How to Start Antiretroviral Therapy

What is the ideal initial therapy for an antiretroviral naïve patient? Because patients may not be as likely to experience reductions in viral loads below quantifiable levels after experiencing treatment failure with their initial combination, choosing the right first combination for a patient may be the most important treatment decision that a patient and his or her practitioner make. Several studies discussed on the first day of the conference were designed to address this question. Three of the following studies compared drug combinations in head-to-head comparisons that allow us to fully evaluate the potential of a combination and compare its successes and limitations with other gold standard options. The fourth study I'll mention compared a three-drug combination to a known two drug combination with inferior activity, but has given us a glimpse into another active combination that may be an option for initial therapy.

Two trials, START I and II, with a similar design are seeking to establish whether the initial combination of reverse transcriptase inhibitiors is important in achieving durable suppression of viral replication. Patients in these studies were antiretroviral naïve, and had pre-treatment CD4+ counts above 200 cells/mm3 and viral loads above 10,000 copies HIV-1 RNA/mL. 200 subjects were enrolled into each study. In the two presentations at this meeting the investigators report on the first 100 patients to be enrolled, with a focus on the proportion of patients with viral loads less than 500 copies/mL (bDNA) and less than 50 copies/mL (National Genetics Institute assay) at study weeks 12 and 24.

Poster Session 12223: An open-label, randomized, comparative study of d4T + 3TC + IDV versus ZDV + 3TC + IDV in treatment naïve HIV-infected patients (START I)

In the START I trial patients were randomly assigned to receive open label treatment with either:

  • stavudine (d4T) + lamivudine (3TC) + indinavir (IDV) or
  • zidovudine (ZDV) + 3TC + IDV.

Baseline viral loads were approximately 46,000 copies/mL in the group receiving d4T + 3TC, and 27,000 in those on ZDV + 3TC. With respect to tolerability, 7 of 49 (14%) patients on d4T + 3TC and 11/51 (22%) on ZDV + 3TC left the study prior to week 24. Using the ultrasensitive viral load assay (limit of detection of 50 copies/mL) to quantitate viral load changes at week 24, patients on d4T + 3TC + IDV had viral load reductions of -2.9 log, while those on ZDV + 3TC had viral load reductions of -2.6 log. After 24 weeks of therapy viral loads were < 500 copies/mL in 34/39 (87%) of patients on d4T + 3TC + IDV and 32/40 (80%) of patients on ZDV + 3TC + IDV; and viral loads were < 50 copies/mL in 19/39 (49%) on the d4T containing limb and 24/40 (60%) on the ZDV containing limb. CD4+ count increases of 150 cells/mm3 were seen in both treatment groups.

Poster Session 12225: An open-label, randomized, comparative study of d4T + ddI + IDV versus ZDV + 3TC + IDV in treatment naïve HIV-infected patients (START II)

In the START II trial patients were assigned to open label:

  • d4T + didanosine (ddI) + IDV or
  • ZDV + 3TC + IDV.

Baseline viral loads were approximately 28,000 copies/mL in patients receiving d4T + ddI and 37,000 copies/mL in patients on ZDV + 3TC. Although ddI and IDV can be difficult to take together because of the requirement to take both drugs without food and separate them over time, 11/50 (22%) of patients left the study prior to week 24 among those assigned to d4T + ddI + IDV, while 12/50 (24%) on ZDV + 3TC + IDV left prior to week 24. After 24 weeks, 23/34 (68%) of patients on the d4T + ddI combination had viral load < 500 copies/mL and 12/34 (35%) had viral loads < 50 copies/mL, compared to 23/30 (77%) and 15/30 (50%) with viral loads < 500 and 50 copies/mL on the ZDV + 3TC limb. Again, CD4+ cell increases were similar between the two groups.

