Nevirapine Plasma Concentrations Are Associated with Virologic Response and Hepatotoxicity in Chinese Patients with HIV Infection

Background Limited information is available on the relationship between nevirapine plasma concentrations and virologic response or liver toxicity in Chinese patients with HIV infection. The objective of this prospective study was to test this relationship and to determine the minimal therapeutic trough concentration of nevirapine for Chinese patients. Methods A total of 227 HIV-infected, treatment naïve patients were enrolled into this study. Blood samples were taken at Ctrough (12 hr postdose) and C2 (2 hr postdose) for measurement of nevirapine concentrations 6 months after treatment initiation. Therapeutic outcomes, viral load and CD4 cell count, were assessed at 3 and 6 months after starting therapy, while the evaluation of hepatotoxicity was undertaken 12 months after nevirapine treatment. Results A significant correlation between nevirapine trough concentrations and viral load was noticed after 6 months of treatment, particularly in patients with partial response and viral failure (p<0.01). The therapeutic Ctrough of nevirapine for Chinese patients was determined to be 3.9 µg/ml using the receiver operating characteristic curve. Virologic failure was observed in 21% (6/29) of patients with low nevirapine concentrations (<3.9 µg/ml) versus 5% (4/87) in patients with concentrations higher than 3.9 µg/ml (p = 0.015). Hepatotoxicity was significantly associated with the median nevirapine trough concentrations among male patients (8.20 vs. 5.48 µg/ml, p = 0.015) and hepatitis C virus co-infection (p = 0.039). Conclusions Among Chinese patients with HIV infection, the therapeutic Ctrough of nevirapine was 3.9 µg/ml, higher than the recommended 3.0 µg/ml. The correlation between nevirapine concentrations, efficacy and hepatotoxicity suggests the benefit of dosage adjustment based on therapeutic drug monitoring among Chinese HIV-infected patients to optimize nevirapine containing antiretroviral therapy.


Introduction
Highly active antiretroviral therapy is considered an effective approach for the management of HIV/AIDS. Due to its low cost and high efficacy, nevirapine -containing regimens are often preferable to others in resource-limited countries [1]. Approximately 80% Chinese patients receiving antiretroviral therapy are currently on nevirapine-containing regimens. The pharmacological characteristics of nevirapine make it an attractive candidate for therapeutic drug monitoring, as previous studies have indicated a significant relationship between nevirapine trough concentrations and virologic response [2,3,4]. Also, high concentrations are linked to an increased risk of liver toxicity and hepatic injury in patients with chronic hepatitis C [5,6,7,8]. Inter-individual variability of nevirapine plasma concentrations among HIVinfected adults was common in routine clinical practice (,50%) [9]. The variability could be partially explained by the differences in ethnicity, gender, polymorphisms in enzyme and transporter genes, hepatitis virus co-infection, and concomitant medications [10,11,12,13].
As the majority of the previous studies were performed among Caucasian subjects, there is generally a lack of information regarding the efficacy and safety of nevirapine in Chinese HIVinfected patients. A number of studies exploring the association of exposure to nevirapine with virologic response yielded a target trough concentration of ,3.0 mg/l [2,4]; however, this value has not been confirmed among Chinese patients. In addition, the relationship between liver toxicity and nevirapine concentrations has not been established conclusively, though several studies indicate transaminase elevations are related to nevirapine exposure [6,8]. Therefore, the overall objective of the present study was to investigate the relationship between nevirapine concentrations, virologic response and liver toxicity in Chinese HIV-infected patients.

