CONFERENCE NEWS

LIPODYSTROPHY AND METABOLIC
COMPLICATIONS IN HIV INFECTION

This ICAAC symposium consisted of four presentations, which reviewed our current understanding of lipodystrophy and metabolic complications. Similar reports were included in the pre-conference workshop reviewed below. The lipodystrophy workshop is discussed in a separate paper later in this issue.

Mitochondrial Toxicity

Mitochondrial toxicity due to nucleoside reverse transcriptase drugs (NRTIs) is thought to be associated with lipodystrophy. At the conference, Dr. William Lewis discussed mitochondrial toxicity due to NRTIs, focusing on AZT-associated cardiomyopathy. NRTIs may also play a role as a cause of liver dysfunction and lactic acidosis. An animal study was performed to assess the relationship of cardiomyopathy to AZT. Using a high dose of 100 mg/kg/day, Lewiss group demonstrated evidence of cardiomyopathy with abnormal mitochondria ultrastructurally. When the study was repeated looking at skeletal muscle and mitochondrial DNA, (mtDNA), the investigators found a 50% decrease in mtDNA. Presumably, AZT acts as both a competitive and noncompetitive inhibitor of mitochondrial DNA polymerase. Of interest, didanosine (ddI) was not found to have the same properties in cardiac muscle, suggesting that not all drugs in the NRTI class are equal in terms of mitochondrial toxicity.

Lewis W. Mitochondrial toxicity: Potential mechanisms and experimental evidence. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, Ontario, Canada. September 1720, 2000. Abstract 1372.

Insulin Resistance

Dr. Steven Grinspoon began his presentation with an overview of the body-shape changes associated with lipodystrophy, sometimes called HIV-associated adipose redistribution syndrome (HARS). He then presented an overview of glucose metabolism in HIV infection in the era of HAART. The evidence for insulin resistance in HAART-treated patients includes fasting hyperinsulinemia, decreased sensitivity to insulin during insulin tolerance tests, and decreased glucose tolerance during oral glucose tolerance tests.

The long-term sequelae of insulin resistance in HIV infection are uncertain. In addition to its association with increased cholesterol and triglycerides and other cardiac risk factors, hyperinsulinemia can also promote microvascular disease. The chance of progression from insulin resistance to impaired glucose tolerance and to diabetes mellitus in HIV infection also is unclear, although this has been seen in HIV-uninfected subjects over the course of many years.

The etiology of insulin resistance is multifactorial. A direct effect of antiretroviral agents, especially protease inhibitors (PIs), has been shown in cross-sectional studies, and in studies of PI administration (e.g., indinavir) to HIV-seronegative subjects. However, insulin resistance may also be found in patients on PI-sparing regimens and in those from the pre-HAART era.

Body composition changes, including increased visceral adipose tissue (VAT) and decreased subcutaneous adipose tissue (SAT), are indirectly related to insulin resistance. Excess VAT appears to increase insulin resistance in the liver via release of free fatty acids into the portal vein, impacting insulin signaling. Depletion of SAT affects insulin resistance either through loss of a depot for insulin or due to the lack of an endocrine signal from adipose tissue to increase tissue insulin sensitivity.

Therapy for insulin resistance in HIV infection includes switching antiretrovirals (see the discussion of Dr. Powderlys presentation below) and changes in diet, exercise, or insulin-sensitizing agents. Combination therapy with metformin and rosiglitazone has been used safely in non-HIV conditions and appears to be more effective than either drug alone. There are no data as yet on this combinations role in HIV-associated insulin resistance.

Dr. Grinspoon concluded that insulin resistance is common in treated, HIV-infected patients, that it is related to cardiac risk, and that treatment with insulin-sensitizing agents decreases insulin resistance and should decrease cardiac risk.

Grinspoon S. Mechanism and treatment of insulin resistance in HIV lipodystrophy. Program and Abstracts of the Second International Workshop on Adverse Drug Reactions and Lpodystrophy in HIV. Toronto, Ontario, Canada. September 13-15, 2000. (No abstract published.)

