Once- or twice-daily dosing of nevirapine in HIV-infected adults: a ...

jac.oxfordjournals.org

Once- or twice-daily dosing of nevirapine in HIV-infected adults: a ...

Journal of Antimicrobial Chemotherapy (2008) 62, 784–792

doi:10.1093/jac/dkn268

Advance Access publication 30 June 2008

Once- or twice-daily dosing of nevirapine in HIV-infected

adults: a population pharmacokinetics approach

José Moltó 1,2 *, Marta Valle 2,3 , Cristina Miranda 1 , Samandhy Cedeño 4 , José Miranda 1 ,

José Ramón Santos 1 , Eugenia Negredo 1 , Josep Vilaró 5 , Joan Costa 6 and Bonaventura Clotet 2,4

1 ‘Lluita contra la SIDA’ Foundation, Hospital Universitari Germans Trias i Pujol, Badalona, Spain; 2 Universidad

Autónoma de Barcelona, Barcelona, Spain; 3 Centre d’Investigació del Medicament, Institut de Recerca de

l’Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; 4 ‘IrsiCaixa’ Foundation, Hospital Universitari

Germans Trias i Pujol, Badalona, Spain; 5 Department of Internal Medicine, Hospital de Vic, Vic, Spain;

6 Department of Clinical Pharmacology, Hospital Universitari Germans Trias i Pujol, Badalona, Spain

Introduction

Received 1 April 2008; returned 19 May 2008; revised 30 May 2008; accepted 3 June 2008

Objectives: The aim of this study was to develop and validate a population pharmacokinetic model for

nevirapine in a population of HIV-infected adults and to evaluate the influence of nevirapine dosing

regimen and patient characteristics on nevirapine trough concentration.

Methods: HIV-infected adults receiving oral nevirapine for at least 4 weeks were included. A

concentration–time profile was obtained for each patient, and nevirapine concentrations in plasma

were determined by HPLC. Pharmacokinetic parameters, inter-individual variability and residual error

were estimated using non-linear mixed effects modelling. The influence of patient characteristics on

the pharmacokinetics of nevirapine was explored, and the predictive performance of the final model

was evaluated in an external data set of observations.

Results: Totals of 40 and 18 Caucasian patients were included in two data sets for model building

and model validation, respectively. A mono-compartmental model with first-order absorption and elimination

best described the pharmacokinetics of nevirapine. Body weight influenced oral clearance

(CL/F) and volume of distribution (V/F). The estimated population pharmacokinetic parameters (interindividual

variability) for an individual weighing 70 kg were CL/F 2.95 L/h (24%), V/F 95.2 L (30%) and ka

1.8 h 21 (96%). The final model predicted nevirapine concentrations in the external model-validation

data set with no systematic bias and adequate precision. Bayesian estimates of nevirapine trough

concentrations were lower when nevirapine was administered once instead of twice daily, and nearly

half of the patients weighing 90 kg had drug concentrations 70 kg.

Keywords: antiretroviral agents, non-linear mixed effects models, suboptimal concentrations, body weight

Nevirapine is a non-nucleoside reverse transcriptase inhibitor

whose efficacy and safety in the treatment of HIV-infected

patients has been shown in several clinical trials. 1–3 The approved

dosage of nevirapine in HIV-infected adults is 200 mg twice

daily. 4 However, based on its long plasma half-life and on previous

pharmacokinetic data comparing nevirapine exposure under

once- or twice-daily nevirapine therapy, 5,6 once-daily dosing of

nevirapine (400 mg once daily) is also used in clinical practice.

.....................................................................................................................................................................................................................................................................................................................................................................................................................................

*Correspondence address. Fundació Lluita contra la SIDA, Hospital Universitari Germans Trias i Pujol, Ctra de Canyet, s/n, 08916

Badalona, Barcelona, Spain. Tel: þ34-93-497-88-87; Fax: þ34-93-465-76-02; E-mail: jmolto@flsida.org

.....................................................................................................................................................................................................................................................................................................................................................................................................................................

