Metabolic Complications of HIV and Its Therapies

July 16, 2004


Once every 2 years, the world's leading scientists and clinicians dedicated to combating the HIV epidemic meet to discuss their work, share the data they have generated, and chart a course intended to lead to significant advances in time for the next International AIDS Conference. But the conference has become an arena of combat between people vying for billions of dollars earmarked for AIDS. How this noble concept became subverted is ironically rooted in the magic of the 13th International AIDS Conference.

Four years ago, thousands came together in Durban, South Africa, and left awakened to the plight of the African people. Many resolved to change the inequalities in HIV care on our planet. In the wake of Durban, people changed careers, set up treatment programs in developing nations and lobbied for increased funding to fight HIV worldwide. Political leaders took notice, and ambitious funds were established to provide treatment to the poorest of the poor living with HIV. It is these funds that were topic A in Thailand this past week, which left the clinicians and researchers who had traveled to the kingdom in search of cutting-edge data asking, "Dude, where's my science?"

The data presented at this year's conference was generally lackluster and added little to what has been reported at previous conferences -- in fact, some of it had been presented at other conferences. In the area of metabolic complications, this was painfully the case.

Oral presentations at the conference, once prestigious, were sadly bereft of results that would advance our understanding of how to prevent and treat metabolic complications. Below are highlights of studies that were presented regarding these complications. They are at least thought provoking, and, at best, provide new information to take back to the clinic.

The Metabolic Profile of Atazanavir When Combined With Ritonavir

Atazanavir (ATV, Reyataz) has transformed the protease inhibitor (PI) playing field. When combined with low-dose ritonavir (RTV, Norvir), atazanavir is surprisingly potent and well tolerated, with convenient once-a-day administration and low pill count. These attractive features have made the agent into an increasingly popular first-line or first-PI therapy. However, there are few data regarding the metabolic effects of atazanavir when it is combined with ritonavir.

A study in which patients on PI- or non-nucleoside reverse transcriptase inhibitor (NNRTI)-based regimens switched to boosted atazanavir or lopinavir/ritonavir (LPV/r, Kaletra) found declines in lipid parameters in the atazanavir arm relative to lopinavir, but could not determine whether lipids returned to pre-treatment baseline values.1 It is not known what happens to lipids during the treatment of a naive patient with administration of 300 mg of atazanavir and 100 mg of ritonavir. Even less is known about the effect of boosted atazanavir on glucose metabolism.

Mustafa Noor and colleagues at the University of California, San Francisco, have previously studied the metabolic consequences of a number of PIs in vivo and in vitro. Noor, now an employee of Bristol-Myers Squibb, presented previous data regarding the effects of unboosted atazanavir on the uptake of glucose by cultured myocytes, and on insulin sensitivity using the euglycemic-hyperinsulinemic clamp in human volunteers. This study, currently in press, found no change in insulin-stimulated glucose disposal in cells exposed to atazanavir alone.

To determine whether the addition of low-dose ritonavir modulates the relatively glucose-neutral effects of atazanavir, Noor examined insulin-stimulated glucose uptake in adipocytes exposed to atazanavir and ritonavir at concentrations approximating those found in plasma at doses of 300 mg and 100 mg, respectively. Reassuringly, low-dose ritonavir alone had a negligible effect on glucose uptake in the model. The combination of the 2 PIs at increasing concentrations led to decreases in glucose uptake that were identical to those seen when atazanavir was applied alone. Again, there was no impact on uptake at clinical concentrations (Figure 1). In contrast, lopinavir/ritonavir had a dramatic effect on glucose uptake at concentrations seen clinically.

Figure 1: Effect of Atazanavir + Ritonavir on Glucose Uptake by Adipocytes

Figure 1: Effect of Atazanavir + Ritonavir on Glucose Uptake by Adipocytes

These in vitro data are reassuring and suggest that low-dose ritonavir, insofar as glucose metabolism is concerned, does not threaten the metabolic-friendly profile of this PI. Noor stated he is putting boosted atazanavir through the same clinical evaluation to which he previously subjected other PIs; data on glucose and lipid parameters in humans is expected shortly. Until then, we can feel more comfortable that even with low-dose ritonavir, atazanavir is unlikely to threaten glycemic control in HIV-infected patients requiring a PI. The effects on lipids remain to be fully explored.