Cardiovascular Risk in HIV Infection

Dr. Matthias Egger presented an epidemiologic analysis of the risks and benefits of HAART. He reviewed the relative risks of cardiovascular disease progression as well as the benefits of antiretroviral therapy (i.e., decreased risk of disease progression).

Cardiovascular disease is multifactorial, and individual risk varies with environmental, genetic, and metabolic factors, and with age, body fat distribution, smoking, etc. Dr. Egger discussed cardiovascular risk in HIV infection with respect to various combinations of hyperlipidemia and insulin resistance or diabetes mellitus. He then compared results from other cohorts including the Caerphilly cohort of middle-aged men. Subjects without lipodystrophy were metabolically similar to younger men from the Caerphilly cohort, while those with moderate lipodystrophy were metabolically similar to older men from the Caerphilly cohort.

Risks and benefits were related to the number of subjects treated in order to produce harm (risk) or help (benefit). His analysis incorporates both relative risk and underlying prevalence of the factor being assessed.

In HIV, the risk of progression is related to CD4+ lymphocyte count, plasma viral load, and treatment with HAART. The risk of progression for a CD4+ count <200 cells/mm3 and a plasma viral load >105 copies/mL in the absence of treatment was more than 90% in the Multicenter AIDS Cohort Study (MACS). In this setting, the number needed to treat with HAART in order to produce a benefit is 1 to 2 patients. In the era of HAART, using data from the Swiss HIV Cohort, the average 3-year risk of progression is about 20%, while it is only about 2% for someone with a CD4+ count greater than 500 cells/mm3 and a plasma viral load less than 104 copies/mL.

Calculation of the number of patients one would need to treat in order to produce harm ranges from 10 to 200 subjects, based upon different pre-existing metabolic characteristics. Egger was clear that this analysis favors treatment. However, there are circumstances in which the riskbenefit ratio may be less beneficial, such as in the setting of high atherosclerotic comorbidity and high CD4+ count or slow disease progression.

Dr. Egger said that lifestyle changes reduce cardiovascular risk significantly; as just one example, a substantial proportion of the added risk to any patient could be reduced through smoking cessation.

Dr. Egger cited these web sites for more information on quantitating cardiovascular risk: <http://www.hbroussais.fr/Scientific/fram.eng.html> and <http://www.chd-taskforce.de>

Egger M. Cardiovascular complications of HAART: Need for perspective. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, Ontario, Canada. September 1720, 2000. Abstract 1374.

Antiretroviral Switch Studies

Dr. Bill Powderly reviewed the current status of antiretroviral switching. Reasons for switching drugs include treatment failure, management of side effects, or the need to simplify a regimen. Most studies involve switching away from PIs, because they were felt initially to be responsible for lipodystrophy.

Dr. Powderly discussed studies that included at least 6 months of followup. He concluded that switching was safe in terms of maintaining virologic suppression. In patients with fully suppressed viral load at the time of switch, suppression is maintained in most when PIs are switched to NNRTIs (nevirapine, efavirenz). However, in one study, switches to abacavir showed almost double the incidence of lost virologic control, possibly due to prior NRTI use with multiple mutations in the RT gene.

For managing metabolic abnormalities, improvement in insulin sensitivity has been seen when switching from PIs to NNRTIs or abacavir. This response may vary, however, since some patients have insulin resistance in the absence of PI therapy. The responses of triglycerides and cholesterol are less predictable and may depend on the specific drugs involved and other factors.

There are no data showing that fat redistribution is affected by switching antiviral agents. The effects of switching on body composition are difficult to interpret as subjects are not well characterized at baseline, quantitative measurements are not done consistently, and patient self-report, when used as an endpoint, is unreliable quantitatively. Given these caveats, a few studies have examined switching stavudine to zidovudine or abacavir. One study showed that such a switch led to a decrease in serum triglycerides, free fatty acids, and lactate, but no change in cholesterol or insulin concentrations. Subcutaneous adipose tissue (SAT) increased while visceral adipose tissue (VAT) did not.

