly responsible for the adipocyte damage we see in lipoatrophy.”
The potential role of macrophages in lipoatrophy was hypothesized by Dr. Shikuma’s group. “Dr. Shikuma explained that it’s not just the nucleosides that are causing lipoatrophy, it’s also the macrophages,” Dr. Kotler summarized. “The nucleosides caused mitochondrial and cellular dysfunction in the adipose tissue, eventually causing cells to become sick and sometimes die. That death causes a recruitment of macrophages. These macrophages then become active. This activation gives rise to inflammation, resulting in further cellular dysfunction and death, which keeps going on and on and on.” In turn, Dr. Kotler argues, switching off of thymidine analogues may not be the end-all, be-all solution; dealing with macrophage infiltration may be another piece of the reversal puzzle.
| Ritonavir-Boosted Atazanavir: Still Metabolic Friendly? | Top of page |
At the 7th International Workshop, Dr. Noor reported these anticipated data (Noor, 2005). The study involved HIV-negative subjects. Upon enrollment, all patients underwent oral glucose tolerance testing (OGTT) and a euglycemic clamp analysis. They were then randomized to take atazanavir (300 mg) plus ritonavir (100 mg) once daily or lopinavir/ritonavir (400 mg/100 mg) twice daily for ten days. A second round of OGTT and euglycemic clamp evaluations were then performed, followed by a washout period of at least seven days. Then, for another ten days, patients originally randomized to atazanavir/ritonavir received lopinavir/ritonavir, and patients originally randomized to lopinavir/ritonavir were switched to atazanavir/ritonavir. A third round of OGTT and euglycemic clamp testing ended the study.
At baseline, the mean glucose disposal rate was 11.5 mg/kg/min per IU/mL of insulin. After ten days of atazanavir/ritonavir therapy, the median glucose disposal rate was 10.4, a decrease that was not statistically significant compared to baseline. After ten days of lopinavir/ritonavir, however, the median glucose disposal rate was 8.6, a decrease that was statistically significant compared to mean averages seen at both baseline and after ten days of atazanavir/ritonavir dosing. As for the percentage change from baseline in terms of insulin sensitivity by euglycemic clamp, there was a 9% decrease after ten days of atazanavir/ritonavir use, compared to a 25% decrease after ten days of lopinavir/ritonavir use. Compared to baseline averages, the decrease in the insulin-stimulated glucose disposal rate after atazanavir/ritonavir treatment was not statistically significant, whereas the decrease after lopinavir/ritonavir treatment was, statistically, highly significant. Furthermore, the difference in insulin resistance during the boosted atazanavir and lopinavir/ritonavir treatment periods also was statistically significant.
As for lipids, total cholesterol averaged 159 mg/dL at baseline, remained steady at 159 mg/dL after atazanavir/ritonavir therapy, and increased to 169 mg/dL after lopinavir/ritonavir therapy (P=0.003 comparing lopinavir/ritonavir to baseline). LDL cholesterol averaged 92 mg/dL at baseline; it averaged 91 mg/dL after ten days of atazanavir/ritonavir and 97 mg/dL after ten days of lopinavir ritonavir (p=0.02 comparing atazanavir/ritonavir to lopinavir/ritonavir). There was a statistically significant decrease in HDL cholesterol levels, compared to baseline, after both atazanavir/ritonavir and lopinavir/ritonavir treatment (from approximately 45 mg/dL to 40 mg/dL after treatment with both drug combinations). Triglycerides increased after treatment with both combinations, from 108 mg/dL at baseline to 140 mg/dL after atazanavir/ritonavir and to 161 mg/dL after lopinavir/ritonavir, with statistically significant differences between the three averages.
“These are results that we would expect,” Dr. Kotler said. “Atazanavir appears to be metabolically neutral. Ritonavir is not metabolically neutral, but it’s being used as a booster, so we can expect some mild effect on insulin resistance and lipids, predominantly triglyceride levels. However, the effects we see with boosted atazanavir are not to the levels seen with Kaletra.”
| III. Treatment/Management Issues | Top of page |
| Uridine | Top of page |
The study presented at the 7th International Workshop set out to determine the effect of uridine on SAT during unchanged antiretroviral therapy containing either stavudine or zidovudine (Sutinen, 2005). The study also assessed the effects of uridine on insulin resistance, along with the safety of uridine in patients receiving antiretroviral therapy. Twenty patients were randomized to receive either NucleomaxX (36 g three-times-daily for 10 days/month) or placebo for three months.
At baseline, patients in the NucelomaxX group had 3.3 kg limb fat, compared to 3.1 kg limb fat in the placebo group. After three months, patients in the NucleomaxX group had 4.2 kg limb fat, compared to 3.3 kg in the placebo group (see Figure 3). This difference was statistically significant, as were differences between leg-fat increases, arm-fat increases, and truncal-fat increases.