In a poster presentation, Noor also provided updated data from the BMS 034 study, a large comparative clinical trial in which treatment-naive subjects were randomized to atazanavir (unboosted) or efavirenz (EFV, Sustiva, Stocrin) in combination with zidovudine (AZT, Retrovir) plus lamivudine (3TC, Epivir). Previously, the results of the trial's metabolic sub-study found no significant differences in body shape between the 2 study arms, as evaluated by L4-5 computed tomography and dual-energy X-ray absorptiometry (DEXA) scans at week 48; both treatment arms experienced increases in fat compartments.2

To identify predictors of body fat changes during the trial, the investigators performed a multivariate analysis of baseline characteristics, including demographic and metabolic parameters. Data were available on 163 study participants. Body mass index (BMI), insulin sensitivity and age were found to be associated with changes in body fat compartments. Underweight subjects (BMI <22 kg/m2) gained more weight than those who were larger at baseline. Fat tissue increased in those with relative insulin sensitivity. Lastly, age >40 years was associated with increases in visceral adipose tissue. Unfortunately, the study did not examine whether these predictors held when looked at by treatment arm. It is clear that in this study, body fat generally increased in both study arms -- a finding that has been interpreted as evidence of the improved overall well-being of the cohort, rather than an indication of metabolic toxicity. Longer-term follow-up of patients on atazanavir, particularly when boosted with low-dose ritonavir, is required before it can be stated that this therapy has minimal or no body shape effects.

Men and Women Really Are Different: Disparate Metabolic Results From the Gilead 903 Study and New Data From the WIHS

The results of the Gilead 903 study have been presented so often that the study has become like one of those overplayed '70s songs that you cringe upon hearing (although after a while, you can't help humming along). To be fair, however, this is a long-term study, and updated data reports are to be expected.

The Gilead 903 study compared the efficacy and safety of tenofovir (TDF, Viread) and stavudine (d4T, Zerit) when combined with lamivudine and efavirenz in treatment-naive patients. The latest presentation, by De Ruiter et al, concerns gender differences in the metabolic parameters measured during the study. It was previously demonstrated that subjects randomized to the stavudine arm experienced excessive metabolic toxicity relative to the tenofovir arm, including higher rates of lactic acidosis, hypertriglyceridemia, hypercholesterolemia and lipoatrophy.3 When examined by gender, however, a more complex picture emerges.

Almost a quarter (26%) of the 600 subjects enrolled in this study were women, and they were equally distributed between the 2 study arms (79 in the tenofovir arm and 76 in the stavudine arm). The women in the 2 arms had similar baseline characteristics, including median weight (~65 kg), mean CD4+ cell count (288-295 cells/mm3) and mean viral load (4.7 log10 copies/mL).

After 144 weeks, almost identical proportions of the women in each arm had viral loads <50 copies/mL (67% for tenofovir versus 64% for stavudine) and increased CD4+ cell counts (314 for tenofovir vs. 320 for stavudine). Men tended to have higher rates of viral suppression (75% and 71% for the tenofovir and stavudine arms, respectively) and smaller increases in CD4+ cell count (246 and 272).

A similar overall proportion of women and men developed serious clinical adverse events, but women in the stavudine arm had fewer such events (17%) than women in the tenofovir arm (28%). However, women in the stavudine arm experienced a higher rate of serious (grade 3/4) laboratory toxicities (41% vs. 29%), and a significant proportion of the excess toxicity was due to increases in amylase.

Interestingly, elevations of creatinine phosphokinase were much more common in the men than the women, which may be due to higher rates of exercise and physical activity among men.

Fasting total cholesterol, triglycerides and low-density lipoprotein (LDL) cholesterol all increased at week 144 among women, but the magnitude of the elevations in triglycerides was greater in men than in women (Table 1).

Table 1: Change From Baseline in Lipid Parameters (mg/dL) at Week 144
 Total CholesterolTriglyceridesLDLHDL

Women in the stavudine arm had a reduction in limb fat between weeks 96 and 144 (times at which limb fat was assessed by DEXA), as did the men in this arm, compared to the lack of change in the tenofovir group. Women in general, however, had much higher levels of limb fat relative to men. Likewise, bone mineral density at the hip and spine declined in all arms across gender, albeit significantly more so in the tenofovir arm, but was generally higher among women than men.

In all arms and genders, bone mineral density reached a plateau early in the study, perhaps suggesting that pre-treatment bone loss continued during the first several weeks of the study and then abated with HIV therapy. Sixteen fractures occurred during the study, all traumatic except one. None occurred among women.

The data regarding gender differences in stavudine-associated dyslipidemia are interesting and mirror those reported by Kumar and colleagues,4 who compared stavudine and zidovudine when combined with lamivudine and nelfinavir (NFV, Viracept) in a cohort of subjects, half of whom were women.

