The initial decline in plasma HIV-1 RNA concentration (pVL) after the start of antiretroviral therapy (ART) is considered to be a marker of early antiretroviral efficacy.1-4 A rapid decline shortens the duration and the level of ongoing viral replication in the presence of drugs and thereby supposedly limits the potential evolution of drug-resistant virus. Studies in patients starting highly active antiretroviral therapy (HAART) have shown that the level of pVL after 8 weeks,5 16 weeks,6 or 6 months7 is predictive for long-term treatment efficacy. The same was shown earlier in treatment-experienced patients.8
Whether rates of pVL decline in the initial weeks of treatment rather than pVL estimates at specific time points are also predictive of late virologic efficacy remains uncertain. This predictive characteristic would enable the clinician to adjust therapy early after the start of treatment. Such an association was reported by Pollis et al9 in a population of patients starting a combination of lamivudine (3TC), zidovudine (ZDV), indinavir (IDV), and nevirapine (NVP). Similarly, it was reported for patients on nelfinavir monotherapy.10 In contrast, Wu et al11 did not find a significant association between the viral decay rate and late virologic response in patients treated with abacavir (ABC) and 1 of 5 different protease inhibitors (PIs). The same was reported for patients treated with ritonavir (RTV) for 10 days and additional 3TC and ZDV thereafter.12
The 2NN study is the first large, randomized, controlled trial to assess the efficacy and safety of the nonnucleoside reverse transcriptase inhibitors (NNRTIs) NVP and efavirenz (EFV) in addition to stavudine (d4T) and 3TC.13 There were 2 reasons to assess the rate of pVL decline as a marker for early efficacy. The first reason was to establish that patients who were randomized to the NVP + EFV treatment group did not receive a suboptimal drug regimen. There were some data that the 2 drugs might have an antagonistic antiviral effect.14 We conducted an interim analysis based on the data from the first 400 patients to assess possible differences in pVL decline between the treatment groups. The Data Safety and Monitoring Board did not see a reason to stop randomization to the NVP + EFV treatment group based on the results of this interim analysis. The second reason was to characterize further the antiviral potency profile of NVP and EFV. This potency is most clearly assessed by analyzing the direct effects of the drugs on the pVL at the start of ART.
Participants and Treatment Allocation
The 2NN study has been described elsewhere.13 In brief, patients were adults and ART naive and had a pVL of at least 5000 copies/mL. All patients received d4T (40 mg twice daily or 30 mg twice daily when they weighed <60 kg) and 3TC (150 mg twice daily) and were randomly allocated to 400 mg of NVP once daily, 200 mg of NVP twice daily, 600 mg of EFV once daily, or 400 mg of NVP plus 800 mg of EFV once daily. Patients were included from 65 different study sites in 17 countries in Asia, Australia, North America, South America, South Africa, and Europe. The present study included patients who remained on their allocated treatment (all components of the regimen used at least 95% of the time) during the first 2 weeks of treatment and had at least 2 pVL measurements during this period. Patients could change d4T and/or 3TC for reasons of toxicity as long as 2 nucleoside analogue reverse transcriptase inhibitors were used. The NVP once-daily and twice-daily groups were combined for this analysis, because the dose taken (200 mg once daily) in the first 2 weeks was the same as a result of the recommended dose escalation when starting NVP.
The pVL was measured at days 0, 3, 7, and 14 at a central laboratory (LabCorp; Research Triangle Park, NC) using the Ultra Sensitive Roche Amplicor 1.5 assay (Roche Diagnostics, NJ) with a lower limit of quantification (LLQ) of 50 copies/mL. Samples up to and including the first sample with a pVL below the LLQ were included in the analyses. This first sample below the LLQ was regarded as 50 copies/mL. Plasma concentrations of NVP and EFV were quantitatively assessed at days 3, 7, and 14 by a validated high-performance liquid chromatography with ultraviolet detection method.15 Bayesian trough concentrations (Cmin) were calculated using random plasma concentration time points and nonlinear mixed-effect modeling (NONMEM).16 It turned out that for NVP and EFV, the clearance rate for the samples taken between days 3 and 14 did not differ. This means that all samples would give comparable trough concentrations based on time between drug intake and blood draw as well as these identical clearance rates. We choose to use the samples of day 7 to estimate the Cmin and use these estimates in the current study.
The primary outcome was the viral decay constant (VDc) during the first 2 weeks of treatment using the nonlinear function V(t) = V(0) · e(−kt), where V(0) and V(t) are the pVL at baseline and time t, respectively, and k is the VDc. The VDc denotes the log10 copies of virus cleared per day. It was calculated separately for weeks 1 and 2 and for both weeks combined.
