Efficacy and safety of switching from nevirapine immediate-release twice daily to nevirapine extended-release once daily in virologically suppressed HIV-infected patients: a retrospective cohort study in Taiwan

Nevirapine immediate-release (NVP-IR) has been used with a high level of efficacy and safety since its introduction in 1996 in the United States and 1998 in Europe as a potent non-nucleoside reverse transcriptase inhibitor (nNRTI) used in combined antiretroviral therapy (cART) [1‐3]. However, efforts to simplify the regimens in terms of pill counts and frequency have continued to improve adherence to HAART [4, 5]. Therefore, although the licensed dosage for NVP-IR is 200 mg twice daily after a 200 mg daily lead-in period, an off-label maintenance dose of 400 mg (2 × 200 mg) NVP-IR has been adopted to improve compliance [6]. Clinical trials comparing a single daily dose of 400 mg NVP-IR versus 200 mg twice daily demonstrated non-inferior efficacy for the new treatment [6‐8]. To further simplify treatment regimens, a new extended-release (XR) formulation was developed [9, 10]. In phase III clinical trials, 400 mg NVP-XR once daily was found not inferior to 200 mg NVP-IR twice daily in terms of virological suppression in treatment-naïve patients (VERxVE study) [11] and in virologically suppressed patients who were switched from NVP-IR to NVP-XR (TRANxITION study) [12].

Nevirapine is pharmacologically characterized by a strong relationship between NVP trough plasma concentrations and virological response, [13, 14] and a target therapeutic trough concentration of 3.0 mg/L has been proposed as a minimum effective concentration [15] because the risk of virological failure increases 5-fold (relative risk 5.0, 95% CI 1.8–13.7) with NVP C_trough ≤ 3 mg/mL compared with patients who had higher trough concentrations [13, 14].

However, in routine clinical practice, optimal plasma concentrations of NVP are reached in only 27.8% patients [16]. The significant inter-individual variability in plasma concentrations of nNRTI could be partially explained by differences in ethnicity, polymorphisms in enzymes, drug–drug interactions, and body weight [17‐19]. Polymorphisms of CYP2B6 may influence the metabolism of NVP, with the variant CYP2B6 516 GT polymorphism resulting in reduced clearance of NVP compared with CYP2B6 516 GG [20]. In a 2NN pharmacokinetic (PK) substudy, NVP clearance was reduced by 19.4% in patients from Thailand and South Africa compared with Caucasian and Hispanic patients [20]. This difference has been related to CYP2B6 516 GT polymorphisms, resulting in a 15.3% reduced clearance of NVP compared with patients with CYP2B6 516 GG [20]. However, a recent study of 171 HIV-infected patients in northern Taiwan showed a predominance of homozygous 516 GG alleles (66.1%), [21] which may lead to reduced NVP C_trough compared with patients with CYP2B6 516 GT. Furthermore, a recent prospective study of 227 HIV-infected, treatment-naïve patients conducted in China recommends a higher target therapeutic C_trough (3.9 μg/mL) of NVP for Chinese patients than the currently recommended level (3.0 μg/mL), which is predominately from the results of studies among Caucasian and African American patients [22].

In addition, previous PK studies of NVP have demonstrated a lower C_trough with 400 mg NVP-XR once daily compared with 200 mg NVP-IR twice daily [11, 12, 23]. Therefore, the question remains whether the non-inferior efficacy and safety results of switching virologically suppressed HIV-1-infected patients from NVP-IR to NVP-XR demonstrated in the TRANxITION study conducted in Europe and North America [12] are also applicable to virologically suppressed HIV-infected patients in Taiwan.

To the best of our knowledge, this small retrospective cohort is the first to investigate the safety and efficacy of switching to NVP-XR from NVP-IR among virologically suppressed HIV-infected patients in Taiwan. Although the TRANxITION study demonstrated non-inferior efficacy and safety of switching virologically suppressed HIV-1-infected patients from NVP-IR to NVP-XR, the results were derived mainly from a Caucasian population. In this retrospective cohort study of virologically suppressed HIV-infected Taiwanese patients, we found that the NVP-XR group was similar to the NVP-IR group with respect to maintaining a virological response. The difference may not have been significant because the numbers of patients were relatively small and the study lacked the power to demonstrate a statistically significant difference. Despite methodological differences, the results of our comparative analyses of virological response indicated worse results compared with those of the TRANxITION study [12]. During NVP-IR therapy, 85.7% and 92.6% of patients in the current and TRANxITION study, respectively, were virologically suppressed. During NVP-XR therapy, 82.9 and 93.6% of patients in the current and TRANxITION studies, respectively, were virologically suppressed. The lower rate of virological response in our study may result from a longer follow-up period, as the mean follow-up times for the current and TRANxITION studies were respectively 18.4 and 6 months. Furthermore, the current study was conducted outside a well-controlled trial setting. Therefore, the efficacy of NVP-IR and NVP-XR in maintaining virological response may be underestimated here owing to a relatively high proportion of LTFU treatment failures in both groups (NVP-IR vs. NVP XR: 50% (3/6) vs. 57% (4/7)), which is inherent to retrospective studies.

