HIV Articles  
High rate of early virological failure with the once-daily tenofovir/lamivudine/nevirapine combination in naive HIV-1-infected patients
  Journal of Antimicrobial Chemotherapy Feb 2009 63(2):380-388;
The study was financially supported by Boehringer-Ingelheim.
D. Rey1,*, B. Hoen2, P. Chavanet3, M. P. Schmitt4, G. Hoizey5, P. Meyer6, G. Peytavin7, B. Spire8, C. Allavena9, M. Diemer10, T. May11, J. L. Schmit12, M. Duong3, V. Calvez13 and J. M. Lang1
1 COREVIH, Hopitaux Universitaires, Strasbourg, France 2 Service des Maladies infectieuses et Tropicales, CHU Besancon, France 3 Service des Maladies infectieuses et Tropicales, CHU Dijon, France 4 Laboratoire de Virologie, Hopitaux Universitaires, Strasbourg, France 5 Laboratoire de Pharmacologie et Toxicologie, CHU de Reims, France 6 Unite de Biostatistiques, Faculte de Medecine, Strasbourg, France 7 Laboratoire de Toxicologie et Pharmacocinetique, Hopital Bichat, Paris, France 8 INSERM U912, Marseille, France 9 Service des Maladies infectieuses et Tropicales, CHU Nantes, France 10 Service de Medecine Interne, Hopital Lariboisiere, Paris, France 11 Service des Maladies infectieuses et Tropicales, CHU Nancy, France 12 Service des Maladies infectieuses et Tropicales, CHU Amiens, France 13 Laboratoire de Virologie, Hopital Pitie Salpetriere, Paris, France
"In conclusion, the DAUFIN trial showed an unexpected and high rate of early virological failures in patients treated with tenofovirDF + lamivudine and nevirapine once daily, with a high incidence of resistance mutations to NNRTIs, and K65R conferring broad cross-resistance to nucleoside analogues. At baseline, failing patients had higher viral loads and lower CD4 cell counts than subjects with success. But adherence to treatment was similar in the two groups and we failed to identify any pharmacological explanation for virological failures. Whatever the mechanism that led to the high viral failure rate, we deem it essential to alert that the once-daily combination of tenofovirDF, lamivudine and nevirapine should not be given as a first-line ARV therapy."

Background: The combination of one non-nucleoside reverse transcriptase inhibitor (NNRTI) with two nucleoside reverse transcriptase inhibitors is a validated first-line antiretroviral (ARV) therapy. The once-daily combination of lamivudine, tenofovirDF and nevirapine has not been evaluated in a clinical trial.
Methods: Randomized, open-label, multicentre, non-inferiority trial comparing lamivudine, tenofovirDF and nevirapine once daily (Group 2) with zidovudine/lamivudine and nevirapine twice daily (Group 1), in naive HIV-1-infected patients with a CD4 count <350/mm3. We planned to enrol 250 patients.
Results: As of May 2006, 71 patients had been enrolled (35 in Group 1 and 36 in Group 2) and an unplanned interim analysis was done. The groups were comparable at baseline: median CD4 count was 195 and 191/mm3 and median plasma viral load was 4.9 log10 and 5.01 log10, respectively, in Groups 1 and 2. Eight early non-responses (22.2%) were observed, all in Group 2, while two later viral rebounds occurred. Resistance genotypes for the nine Group 2 failing patients showed the mutations M184V/I (n = 3), K65R (n = 6), one or more NNRTI resistance mutations in all cases. At baseline, the nine Group 2 patients who failed had higher median plasma viral load (5.4 log10) and lower median CD4 count (110/mm3) than the other Group 2 patients (4.7 log10, P = 0.002 and 223/mm3, P = 0.004). Nevirapine trough concentrations were not different between the two groups, nor between patients with full viral suppression or those who failed in Group 2. Due to slow recruitment, and those results, the steering committee decided to stop the trial at 12 months.
