Differentiating within and between drug classes: sometimes it’s toxicity, sometimes its efficacy

Dr Graeme Moyle MD, MBBS, Chelsea & Westminster Hospital, London for NATAP

The new DHHS guidelines include additions to the list of strongly recommended agents and nucleoside combinations but nothing has been removed from the list since the last major revision.

The additions include 3 boosted PI regimens, low dose ritonavir (RTV) with either saquinavir, indinavir or lopinavir (combined tablets Kaletra). However, 2 unboosted PIs, nelfinavir and indinavir remain on the strongly recommended list despite inferior antiviral efficacy being observed in randomised clinical trials; nelfinavir being inferior to Kaletra, indinavir to efavirenz. Further efforts to compare the combinations most widely used were reported during CROI.

Are the nucleoside combinations different as backbones?

Studies form the dual therapy era suggested that, at least for the most widely used (and DHHS recommended) regimens of ZDV or d4T plus either ddI or 3TC the answer was no difference. Thus the choice between these backbones would be made on issues such as impact of future treatment options via resistance, frequency and severity of adverse effects, impact on body shape and convenience of administration with the third agent in the regimen. Evidence to clearly inform these debates are limited, but each approach has its pros and cons.

Data regarding body shape, for example, vary substantially between studies. In patients who have only received ZDV or d4T backbones, one study (consistent with previous data) found similar rates of fat loss and fat accumulation between these drugs [1]. Whereas results from a subset of patients drawn from a randomised study of d4T vs. ZDV based regimens (in combination with 3TC + indinavir) in individuals with substantial prior ZDV experience found similar efficacy and tolerability. However, in a cross sectional survey of 96 of the 170 randomised subjects, higher rates of lipoatrophy were found in those treated with d4T [2].

The reasons for the differences between studies remains unclear, and could relate to methodological issues, but raise the question as to whether prior ZDV use predisposes d4T recipients to lipoatrophy. Efficacy comparison between ZDV+3TC (n = 206) and d4T + 3TC (n = 101) or ddI chewable tablets (n = 102) (all with indinavir) were provided from the na•ve patients in the START study [3]. Overall the median baseline viral load (VL) and CD4 counts were 4.53 log10copies/mL (34,000 copies/ml) and 407 cells/mm3, and were similar between groups. At week 72, the proportions of available patients remaining on initial therapy with HIV VL <500 copies/mL were for ZDV/3TC 72% (48/67), d4T/ddI 69% (24/35), and d4T/3TC 80% (35/44) with no significant difference between groups (p > 0.20 for all comparisons). This suggests that these backbone nuke combinations each provide similar antiviral efficacy.

The main drawback to ddI use has been the mostly mild but relatively common gastrointestinal (GI) side effects observed with the chewable tablets and the need for separation from fat and protein containing meals. The recently approved EC formulation, which is thought to avoid these GI effects, was compared in a randomised trial of 138 treatment-naive patients to the older chewable tablet, each once per day, in combination with d4T and nelfinavir [4].

Median baseline VL levels (4.73, 4.59 log10 copies/mL, 54,000 vs 39,000) and CD4 counts (382, 381) cells/mm3 in the EC and tablet arms respectively were similar between groups. Mean HIV RNA levels over 48 weeks showed similar profiles for both regimens, with a slightly larger decrease in EC recipients (-2.62 and -2.35 log10c/mL for EC and tablet, respectively). CD4 increases were about the same in both groups. The tolerability data indicated that significantly more subjects discontinued for adverse events from the tablet regimen (20%) compared with the EC regimen (7%; p = 0.04), suggesting better tolerability the new regimen provides better tolerability. Previously reported data with efavirenz, nevirapine, indinavir and RTV boosted saquinavir all indicate that ddI EC does not need to be separated from the dosing of these agents.

In a separate trial ddI EC plus d4T was compared with Combivir (ZDV+3TC combined tablets) in an open-label, randomised study [5]. Data from 333 of 511 treatment-naive subjects with therapy through 48 weeks were analysed. Similar proportions of subjects on both regimens had VL <400 copies/mL at week 48 (ddI EC/d4T 57%, Combivir 55%) and <50 copies/mL (difference 0.2%). Median CD4 counts also increased similarly in both regimens. As these two combinations appear similar in antiviral efficacy the choice for each individual will be based on adverse events rates. Between the ddI EC/d4T and Combivir arms rates of all grades of adverse events reported were diarrhoea 54% and 56% (presumably mostly secondary to nelfinavir), nausea 21% and 35%, vomiting 13% and 18%, and peripheral neurological symptoms (all grade I-II) in 17% and 8%.