784

# The Author 2008. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.

For Permissions, please e-mail: journals.permissions@oxfordjournals.org

Downloaded from

http://jac.oxfordjournals.org/ by guest on March 1, 2013


Several studies have examined the relationship between

exposure to nevirapine and virological response. In a retrospective

analysis of the INCAS trial, 7 nevirapine concentrations

in plasma were related to the virological response in the short

(2 weeks) as well as in the long (52 weeks) term. In addition,

de Vries-Sluijs et al. 8 showed that the risk of virological failure

in patients who were receiving nevirapine-based antiretroviral

therapy increased 6-fold when nevirapine plasma concentrations

were ,3.0 mg/L compared with patients with higher concentrations.

As a result, current guidelines for therapeutic drug

monitoring of antiretroviral agents recommend 3.0 mg/L as the

cut-off value for nevirapine trough concentrations. 9 Nevertheless,

this cut-off has been questioned based on results observed in the

2NN study, where no evident relationship between nevirapine

concentrations and virological response was found. 10 On the other

hand, although some authors have proposed a role for drug

exposure in driving toxicity, 11 such a role has not been confirmed

in other studies and no nevirapine concentration cut-off value that

can predict drug toxicity has been clearly established to date. 12

The pharmacokinetic profile of nevirapine is characterized by

high oral bioavailability, moderate binding to plasma proteins

and prolonged elimination. 4,5 In addition, nevirapine concentrations

in plasma seem to vary greatly in HIV-infected individuals

in clinical practice. 13 Moreover, although daily exposure

(AUC 24) to nevirapine with the 400 mg once-daily regimen was

similar to that observed with the 200 mg twice-daily dosing

regimen in the study by van Heeswijk et al., 6 nevirapine trough

concentration was significantly lower for the once-daily

regimen. As a consequence, a significant proportion of patients

may not achieve high enough drug concentrations to effectively

suppress viral replication. Moreover, viral mutations conferring

high-level resistance to nevirapine and other non-nucleoside

reverse transcriptase inhibitors are easily selected when viral

replication occurs in the presence of the drug, 14 – 16 which

highlights the importance of maintaining optimal nevirapine

concentrations throughout the dosing interval.

Knowing individual factors affecting variability in nevirapine

pharmacokinetics could be very useful for achieving a desired

target drug concentration in each patient. Therefore, the objective

of the present study was to develop and validate a population

pharmacokinetic model for nevirapine and to identify characteristics

that might explain inter-individual variability in the pharmacokinetics

of nevirapine in a population of HIV-infected adults

routinely attending an outpatient HIV clinic. In addition, the influence

on nevirapine trough concentration of the dosing regimen

and changes in patient characteristics were assessed.

Methods

Population pharmacokinetics of nevirapine

Patients included in this cross-sectional study were HIV-infected

and aged 18 years or older. They had all been receiving stable

antiretroviral therapy with oral nevirapine (Viramune w , Boehringer,

Ingelheim GmbH, Ingelheim am Rheim, Germany) in routine clinical

practice for at least 4 weeks. Self-reported treatment adherence

,90% within the previous 2 weeks, active alcohol consumption

(.50 g/day) and concomitant administration of other drugs known

to affect nevirapine pharmacokinetics were considered exclusion

criteria. 4 The study was approved by the Ethics Committee of the

reference hospital (Hospital Universitari Germans Trias i Pujol), and

all subjects gave their written consent before enrolment.

785

Demographic and clinical variables, including age, sex, body

weight, height, time since HIV diagnosis, hepatitis C virus (HCV)

antibodies, hepatitis B virus surface antigen and concomitant medications

including over-the-counter drugs, were recorded for each

participating patient. In addition, a complete blood cell count,

analysis of serum chemistry (including creatinine, total protein,

albumin, total bilirubin, aspartate aminotransferase and alanine

aminotransferase), CD4+ T lymphocyte count and HIV-1 RNA load

were determined/performed.

Sample collection and analytical determination

A full concentration–time profile was obtained for each participating

patient, and blood samples were collected immediately before

(0), and 1, 2, 4, 6, 8, 10 and 12 h after a witnessed morning dose of

nevirapine. In addition, 62 samples were drawn between 7 and 24 h

after drug intake from eight patients who received nevirapine once

daily at night. Patients were asked to record the time they had last

taken a nevirapine dose on the day before the visit, and the exact

times of the nevirapine dose and blood sampling during the visit

were recorded.