It is unclear why relatively less hypertriglyceridemia among women was seen. Clearly, these results do not mean that women are completely resistant to the lipid-elevating effects of stavudine in combination with other antiretrovirals, and monitoring of lipid parameters in women remains essential.

The data do raise questions about the mechanism of lipid increases on therapy, and have implications for the analysis and comparison of metabolic studies. The study also points out that osteopenia is not only a "women's issue," especially when it comes to HIV. Men had lower bone mineral densities, and the only subject with a vertebral fracture was a man. Lastly, Gilead, who was the sponsor of the trial, should be acknowledged for putting in the effort to analyze and present these data.

Dyslipidemia Among HIV-Infected Women

Kathryn Anastos et al conducted an interesting examination of the prevalence of dyslipidemia among HIV-infected women, and the contribution of HIV and its therapies to the disorder. This was part of the Women's Interagency HIV Study (WIHS), a bicoastal cohort study of women who are HIV infected, or HIV uninfected but at risk. This sub-study compared fasting lipids of 183 HIV-infected women (83 not on highly active antiretroviral therapy [HAART]) and 89 HIV-uninfected women.

About half the participants were African-American, with a median age of approximately 40 years. More than half were smokers. Among the HIV-infected women, those not on HAART had lower CD4+ cell counts, but higher nadir CD4+ cell counts, than those receiving HIV therapy. The women on HAART were receiving a variety of regimens -- 29% were taking stavudine, 26% were receiving an NNRTI and approximately 40% were on a PI (14% nelfinavir and 7% lopinavir/ritonavir).

The median total cholesterol levels of HIV-uninfected women and those not on HAART were not significantly different (172 vs. 162 mg/dL). HAART subjects' median total cholesterol (182 mg/dL) was significantly higher than women who were not on HAART (P = .001), but was not different than those who were HIV uninfected.

The same pattern was seen for LDL cholesterol: women on HAART had higher levels than those who were HIV infected and not on treatment, but similar levels to uninfected women. Median high-density lipoprotein (HDL) cholesterol was higher in the HIV-uninfected women (54 mg/dL) than in untreated HIV-infected women (44 mg/dL), although this was not significantly different from the median level of HAART recipients (50 mg/dL). Triglyceride levels were highest in the HAART group (149 mg/dL); this was greater than among the HIV-uninfected women (118 mg/dL) but not the untreated group (131 mg/dL).

These data nicely place in context the lipid elevations seen subsequent to HIV infection and during antiretroviral therapy. In this cohort, HAART, but not HIV infection, was associated with elevated triglycerides and total and LDL cholesterol. HIV was associated with lower HDL compared to what was seen in HIV-uninfected women, but this appeared to be reversed with HAART.

The authors conclude that the HDL-raising effect of HAART was unique to this study, and may suggest a gender difference in response to therapy. However, over a quarter of the subjects on HAART were receiving an NNRTI, and the data presented indicated that the lion's share of the increase in HDL seen in the HAART group occurred among those taking efavirenz, nevirapine (NVP, Viramune) and, surprisingly, lopinavir/ritonavir (although use of lipid-lowering therapy was not reported, and this and other PIs could have led to lipid increases that required drug treatment).

Certainly, other predominantly male studies, including the 2NN,5 Dupont 0066 and Gilead 903 (see above) studies, have demonstrated an increase in HDL cholesterol after NNRTI treatment. Therefore, these results are congruent with what has been seen in men. This study, rather than identifying gender differences, supports a similar lipid response in men and women after HIV infection and HAART initiation. Riddler and colleagues recently reported on the lipid dynamics of a cohort of patients who had lipid data available before HIV infection, after infection but prior to therapy, and then during HAART.7 The results were similar to the WIHS data. The subjects were from the Multicenter AIDS Cohort (MACS), a study that -- you guessed it -- only enrolled men.

What Lopinavir/Ritonavir Monotherapy Does for Your Lipids

There have been a number of fascinating reports regarding the use of lopinavir/ritonavir monotherapy. To date, these small studies8 have demonstrated impressive rates of virologic suppression, particularly among patients with viral loads <100,000 copies/mL. Yet, there have been no data on the metabolic consequences of such simplified therapy. Although this PI is known to raise total and LDL cholesterol, as well as triglycerides, research suggests that the effects of PIs on lipids are heightened when combined with nucleoside analogues. Three studies at this conference described lipid parameters after the switch from a traditional multidrug regimen to lopinavir/ritonavir alone.

