Report from the 40th ICAAC
     Written for NATAP by Robert Murphy, MD, Northwestern University in Chicago, and ACTG researcher
  1. ìSimplificationî Studies: Switching from Protease Inhibitor-Based Therapies
  2. New Drugs

Considerable attention has been focused upon studies that are designed to change or switch a patientís therapy from one ìHAARTî regimen to another. The original intent was to see if removing the protease inhibitor would reverse ìlipodystrophyî, a syndrome still without a formal definition. Early on, many assumed that the protease inhibitors alone were responsible for the morphologic and metabolic changes that were being observed in persons with good virologic control on protease inhibitor-based HAART. The morphologic changes were primarily abdominal fat accumulation and the metabolic changes were primarily related to serum lipids. We now know that the physical and metabolic changes associated with therapy are much more complex and include subcutaneous fat atrophy, visceral fat accumulation, insulin resistance, impaired glucose tolerance, osteopenia, hypertriglyceridemia, hypercholesterolemia, and decreased HDL cholesterol. The causes or associations include not only protease inhibitors but the nonnucleoside and nucleoside reverse transcriptase inhibitors, time on therapy, disease stage, age of the patient and degree of virologic control.  In addition to switching therapy because of toxicity, switching is also being done to ìsimplifyî a treatment regimen by reducing the number of pills, frequency of dosing and food restrictions and therefore, presumably, improve adherence and long term virologic outcome.

Bill Powderly from Washington University in St. Louis gave a review of the ìswitchî studies reported on to date [abstract 1375]. So far there have been reports from 11 cohort and observational studies and 5 randomized trials. Although the results are mixed, in general, when switching from a protease inhibitor to either of the nonnucleosides efavirenz or nevirapine or to a third nucleoside, abacavir, triglyceride levels and cholesterol decrease, HDL cholesterol rises, insulin resistance decreases and there is little if any change in body fat composition or distribution. Virologically the switch studies have been safe in that there have not been excessive treatment failures following the change in therapy with the exception that changing to abacavir is associated with the 2-3 times the virologic failure rate particularly in subjects with prior nucleoside-only treatment history.

(Ed comments from Jules Levin: for the patient with treatment experience switching prior to the regimen they are currently taking is more risky & complicated, as also mentioned below by Dr. Murphy. Previous NRTI experience may increase risk of virologic failure. For example, in the switch studies from a PI regimen with 2 nukes, switching the PI to abacavir while maintaniing the other two NRTIs increased risk for failure. If a person is on a double PI regimen with 2 NRTIS and has NRTI experience prior to their current regimen switching the PI therapy to a NNRTI may not be adequate to hold viral load below detection).

Clotet et al reported on the results of a study in 77 nonnucleoside-naÔve (bur NRTI experienced prior to current regimen) subjects who were on a stable protease inhibitor-based regimen for at least 12 months with recent viral load measurements below 50 copies/mL.[abstract 473] Subjects were randomized to remain on their protease inhibitor-based therapy or substitute the protease inhibitor with nevirapine or efavirenz. The follow up time reported was approximately 9 months.  Sixty percent of subjects were on stavudine with lamivudine primarily and 30% were receiving zidovudine/lamivudine as their nucleoside backbone therapy. Fifty-six percent of subjects had clinical lipodystrophy at the start of the study. The patients who switched to nevirapine or efavirenz did well in terms of virologic control and improved ìquality of lifeî.  The nevirapine group had significant decreases in total cholesterol, LDL cholesterol and triglycerides.  The nevirapine group had 5 cases of chemical hepatitis however only 1 subject had to discontinue therapy. Most of these patients were coinfected with hepatitic C virus. Another nevirapine subject discontinued because of rash. Nine subjects in the efavirenz group had significant central nervous system symptoms, 3 of whom discontinued therapy. There were no differences in clinical lipodystrophy, DEXA scan results, or anthropometric measurements between any of the groups.  In this relatively small trial, switching to either of the nonnucleosides was safe, however the only metabolic changes of benefit were lipid related in the nevirapine group. Treatment-related toxicities occurred in approximately 10%, particularly those coinfected with hepatitis C receiving nevirapine. These patients need to be monitored closely.