Blood samples for the determination of drug plasma concentrations

were collected in potassium and ethylenediaminetetraacetic

acid-containing 10 mL tubes. Plasma was isolated by centrifugation

(4000 rpm for 15 min) and stored at 2208C until analysis.

Nevirapine concentrations were determined using 200 mL ofplasma

by HPLC with a photodiode array detector (2996 Waters, Barcelona,

Spain). The analytical column was a NovaPak C18 3.9 150 mm

with a NovaPak C18 column guard (Waters). The method involved

precipitation of proteins with 3% HClO4 (200 mL). The supernatant

was injected and the drug was resolved by isocratic elution with

phosphate-buffered acetonitrile containing 0.1% triethylamine (pH 6).

The method was linear over a concentration range of 0.1–10 mg/L.

The inter-day coefficients of variation for nevirapine concentrations

of 0.2, 3 and 7 mg/L were 8.7%, 1.8% and 3.1%, respectively. The

intra-day coefficients of variation for nevirapine concentrations of 0.2,

3 and 7 mg/L were 2.1%, 0.7% and 1.4%, respectively. The assay

was externally controlled by the International Interlaboratory Quality

Control Program for Therapeutic Drug Monitoring in HIV Infection

(KKGT, Nijmegen, The Netherlands). 17

Pharmacokinetic analysis

Non-linear mixed effects modelling was performed with the

computer program NONMEM (version V, Globomax LLC,

Hanover, MD, USA), 18 using a Fortran compiler (Compaq Visual

Fortran Version 6.0, Compaq Computer Corporation, Houston, TX,

USA). NONMEM uses mixed effects (fixed and random) regression

to estimate population means and variances of the pharmacokinetic

parameters and to identify factors that may influence them. All

models were fitted using the first-order conditional estimation

procedure with interaction between inter-individual and residual

variability. Relative standard error of the parameter estimates was

calculated using the COVARIANCE option in NONMEM. The

POST-HOC option was used to obtain individual predictions.

Structural model

Disposition characteristics of nevirapine were determined by fitting

mono- and bi-compartmental models, with linear and non-linear

elimination, to the data. To describe drug absorption, models

assuming either a first- or zero-order rate of absorption with and

without absorption lag-time were tested.

Downloaded from

http://jac.oxfordjournals.org/ by guest on March 1, 2013


Statistical model

Exponential errors following a log-normal distribution were assumed

for the description of inter-individual variability in pharmacokinetic

parameters, as shown in the following equation:

ui ¼ u expðh iÞ

where u i is the pharmacokinetic parameter of the ith individual, u is

the typical population value and h i is the inter-individual random

effect with a mean 0 and a variance v 2 .

The validity of the inter-individual variability model was assessed

by evaluating correlations between individual random effects (h) for

all the pharmacokinetic parameters. When substantial correlation was

present, covariance between these parameters was included in the

model.

Residual variability (reflecting the difference between the observed

and model-predicted concentrations) was modelled initially with a

combined error model. If one of the components (additive or proportional)

of the residual error was negligible, it was deleted from the

model.

Covariate model

Once a model providing an adequate description of the data without

the incorporation of covariates was selected, the influence of

patients’ characteristics on the variability of the pharmacokinetic

parameters was explored for significance by implementing the

generalized additive model (GAM) approach in the software Xpose,

version 3.1 (Uppsala University, Uppsala, Sweden). 19 This approach

enables straightforward display of the role of the various covariates.

It uses a stepwise method, where each covariate is introduced in the

model through specified linear or non-linear representations. At each

step, the model is improved by the addition of the single term that

results in the largest decrease in the Akaike information criterion

(AIC). The search stops when AIC reaches a minimum value. The

covariates initially selected during the GAM analysis were further

tested for significance in NONMEM, using the forward inclusion

and backward elimination approach.

Selection between models was based on the precision of the

parameter estimates, goodness-of-fit plots and the minimum value

of the objective function (OF) provided by NONMEM. A covariate

was introduced into the model when its inclusion decreased the OF

by at least 3.84 points (P , 0.05) and reduced the inter-patient

variability in the parameters. During the backward elimination

procedure, a covariate was only retained in the model when its

influence was statistically significant (P , 0.01). An OF with an

approximate x 2 distribution with 1 degree of freedom was the basis

for these criteria.