What can we conclude from these two trials? These studies focus on the three most commonly prescribed combinations of reverse transcriptase inhibitors: ZDV/3TC, d4T/3TC, and d4T/ddI together with IDV, the most commonly prescribed protease inhibitor. The combinations appear to be well tolerated, with fewer patients on d4T + 3TC dropping out of the study early compared to AZT + 3TC. Despite being a difficult regimen from an adherence perspective, patients on d4T + ddI were no more likely to develop an adverse experience as ZDV + 3TC recipients. No combination stands out as being the most active. It must be acknowledged that these reports are preliminary in that the experiences of only half of the enrolled patients are being reported. It may take longer than 24 weeks for a patient's viral load to fall below 50 copies/mL, so the proportions of patients with viral loads below this threshold may increase with more prolonged observation. If so, one of the combinations may be shown to be more active than the others. However, we should also view these results with some concern. Studies have suggested that unless viral loads are reduced below those levels that can be measured by the ultrasensitive assays (< 20 or 50 copies/mL) patients on protease inhibitors are likely to eventually experience a rebound in their viral load with resistant virus. Therefore, if only half of patients have viral loads < 50 copies/mL after six months of therapy, a significant proportion of patients may not be experiencing active enough therapy to maintain durable suppression of viral replication.

Poster Session 22336: A phase II, multicenter, randomized, open-label study to compare the antiretroviral activity and tolerability of efavirenz (EFV) + indinavir (IDV), versus EFV + zidovudine (ZDV) + lamivudine (3TC), versus IDV + ZDV + 3TC at 24 weeks [DMP 266-06]

The big antiretroviral story of day 1 at the Geneva Meeting was the first report of the pivotal efavirenz trial (DMP 266-06). Many practitioners have experience with this agent as part of salvage therapy, but little data in naïve patients has been presented. Today we had our first glimpse of the activity of efavirenz as part of a 2- and 3-drug combination therapy in comparison to a standard 3-drug regimen. This study put an NNRTI, efavirenz (EFV), toe-to-toe with indinavir in an attempt to clearly establish a 3-drug combination including an NNRTI as a combination with comparable activity to 3-drugs including a protease inhibitor.

The study included 450 patients who were naïve to 3TC, NNRTIs, and protease inhibitors, and had viral loads > 10,000 copies/mL and CD4+ counts > 50 cells/mm3. Pretreatment viral loads were about 70,000 copies/mL for the entire group and CD4+ counts were 345 cells/mm3. Patients were randomized to one of three open-label treatment arms:

  • ZDV + 3TC + EFV 600 mg qd
  • EFV 600 mg qd + IDV 1000 mg q8H or
  • ZDV + 3TC + IDV 800 mg q8H

Indinavir was dosed as 800 mg q8H as part of the 3-drug combinations, but was given as 1000 mg q8H when used with EFV as part of the 2-drug combination because EFV decreases serum concentrations of IDV. Data up to study week 24 were presented.

Viral load reductions of -2 logs or more were seen with each group. Reductions of viral load below 400 and 50 copies/mL were analyzed by two intent-to-treat analyses and by an on-treatment analysis (counting only those patients still being followed on assigned medication). In the most stringent intent-to-treat approach, patients who dropped out of the study for any reason were considered as treatment failures (i.e., viral loads not less than 400 or 50 copies/mL). In the second intent-to-treat analysis, the last observation of the patient was carried forward (LOCF) and considered as the data point for week 24 analysis.

At week 24, 95% of patients on ZDV + 3TC + EFV, 87% of EFV + IDV recipients, and 89% of ZDV + 3TC + IDV recipients had a viral load < 400 copies in the on-treatment analysis. Using the most stringent criteria (drop-out = failure), 75% of patients assigned ZDV + 3TC + EFV had viral load < 400 copies/mL, compared to 63% of those assigned to EFV + IDV, and 56% of those assigned to ZDV + 3TC + IDV. The differences between ZDV + 3TC + EFV and ZDV + 3TC + IDV were statistically significant (p < 0.05).

Some of the differences seen with the intent-to-treat analysis can be explained by the higher drop-out rate among patients on indinavir. Adverse events leading to treatment discontinuation occurred in 26/148 (18%) of patients on ZDV + 3TC + IDV, compared to 7/148 (5%) patients on EFV + IDV, an 10/154 (6%) patients on ZDV + 3TC + EFV. Commonly reported grade 2 (moderate) or more severe adverse events included nausea and vomiting which were most common in patients on ZDV + 3TC + IDV, rash occuring in approximately 12% of EFV recipients, and the central nervous system symptoms of headache and dizziness reported in about 5% and 7% of EFV recipients, respectively.