A. Virologic Response
Of 227 HIV-positive antiretroviral-naive patients screened, 173 were treated by a fixed combination antiretroviral therapy containing nevirapine for at least 24 weeks and included in the virologic response analysis (Figure 1). Demographic characteristics of these patients are summarized in Table 1. There were no statistically significant differences among the three groups (complete response, partial response and viral failure) at baseline. Plasma nevirapine concentrations were measured 24 weeks after treatment initiation, including C trough (n = 116) and C 2 (n = 49).
The overall incidence of virologic failure was 10% (18/173). Nevirapine C trough and C 2 in the three groups are summarized in Table 2. The median C trough in patients with virologic failure was significantly lower than that in partial response and complete response groups (3.81 vs. 5.65 and 5.18 mg/ml, respectively, p = 0.018), whereas no significant differences were noted between partial and complete response groups. In contrast, no significant difference in median C 2 was noted among these groups (p = 0.608). A relatively large inter-patient variability in nevirapine C trough was noted ( Figure 2, Table 2). Linear regression analysis indicated a significant linear relationship between viral load and C trough (r = 0.327, p,0.05) for all patients, particularly when limited to patients with a partial response or viral failure (r = 0.619, P,0.01).
In order to determine the nevirapine C trough threshold for therapeutic drug monitoring in Chinese HIV-infected patients, the receiver operating characteristic curve (ROC curve, Figure 3) was used to evaluate the sensitivities and specificities of different cut-off values as shown in Table 3. While the recommended threshold of 3.0 mg/ml corresponded to a low sensitivity of 40% and a high specificity of 88% for viral failure, a new threshold value of 3.9 mg/ml would have a significantly increased sensitivity of 60% (p = 0.005) and an acceptable specificity of 78% (p = 0.06). Virologic failure was observed in 21% of patients with concentrations ,3.9 mg/ml versus 5% with concentrations higher than 3.9 mg/ml (p = 0.015) ( Table 4).

B. Nevirapine Related Liver Toxicity
Of the 227 patients screened, 191 patients were qualified for the liver toxicity analysis (Figure 1). The incidence of liver toxicity was 71% (136/191), including 43 patients (23%) with severe liver toxicity within 12 weeks of starting therapy. Demographic characteristics of those with and without severe liver toxicity are summarized in Table 5. There were no statistically significant differences between these groups at baseline, except for the body weight.
Nevirapine C trough and C 2 in different groups are summarized in Table 6. There was no significant difference of C trough between these groups (median 5.91 vs. 5.54 mg/ml, p = 0.290). Previous  studies suggest that gender, hepatitis B or C virus co-infection and baseline liver function are independent risk factors for nevirapine related liver toxicity. Thus, the impact of gender, baseline liver function and presence of hepatitis C virus antibody on C trough were evaluated. A similar analysis for C2 was not possible due to the small number of C2 samples. A significantly higher median C trough was noted in males with severe liver toxicity in comparison to those without severe toxicity (8.20 vs. 5.48 mg/ml, p = 0.015). No significant differences were observed in C trough medians in other subgroups with and without severe hepatotoxicity (Table 6). Baseline CD4+ T cell count was not statistically associated with the incidence of severe liver toxicity among female Chinese patients, 29% (9/31) and 16% (10/63) in the patients with the baseline CD4+ T count .250/mm 3 , and #250/mm 3 , respectively (p = 0.184). The hepatitis C virus co-infection rate was 29% (56/191) in our study population. Active HCV co-infection, defined as HCV RNA .10 3 copies/ml, led to a significantly higher incidence of liver toxicity, 44% (10/23) compared to 18% (6/33) among patients without active HCV co-infection (p = 0.039, Table 7).