Discontinuations for adverse events were similar in number, occurring early in Combivir and later in the ddI EC/arm. Lipodystrophy events were not reported. These data suggest the choice between these combinations will be one of risk of mild to moderate peripheral neurological symptoms but better initial tolerability with ddI EC/d4T versus more nausea and vomiting with Combivir.

Once daily regimens

The number of potential or approved once daily regimens is gradually increasing making the arrival of routine once daily therapy increasing a reality. Data on FTC and tenofovir, new once daily NRTIs are included in other reports from this conference. Both efavirenz and ddI are approved for once daily use and interest is growing in using 3TC and abacavir as once daily drugs. A 24-week comparison confirmed the feasibility of once daily 3TC [6]. This prospective, randomised, trial compared the efficacy and tolerability of a switch to 3TC 300 mg QD vs. continued standard dosing of 3TC 150 mg BID in subjects with virologic suppression <400 copies/mL for >3 months) on >6 months 3TC, d4T and indinavir or nelfinavir. Of 81 dosed patients, 78 subjects completed the study. Not surprisingly, a high rate of virologic suppression (<400 copies/mL) as sustained through 24 weeks by both regimens (once daily 95%, BID 90% by ITT analysis). For <50 copies/mL 82% of subjects on QD regimen had VL and 81% on BID regimen remained undetectable. No virologic failures (for some obscure reason defined as >1265 copies/mL on 2 separate occasions >2-4 weeks apart) occurred and CD4 increases were similar for both dosing regimens. Both dosing regimens were well tolerated with no drug-related serious adverse events were reported.

Given these and other data it is not surprising that data from two non-comparative evaluations of once daily ddI + efavirenz with 3TC [7] and FTC [8] reported impressive levels of viral suppression and tolerability.

Efficacy differences between NNRTIs?

Concerns regarding liver toxicity with nevirapine had been raised from the FTC studies (see my previous NATAP report from this conference). A further presentation at the conference raised concern that this agent may also have lower efficacy relative to efavirenz. This report was a retrospective analysis of 1932 patients given their first NNRTI regimens containing either nevirapine (n=1202) or efavirenz (n=730), derived from the large EuroSIDA cohort [9].

EuroSIDA is a pan-European, clinic-based project that collects data from 64 clinics. Patients who began a regimen with nevirapine or efavirenz after July 1997 were studied. Virologic failure was defined as 2 consecutive VL >500 copies/mL at least 6 months after starting the regimen. Baseline characteristics were similar with a viral load of 3.7 vs. 3.9 log10 copies/mL, for EFV and NVP groups, respectively), and baseline CD4+ cell count (269 vs. 266 cells/mm3). Prior AIDS diagnosis was slightly more common in EFV recipients 41% relative to 34%of NVP recipients. Only 5% of patients were NRTI-naive, and over 75% had received at least 3 NRTIs, while almost half of the patients had received at least 2 protease inhibitors (PIs).

Factors associated with clinical or virologic failure included number of previous NRTIs or PIs, CD4+ cell count nadir or previous AIDS diagnosis, baseline viral load and number of NRTIs in the regimen. Virologic rebound at 12 months were 48% and 65% for EFV and NVP, respectively. The chance of virological rebound or clinical progression was significantly lower with EFV regimens. The relative hazard (adjusted for baseline characteristics) of viral rebound at 12 months was 0.57 (CI 0.47-0.69, p=0.001) representing a 43% lower chance of this event relative to NVP. For clinical progression the relative hazard was 0.49 (CI 0.33-0.74, p=0 .0005), a reduction of 51% of progression. A randomised comparison of these drugs in treatment na•ve patients is currently recruiting and may answer the issue as to relative efficacy in these circumstances. A direct comparison in NRTI, PI experience patients, based on these data, is urgently warranted.

NNRTI versus PI in initial therapy

The DPC-006 study established efavirenz as a first-line NNRTI option demonstrating superiority of 8-hourly dosed indinavir both in combination with ZDV and 3TC. Longer-term (144-week) follow-up data indicated that these advantages over indinavir, and viral suppression, are sustained. Only 8% of patients randomised to EFV failed to achieve viral suppression to <50 copies/mL in the study with the rebound rate being 8% by 52 weeks, with a further 7% between weeks 53 and 96. Incomplete suppression rates of approximately 16% and rebound rates of 16% by week 52 and a further 6% between 52 weeks and 96 weeks occurred in the IDV arm.