Raffi et al reported on the results of switching from a protease inhibitor-based therapy to nevirapine (n=63) or efavirenz (n=10) in nonucucleoside-naÔve patients with viral load below 200 copies/mL for at least 6 months.[abstract 474] Most subjects changed therapy for ìpersonal desireî;18 had lipodystrophy, primarily fat accumulation. Overall, 10/63 (13.7%) experienced virologic failure as defined by greater than 200 copies/mL.  The median viral load at the time of rebound was 915 copies/mL and in all 10, when switched back to their protease inhibitor, they returned to undetectable levels. The CD4 counts increased by 151 cells/mm3.  There were no significant changes in cholesterol, triglyceride levels decreased significantly compared to baseline (P=0.3), more were likely to have virlogic rebound if they had failed more than 1 previous regimen, but this was not significant (P=0.9). The rash rate in those receiving nevirapine was quite low, only 1/63.  Again, no changes in body fat composition were noted. This results from this observational study suggests that changing to the nonnucleosides is well tolerated but not associated with significant metabolic or body fat changes except for normalization of triglyceride levels.

(Additional information from Jules Levin: Raffi reported that there was a trend towards a higher viral rebound rate for individuals with NRTI experience prior to initiating their PI HAART therapy compared to persons without any prior NRTI experience 19.2% vs 6.5%). He is examining these patients blood samples to see if they had pre-existing resistance).

Rachlis et al reported on the results from the DMP 266-049 study, a prospective, randomized, multicentered open label trial of switching to an efavirenz-based therapy versus continuing protease inhibitor-based therapy in 310 subjects with viral load less than 50 copies/mL on their first protease inhibitor regimen. [abstract 475] Subjects were randomized 2:1 into the efavirenz arm. At 48 weeks, the proportion of subjects with less than 50 copies/mL was 81% in the protease inhibitor group versus 89% in the efavirenz switch group (P=0.05). The number of adverse events was higher in the efavirenz arm (30% versus 17%), however, the reported events were generally mild and not treatment limiting.

(Additional info from Jules Levin: Through 24 weeks, 9 (4%) and 15 (12.5%) patients were randomized but not dosed in the EFV and PI arms, respectively. Discontinuing for any reason were 6.9% (15) in the EFV arm and 12.4% (13) in the PI arm. Discontinuing for adverse events were 2.8% (6) in the EFV arm and 2.0% (2) in the PI arm.

HDL cholesterol (good cholesterol) increased significantly in the efavirenz group (P=0.05). Adherence was better in the efavirenz group and the authors believe that this is what drove the positive results favoring the nonnucleoside. In this large prospective study, adherence to a simplified regimen including efavirenz resulted in an improved virologic outcome and increases in HDL cholesterol. 

Two studies looked at the outcome associated with changing the protease inhibitor to a third nucleoside, abacavir. In the study CNA 30017 reported by Montaner et al, 211 patients were randomized to either remain on their protease inhibitor or switch to the combination abacavir, zidovudine, lamivudine. [abstract 477] Patients were treatment-naÔve prior to regimen they were on when they entered study. Four patients in the abacavir arm experienced virologic failure as defined by an increase in viral load to greater than 400 copies/mL. In 3 of the 4, the M184V mutation was observed. Only 2 subjects failed in the protease inhibitor arm. The time to virologic failure was longer in those randomized to abacavir (P<0.028). In 15 patients remaining on their protease inhibitors, therapy was discontinued due to adverse events compared to only 5 in the abacavir arm. In the study presented by Opravil et al, the results were dissimilar and of particular concern.[abstract 476] In this trial, subjects on their original protease inhibitor therapy and with a viral load of less than 50 copies/mL were randomized to remain on their protease inhibitor-based therapy or switch to abacavir, zidovudine, lamivudine.  Subjects had to be negative for the 215 codon mutation by amplified PBMC DNA. In the intent-to-treat analysis after a median of 68 weeks, 15% of subjects experienced virlogic failure in the abacavir arm compared to 6% in the control group. The on-treatment results were similar with 13 persons in the abacavir arm compared to 2 persons in the control group experiencing virologic failure. Prior zidovudine therapy without a full HAART regimen was a significant predictor of failure. In those subjects with prior zidovudine without HAART, 29% had virologic failure on abacavir compared to 10.5% who remained on their protease inhibitor therapy. In subjects without prior non-HAART zidovudine, 7.5% on abacavir compared to 2.4% remaining on their protease inhibitors experienced virologic failure. Total and LDL cholesterol levels declined significantly and adherence was better in those treated with abacavir. These studies point out the problems associated with the simplified triple nucleoside abacavir-based therapy. Although generally effective, this approach may not be as successful in subjects with prior nucleoside therapy, particularly if they received zidovudine outside of a HAART regimen. Metabolic changes were primarily related to decreases in LDL and total cholesterol levels.