Gerald Pierone and colleagues in Florida enrolled 18 patients who were on NNRTI-based therapies, had HIV RNA <75 copies/mL and no prior PI exposure, and had switched the NNRTI for lopinavir/ritonavir. The pre-study nucleoside reverse transcriptase inhibitors (NRTIs) were continued for 2 weeks, then discontinued.

To date, 14 of the subjects have reached 24 weeks after the start of lopinavir/ritonavir (3 subjects discontinued due to diarrhea and 1 subject due to virologic failure). Prior to the switch, 6 subjects were on lipid-lowering therapy. Three additional subjects added lipid-lowering drugs while on lopinavir/ritonavir monotherapy, and 2 others who entered on such medication required a dose increase. Among the 9 subjects not on lipid-lowering therapy, the mean baseline total cholesterol was 167 mg/dL, and mean baseline triglycerides were 109 mg/dL. Over 24 weeks, this increased to a mean total cholesterol of 176 mg/dL and a mean triglyceride level of 159 mg/dL. The 5 subjects on lipid-lowering therapy during the study had an increase in mean total cholesterol from 194 to 204 mg/dL, and an increase in mean triglycerides from 175 to 218 mg/dL. Two subjects with baseline hyperglycemia developed diabetes, which was controlled with diet and oral hypoglycemic agents.

On the other side of the country, in Los Angeles, Peter Ruane and colleagues conducted a similar study among 19 HAART-receiving individuals with undetectable HIV RNA. As in the above study, lopinavir/ritonavir was added to pre-study therapy for 2 weeks. At that point, lopinavir concentration was measured, and those with a trough of <3µg/mL were not enrolled unless they increased adherence and re-screened with sufficient lopinavir levels. Those with lopinavir levels >3µg/mL discontinued all HIV medications except for the lopinavir/ritonavir. After 24 weeks on lopinavir/ritonavir monotherapy, 15 subjects remained on treatment, 13 with a viral load <75 copies/mL. During the 24 weeks, total cholesterol increased by 7%, LDL cholesterol by 3%, HDL cholesterol by 5% and triglycerides by 90%.

Lastly, Frederico Pulido and colleagues in Spain conducted a randomized study in which 42 patients receiving lopinavir/ritonavir and 2 NRTIs, with viral loads <50 copies/mL, were assigned to discontinue or maintain their NRTI agents. The median baseline (BL) and on-study lipid values are listed in Table 2.

Table 2: Median Baseline and 24-Week Lipid Values, With or Without NRTIs
 Total Cholesterol (mg/dL)LDL (mg/dL)HDL (mg/dL)Triglycerides (mg/dL)
 BLWeek 24BLWeek 24BLWeek 24BLWeek 24
LPV/r Alone17620184102445186176
NRTIs Continued1931951041074345208226

Interestingly, stopping the NRTIs led to an increase in total, LDL and HDL cholesterol, but a small decrease in triglycerides. No statistical analyses were performed to indicate whether these changes were statistically significant, although one would have expected lipids to improve after NRTI discontinuation (approximately a third of the subjects in each arm were on stavudine at baseline). Further study will determine whether this result was a fluke.

These 3 studies all indicate that lopinavir/ritonavir monotherapy is surprisingly able to maintain the virologic suppression usually achieved with combination antiretroviral therapy, but that increases in lipids are common. The convenience of this novel strategy must therefore be balanced with the potential need to add lipid-lowering therapy and the attendant increase in pill counts, co-payments and potential toxicity.

Bone Cement for Facial Fat Wasting

In the first ever infomercial to be permitted at an International AIDS Conference, Luiza Oyafuso from Brazil described his experience injecting polymethylmethacrylate (PMMA) into the faces of patients with facial fat wasting. PMMA, according to the Encyclopedia of Medical Imaging, is "a resinous acrylic bone graft material used in fracture fixation and bonding of a prosthesis." In other words, surgeons use this stuff when replacing joints.

The presenter reported that over 500 patients had been treated with PMMA, with follow-up ranging from 1 month to 5 years. The results? PMMA was "lasting," "safe," "improved quality of life" and "produced excellent esthetic results." Cool, but we had to take Oyafuso's word for it, as he presented a series of anecdotes but no real data.

Several impressive "before and after" pictures were presented, but there was no aggregate analysis of objective or even subjective measures of body image or toxicity. We were informed that "some" patients needed repeat treatments 12 or more months after the initial procedure, but there were no data backing these assertions.