Factors assessed for an association with a high VDc (>75th percentile) were age (per year increase), sex, region (Asia/Australia, South Africa, South America, and Europe/United States/Canada), baseline pVL (≤100,000 vs. >100,000 copies/mL), baseline CD4 count (<50 vs. 50-200 vs. >200 cells/mm3), and a Centers for Disease Control and Prevention (CDC) event of category C17 at baseline (present vs. absent). Finally, we examined the association of the VDc with the Cmin at day 7 in the NVP and EFV treatment groups and with the occurrence of virologic failure on or before week 48 (never reaching a pVL <50 copies/mL or a rebound to 2 consecutive pVLs >50 copies/mL).
The VDcs of the different treatment groups were compared using the Kruskal-Wallis test. The correlation between the VDc in weeks 1 and 2 was calculated as the Spearman correlation coefficient. Factors associated with a high VDc were assessed by univariable and multivariable logistic regression analysis. The multivariable model included the variables “treatment group,” all variables associated with the outcome with a P value <0.1 in the univariable model, and possible interaction terms between treatment group and other variables. The relation between VDc and Cmin was assessed by linear regression, and that between VDc and virologic failure was assessed by Cox regression analysis, where patients were censored at change of allocated treatment. A 2-sided P value <0.05 was considered statistically significant in the final analyses. SAS statistical software (version 8.02; SAS Institute, Cary, NC) was used for all analyses.
Of the 1216 patients enrolled in the 2NN study, 1111 (91%) met the inclusion criteria for the present study. The baseline characteristics of these patients did not differ between the treatment groups and were comparable with the characteristics of the patients in the main study (Table 1).
The median VDc (log10 copies per day; interquartile range [IQR]) in the first week was 0.47 (0.37-0.59) for NVP, 0.51 (0.43-0.60) for EFV, and 0.49 (0.39-0.59) for NVP + EFV (P = 0.002). In the second week, the clearance of virus occurred at a lower rate of 0.15 (0.09-0.23) for NVP, 0.13 (0.07-0.21) for EFV, and 0.15 (0.10-0.21) for NVP + EFV (P = 0.036). A high VDc in week 1 was correlated with a low VDc in week 2 (Spearman coefficient: −0.22; P < 0.001). This indicates that the VDc in week 2 was dependent on the amount of virus cleared in week 1. Therefore, a separate VDc for week 2 is a rather arbitrary estimate. We combined these 2 time periods and calculated the VDc between weeks 0 and 2. This new estimate (IQR) was 0.30 (0.25-0.36) for NVP, 0.31 (0.27-0.37) for EFV, and 0.30 (0.27-0.36) for NVP + EFV (P = 0.109).
Factors in the multivariable analyses that were independently associated with a high VDc (>75th percentile) are summarized in Table 2. Patients with a baseline pVL >100,000 copies/mL were 8.7 times (95% confidence interval [CI]: 6.2-12.3; P < 0.001) more likely to have a high VDc compared with patients with a baseline pVL ≤100,000 copies/mL. Compared with patients in Europe/United States/Canada, patients in Asia/Australia had a slightly higher chance of having a high VDc (odds ratio [OR] = 1.3, 95% CI: 0.8-2.0), whereas patients from South Africa had a significantly lower chance (OR = 0.7, 95% CI: 0.4-1.0). Within each region however, there were no differences in effect between the 3 treatment groups (interaction group and region: P = 0.539). Sex, age, baseline CD4 count, and disease stage were not independently associated with VDc; neither was there evidence of interaction between any of these variables and treatment group. After adjusting the VDc in the first 2 weeks of treatment for baseline pVL and region, the estimates (IQR) were identical in all 3 treatment groups, being 0.30 (0.28-0.34) for NVP, 0.30 (0.29-0.35) for EFV, and 0.30 (0.29-0.34) for NVP + EFV.
The relation between high VDc (>75th percentile) and virologic failure was assessed by Cox proportional hazard analyses. The proportional hazard assumption was tested and found valid. It was shown earlier in the 2NN trial that a high baseline pVL was associated with virologic failure on or before week 48.13 We therefore included this parameter in the Cox proportional hazard analysis. There was no statistical evidence that patients with a high VDc were less likely to have virologic failure while on treatment compared with patients with a lower VDc (hazard ratio [HR] = 0.8, 95% CI: 0.6-1.2; P = 0.253).
The median (IQR) Cmin at day 7 was 2.8 (2.2-3.6) mg/L for NVP and 1.3 (0.9-2.1) mg/L for EFV. All patients had a Cmin at day 7 above the 95% coefficient inhibition (IC95) (710 nM for NVP, 25nM for EFV). There was no association between the Cmin of NVP or EFV with the VDc (regression coefficient = 0.004; P = 0.230 and 0.006, P = 0.142, respectively (Fig. 2). The Cmin of NVP and EFV differed significantly between regions, with patients from Asia/Australia having the highest concentrations and patients from Europe/United States/Canada having the lowest (Table 3). Repeating the linear regression analysis adjusted for region did not change the results (regression coefficient = 0.001; P = 0.718 and 0.006, P = 0.135 for NVP and EFV, respectively).