Because previous studies have demonstrated a significant relationship between NVP trough plasma concentrations and virological response, [13, 28] we further investigated the NVP trough concentrations in both NVP-IR and NVP-XR groups. The median of trough plasma concentrations was 6775.0 ng/mL for NVP-IR and 5145.0 ng/mL for NVP-XR in our study (Additional file 3). The plasma concentrations of NVP in both groups are higher than the C_trough of 3.0 mg/L recommended in most guidelines [15] or C_trough of 3.9 μg/mL for Chinese patients in a recent study [22]. The non-significant difference of NVP trough concentrations between these two groups may be owing to the small sample size in each group. This finding is compatible with previous studies, which demonstrated a lower C_trough with NVP-XR therapy compared with 200 mg NVP-IR twice daily [11, 23].

We believe that there was no obvious selection bias between patients receiving NVP-XR and NVP-IR who also underwent therapeutic drug monitoring of steady-state plasma NVP concentration in the current study. Compliance, drug–drug interactions, and differences in pharmacokinetics are critical factors that influence the plasma concentrations of antiretroviral drugs [29]. Our patient population was assumed to be reasonably compliant because they had maintained a stable virological response to NVP-containing cART prior to entering the study, and all participants remained virologically suppressed for at least 1–4 months (duration of continued virological suppression at entering this study, days (±SD): NVP-XR vs. NVP-IR, 1042 (565) vs. 959 (875); P = 0.60). Additionally, these patients only received antiretroviral agents and did not take any other prescribed medicine concomitantly. Finally, genotype analysis of CYP2B6 516 in 15 patients also revealed no significant difference in CYP2B6 516 polymorphisms between these two groups.

Surprisingly, however, our study demonstrated higher C_trough levels in both groups compared with other studies conducted in the Netherlands, [30] Germany, [31] and Canada [32]. This difference may result from a higher prevalence of CYP2B6 516 GT in our study population (86.6%) compared with other reports of Han Chinese, [33] Caucasians, [29] and Taiwanese in northern Taiwan [21]. Our finding of a high prevalence of CYP2B6 516 GT in the current study requires further verification in a larger Taiwanese population.

Maximal and durable suppression of plasma HIV viral load has now been established as the primary aim of antiretroviral treatment for both treatment-naïve and treatment-experienced patients [24]. However, viral replication is still not fully controlled in all patients at all times, and transient viremia, often between 50 and 400 copies/mL (a virological blip) has been frequently reported [34]. However, the impact of HIV RNA blips on virological failure remains controversial [35, 36]. In the ACTG 343 and Merck 035 trials, treatment-experienced patients who had virological blips did not have an increased risk of virological failure compared with patients with continuing virological suppression [35]. However, the Swiss HIV Cohort and Frankfurt HIV Clinic Cohort studies showed a slightly increased risk of virological failure for patients with one or more virological blips compared with viral suppression (HR 2.01, 95% CI 1.51–2.91; P < 0.0001) [36]. In our cohort study, six of 84 patients (7.1%) experienced a virologic blip during the 18.4-month follow-up. In Cox regressions adjusted for age, sex, history of AIDS, cART regimen, and history of virological suppression, the HR of viral blips was 1.023 (P = 1.00) for the NVP-XR group compared with the NVP-IR group. The relationship between viral blip and virological failure in Taiwanese patients requires further investigation in a larger population.

The current study has several limitations and biases inherent to retrospective cohort studies, the first being selection bias. Second, there were significant differences in cART backbone drugs between the NVP-XR and NVP-IR groups. However, the CNA30024 study comparing the durability of viral effect and the safety profile of triple therapy with either abacavir or zidovudine, combined with lamivudine and efavirenz, demonstrated similar virological suppression (HIV RNA ≤ 50 copies/mL) in treatment-naïve patients at week 48 (abacavir group vs. zidovudine group: 70% vs. 69%) [37]. Third, it is difficult to determine patient compliance with medications from retrospective chart reviews. This is important because adherence to treatment is closely related to sustained viral suppression [38]. As mentioned above, we believe that our patient population was reasonably compliant because only patients with virological suppression for at least 1–4 months were included. Finally, the frequency of adverse reactions to NVP in both groups may be underestimated. Although the frequency of moderate to severe liver function abnormalities is similar to that of the TRANxITION study, it is likely that we missed some low-grade skin rashes and gastrointestinal disorders that did not require modification of the antiretroviral regimen.

In this retrospective study, we found that switching from cART regimens with twice-daily doses of 200 mg NVP-IR to a once-daily dose of 400 mg NVP-XR appeared nearly as safe and effective as continuing NVP-IR in virologically suppressed HIV-infected Taiwanese patients. Our findings of higher C_trough levels in both groups compared with other studies conducted among Caucasian populations may be due to the high prevalence of CYP2B6 516 intermediate/low metabolizers in the present study.