After 96 weeks with fewer patients available for analysis, rebound occurred in 3.5% of EFV and IDV patients. Using a mathematical model based on rebound rates in years 1 and 2 it was estimated that the median time to virologic rebound with the EFV based regimen will exceed 6 years [10]. If the lower rate of rebound in years 3 were considered, it is possible this estimate of durability may have been longer.

Data from the COMBINE Study comparing NVP with nelfinavir (both dosed BID with Combivir) in a randomised, open label study of 142 naive patients reported 36-week data at the conference [11]. One difficulty with interpretation of the results of this study lies in modest baseline differences, which occurred by chance at randomisation, but may have favoured the NVP arm. About 40% in each arm were HIV-infected by IVDU. This may have affected adherence, as adherence to NFV was less than to NVP. It’s also possible that methadone or illegal drug use may have affected blood levels of HIV drugs. As well, the percentages of heterosexuals were 31% NFV vs 20% NVP, and homosexuals were 18% NFV vs 37% NVP.

Data on the poster on a selected subgroup of 15 patients on NFV who had <20 copies/ml showed good blood levels. Indeed, 10 patients did not return for follow-up after randomisation, 8 in NFV and 2 in NVP arms, this difference affecting the ITT analysis in the NFV arm by >10% but by <5% in the NVP arm. At 1 month into study 95% adherence was 59% for NFV vs 75% for NVP (p=0.03). Other baseline characteristics included a mean CD4 count of 359 (range 10-908) cells/mmm3 and median viral load of 4.78 (range3.2-6.2) log10 Ñabout equal in both arms.

In the 36-week ITT analysis, 55.7% of the NFV versus 70.8% of the NVP recipients had VL <200 copies/mL (p=0.06). The on-treatment analysis showed 78% in NFV arm versus 83.7% in the NVP arm were <200 copies/mL (p=0.50). Using a <20 copies/mL assay, 38.6% in the NFV arm versus 66.7% in the NVP arm achieved this level of suppression in an ITT analysis (p<0.001). The OT analysis to <20 copies/mL showed 56.1% in the NFV arm versus 79.6% in the NVP arm (p=0.02). There were significant numbers of discontinuations due to side effects in both arms, 13/70 in NFV arm and 16/72 in NVP arm including several discontinuations on NVP due to hepatotoxicity (5/72), and due to NFV GI symptoms (10/70). Given the problems with these data it is difficult to say this study represents the final word in comparison between these two drugs. However, as both NVP and NFV are increasingly less common treatment choices amongst HIV clinicians, the definitive study is unlikely to be done. In the ITT analysis of patients with >100,000 viral load at baseline, 15.4% (4/26) taking NFV vs 61.9% (13/21) taking NVP had <20 copies/ml (p=0.001). 53% (14/26) for NFV vs 71% (15/21) had <200 copies/ml but this was not statistically significant (p=0.22).

Booster PIs outshine standard dosing: Kaletra

The combination of ritonavir 100mg bid with other PIs is increasingly viewed as the best way to use these agents, reducing pill load, boosting trough values albeit sometimes at the cost of (mostly) modest lipid increases. The co-formulation of lopinavir and ritonavir is the most recent example of this approach. Unlike boosted saquinavir and indinavir, this combination has the drawback of requiring dosing with food. The key trial for the approval of lopinavir/ritonavir (Kaletra) is a double blind, randomised trial of twice-daily stavudine/lamivudine plus either twice-daily lopinavir/ritonavir or 3-times-daily nelfinavir (changed to twice daily after 24 weeks) in patients essentially naive to antiretroviral therapy [12].

The study involved 653 patients showed superiority for the Kaletra arm over NFV, which became most evident at week 36 and 48 time-points. At 48 weeks, VL was <400 copies/mL in 75% of patients in the lopinavir/ritonavir arm and 63% of those in the NFV arm (P < .001) and <50 copies/mL in 67% and 52% of these arms, respectively (P < .001). The analysis presented at this conference looked at the time to treatment success defined as <50 copies/mL. The drawback of this analysis was that the <50 copy assay was not used on samples until week 24. At this time-point approximately similar proportions, 65% vs. 60% (p= not significant) of patients had VL <50 copies/mL. Amongst week 24 incomplete responders (>50 copies/ml at week 24) who remained on study, however, 88% of lopinavir/ritonavir-treated patients subsequently responded whereas only 41% of NFV patients subsequently achieved <50 copies/mL (P < .001).