The problems with interpreting these studies is that the reason for switching varies between studies and even within the same trial. Patients have been enrolled in these trials primarily because of clinical lipodystrophy, a formally undefined syndrome, or because of ìpersonal desireî suggesting that they want something different. This mixture of reasons for entering the studies effects the non-virologic endpoints. More focused studies are needed to help clarify the effects on the metabolic and fat redistribution endpoints.

New Antiretroviral Drugs

Clinical data on new drugs in development included updated reports on the protease inhibitors, ABT-378/ritonavir and BMS 232,632, the nucleoside DAPD and the nucleotide tenofovir.

The results of 2 phase III studies of ABT-378/ritonavir were reported by Walmsley and Becker [abstracts 693 and 697]. Their presentations were upstaged somewhat by the FDA approval of ABT-378/ritonavir for use in the United States market 2 days prior to ICAAC. Study M98-863 is a placebo-controlled, randomized comparison of ABT-378/ritonavir versus nelfinavir in 653 essentially treatment-naÔve subjects who also received stavudine and lamivudine. At entry, viral load had to be greater than 400 copies/mL and there were no CD4 count restrictions. At baseline viral load was 4.9 log10 copies/mL and CD4 count was 259 cells/mm3.

In the intent-to-treat analysis at 40 weeks, the proportion of subjects with a viral load less than 400 copies/mL was 79% in the ABT-378/ritonavir group versus 64% in the nelfinavir group (P<0.001). The proportion with less than 40 copies was 70% and 54% respectively (P<0.001). At 24 weeks, there was no significant difference in proportion of subjects with less than 40 copies/mL. In the subjects who had a viral load greater than 400 copies/mL and a sequenced genotype at week 24, none in the ABT-378/ritonavir group and 14 in the nelfinavir group had mutations in the protease gene.

The proportion of subjects discontinuing therapy because of virlogic failure was greater in the nelfinavir group (8% versus <1%).  Hypertriglyceridemia (without regard to fasting) was more common in the ABT-378/ritonavir group compared to the nelfinavir group (P<0.001), otherwise there was no difference in reporting of adverse events. In the M98-957 study, 57 nonnucleoside naÔve patients had received at least 2 protease inhibitors in the past and were currently receiving one and had a viral load of greater than 1000 copies/mL.

Patients received either ABT-378/ritonavir at the 400/100 mg dose plus efavirenz and a nucleoside or they received the same regimen but increased the dose to 533/133 at day 15. Efavirenz is known to increase the metabolism of most protease inhibitors. Subjects on the 400/100 mg dose of ABT-378/ritonavir (3 capsules twice daily) had a 33% decrease in the trough concentrations and 25% decrease in the area-under-the-curve (AUC) compared to the 533/133 mg dose (4 capsules twice daily).

The proportion of subjects with viral load less than 50 copies/mL at week 24 were similar, 62 and 64% respectively. Only 3 subjects experienced virologic rebound, 2 in the low dose and 1 in the higher dose group. Virologic response defined by obtaining a viral load less than 50 copies/mL was correlated with the number of genotypic mutations at baseline.  Individuals with 0-5 mutations had an 88% response, those with 6-7 at 57% response and those with 8-10 a 17% response. The phenotype resistance in general held to 40-fold decrease in susceptibility relative to wild type.  Diarrhea (14 versus 7) , asthenia (14 versus 7) and grade 3 or 4 hypertriglyceridemia (36% versus 31%) or hypercholesterolemia (36% versus 28%) were more common in the higher dosed group.