Conclusions: In ARV-naive HIV-1-infected patients, the once-daily lamivudine, tenofovirDF and nevirapine regimen resulted in a high rate of early virological failures. The reasons for the failures remain unclear.
The choice of the drugs included in the first-line therapy offered to HIV-infected patients is a critical issue, since early virological response is key to limit the risk of emergence of resistance mutations,1 and thus to maintain a durable suppression of viral replication. Adherence to and convenience of antiretroviral (ARV) therapy are also essential for long-term antiviral efficacy.2
According to the international3 and French4 guidelines that prevailed when our trial was built up and initiated, the combination of a non-nucleoside reverse transcriptase inhibitor (NNRTI) with two nucleoside reverse transcriptase inhibitors (NRTIs) was one of the options recommended for first-line ARV therapy. Actually, NNRTIs are as effective as a non-boosted protease inhibitor in combination with two NRTIs in ARV-naive HIV-infected patients.5-7 More recently, efavirenz was shown to be equivalent to lopinavir/ritonavir treatment in an observational study,8 and to have virological superiority to lopinavir/ritonavir in a randomized trial (ACTG 5142),9 when combined with two NRTIs.
Nevirapine was first thought to be inferior to efavirenz.10-12 The only large randomized study comparing the two NNRTIs showed a lower proportion of patients with treatment failure in the efavirenz group (37.8%) compared with nevirapine (43.6% and 43.7% in the once- and twice-daily groups, respectively). The difference between nevirapine twice daily and efavirenz was not significant, but equivalence within the 10% limits could not be demonstrated.13 Nevirapine and efavirenz are therefore considered to have similar virological potencies.
The combination of tenofovirDF, lamivudine and efavirenz was found highly effective in a recent study.14 However, once-daily nevirapine in combination with lamivudine and tenofovirDF has so far not been evaluated in a prospective clinical trial.
This study was designed to compare the efficacy and tolerability of a twice-daily combination of nevirapine and zidovudine/lamivudine versus a once-daily combination of nevirapine, lamivudine and tenofovirDF, and has been named the DAUFIN study (Trial Registration: number: NCT 00199979).
This randomized, multicentre study showed a high rate of early virological failures with a once-daily combination of lamivudine, tenofovirDF and nevirapine, in ARV-naive HIV-infected patients, while the validated comparator (twice-daily combination of zidovudine/lamivudine and nevirapine) induced a virological response rate in the expected range. Moreover, the broad-spectrum resistance profile, to both NNRTIs and NRTIs, which was selected, was cause for concern. These results were highly unexpected.
To our knowledge, the DAUFIN study is the first to show such a high and early proportion of failures, with this once-a-day regimen. However, Towner et al.21 showed comparable results in a small group of 23 ARV-naive patients, in a study only presented in a meeting: seven virological failures (30%) occurred, five of whom developed the Y181C mutation. More recently, an Italian study22 also showed a high rate of early viral failures (3/7) with tenofovirDF/emtricitabine once a day combined with nevirapine twice a day. All three patients developed NNRTI mutations, including Y181C, plus M184V (n = 2) and K65R (n = 1).
Thus, two studies recently showed similar poor outcomes with nevirapine + tenofovirDF and lamivudine or emtricitabine, as reported in our trial. Although the reasons for the high rate of virological failure of this once-daily regimen are unclear, several potential hypotheses may be considered: pre-existing resistance mutations, high baseline viral load, non-B subtypes, underexposure to one of the studied drugs and, lastly, poor adherence despite a simplified regimen.
Regarding the first hypothesis, only one patient harboured a pre-existent mutation (K101E). One could hypothesize that some of these patients had archived NNRTI-resistant viruses, non-detectable with standard genotyping, but even with a small number of patients, this should also have occurred in the other group.