In this study, the investigators saw patients entering with approximately 3 kg of leg fat and were able to increase their leg fat by 1 kg after three months,” Dr. Kotler explained. “However, we really want to see legs with 8 kg of fat, like the controls in the MITOX study, so a jump of 1 kg isn’t all that terrific. However, these really are impressive data, especially since the underlying antiretroviral regimen was continued and the study duration was only three months. It’s possible that they would have seen even more pronounced fat gains if they switched therapies and began uridine supplementation, continuing it for more than three months. Unfortunately, though, nothing else changed. Patients’ insulin levels didn’t come down, nor did their lactate levels or liver fat. Hopefully we’ll see more data involving this product soon.”
| Creatine Monohydrate | Top of page |
The study reported at the 7th International Workshop set out to evaluate changes in strength in response to creatine-aided resistance training, along with changes in body composition, mitochondrial energy metabolism, and safety measures. Twenty sedentary patients were randomized to creatine (20 g/day loading dose for five days, followed by 4.8 g/day maintenance therapy) and 20 sedentary patients were randomized to placebo. For the first two weeks, patients only received creatine or placebo, without resistance training. For weeks 3 through 24, resistance training was added.
After two weeks of supplementation alone, lean body mass increased by approximately .4 kg in the placebo group and .9 kg in the creatine group. This difference was not statistically significant. By week 14—12 weeks after the addition of resistance training—a significant difference in lean body mass gains was seen: approximately 1.4 kg gain over baseline in the placebo group, compared to an approximate 2.3 kg gain over baseline in the creatine group.
“While people in the creatine group definitely got bigger,” Dr. Kotler noted, “the study also demonstrated that muscle strength increased in all subjects following resistance training, without statistically significant differences between groups. There was a 44% increase in muscle strength in the creatine group, compared to a 42% increase in muscle strength in the placebo group. And the fact of the matter is that people in the creatine group actually got bigger before they even started exercise.”
| Niacin | Top of page |
Niacin, a B vitamin, has been shown to lower both cholesterol and triglycerides in HIV-negative study volunteers. It has been shown to reduce total cholesterol by approximately 20% to 30%, lower triglycerides by 35% to 55%, and increase HDL cholesterol by 20% to 35%. However, the side effects of niacin are noteworthy and include nausea, diarrhea, vasodilatory symptoms, insulin resistance, and increased liver enzymes. As a result, niacin has generally been avoided in HIV.
To get a clearer sense of the effectiveness and safety of niacin in HIV-infected patients, the AIDS Clinical Trials Group conducted a study (A5148) evaluating Niaspan (an extended release formulation of niacin) involving an open-label, dose escalation, single-arm protocol (Dubé, 2005). A5148 was a 48-week study and enrolled 33 patients. “Patients were started at 500 mg daily,” Dr. Kotler said. “They were then increased to 1,000 mg, 1,500 mg, or 2,000 mg, depending on their lipid response.”
At baseline, the median total cholesterol was 253 mg/dL. Also at baseline, the median HDL cholesterol was 34.5 mg/dL, the median non-HDL cholesterol was 217 mg/dL, and the median triglyceride level was 478 mg/dL. After 48 weeks of treatment, total cholesterol decreased by 8 mg/dL, HDL cholesterol increased by 5 mg/dL, non-HDL cholesterol decreased by 19 mg/dL, and triglycerides decreased by 153 mg/dL.
As for safety and tolerability, there were one grade 3 and two grade 2 increases in ALT/AST. There were three grade 2 and three grade 3 reports of flushing. “Increased insulin resistance was seen in this study,” Dr. Kotler added. “What we gain in terms of lowering lipids may actually worsen insulin resistance, meaning that we may be trading one toxicity for another. While HOMA-IR testing demonstrated that insulin resistance increased and stayed increased through to 48 weeks, an oral glucose tolerance test demonstrated insulin resistance decreased to near baseline levels after 48 weeks. So the jury is still out with niacin. Despite demonstrating evidence of benefit in dyslipidemia in HIV-infected individuals, these data kind of put us back where we started.”
| Treatment Switching and Facial Lipoatrophy | Top of page |
But what about facial lipoatrophy? To study the effects of switching on this very important parameter, the investigators conducted a substudy involving 47 patients who underwent three-dimensional laser assessment of the face (Benn, 2005). “It sounds really fancy,” Dr. Kotler wryly noted, “but companies are now using the exact same technology to fit people for suits and dresses. It’s a topography measure. Essentially, somebody sits in a chair and a laser is beamed at the face. They eyes are covered. The data is fed into a computer and a topographical image is rendered. This is done two different times and the images are then placed on top of one another, figuratively, to detect additions and subtractions in volume.”
The investigators noted little change in forehead fat 48 weeks after switching to abacavir or tenofovir. Combined cheek-fat gains—meaning fat increases in both the left and right cheeks—averaged 2,812 mm3 in the tenofovir group and 2,208 mm3 in the abacavir group. These differences were statistically significant compared to baseline, with no statistically significant difference between the two groups. “The cheeks actually increased by approximately a gram each,” he commented. “The total increase in cheek volume was somewhere around 2 ½ to 3 ccs. The investigators also compared the difference in the RAVE patients to a group of patients who received collagen injections. Essentially, they saw the same level of improvement. Finally, the study also showed a positive correlation between an increase in limb fat, measured by DEXA, and an increase in cheek fat, measured using the three-dimensional imaging. “They really change together,” Dr. Kotler said. “It’s not like you’re going to get your legs back and your face is not going to improve. Where there is improvement in one parameter, there should be an improvement in the other.”
| References | Top of page |
