Effect of twice-daily nevirapine on adherence in HIV-1-infected patients: a randomized controlled study

Objective: For optimal adherence, once-daily dosing is best. Whether this applies to antiretroviral therapy is unknown. We thus aimed to determine the effect of once-daily dosing on adherence to nevirapine.

Design: A three-phase (3-month observational, 4-month randomized, 5-month interventional) open-label, clinical trial at four French academic medical centres during 2005-2006 among 62 chronically HIV-1-infected subjects with long-lasting viral suppression under a twice-a-day nevirapine-based antiretroviral combination.

Methods: Adherence was measured using electronic monitoring devices and validated by sequential plasma drug levels. Participants were randomly assigned to switch to nevirapine 400 mg once-daily (n = 31) or continue nevirapine 200 mg twice-a-day (n = 31). After the randomized phase, participants had an opportunity to choose their antiretroviral dosage. Primary outcome was the mean percentage of adherence.

Results: Fifty-two patients qualified for electronic data analysis. During the randomized phase, the mean adherence rate was non-significantly superior by 0.5% in once-daily versus twice-a-day dosing (P = 0.68), adjusting for previous twice-a-day adherence rate (P < 0.0001). Once-daily group increased days without dose [odds ratio (OR) 1.7; 95% confidence interval (CI) 1.0, 2.8; P = 0.04], adjusting for previous drug interruptions (P < 0.0001). In the longitudinal analysis, once-daily dosing was significantly associated with at least two consecutive days without dose (OR 4.4; 95% CI 1.9, 10.3; P < 0.001).

Conclusion: Changing from twice to once-daily nevirapine does not improve adherence. Supporting continuous adherence to antiretroviral therapy in the 'once-a-day era' remains a challenge, even if more potent regimens can achieve viral suppression at lower adherence levels.

Introduction

High adherence rates are necessary to control HIV ^[1], prevent opportunistic infections and death ^[2]. In addition, HIV can become resistant to drugs in the case of imperfect adherence ^[3,4]. In an effort to simplify HIV treatment, once-daily dosing has been developed. Although frequently considered a fact, evidence of improved adherence with once versus twice-daily dosing is, however, limited.

We hypothesized that once-daily dosing improves adherence in patients with chronic HIV-1 infection and viral control under nevirapine twice a day. This study compared objective measurements of adherence to nevirapine once versus twice-daily dosing as the primary outcome. Secondary outcomes were safety, efficacy and patient's preference.

Methods

Study population

This trial was a multicentre randomized study conducted at four university teaching hospitals in France. Patients were chronically HIV-1-infected adults, receiving nevirapine-based antiretroviral therapy with RNA-HIV levels less than 400 copies/ml for more than 6 months and without liver enzyme abnormality (asparate aminotransferase or alanine aminotransferase > 2.5 N and > 1.25 N if hepatitis virus B or C were negative and positive, respectively; see details at: http://clinicaltrials.gov/ct/show/NCT00466180). The Côte de Nacre University Institutional Review Board approved the study in April 2004 and all patients provided written informed consent.

Study design and study monitoring

The POSOVIR study consisted of three phases with a total follow-up period of 12 months. Adherence to nevirapine was prospectively evaluated by electronic monitoring devices [Medication Event Monitoring System 6 (MEMS) caps; Aardex Ltd., Switzerland], as previously described ^[1] and transversally by the nevirapine plasma level at the end of each phase. During the first 3-month period (phase 1, prospective, observational study), all the patients received nevirapine twice a day. At the end of month 3, patients were 1: 1 randomly allocated to continue nevirapine twice a day or switch to nevirapine once-daily for a 4-month period (phase 2, randomized controlled trial). At the end of month 7, patients were monitored (phase 3, prospective, interventional study) for five additional months under their preferred dosage. Electronic adherence data showing no event (i.e. no nevirapine box opening) within the 14 preceding days of a plasma sample with detectable nevirapine levels (i.e. with evidence that nevirapine was taken) were considered invalid and the overall phase was suppressed. In addition, electronic data were validated by patient's self-report to detect pocket-dose (nevirapine taken without opening the MEMS cap) and box-filling (MEMS cap opened without taking nevirapine) in order to adjust events.

If once-daily nevirapine was started, liver enzymes were monitored every 2 weeks for 2 months. All patients were evaluated at weeks 12, 28 and 52 for clinical, biological and immunovirological parameters. Safety monitoring used the Agence Nationale de Recherche sur le SIDA toxicity grading.

Endpoints

The primary outcome was the mean percentage of adherence, which is calculated as the number of doses taken over the total number of doses prescribed. Secondary adherence endpoints were the monthly count of drug interruptions for less than 2 days (day without dose) and for at least two consecutive days (drug holidays). Other secondary endpoints included safety, efficacy and patient's preference.