The most impressive testimony to efavirenz's activity was presented in a subanalyis of 117 patients with baseline viral loads > 100,000 copies/mL. In an on-treatment analysis, a remarkable 90% of these patients had viral loads < 400 copies/mL after 24 weeks of therapy, compared to 78% of patients on the 2-drug EFV + IDV combination, and 73% of ZDV + 3TC + IDV recipients.

This study creates a clear protease-sparing option for initial therapy. The results are still somewhat limited by the short duration of follow-up. Clearly, durable suppression of viral replication is the goal of therapy and a more prolonged follow-up period may alter the outcomes to some extent. Studies with nevirapine and delavirdine as part of triple combinations demonstrated that patients were most likely to have viral loads below the limit of quantification after 24 weeks of therapy, with a higher failure rate over developing over time. Fewer mutations are required for virus to become resistant to the NNRTIs compared to the protease inhibitors, so more failures over the next 24 weeks may be seen in the EFV arms. However, many studies have shown that the surest way to prevent resistance is to reduce viral loads to the lowest possible levels, and efavirenz clearly passes that test. A possible point of contention for study-design purists is that the study was not placebo controlled. This may have biased the reporting of adverse experiences because patients may have been familiar with the possible toxicities associated with each agent.

Session 127/12230: Safety and Activity of Abacavir (ABC, 1592) with 3TC/ZDV in Antiretroviral Naïve Subjects

Another option for a protease sparing regimen was studied in a placebo-controlled study comparing three nucleoside reverse transcriptase inhibitors, ZDV + 3TC + abacavir (ABC) 300 mg bid versus ZDV + 3TC in antiretroviral naïve patients with CD4+ counts > 100 cells/mm3. There was no viral load entry criteria. Median baseline viral loads were approximately 40,000 - 50,000 copies/mL for both groups and median CD4+ counts were 420 - 470 cells/mm3.

The analysis was conducted up until week 16, at which time patients without viral loads < 400 copies/mL could take open label ABC together with the nucleoside reverse transcriptase inhibitors of the investigator's choice. In this way, patients on ZDV + 3TC would be assured of the opportunity of receiving a 3-drug combination if there were any evidence of viral load rebound by week 16 or beyond.

The study wasn't a fair comparison -- comparing a 3-drug combination with a standard 2-drug regimen is not really a contest, and fails to put a combination into a meaningful context. In the on-treatment analysis, 47/70 (67%) of subjects on the 3-drugs had viral loads < 50 copies/mL, compared to 13/68 subjects on ZDV + 3TC. In the intent-to-treat analysis, 47/87 (54%) of patients on ABC had viral loads < 50 copies/mL, compared to 13/86 (15%) on the placebo.

Achieving viral loads less than quantifiable on each arm was related to the patients' baseline viral load. In an intent-to-treat analysis focusing on the patients who received the three drugs, 62% of patients with baseline viral loads < 10,000 copies/mL had their viral loads go below quantifiable levels, compared to 58% for patients with baseline viral loads of 10,000 to 100,000 copies, and 33% with baseline viral loads > 100,000.

Adverse experiences in the patients on triple therapy included headache in 31%, malaise or fatigue in 34%, nausea in 47%, diarrhea in 12%, nausea and vomiting in 16%, and insomnia and other sleep disorders in 7%. Most of these toxicities were not severe in nature, and did not result in patients discontinuing treatment. The abacavir hypersensitivity reaction characterized by nausea, fever and other constitutional symptoms with or without a rash was seen in 2 patients. This is a reaction that all clinicians and patients should be aware of, because abacavir must not be continued if a patient develops this syndrome, and the deaths have occurred when patients have been rechallenged with the drug after its cessation.

We tend not to think of a three nucleoside combination when selecting initial therapy. However, abacavir, as the only guanosine analog reverse transcriptase inhibitor, appears to represent another option together with efavienz or a protease inhibitor as part of initial therapy. Thus far, we have only seen data on abacavir combined with ZDV + 3TC. We need to see data showing the effects of abacavir with the other reverse transcriptase inhibitors in order to know how best to use this drug.

The debate over the proper initial therapy will continue for a long time. Does convergent therapy (all agents within the same class) or divergent therapy (agents from several classes) provide more durable viral load suppression and maintain options for second line treatment in patients who experience viral load rebounds with their initial therapy? What is the proper way to sequence the nucleoside reverse transcriptase inhibitors and protease inhibitors to maintain future options? Insights into some of these questions will be addressed in tomorrow's session.