Discussion
This study found a significant correlation between nevirapine trough concentrations, virologic response and liver toxicity among Chinese HIV-infected patients. While patients with low nevirapine C trough have a significantly higher risk of virologic failure, those with a high C trough are at greater risk of severe liver toxicity, especially male patients. Our data suggest that for the optimal virologic outcomes, a trough nevirapine concentration of $3.9 mg/ml should be sustained throughout the treatment course for Chinese patients with HIV infection. To our knowledge, this is the first report of an association between nevirapine concentration and clinical outcomes in Chinese patient population. This adds to the evidence supporting the usefulness of therapeutic drug monitoring in optimizing treatment of this population.
The relationship between nevirapine concentrations and virologic response has been explored in previous studies, primarily including Caucasians and African Americans. The INCAS trial suggested that a nevirapine plasma concentration range of 3.45-3.88 mg/ml at week 12 was predictive of virologic success after 52 weeks of therapy [4]. A more recent clinical trial of 74 patients confirmed that patients with virologic success had higher concentrations than those with virologic failure (mean 6 SD, 4.5562.08 vs. 2.5761.64 mg/ml, p = 0.003) [3]. Our results in Chinese patients with HIV infection are consistent with the previous results obtained from clinical trials on Caucasians and Africans, higher nevirapine trough concentrations were associated with a lower rate of virologic failure. The median levels of C trough among Chinese patients with partial and complete response after a  Figure 3. Evaluation of nevirapine trough concentrations as a prognostic factor for virologic failure using ROC curve. The receiver operating characteristic (ROC) curve was utilized to evaluate nevirapine trough concentrations as a prognostic factor for virologic failure accommodating both sensitivity and specificity. A cut-off value of 3.9 mg/ml was determined for Chinese HIV+ patients with a significantly improved sensitivity of 60% in comparison to the original cut-off of 3.0 mg/ml (p,0.05,  6-month treatment were significantly higher than that among patients with viral failure ( Table 2). The regression analysis revealed a significant linear relationship between C trough and viral load ( Figure 2). However, no relationship between C 2 and treatment outcomes was detected in our study population.
No target trough concentration of nevirapine has been established for treatment success in Chinese patients. Although treatment guidelines recommend various target thresholds for nevirapine trough concentrations, these data are predominately from studies in Caucasian and African American patients, which may not be applicable to patients of Asian descent. Additionally, the current recommendations for nevirapine C trough thresholds are not completely in agreement. For instance, an observational cohort analysis [2] suggested 3.0 mg/ml as the minimal C trough because risk of virologic failure in patients with nevirapine concentrations #3 mg/ml was significantly increased (relative risk 5.0, 95% CI 1.8-13.7). Another study of 178 patients identified a mean plasma concentration of 4.3 mg/ml to be an independent factor for virologic failure [14]. Although the cut-off of 3.0 mg/ml was tentatively adapted to clinical practice in China according to the 2006 Chinese HIV Treatment Guidelines, this level does ensure treatment success as virologic failures often occur among patients with concentrations .3.0 mg/ml. This phenomenon is possibly explained by racial or ethnicity variability in the effective concentration needed to ensure therapeutic success and an underrepresentation of Chinese patients in previous studies. According to our data, the currently accepted threshold for nevirapine C trough (3 mg/ml) had an unfavorable sensitivity (40%) and may lead to a high incidence of virologic failure. The sensitivity would be improved to 60% without a marked decrease of specificity using 3.9 mg/ml. Therefore, to minimize the risk of virologic failure, we believe a more suitable nevirapine C trough threshold is 3.9 mg/ml in Chinese patients, a 30% increase from the current nevirapine C trough recommendation (3 mg/ml).
Liver toxicity is the most common adverse effect associated with nevirapine treatment while results from previous studies on the relationship of nevirapine C trough and liver toxicity are not in agreement. González de Requena et al [6] investigated the effect of nevirapine plasma exposure on liver enzyme elevations and observed that among patients with chronic hepatitis C coinfection, nevirapine concentrations .6 mg/ml were associated with a 92% risk of liver toxicity. Therefore, monitoring nevirapine levels, especially in individuals with chronic hepatitis C coinfection, might be warranted. A multivariate linear regression analysis [15] also revealed that high C trough at week 24 was an independent predictor for liver enzyme elevations. However, a prospective population pharmacokinetic study [5] and the 2NN study [11] suggested no correlation between nevirapine trough concentrations and liver enzyme levels, after corrected for known covariates such as gender, CD4 cell count at baseline, and hepatitis C coinfection.