Together, these studies show the relative potency of the recently approved protease inhibitor ABT-378/ritonavir in both treatment-naÔve and experienced patients.  Patients receiving concomitant efavirenz and presumably nevirapine, another metabolic inducer, will require an increase in the dosage of ABT-378/ritonavir to 533/133 twice daily. This will come at some cost in terms of side effects and drug purchasing expense. 

Phase II data from the AI424-007 study of the protease inhibitor BMS 232,632 were presented by Sanne et al. from South Africa.[abstract 691] This new protease inhibitor is potent, with an IC50 of 2-5 nM, which is 2-19 fold more potent than other protease inhibitors in vitro. The half life is long, 4-7 hours, which has lead to single daily dosing studies. In the doses tested so far, the Cmin has exceeded the IC90 (8-12 nM) at all time points up to 48 hours; however drug exposure is significantly enhanced if it is taken with food.

In this 324 patient dose comparison study of 200 mg, 400 mg, and 500 mg of BMS 232,632 plus a 4th arm of nelfinavir 750 mg three times daily, the protease inhibitor was given as a monotherapy for 2 weeks, followed by the addition of stavudine and didanosine. The proportion of subjects with viral load less than 50 copies/mL at 24 weeks ranged from 30-80% favoring the higher doses. CD4 cells counts rose 80-124 cells/mm3. Of great interest was that no changes in cholesterol were noted. The only significant problem with BMS 232,632 was hyperbilirubinemia, which typically was not treatment-limiting. The interest in this drug is high for several reasons:  it may be the first protease inhibitor that can be dosed once daily, adequate concentrations are achieved without pharmacologic enhancement with ritonavir, and it apparently does not affect serum lipid values.

Other potential effects on metabolic parameters remain to be determined. The potential problems include that hyperbilirubinemia is relatively common but of little or unknown concern. The drug must be taken with food which is relatively easy for patients, but will preclude a once daily dosing regimen that includes the new expanded release formulation of didanosine which must be taken without food. Overall development prospects look promising.

DAPD, the new nucleoside analogue, looks very promising if early phase II data hold up. DAPD is active against most resistant strains of HIV including those with the 69SS insertion but not against those with the 65 and 74 mutations (this may be important in patients with didanosine resistance). In a dose ranging study in treatment naÔve subjects of DAPD alone for 15 days, viral load dropped 1.5 log10 copies/mL in the 300 and 500 mg twice daily groups.

 In 18 treatment-experienced subjects with considerable nucleoside use, subjects were treated with 200, 300 and 500 mg DAPD twice daily after a treatment wash-out or as an add-on therapy. In those that had stopped their prior treatment, the viral load declined by 1.0 log10 copies/mL and in those who added DAPD to their failing regimen, the viral load declined 1.9 log10 copies/mL in the 500 mg twice daily group. No significant serious adverse events were reported and no new genetic mutations were noted in the 5 subjects who experienced virologic increases. [abstract 690] DAPD development will be watched very closely and may become an extremely important drug particularly for subjects with multiple nucleoside-reverse transcriptase mutations.

Tenofivir is a nucleotide analogue that has a favorable resistance profile and long intracellular half-life of approximately 30 hours.  In a placebo controlled dose-ranging study, 186 subjects on stable antiretroviral therapy were treated with tenofovir 75 mg, 150 mg, 300 mg or placebo. At 48 weeks, the decline in viral load in the 300 mg group was 0.6 log10 copies/mL. There were no differences in change from baseline in CD4 cell counts despite the viral load decline. Of importance, no significant renal dysfunction was noted, unlike its nucleotide predecessor adefovir. Grade 3 and 4 adverse events occurred in 24-29% of subjects but were generally not treatment-limiting. The same was true for grade 3 and 4 laboratory events which occurred in 39-45% of subjects. [abstract 692]  As a treatment intensification, tenofovir does not appear to be very effective in this heavily nucleoside experienced population. There is a virologic effect noted at the highest dose, however no corresponding increase in CD4 count was observed. It will be of great interest to see how this drug works in other clinical populations. This study at least has very carefully put to rest the issue of renal toxicity for the class of nucleotides.