In our study, the only significant differences shown in failing subjects, compared with those with success, were a higher baseline viral load, as well as lower initial CD4 cell counts. The same observation was made by Lapadula et al.22 It has already been shown that such factors (HIV RNA >100 000 copies/mL, and CD4 <200/mm3) are strong predictors of clinical evolution to AIDS or death,23 and more recently of increased risk of virological failure with both efavirenz and nevirapine in the 2NN study.24
Virological response to the first-line ARV therapy is thought to be not significantly different for B and non-B subtypes.25,26 However, the V106M mutation may be selected following NNRTI therapy in subtype C viruses.27 A rapid selection of K65R has also been described in subtype C following tenofovirDF treatment,28 but only in vitro, and has not been found in vivo.29 As 67% of the patients who failed in Group 1 of the DAUFIN study were infected with a B subtype, it is highly unlikely to be the explanation of the high rate of virological failures.
From a pharmacokinetic standpoint, nevirapine trough plasma concentrations of 4 mg/L or higher are accepted as appropriate for wild-type viruses.4,30 In this study, mean nevirapine trough concentrations were above the 4 mg/L threshold in both groups and not significantly different between the two groups. Moreover, nevirapine concentrations were not lower in subjects who failed than in those with full viral suppression of the same once daily group. In conclusion, underexposure to nevirapine is very unlikely to explain our findings.
The expected tenofovir trough plasma concentration following tenofovirDF 300 mg once daily is 66 ng/mL.31 This concentration was reached in our Group 2 patients and, although lower in patients who failed, the concentrations were not significantly different when compared with subjects with viral suppression. Analysis of variance for repeated measures was used for this analysis, which has a strong power (a patient is withdrawn from analysis when a single measure is lacking). As for nevirapine, no underexposure to tenofovirDF was identified.
Also, global adherence could appear to be low in our study, as 77% was the best result achieved. Mean adherence to treatment, for example, was better in the 934 trial with efavirenz:32 90% among patients receiving tenofovirDF + emtricitabine and efavirenz, and 87% among those receiving zidovudine/lamivudine and efavirenz (P = 0.04 in favour of the once-daily combination), but adherence was assessed on the basis of pill counts at each visit, and potential variation from visit to visit was not indicated. In addition, we used a very strict definition for good adherence, as a patient was considered to be non-adherent if a single question, out of four, identified suboptimal adherence. It has also been shown that <95% adherence to NNRTI therapy (but >75%) could lead to viral suppression.33 Even if some of the failures observed in our study clearly appeared to be the consequence of poor adherence (complete lack of viral load decrease in three subjects), it is highly unlikely that suboptimal adherence difficulties would solely explain our results (as no differences were observed between the two groups).
Once-a-day nevirapine appeared to be rather safe, at least as well tolerated as twice-a-day administration. In the DAUFIN study, we did not observe more cutaneous or hepatic side effects compared with twice-a-day intake. The 2NN trial showed higher hepatic toxicity with nevirapine once a day, compared with a twice-a-day regimen,13 but in the NEFA study,34 patients with high CD4 cell counts were switched to a nevirapine-containing regimen without hepatotoxicity increase.
The main limitation of the DAUFIN trial is the small number of patients. It has been claimed that some of our results could be attributable to chance, and there were criticisms relating to pharmacological (low nevirapine plasma trough concentrations) or virological (non-B subtypes) grounds.35 We have ruled these out as discussed earlier. We acknowledge that we were far from the initial enrolment goal, and we had neither planned an interim analysis nor predefined criteria for study cessation. However, the steering committee unanimously decided to stop the trial, and we strongly think that continuing the trial would have been unethical. Actually, the failures we observed were associated with mutations that dramatically reduced the remaining therapeutic options.