Interpreting the results of switching agents in terms of cause and effect is also difficult. If a change occurs, the antiretroviral agent in question may be the cause of the abnormality, or the change may be merely coincidental and due to lifestyle changes, dieting, or other therapy. If a change does not occur, the antiretroviral agent may not have been the cause, the change may be irreversible, or, more likely, the cause is multifactorial, other factors are involved, or followup may be too short to detect the change.

Dr. Powderly summarized by stating that switching is safe, with a note of caution about abacavir in heavily pretreated subjects. Switching from PI therapy can be expected to lead to decreased insulin resistance. Triglyceride levels may or may not change, and total cholesterol, HDL cholesterol, and body fat redistribution are less likely to change.

Powderly WG. The strategy of antiretroviral switch studies: A review. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, Ontario, Canada. September 1720, 2000. Abstract 1375.

Antiretrovirals

Another fixed combination of antiretrovirals was approved just prior to the conferences opening. The drug Kaletra® (lopinavir, formerly ABT-378, 400 mg with ritonavir 100 mg twice daily) was described in treatment-naïve patients who had viral loads of at least 5000 copies/mL.

After 96 weeks, 83% of subjects in a Phase II study had plasma HIV RNA <400 copies/mL and 78% had <50 copies/mL. During the 96-week study, 2 subjects discontinued therapy due to adverse events attributed to the study drug. Cholesterol levels greater than 300 mg/dL and triglyceride levels greater than 750 mg/dL were seen in 14% and 12% of patients, respectively. These values are difficult to interpret since many patients had abnormal lipids at the start of the trial and fasting levels were not available.

The effects of lopinavir/ritonavir on serum lipids will be more clearly defined as we gain more experience with this new co-formulation.

Another study in treatment-naive patients was presented. This Phase III study compared lopinavir/ritonavir to nelfinavir; both were combined with lamivudine and stavudine. Subjects had viral loads >400 copies/mL; there was no CD4+ count requirement.

653 patients with similar baseline characteristics were randomized, 326 to the lopinavir/ritonavir arm and 327 to the nelfinavir arm. The median baseline viral load was high (4.9 log10 copies/mL) and median CD4+ count was 259 cells/mm3. At 24 weeks, using an intent-to-treat (ITT) analysis in which noncompleters (NC) were considered failures (F), (ITT NC = F), 79% of subjects in the lopinavir/ritonavir had viral loads <400 copies/mL compared with 71% in the nelfinavir arm. This difference increased significantly over time, to 79% and 64%, respectively, at 40 weeks (P < .001).

The results for viral load <50 copies per mL were similar: there was no significant difference between the arms at 24 weeks. However, by 40 weeks, the proportion with undetectable viral loads in the lopinavir/ritonavir arm continued to increase, while the proportion in the nelfinavir arm did not (70% vs 54%). The CD4+ cell count increases were large and similar in both groups, reaching +190 and +177 cells/mm3 at 40 weeks.

When patients with virologic failure were studied, there was a significantly lower prevalence of protease-inhibitor resistance in the lopinavir/ritonavir group.

These two studies support the use of Kaletra in treatment-naïve individuals. In addition, the Walmsley study begins to position this co-formulation in terms of its potency, resistance profile, and safety.

Benson C, Brun S, King M, et al. Two-year followup of ABT-378/ritonavir (ABT-378/r) in antiretroviral naive HIV+ patients. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, Ontario, Canada. September 1720, 2000. Abstract 546.

Walmsley S, Badley A, Beall G, et al. Efficacy of ABT-378/r vs. nelfinavir (nelfinavir) in antiretroviral (ARV)-naive subjects: Results of a Phase III blinded randomized clinical trial. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, Ontario, Canada. September 1720, 2000. Abstract 693.

The lopinavir/ritonavir fixedcombination was also described in treatment-experienced patients, resulting in good outcomesin this case, undetectable viral loads.