Although region was significantly associated with the VDc and the Cmin for NVP and EFV on day 7, the pattern of association clearly differed. Patients from Asia/Australia had the largest probability of having a high VDc and showed the highest Cmin for NVP and EFV. Patients from South Africa likewise had a higher Cmin for NVP and EFV compared with patients from Europe/United States/Canada, but the probability of a high VDc was significantly lower. This illustrates that it is not possible to interpret the VDc within a specific region as a simple function of plasma drug concentrations.
The present study showed a similar rate of initial decline of pVL for NVP, EFV, and NVP + EFV after adjusting for baseline pVL and region. The VDc was not predictive for longer term virologic efficacy and was not associated with NNRTI plasma concentrations. Combining NVP and EFV does not have an additional effect on early efficacy. The results of the present study are consistent with the overall findings of the 2NN study.13 These showed no statistically significant differences between NVP and EFV in the proportion of patients with virologic failure or the proportion of patients with a pVL <50 copies/mL at week 48.
The evidence for a relation between the initial pVL decline and longer term treatment success is inconclusive.9-12 The absence of such a relation in the present study might be related to the potency of the treatment regimens used. This potency can be implied from the strong relation between baseline pVL and VDc. The more HIV-1 RNA is present at the start of treatment, the larger is the VDc in the first 2 weeks of treatment. Another reason might be that long-term efficacy is not only based on the intrinsic potency of the antiretroviral drugs used but relies heavily on treatment adherence by the patients.
The initial decline in pVL after the start of ART is a result of the elimination of free virus particles and the decline in productively infected CD4+ T lymphocytes (“first phase”), followed by a much slower decline representing the loss of long-lived, infected, CD4+ T lymphocytes (“second phase”).18-20 Because of the absence of information on the turnover of the T-lymphocyte pool in the present study, we interpreted the observed decline in pVL as a compound effect of both mechanisms over a 2-week period, as described earlier.21 Reported viral decay rates vary widely between studies as a result of analytic differences (eg, sampling frequency, formula used) and differences in populations studied.12,22 Cross-study comparisons of the viral decay rate are therefore not useful.
We used a formula for estimating the VDc that disregards the turnover of the T lymphocytes and assumes that ART completely blocks the production of new HIV-1 virions. The resulting estimate therefore represents only a lower boundary of the VDc.2 Because the main aim of the study was to compare treatment regimens within the same study, this assumption is unlikely to affect the relative effectiveness of the treatment regimens.
The positive association between baseline pVL and VDc reported in the present study has been described previously by Notermans et al18 Others have described a negative correlation between baseline pVL and the first-phase VDc and a positive correlation with the second-phase VDc.11,12 It is unclear what the duration of the phases were in these studies, and it is therefore impossible to say whether the present relation between baseline pVL and VDc is in concordance with these findings. More importantly, the effect of a high baseline pVL on the VDc was similar in all 3 treatment groups. This is contrary to the general perception that the early efficacy of NVP is inferior to that of EFV in patients who start treatment with a high baseline pVL. This perception arose from observations in clinical trials that the difference in efficacy between patients with a high or low baseline pVL was larger in NVP trials compared with EFV trials.
An estimated 1% of the patients in the 2NN study had drug-resistant virus at baseline.23 The presence of genotypic resistance for the study drugs used did not affect the VDc, which was 0.29 (95% CI: 0.23-0.35) for patients with resistant virus and 0.31 (95% CI: 0.29-0.37; P = 0.181) for patients without genotypic resistance at baseline.
With all patients having a Cmin at day 7 greater than the IC95, it is not surprising that there is no association with the VDc. It indicates that NVP and EFV exert their maximal antiviral effect. Even in patients with a baseline pVL >100,000 copies/mL, the achieved plasma concentrations were high enough to clear free virus. To our knowledge, this is the first time that VDc has been analyzed directly in relation to NNRTI drug concentrations. Earlier studies examining the efficacy of high-dose NVP have focused on the steady-state period after 4 weeks of NVP treatment.24,25
The findings of the present study of no difference in early virologic efficacy between the treatment groups expand the overall results of the main 2NN study. NVP and EFV demonstrate comparable efficacy early and later during treatment, especially when comparing their direct effect on pVL. Combining these drugs does not have an additional effect; therefore, there is no rationale for the use of this combination.