In conclusion, the DAUFIN trial showed an unexpected and high rate of early virological failures in patients treated with tenofovirDF + lamivudine and nevirapine once daily, with a high incidence of resistance mutations to NNRTIs, and K65R conferring broad cross-resistance to nucleoside analogues. At baseline, failing patients had higher viral loads and lower CD4 cell counts than subjects with success. But adherence to treatment was similar in the two groups and we failed to identify any pharmacological explanation for virological failures. Whatever the mechanism that led to the high viral failure rate, we deem it essential to alert that the once-daily combination of tenofovirDF, lamivudine and nevirapine should not be given as a first-line ARV therapy.
Study population and disposition

The CONSORT flow diagram is shown in Figure 1. Baseline characteristics of the patients are summarized in Table 1. The two treatment groups were well balanced, especially for immunological and virological characteristics.
Final analysis of the reasons for treatment interruptions revealed eight early viral failures that occurred during the first 12 weeks of treatment, including three non-responses and five rebounds after significant plasma HIV RNA decrease, all in the once-a-day treatment arm, and two later viral rebounds (weeks 24 and 36), one in each group.


Genotypic analysis
Table 2 shows drug resistance mutation results, viral load change and viral subtype in the 10 patients who experienced virological failure. Six patients were infected with a B subtype, while four non-B subtypes were identified (C subtype in one case, 'CRF 01' in two cases and 'CRF 06' in one case).



The only failure in Group 1 clearly appears to be the result of poor adherence, as no resistance mutation was detected, and viral increase occurred after a period of viral suppression. In Group 2, several mutations were already present at week 4 of treatment, conferring resistance to NNRTIs and NRTIs.
In order to explain the high rate of viral failures in Group 2, we compared baseline immuno-virological characteristics of patients with full viral suppression (n = 27) and those who failed (n = 9): the former had a statistically significant lower viral load and higher CD4 cell counts than the latter, with 51 189 and 262 747 copies/mL (P = 0.002), and 223 and 110/mm3 (P = 0.004), respectively.
Drug concentrations
Mean nevirapine trough plasma concentrations were compared between the two groups. In both groups, an increase in nevirapine concentration was observed after the lead-in period of 2 weeks, followed by a stabilization of concentration. As expected, Group 1 subjects (twice-a-day nevirapine) had higher nevirapine trough concentrations, but the difference was not significant with analysis of variance for repeated measures, compared with Group 2 subjects (Table 3 and Figure 2).
Figure 2. Comparison of nevirapine (NVP) trough concentrations between treatment arms using box-plot (variability around the median).



Both groups of patients showed nevirapine mean trough concentrations above the 4 mg/L threshold, even with estimated 24 h levels in the once-a-day group (as some blood samples were drawn around 12 h after the last intake, thus 12 h before the next dosing, according to the patient's decision to take the once-a-day regimen in the morning or in the evening, reflecting 'real life').
Trough tenofovir plasma concentrations were lower in patients who failed (n = 7), compared with those with full viral suppression (n = 17), but the difference was not statistically significant (P = 0.094): respectively, 48.5 and 63.1 ng/mL at week 2, 44 and 56.7 ng/mL at week 4, and 51.3 and 67.2 ng/mL at week 12 (Figure 4). Comparison of nevirapine plasma concentrations, restricted to these 24 patients, was also not different (data not shown).
Figure 4. Comparison of tenofovir (TFV) trough concentrations between patients with viral suppression (0) and those who failed (1) in Group 2, using box-plot (variability around the median).


Grade 3 or 4 adverse events, leading to study discontinuation, were more frequent in Group 1 compared with Group 2, with 10 and 6 subjects, respectively. Hepatic or cutaneous toxicity was similar in the two groups: two grade 3 or 4 transaminase increases in each group, and three and four skin rashes in Groups 1 and 2, respectively. The difference was mainly due to zidovudine-related anaemia (four in Group 1 versus none in Group 2).