Consistent with a similar analysis from the DPC-006 study with EFV, time to VL <50 copies/mL took longer in patients with baseline VL above100,000 copies/mL. Differences in the performance of the two drugs was most evident at this levels with 84% of patients randomised to lopinavir/ritonavir, compared with 60% of those on NFV eventually responding. These data also show that some individuals take longer than 24 weeks to reach <50 copies/ml. In an analysis using data from both treatment groups, patients with baseline <100,000 copies/ml were compared to those with >100,000 copies/ml. Patients with lower baseline viral load achieved <50 copies/ml more quickly than those with higher viral loads (p<0.001), regardless of whether patient received NFV or Kaletra.

However, there was a difference between NFV & Kaletra in terms of getting below 50 copies/ml if baseline viral load was >100,000. In the Kaletra group, similar proportions of patients with VL >100,000 copies/ml at baseline (84%) or >100,000 (85%) reached <50 copies/ml. In the NFV group, a significantly lower proportion of patients with VL >100,000 copies/ml (60%) than those with <100,000 copies/ml reached <50 copies/ml. People with higher baseline viral load can take longer to reach <50 copies/ml. Within each treatment group this was reflected. Ultimately, 85% for Kaletra vs 71% for NFV ever achieved <50 copies/ml, and a difference between the two arms was not statistically seen until week 24. Abbott suggested week 48 as a more appropriate time-point to evaluate I reaching <50 copies/ml.


  1. Rubio R, M. Torralba M, Antela A, et al. Body Shape Abnormalities in a Cohort of HIV- Infected Patients on First-Line HAART. Abs 646.
  2. Joly V, Flandre P, Meiffredy V, et al. Assessment of Lipodystrophy in Patients Previously Exposed to AZT, ddI or ddC, but Naive for d4T and Protease Inhibitors (PI), and Randomised Between d4T/3TC/ Indinavir and AZT/3TC/Indinavir (NOVAVIR Trial). Abstract 539
  3. Murphy R, Santana J, Squires K, et al. START Observational Study: Longitudinal Follow-Up of Virologic and Immunologic Responses in START I and START II Patients. Abstract 314
  4. Schrader S, Sharma S, Seekins D, et al. Comparison of HIV RNA Suppression Produced by Triple Regimens Containing either Didanosine Enteric-Coated or Didanosine Tablet Formulations Each Administered Once Daily. Abstract 318
  5. Gathe J, Badaro R, Grimwood A, et al. Comparison of a Triple Combination Regimen Containing an Enteric-Coated Formulation of Didanosine Administered Once Daily Versus a Regimen of Combivir Plus Nelfinavir. Abstract 319
  6. Sension M, Bellos N, J. Johnson J, et al. Efficacy and Safety of Switch to 3TC 300 mg QD vs. Continued 3TC 150 mg BID in Subjects with Virologic Suppression on Stable 3TC/d4T/ PI Therapy (COLA4005): Final 24- Week Results. Abstract 317.
  7. Maggiolo F, Migliorino M, Maserati R, et al. Once-a-Day Treatment for HIV Infection: Final 48-Week Results. Abstract 320.
  8. Molina JM, Perusat S, Ferchal F, et al. Once-Daily Combination Therapy with Emtricitabine, Didanosine and Efavirenz in Treatment-Naive HIV-Infected Adults: 64-Week Follow-Up of the ANRS 091 Trial. Abstract 321
  9. Phillips AN, Pradier C, Lazzarin A, et al. Virological and clinical outcome of NNRTI-containing regimens for 1932 patients in EuroSIDA. Abstract 324.
  10. Levy R, Labriola D, Ruiz N. Low two year risk of virologic failure with first regimen HAART. Abstract 325
  11. Podzamczer D, Ferrer E, Consiglio E, et al. A Randomised, Open, Multicenter Trial Comparing Combivir plus Nelfinavir or Nevirapine in HIV-Infected Naive Patients (The Combine Study). Abstract 327
  12. King M, Bernstein B, Kempf D, et al. Comparison of time to achieve HIV RNA <400 copies/mL and <50 copies/mL in a phase III, blinded randomised clinical trial of Kaletra vs. nelfinavir in ARV-naive patients. Abstract 329

Links to other websites are current at date of posting but not maintained.