Sample size and statistical analysis

Assuming a baseline mean adherence of 85% and a SD of ±10% in the twice-a-day control group, 90% power and two-sided type I error of 0.05, the inclusion of at least 60 patients would be necessary to demonstrate a 10% difference ^[5] in the once-a-day experimental arm.

Differences were assessed using Student's t-test (continuous variables) and Fisher's exact test or the MacNemar test (categorical variables) for independent and dependent samples, as appropriate. The impact of the intervention was assessed in two ways. First, we computed generalized linear models of the phase 2 adherence outcome (covariance analysis), which is robust in the case of non-normal distributions ^[6]. For continuous, count and binary outcomes, we used linear, Poisson and logistic models, respectively. Second, we used the concept of longitudinal analysis with repeated adherence measurements by computing generalized estimating equations. We used PowerView 2.3.3 (Aardex Ltd.) and SAS version 9.1 (SAS Institute, Cary, North Carolina, USA) for data analysis. All tests were two-sided and a P value less than 0.05 was considered to denote statistical significance.

Results

Baseline characteristics

Sixty-two patients aged 24-76 years (mean 48.1) were enrolled in the study. One patient died accidentally during phase 2. Two MEMS caps were lost. Examination of electronic adherence data and nevirapine plasma levels revealed that seven patients did not use their MEMS caps correctly and were excluded from primary endpoint analysis. Among the remaining 52 patients (84%), mean (SD) observational phase 1 twice-a-day adherence rates were balanced between once-daily [93.7 (7.4), n = 27] and twice-daily groups [91.4 (16.5), n = 25], as well as other baseline patient characteristics. Taking other drugs once a day (n = 26) did not influence the baseline twice-daily adherence rate to nevirapine (94.1 versus 94.2 among patients with twice-daily other drugs, P = 0.97).

Effect of once-a-day dosing

Randomized analysis

As shown in Figure 1, once-daily dosing marginally increased adherence [+0.5%, 95% confidence interval (CI) -1.8, 2.9; P = 0.68], adjusting for baseline adherence rate (P < 0.0001). Controlling for the baseline drug interruption count (P < 0.0001 in both models), the once-daily group was significantly and independently associated with increased monthly days without dose [odds ratio (OR) 1.7; 95% CI 1.0, 2.8; P = 0.04] and not significantly with monthly drug holidays (OR 2.0; 95% CI 0.9, 4.6; P = 0.09). Viral control (RNA-HIV level < 400 copies/ml) was 61 out of 61 at the end of phase 2.

Fig. 1

Longitudinal analysis

Merging adherence measurements from the three phases yielded to 52, 52 and 36 observations from phase 1 (all twice a day), phase 2 (27 once-daily) and phase 3 (22 once-daily), respectively.

The adherence rate remained marginally increased by once-daily dosing (+0.5%, 95% CI -2.4, 3.3, P = 0.77). In this analysis, the increased count of monthly days without dose in the once-daily group was not significant (OR 1.3; 95% CI 0.9, 2.0; P = 0.20). The once-daily regimen was associated with an increased monthly count of drug holidays (OR 4.4; 95% CI 1.9, 10.3; P < 0.001).

Adherence and efficacy

Viral suppression at the end of the adherence phases was 11 out of 14 (78.5%) for rates less than 80% and 126 out of 126 (100%) for adherence greater than 80%. A 10% adherence rate increase was significantly associated with RNA-HIV levels less than 400 copies/ml (OR 3.4; 95% CI 2.0, 5.6, P < 0.0001). Two patients had three virological failures, as shown in Figure 2.

Fig. 2

Safety and preference

Ten serious adverse events including one death were reported in seven patients. None were drug related. In particular, no cutaneous or hepatic severe (greater than grade 2) abnormalities occurred after the switch to once-a-day nevirapine.

At the end of phase 2, more patients in the twice-a-day group wished to switch to once-a-day (15 versus eight patients who switched back from once to twice-a-day dosing, P = 0.15). These figures were statistically significant in the subgroup of patients with once-daily other antiretroviral drugs (10 versus two, P = 0.02).

Discussion

This randomized trial is the first, to our knowledge, to compare objective adherence to the same antiretroviral drug taken once daily and twice daily. We found no evidence of an improved adherence rate with once-daily dosing. Furthermore, once-daily dosing was associated with an increased number of drug interruptions. Although omission of a single pill is enough for a day without dose, once-daily dosing was also associated with longer interruption episodes in otherwise adherent patients.