Our data confirmed no significant differences in nevirapine C trough or C 2 between patients with and without severe liver toxicity. Several confounding factors such as gender, HCV coinfection and baseline liver function have been suggested for nevirapine exposure and liver toxicity [16,17], although no association between those factors and the occurrence of severe liver toxicity were detected in our patients. The nevirapine C trough Two patient groups, complete and partial response with viral load ,50 copies/ ml and 50-400 copies/ml, respectively and virologic failure with viral load greater than 400 copies/ml, were included for analysis. doi:10.1371/journal.pone.0026739.t004 was significantly higher in Chinese male patients with severe liver toxicity. However, such relationship was not observed in females and other subgroups, suggesting the nevirapine-related liver toxicity might be gender dependent. While previous studies [17,18] have demonstrated that female gender combined with high CD4+ T cell count significantly increased the incidence of nevirapine-induced liver toxicity, our data found a non-significant, modest increase of liver toxicity in female patients with high baseline CD4+ count (.250/mm 3 ) (29% vs. 16%, p = 0.184). This may be due to the relatively small sample size, along with racial and ethnic factors. Our data confirmed that hepatitis C virus co-infection was significantly associated with an increased risk of severe liver toxicity in Chinese HIV+ patients [19]. A prospective study [16] of 568 patients, primarily Caucasians and African Americans, also demonstrated a higher risk of severe hepatotoxicity among patients with chronic viral hepatitis C and B infections. However, such risk was remarkably reduced by concurrent interferon-based therapy with sustained HCV clearance in patients with HIV/HCV coinfection [20]. In the present study, Chinese HIV+ patients with high HCV RNA levels had a significantly greater risk of severe liver toxicity than those patients with low HCV RNA levels. This suggests that active HCV coinfection, correlated with HCV RNA levels, might be a more accurate predictor of nevirapine related liver toxicity than positive HCV antibody alone. Although the underlying mechanisms of this hepatotoxicity are not fully understood, an association between increased HCV-specific immune response and T cell activation have been suggested in previous studies [21]. Therefore, based on our results, nevirapine containing regimens should be avoided in Chinese patients with active HCV coinfection until sustained control of HCV is achieved.
The use of nevirapine peak concentration as a monitoring parameter was explored in this study, as limited information is available on this topic to date. Nevirapine C 2 was chosen to represent nevirapine peak concentration, based on previous pharmacokinetic results that indicate the mean time to reach nevirapine peak concentration was 2.01 hours [22]. Unfortunately, nevirapine C 2 had no correlation with any clinical outcomes evaluated in this study. This may be due to inter-patient variability, i.e. the exact peak time for nevirapine is different from patient to patient. Thus, it is difficult to collect samples that truly represent peak concentrations. Our data suggest that the clinical use of nevirapine peak concentrations may not be practical. Future studies using other metrics, such as area under the timeconcentration curve derived from population analysis, are warranted to explore the relationship between pharmacokinetics of nevirapine and clinical outcomes in Chinese HIV+ patients.
There were several limitations to this multi-centre prospective study in patients receiving nevirapine-containing therapy. The cohort was relatively small and heterogeneous, including patients with various underlying medical conditions. However, our patients reflected a typical Chinese HIV+ clinical cohort [1], making the results applicable to the study population. Selection bias was prevented by patient randomization. Because only trough and 2hour post dose concentrations were included in our study, other untested pharmacokinetic metrics including exposure and clearance may be of more clinical relevance. The strength of some of  the associations presented in this study should be interpreted with caution, owing to the relatively small sample size. Finally, unrecognized confounders may have influenced nevirapine concentrations, e.g. dosing in relation to food, concurrent medications, or genetic polymorphisms of CYP2B6, which significantly influenced nevirapine metabolism in HIV-infected patients in Uganda [19]. However, genotyping was not routinely available, and dose adjustments based on results of genotypic tests have not been proposed as yet for nevirapine.
In conclusion, this study supports the association between nevirapine concentrations and treatment outcomes in Chinese HIV+ patients. For patients with low trough concentrations, dose adjustments should be considered to achieve and sustain a median nevirapine trough concentration greater than 3.9 mg/ml. Given the inter-patient variability and the lack of an upper pharmacokinetic threshold for nevirapine associated hepatotoxicity, dose modification should be considered only in symptomatic patients. Finally, further studies are warranted, as a therapeutic range for nevirapine in Chinese HIV+ patients remains elusive.  (Figure 1). The original study design was published previously [1].