At week 4 of treatment, 16/27 (59%) patients were adherent in Group 1, compared with 22/31 (71%) in Group 2 (P = 0.41). There was no significant difference at week 8: 15/22 (68%) and 20/26 (77%) adherent subjects in Groups 1 and 2, respectively (P = 0.53). The only difference was seen after 24 weeks of treatment, with a higher number of adherent patients on once-a-day treatment (13/18=72%) compared with twice-a-day treatment (4/13 = 31%, P = 0.03). Due to premature study discontinuation, next visits were not analysed.
Patients and methods
Study population

Eligible patients were HIV-1-infected, naive with regard to ARV therapy, at least 18 years old and had CD4 cell counts below 350/mm3 in men and 250/mm3 in women.15,16 There was no plasma HIV viral load restriction.
Exclusion criteria were the following: the first 6 months after acute primary infection, diagnosis of a new condition defining the acquired immunodeficiency syndrome within 30 days before entry into the study, receipt of agents that interact with one of the studied drugs, or cytotoxic chemotherapy, immunomodulatory agents, and pregnant or lactating women. All patients were required to use effective contraception. Subjects with hepatitis B or C co-infection, alcohol abuse or hepatic insufficiency were not included. Hepatitis C serology (ELISA, commercially available kits) was performed at screening, and HCV RNA was measured in the case of antibody positivity. Patients with detectable plasma HCV RNA were not included. For hepatitis B, HBs antigen positivity was a non-inclusion criterion.
Biological exclusion criteria were: haemoglobinaemia <10 g/dL, absolute neutrophil count <1000/mm3, platelets <50 000/mm3, creatininaemia >2 times the upper limit of the normal range (ULN), serum transaminases >2.5x ULN, serum total bilirubin level >2x ULN and phosphataemia <2 mg/dL (<0.65 mmol/L).
Trial design
The study was a randomized, open-label, multicenter, non-inferiority trial that compared zidovudine, lamivudine and nevirapine twice a day with tenofovirDF, lamivudine and nevirapine once daily. The study protocol was approved by the Comite Consultatif de Protection des Personnes dans la Recherche Biomedicale d'Alsace no. 1 Strasbourg, and all patients gave written informed consent. Enrolment started in May 2005, in 41 centres.
Patients were stratified according to baseline CD4 cell counts (<200 versus 200-350/mm3) and to baseline HIV-1 RNA (below or above 100 000 copies/mL), in a 1:1 ratio, to receive either a fixed-dose combination of zidovudine/lamivudine 300/150 mg twice a day combined with nevirapine 200 mg twice a day (Group 1), or a once-daily regimen of tenofovirDF 245 mg plus lamivudine 300 mg and nevirapine 400 mg (Group 2). In both groups, nevirapine was started with a lead-in dose of 200 mg per day for 14 days, and increased to 400 mg per day if serum aminotransferase levels were normal.
Study procedures
Clinical examinations were performed at the screening visit, the baseline visit (first day of treatment), at weeks 2, 4, 6, 8 and 12, and then every 12 weeks until week 96 of the study. Clinical status, adverse events and concomitant medications were noted at each on-treatment visit. Occurrence of clinical and laboratory adverse events were compared between the two groups, and their severities were graded according to the ANRS scale (version no. 5, 11 June 2002).
CD4 cell counts and plasma HIV-1 viral load were measured at baseline, weeks 4, 12 and 24, and every 12 weeks thereafter. Laboratory tests, including blood cell counts, plasma chemistry profiles, fasting lipid panel and urinalysis, were performed at baseline, weeks 2 and 4, and every 12 weeks thereafter.
Adherence was self-reported using a standardized questionnaire, at weeks 4, 8 and 24, and then every 12 weeks. This questionnaire included several questions about patient's adherence to ARV therapy over the last 4 days or last 4 weeks prior to the interview. Adherence to ARV therapy was assessed using a previously validated dichotomous score.17,18 Briefly, patients were classified as adherent only if they consistently declared: 100% intake of their prescribed daily number of pills for each drug included in their ARV therapy regimen during the 4 days prior to the visit; having 'totally' taken their prescribed doses of ARV therapy, never having modified the prescribed schedule during the same 4 day period; and not having skipped a dose during the last weekend prior to the visit.