The primary outcome measure used prospective objective data with electronic monitoring, validated by other measures combining plasma levels and self-reports to avoid an underestimation of adherence, as previously demonstrated ^[7]. The randomized design avoided channelling bias possibly associated with cohort studies. In addition, the longitudinal analysis, which also assessed the intraindividual effect of once-daily dosing on adherence, found concordant results. The fact that the incidence of drug holidays was significantly higher only in the longitudinal analysis may be the consequence of a switch from 'day without dose' to 'drug holidays' during the third phase in the once-daily nevirapine dosage, as a consequence of adherence decline with time ^[7]. Figure 2 described the dynamics of adherence and illustrated the importance of multiple measures to interpret outcome.

The more than 0.5% increase in the adherence rate with once-daily versus twice-daily dosing is consistent with previous randomized studies in this setting (+1.7% ^[8], +2.0% ^[9]). Using electronic adherence monitoring ^[8], 22 patients switched from twice-a-day stavudine or zidovudine to once-daily stavudine and 21 continued their regimens. Adherence rates were not statistically different between groups (99.4 versus 97.7%, P = 0.24). In the 934 study, mean adherence did not differ significantly between patients receiving once-daily (89%, n = 232) and twice-daily (87%, n = 231) efavirenz-based regimens (P = 0.13), on the basis of pill counts ^[9]. In contrast, Molina et al. ^[10] found that once-a-day therapy increased patient-reported complete adherence, but half of the control group received protease inhibitors three times a day, which is known to impede adherence compared with twice-daily dosing ^[5].

On one hand, periodic lapses in single doses, such as 'day without dose' have some insurance against virological failure and resistance. On the other hand, repeated longer drug interruptions with non-nucleoside reverse transcriptase inhibitor (NNRTI)-based therapy have been repeatedly associated with high-level class resistance ^[3,11-17], except in one small study ^[18]. Our result and the results of other ^[3,11-17] has potential important implications for supporting continuous therapy, including in low-income countries, because antiretroviral drug interruptions ^[19,20] and the emergence of drug resistance ^[21] are both associated with the risk of death.

Previous nevirapine twice-daily exposure with cytochrome P450 isoenzyme-induced metabolism may explain the excellent safety profile of once-daily nevirapine dosing, as compared with naive patients ^[22] or other cohorts with 25% hepatitis co-infection ^[23]. Differences in efficacy were not expected, as demonstrated by the larger 2NN randomized trial ^[22], which compared once (n = 220) versus twice-a-day (n = 327) nevirapine with the same twice-a-day stavudine plus lamivudine nucleoside analogues.

We are aware of limitations. First, our results are compatible with increases in adherence smaller than 10% with once-daily nevirapine, because of the small sample size. Differences in adherence of this magnitude with NNRTI-based therapy may not be as critical as with unboosted protease inhibitors, because a wider range of adherence can achieve viral suppression ^[24-26]. Second, inclusion criteria requiring virological control may have selected patients with previous good adherence to therapy and thus limited room for improvements by a ceiling effect. In addition, a Hawthorn effect related to intensive monitoring in both arms may have biased the results towards no difference, even if the 3-month observation period may limit this to some degree in the randomized analysis. Interestingly, we identified 45% of patients with baseline adherence rates below the recommended 95% threshold for viral suppression. Third, other antiretroviral drugs were not randomized, such that only 26 out of 62 (42%) were administered once daily. We found no adherence increase in the subgroup of patients receiving other drugs once a day (data not shown).

In conclusion, HIV-1-infected patients, like any other individuals, prefer simpler once-daily dosing, as previously shown ^[8,10,27]. When drug interruptions are an issue such as in NNRTI-based therapy ^[3,11-17], however, patients prone to drug holidays should probably take antiretroviral drugs with a higher genetic barrier to resistance. Our findings suggest that the effect of once-daily versus twice-daily dosing does increases the number of drug interruptions and is compatible with modest improvements in adherence rates, below 10%.

Participants in the POSOVIR Study

The POSOVIR study investigators by centres (number of inclusion) were as follows: J.-J. Parienti (11), R. Verdon (6), C. Bazin (3), V. Noyon (1), Hôpital Côte de Nacre, Caen; C. Arvieux (10), C. Michelet (5), A.G. Depatureaux (1), P. Tattevin (1), Hôpital Universitaire, Rennes; G. Le Moal (7), Hôpital Universitaire, Poitiers; V. Reliquet (5), C. Allavena (5), B. Bonnet (3), V. Gagey (2), E. Billaud (1), N. Morineau (1), Hôpital Universitaire, Nantes.

Acknowledgements

The authors would like to thank the patients who agreed to participate, and P. Goubin, S. de Bouard, J.-J. Dutheil, B. de la Gastine, C. Peaucelle, M. Ratajczak, H. Huë, and D. Brosseau for their role in data management and safety reporting in each centre.

Sponsorship: This study was sponsored by an academic grant from the Côte de Nacre University Hospital.

Conflicts of interest: None.