1. Louie M, Hogan C, Di Mascio M, et al. Determining the relative efficacy of highly active antiretroviral therapy. J Infect Dis
2. Perelson AS, Nelson PW. Mathematical analysis of HIV-1
dynamics in vivo. Siam Review
3. Ho DD, Neumann AU, Perelson AS, et al. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1
4. Wei X, Ghosh SK, Taylor ME, et al. Viral dynamics
in human immunodeficiency virus type 1 infection. Nature
5. Cozzi Lepri A, Miller V, Phillips AN, et al. The virological response to highly active antiretroviral therapy over the first 24 weeks of therapy according to the pre-therapy viral load and the weeks 4-8 viral load. AIDS
6. Raboud JM, Rae S, Montaner JS. Predicting HIV RNA virologic outcome at 52 weeks follow-up in antiretroviral clinical trials. The INCAS and AVANTI Study Groups. J Acquir Immune Defic Syndr
7. Chene G, Sterne JA, May M, et al. Prognostic importance of initial response in HIV-1
infected patients starting potent antiretroviral therapy: analysis of prospective studies. Lancet
8. Ledergerber B, Egger M, Opravil M, et al. Clinical progression and virological failure on highly active antiretroviral therapy in HIV-1
patients: a prospective cohort study. Swiss HIV Cohort Study. Lancet
9. Polis MA, Sidorov IA, Yoder C, et al. Correlation between reduction in plasma HIV-1
RNA concentration 1 week after start of antiretroviral treatment and longer-term efficacy. Lancet
10. Mittler J, Essunger P, Yuen GJ, et al. Short-term measures of relative efficacy predict longer-term reductions in human immunodeficiency virus type 1 RNA levels following nelfinavir monotherapy. Antimicrob Agents Chemother
11. Wu H, Mellors J, Ruan P, et al. Viral dynamics
and their relations to baseline factors and longer term virologic responses in treatment-naive HIV-1
-infected patients receiving abacavir in combination with HIV-1
protease inhibitors. J Acquir Immune Defic Syndr
12. Wu H, Kuritzkes DR, McClernon DR, et al. Characterization of viral dynamics
in human immunodeficiency virus type 1-infected patients treated with combination antiretroviral therapy: relationships to host factors, cellular restoration, and virologic end points. J Infect Dis
13. van Leth F, Phanuphak P, Ruxrungtham K, et al. Comparison of first-line antiretroviral therapy with regimens including nevirapine
, or both drugs, plus stavudine and lamivudine: a randomised open-label trial, the 2NN
14. King RW, Klabe RM, Reid CD, et al. Potency of nonnucleoside reverse transcriptase inhibitors (NNRTIs) used in combination with other human immunodeficiency virus NNRTIs, NRTIs, or protease inhibitors. Antimicrob Agents Chemother
15. Kappelhoff BS, Rosing H, Huitema AD, et al. Simple and rapid method for the simultaneous determination of the non-nucleoside reverse transcriptase inhibitors efavirenz
in human plasma using liquid chromatography. J Chromatogr B Analyt Technol Biomed Life Sci
16. Beal S, Sheiner L. NONNMEM User's Guides
. NONMEM Project Group. San Francisco: University of California at San Francisco; 1998.
17. Centers for Disease Control and Prevention. Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep
18. Notermans DW, Goudsmit J, Danner SA, et al. Rate of HIV-1
decline following antiretroviral therapy is related to viral load at baseline and drug regimen. AIDS
19. Perelson AS, Essunger P, Cao Y, et al. Decay characteristics of HIV-1
-infected compartments during combination therapy. Nature
20. Perelson AS, Neumann AU, Markowitz M, et al. HIV-1
dynamics in vivo: virion clearance rate, infected cell life-span, and viral generation time. Science
21. Nowak MA, Bonhoeffer S, Loveday C, et al. HIV results in the frame. Results confirmed. Nature
22. Wu H, Ding AA, De Gruttola V. Why are the decay rates in plasma HIV-1
different for different treatments and in different patient populations? AIDS
23. Hall D, van Leth F, Scherer J. Genotypic analysis of reverse transcriptase in treatment naive HIV-1
patients treated with lamivudine, stavudine, nevirapine
. Presented at the 11th Conference on Retroviruses and Opportunistic Infections, San Francisco, February 2004.
24. de Jong MD, Vella S, Carr A, et al. High-dose nevirapine
in previously untreated human immunodeficiency virus type 1-infected persons does not result in sustained suppression of viral replication. J Infect Dis
25. Havlir D, Cheeseman SH, McLaughlin M, et al. High-dose nevirapine
: safety, pharmacokinetics, and antiviral effect in patients with human immunodeficiency virus infection. J Infect Dis
Keywords:© 2005 Lippincott Williams & Wilkins, Inc.
HIV-1; viral dynamics; nonnucleoside reverse transcriptase inhibitor; nevirapine; efavirenz; 2NN; randomized clinical trial