Becker S, Brun S, Bertz R, et al. ABT-378/ritonavir (ABT-378/r) and efavirenz: 24 week safety/efficacy evaluation in multiple protease inhibitor experienced patients. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, Ontario, Canada. September 1720, 2000. Abstract 697.

Early Treatment and Immune System Preservation/Recovery

An observational, unrandomized study of HIV primary infection was presented in which one group was treated very early in the course of their HIV disease, while therapy was delayed in the second group. The investigators looked at immunologic and virologic response rates in these two groups treated with HAART. The early group began therapy an average of 50 days following infection, and the later group began treatment an average of 750 days after infection. CD4+ cell counts following therapy were significantly higher in the early-treatment group (950 vs 700 cells/mm3), although the change from baseline was not significantly different. All 15 subjects treated early had viral loads <500 copies/mL, while 8 of 30 subjects in the later-treatment group were not suppressed below this level. More impressive was the difference between the groups CD4+/CD8+ ratio: 1.55 in the early group and 0.88 in those treated later.

Although finding and treating HIV patients during primary infection is not accomplished easily, these data support the concept that immune system preservation is associated with earlier treatment. Later treatment with HAART, during chronic infection, is associated with incomplete immune recovery. However, in order to know whether these differences have any clinical importance, we need studies with very long-term followup.

Geise RE, Maenza J, Collier A, et al. Comparison of early therapy versus delayed therapy in patients followed since early HIV infection. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, Ontario, Canada. September 1720, 2000. Abstract 808.

Immune System Reconstitution and IL-2

The CPCRA 059 trial was presented by Abrams et al, who studied 511 patients with CD4+ >300 cells/mm3 either receiving or about to begin antiretroviral therapy.

Previous studies have shown inconsistent results regarding the impact of IL-2 on increasing viral load through activation of CD4+ cells. More recent data suggest that patients responding to aggressive HAART therapy may experience an actual decrease in viral load with IL-2. In this study, the primary endpoint was the proportion of patients with plasma HIV RNA <50 copies/mL after 1 year of study.

Patients were randomized to 1 or 2 doses of IL-2 (4.5 or 7.5 million international units [MIU] for 5 days every 8 weeks) or to no IL-2. This regimen was repeated for 3 cycles and then

IL-2 was dosed as needed to keep CD4+ cell counts above 1000 cells/mm3, or twice baseline levels. 255 randomized patients received no IL-2 and 256 received IL-2 (130 received 4.5 MIU and 126 received 7.5 MIU). Eighty-two percent of patients receiving IL-2 completed at least 3 cycles of therapy.

After 1 year of followup, 64.4% of patients had viral loads <50 copies/mL in the IL-2 arm compared to 61.7% in the control arm (P = .63). In patients with detectable viral loads at entry, viral load steadily decreased during the trial and was not different between the two arms. No increase in viral replication with IL-2 administration was seen during the course of study. Mean CD4+ cell counts increased substantially among IL-2treated patients: there was a mean CD4+ cell count increase of 22 cells/mm3 and 276 cells/ mm3 in the control and IL-2 study arms, respectively (P < .0001).

The authors concluded that in this population of early-stage HIV-infected patients on antiretroviral therapy, IL-2 could produce substantial and sustained increases in CD4+ cell count without significantly altering viral loads after 12 months.

Abrams DI, Bebchuk JD, Denning ET, Markowitz NP. IL-2 therapy produces no change in HIV RNA after one year. Beirn Community Programs for Clinical Research on AIDS. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy Addendum. Toronto, Ontario, Canada. September 1720, 2000. Abstract L-11.

The Value of Resistance Testing

Clotet et al presented a study from Barcelona called the Havanna study, in which the investigators evaluated how genotyping and expert advice influenced the virologic outcomes in 326 heavily pretreated patients on failing antiretroviral therapy. Patients were randomized to four groups:

1. Genotyping with expert advice

2. Genotyping without expert advice

3. Standard of care with advice, and

4. Standard of care without advice

The group with both genotype and advice had the best results. There was no difference between the outcomes of the group with expert advice only and the group with genotyping only. The standard-of-care group without expert advice had the poorest outcomes. Using an intent-to-treat analysis, the percentage of patients with viral loads <400 copies/mL at week 24 was 69% in group 1, 46% in group 2, 49% in group 3, and 36% in group 4 (P < .01).