Additionally, although 5-year follow-up data were promised in the title of this paper, it is unclear what the median follow-up time was. From what was presented, PMMA appeared to work. The compound is commonly available for orthopedic procedures and is therefore not expensive. The use of PMMA for facial fat wasting is an interesting concept that, if found to be safe and effective, may be an alternative to poly-L-lactic acid (Sculptra, New-Fill). Unfortunately, we learned little about the efficacy, durability and safety of PMMA at this conference. Hopefully, this group or others will produce more rigorous data regarding this compound.


Unfortunately, few of us left Bangkok with much more than a renewed appreciation for Thai cuisine, a few tailor-made shirts and pictures in front of a golden Buddha. With so many important HIV-related conferences each year, notable study results are parsed out to CROI (Conference on Retroviruses and Opportunistic Infections), ICAAC (Interscience Conference on Antimicrobial Agents and Chemotherapy) and IAS (International AIDS Society Conference on HIV Pathogenesis and Treatment), as well as specialty meetings such as the Lipodystrophy Workshop. This dilutional effect is certainly one explanation for the paucity of important results to be found at this year's International AIDS Conference. The abundance of empty poster boards where accepted posters were to be placed offers another. That some did not even show up to present their data delivers a stark message that the scientific attractiveness of this conference has waned.

The most well-done studies did help confirm what we had begun to suspect, and this is of value. However, there were no surprises except for how few surprises there were. We will always have Paris, but remember, next time it's Rio.


  1. DeJesus E, Grinsztejn B, Rodriguez C, et al. Efficacy and safety of atazanavir with ritonavir or saquinavir vs lopinavir/ritonavir in patients who have experienced virologic failure on multiple HAART regimens: 48-week results from BMS A1424-045. In: Program and abstracts of the 11th Conference on Retroviruses and Opportunistic Infections; February 8-11, 2004; San Francisco, Calif. Abstract 547.

  2. Jemsek JG, Arathoon E, Arlotti M, et al. Atazanavir and efavirenz have similar effects on body fat distribution in antiretroviral-naive patients when combined with fixed-dose zidovudine and lamivudine: 48-week results from the metabolic subset. In: Program and abstracts of the 2nd International AIDS Society Conference on HIV Pathogenesis and Treatment; July 13-16, 2003; Paris, France. Abstract LB13.

  3. Staszewski S, Gallant JE, Pozniak AL, et al. Efficacy and safety of tenofovir DF (TDF) versus stavudine (d4T) when used in combination with lamivudine and efavirenz in antiretroviral naive patients: 96-week preliminary interim results. In: Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Mass. Abstract 564b.

  4. Kumar P, Rodriguez-French A, Thompson M, et al. Prospective study of hyperlipidemia in ART-naive subjects taking combivir/abacavir (COM/ABC), COM/nelfinavir (NFV), or stavudine (d4T)/lamivudine (3TC)/NFV (ESS40002). In: Program and abstracts of the 9th Conference on Retroviruses and Opportunistic Infections; February 24-28, 2002; Seattle, Wash. Abstract 33.

  5. van Leth F, Phanuphak P, Gazzard B, et al. Lipid changes in a randomized comparative trial of first-line antiretroviral therapy with regimens containing either nevirapine alone, efavirenz alone or both drugs combined, together with stavudine and lamivudine (2NN Study). In: Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Mass. Abstract 752.

  6. Tashima K, et al. Lack of clinical lipodystrophy in patients receiving efavirenz + NRTIs in study 006. In: Program and abstracts of the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy; September 26-29, 1999; San Francisco, Calif. Abstract 1304.

  7. Riddler SA, Smit E, Cole SR, et al. Impact of HIV infection and HAART on serum lipids in men. JAMA. June 11, 2003;289(22):2978-2982.

  8. Gathe JC Jr, Washington MY, Mayberry C, Piot D, Nemecek J. IMANI-1 TC3WP single drug HAART -- proof of concept study. pilot study of the safety and efficacy of Kaletra (LPV/r) as single drug HAART in HIV + ARV-naive patients-interim analysis of subjects completing final 48 week data. In: Program and abstracts of the XV International AIDS Conference; July 11-16, 2004; Bangkok, Thailand. Abstract MoOrB1057.

It is a part of the publication The XV International AIDS Conference.

Please note: Knowledge about HIV changes rapidly. Note the date of this summary's publication, and before treating patients or employing any therapies described in these materials, verify all information independently. If you are a patient, please consult a doctor or other medical professional before acting on any of the information presented in this summary. For a complete listing of our most recent conference coverage, click here.