Laboratory methods
Plasma viral load was measured with the commercial kit in use in each participating centre; all but one had a level of detectability of 50 copies/mL (200 copies/mL for the last one).
In patients experiencing virological failure, HIV-1 resistance genotyping was performed on frozen plasma samples obtained at the time of HIV RNA increase (or on the closer available sample), in parallel with the baseline sample.
HIV-1 reverse transcriptase drug resistance testing was performed in a centralized laboratory, using a Trugene HIV-1 Genotyping Kit (Siemens Diagnostics). Genotypic resistances were interpreted using the ANRS 2006 algorithm ( Reverse transcriptase sequences were subtyped using the NCBI subtyping tool ( For study analysis purposes, all genotypes were performed after trial completion. For individual patient management, local genotyping was performed in each participating centre in real time when viral failure was confirmed, and treatment modification was up to each investigator.
Trough nevirapine plasma concentrations were measured in all patients at weeks 2, 4 and 12, and then every 12 weeks, using a modified HPLC assay with an ultraviolet photodiode-array detector, as previously described.19 The lower limit of quantification was 0.5 mg/mL.
Tenofovir plasma concentrations were measured using a specific and validated HPLC assay with fluorimetric detection after chemical derivatization.20 The lower limit of quantification was 5 ng/mL.
Endpoints and statistical analysis
The primary endpoint was the proportion of patients with HIV RNA <400 copies/mL at all visits through week 96, between the two treatment arms (confirmed viral load >400 copies/mL for a second sample defined virological failure). Secondary endpoints included the CD4 cell count increase at 96 weeks of treatment, the percentage of subjects with viral load <50 copies/mL at all visits through week 96, the time to virological failure, genotypic resistance profile evaluation, measurements of nevirapine plasma trough concentrations, grade 3 or 4 adverse events, and adherence to treatment. Assuming a virological response rate of 80% at week 96 for the twice-daily zidovudine/lamivudine/nevirapine regimen, a type I error of 5% and 90% power, the required sample size of the trial to demonstrate non-inferiority of the tenofovirDF/lamivudine/nevirapine regimen was 244 patients (122 in each treatment arm). The study was therefore planned to enrol 250 subjects. Non-inferiority was defined as the lower boundary of the two-sided 95% CI for the treatment difference being above -15%.
Enrolment in the study started in May 2005. No interim analysis had been planned in the first design. However, some premature and unexpected study discontinuations were noted as soon as in the first months of 2006, during which time 71 patients had been included, which triggered an interim analysis. The first review of the reasons for trial interruptions revealed seven virological failures at week 12 of treatment (viral rebound after initial decrease, or absence or weak HIV RNA response), all in the once-a-day treatment arm. The steering committee therefore decided to stop the trial in May 2006.
Based on these preliminary results, we considered a new definition of virological failure in our study: early viral failure was defined as a <2 log10 decrease in plasma viral load by week 12, or a rebound of more than 1 log10 at week 12, after an initial decrease. Late viral failure was defined as a confirmed (second sample, 4 weeks apart) detectable HIV RNA after reaching undetectability.
Demographic data were compared with a two-tailed non-parametric Wilcoxon rank-sum test (quantitative variable) or {chi}2 test and Fisher's exact test as appropriate.
The analysis of variance for repeated measures was used for the data with normal distribution (tenofovir and nevirapine plasma concentrations).
Comparison of CD4 cell counts and HIV viral load between the two groups, or between Group 2 patients with success and those with failure, and of clinical side effects was done with the Mann-Whitney method. Fisher's exact probability test was used for adherence comparison.
All the tests were two-tailed; a P value of <0.5 was considered to be significant. The SPSS 15.0 statistical software was used.
  iconpaperstack view older Articles   Back to Top