This study adds to our understanding of the benefits of resistance testing and expert advice; when given together, they improve virologic outcomes. Moreover, a genotype test without expert advice is of no more benefit to the patient than simply having expert advice alone. The question that remains is how best to help a practitioner with interpreting genotypic information (see the discussion of Dr. Johnsons presentation below).

Tural C, Ruiz L, Holtzer C, et al. The potential role of resistance decision support software with or without expert advice in a trial of HIV genotyping versus standard of care-The Havanna Trial. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, Ontario, Canada. September 1720, 2000. Abstract L-10.

Johnson et al showed that the virtual phenotype (i.e., phenotype reported based on genotype results compared to a database of resistance) predicted response in a retrospective clinical cohort. This study compared the virtual phenotype vs. expert opinion for predicting virologic response. The experts (2 virologists, 1 clinician, and 1 clinical pharmacologist) all predicted response based on genotype results with accuracy comparable to that of the virtual phenotype used in the Virco® assay. The individual predictions of the different experts were comparable. These results were highly reproducible when same experts repeated their genotype interpretations 2 weeks later. There did not appear to be any specific antiretroviral drug whose effect was more difficult to predict than any other in this study.

Other studies have not shown the same degree of reproducibility. However, in the hands of experienced consultants, the results appear to be as reliable as those obtained using the virtual phenotype. For more information on resistance and resistance testing, visit: <http://www. hivresistance. com/index_main.html>

Johnson VA, Call SA, Westfall AO, et al. Predictive value of genotypic resistance testing as interpreted by experts and a virtual phenotype. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. September 1720, 2000. Toronto, Ontario, Canada. Abstract 1259.

Structured Treatment Interruptions

Saag et al reported on 78 HAART-experienced patients (mean duration of treatment before interruption: 506 days) who had a treatment interruption of 30 days or longer and then resumed HAART for at least another 30 days.

CD4+ cell counts returned to 90% of pre-interruption values in 46 patients (59%); 8 (10%) were unable to achieve 50% of pre-interruption value. Viral-load reduction to within 20% of pre-interruption values was seen in 77% of patients, with 72% having viral-load values less than or equal to pre-interruption values. The authors concluded that the majority of treatment-experienced individuals who interrupt therapy are able to re-establish their pre-interruption levels of HIV viral-load suppression and CD4+ cell counts when therapy is reinitiated.

Saag MS, Chen RY, Westfall AO, et al. Immunologic and virologic consequences of antiretroviral treatment interruption (TI) in clinical practice. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy Addendum. Toronto, Ontario, Canada. September 1720, 2000. Abstract L-18.

Another study reported that treatment interruptions in patients taking NNRTIs might be complicated by the long half-lives of efavirenz and nevirapine. Active drug levels persist after NRTIs have been eliminated, resulting in a period of "NNRTI monotherapy" that may contribute to selection of NNRTI-resistant virus.

The effect of interruption of efavirenz, the NNRTI with the longest half-life of any licensed antiretroviral, was studied retrospectively in patients in the pivotal DMP-006 study who had viral loads <50 copies/mL while receiving efavirenz, who interrupted therapy due to adverse effects, and who later restarted all drugs.

Although there was no statistically significant difference between the interrupters and the noninterrupters, the authors concluded that there was a trend toward less favorable outcome among patients who interrupted therapy in both the indinavir and efavirenz groups, suggesting that there may be some long-term risk associated with this approach.

Ruiz NM, Bacheler LT, Farina DR, Baker DB, Manion DJ, Gregoire V. Temporary Interruption (TI) of SUSTIVA™ (efavirenz, EFV) containing therapy does not impact virological response rates. Program and Abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, Ontario, Canada. September 1720, 2000. Abstract 479.