Study subjects
Nevirapine (Desano Pharma, Shanghai, China) was administered at 200 mg once daily for 2 weeks and 200 mg twice daily thereafter. The bioequivalence of nevirapine in reference to ViramuneH from Boehringer Ingelheim was demonstrated in the previous studies [23]. Medication adherence was monitored by self-report using validated surveys and directly observed therapy (DOT). Main exclusion criteria were acute HIV infection, AIDS-defining illness within 2 weeks of entry, white blood cell count ,2.0610 9 /L, absolute neutrophil count ,1.0610 9 /L, hemoglobin level ,90 g/l, platelet count ,0.75610 12 /L, transaminase and alkaline phosphatase level .3 times the upper limit of the normal range, bilirubin level .2.5 times the upper limit of the normal range, and serum creatinine level .1.5 times the upper limit of the normal range. Patients were followed up prior to starting therapy, then 1, 2, 3, 6, 9 and 12 months after the initiation of treatment. During all of the visits, the clinical assessment was performed and blood samples were taken for liver function and plasma viral load analyses.

Outcome measures
The primary efficacy outcome was virologic failure, defined as a single plasma HIV-1 RNA concentration (viral load) .400 copies/ml at week 24. Complete virologic response was defined as a viral load ,50 copies/ml at week 24, whereas partial virologic response was defined as a viral load between 50 and 400 copies/ml at week 24. The QUANTIPLEX HIV-1 RNA assay was used to measure the plasma HIV-1 RNA concentrations. The limits of detection of the assay, as indicated by the manufacturer, was 50,500,000 copies/ml.
All clinical adverse events and laboratory abnormal results were documented. Severe liver toxicity was defined as laboratory abnormalities greater than grade 3 (WHO, 1979), including a significant increase (.5 times of normal upper limit) of aspartate aminotransferase (AST), alanine aminotransferase (ALT) or total bilirubin, severe clinical symptoms consistent with liver toxicity, or a therapy change caused by liver toxicity during the first 12 months of treatment.

Plasma nevirapine concentration analyses
A plasma sample was collected at 2 h after oral administration of nevirapine (C 2 ) during DOT or prior to the morning dose for the detection of trough concentration (C trough , at least 12-h postdose). The nevirapine concentration in plasma was determined by a validated HPLC assay modified from a previous study [24]. The nevirapine concentration were analyzed on a Shim-pack CLC-ODS column (6 mmID615 cm, 5 mm) with a mobile phase consisting of water-acetonitrile (23:77) at a flow rate of 1 ml/min, and the wavelength for detection was 260 nm. Tegafur was used as an internal standard. The calibration curve of nevirapine was linear in the range of 0.05-10 mg/ml (r = 0.9999), and the limit of detection was 0.05 mg/ml. The RSDs of intra-and inter-run validations were less than 7%. The mean recoveries fell in the range of 90-110% for the high, middle and low concentrations. All plasma samples collected at the 6 th month of nevirapine treatment were assayed for concentrations. Additional plasma samples were collected for concentration assessment if severe liver toxicity presented. For patients without severe liver toxicity, the nevirapine plasma concentrations were determined during the 12month follow-up visit.

Statistical analyses
All statistical analyses were performed using the SPSS 11.5 statistical package. Categorical variables were tested with Chi-Square or Fisher's exact test, and continuous variables were tested with Kruskal-Wallis or student t-test. The cut-off value for the effective range of plasma drug concentration was determined by the receiver operating characteristic curve (ROC curve). For all tests, p,0.05 was considered to be statistically significant.