Model validation

The final model was internally validated by means of Monte Carlo

simulations. One thousand individual concentration–time profiles were

generated for individuals receiving nevirapine 200 mg twice daily or

400 mg once daily using the fixed and random population estimates

obtained from the model. The median profile and the intervals including

90% of the simulated concentrations of nevirapine were plotted

together with the raw data. The agreement between simulations and

observations was judged visually, and the proportion of nevirapine

concentrations outside the 90% prediction interval was calculated.

External validation was performed for the final model based on

qualitative and quantitative evaluations of the predictive performance

Moltó et al.

786

in an external data set of observations collected in another study

where a subset of patients received nevirapine-based antiretroviral

therapy (model-validation data set). 20 The fixed and random estimates

obtained from the models were used to predict individual nevirapine

concentrations, which were plotted and compared with actual concentration

values. Mean prediction error and root mean squared error,

and their 95% confidence intervals (CIs), were calculated as a

measure of bias and precision, respectively. 21

Influence of nevirapine dosing regimen and covariates on

nevirapine trough concentrations

Three thousand individual concentration–time profiles were generated

for each nevirapine dose (200 mg twice daily or 400 mg once daily);

one thousand for individuals with a covariate value corresponding to

the median, one thousand for patients on the 10th percentile of the

range of the covariable and one thousand for patients on the 90th

percentile of the range of the covariable. The median profile of the

simulated concentrations for each category was plotted together.

Mean predicted nevirapine trough concentrations and the proportion

of patients with nevirapine trough concentrations ,3.0 mg/L were

compared between different categories using the t-test and the x 2 test,

respectively.

Results

A total of 319 nevirapine observations from 40 Caucasian patients

was included in the model-building data set. A median (range) of

8 (8–15) samples per subject were available. Nevirapine dose

was 200 mg twice daily in 14 patients, 400 mg once daily in the

morning in 18 patients and 400 mg once daily at night in the

remaining 8 patients. Table 1 summarizes the characteristics of

the patients studied.

Nevirapine concentrations ranged between 2.2 and 10.3 mg/L

[coefficient of variation (CV) 36%] in patients receiving 200 mg

of nevirapine twice daily, and from 2.5 to 16.1 mg/L (CV 32%)

in patients receiving 400 mg of nevirapine once daily (Figure 1).

A mono-compartmental model with first-order absorption

and elimination best described the population pharmacokinetics

of nevirapine. Models with zero-order absorption and two compartments

were also studied but did not adequately fit the data.

The introduction of an absorption lag-time did not improve the

pharmacokinetic model (DOF 23.3). In addition, although the

introduction of inter-individual variability in bioavailability

significantly decreased the OF (DOF 27.1), it increased the

relative standard error of the pharmacokinetic estimates and did

not improve agreement between observed and predicted data.

Finally, as the additive component of the residual error was negligible,

and the OF remained unchanged when that component

was deleted from the model, only the proportional component of

the residual error was retained in the final model. Table 2 lists

the fixed and random effects estimated in the model without

covariates.

The model-building steps for the covariate analysis are summarized

in Table 3. None of the covariates except body weight

improved the model when they were included in the pharmacokinetic

parameters of nevirapine. Inclusion of body weight in

oral clearance (CL/F) resulted in a significant improvement of

goodness of fit (DOF 236.7). Similarly, the introduction of

body weight in volume of distribution (V/F) improved the fit

Downloaded from

http://jac.oxfordjournals.org/ by guest on March 1, 2013


Population pharmacokinetics of nevirapine

Table 1. Demographic characteristics of the patients included in the model-building and in

the model-validation data sets

(DOF 29.9) and reduced the inter-individual variability of V/F

by 13.8%. Finally, when body weight was included in both

CL/F and V/F, there was an additional improvement in goodness

of fit (DOF 29.7). The parameter estimates for the final model

are given in Table 2.

Model validation

Figure 1 shows the average population prediction with the 90%

prediction interval for each nevirapine dosing regimen, and

Figure 2 depicts the overall goodness of fit of the final pharmacokinetic

model to the data. The prediction interval obtained

with the final model encompassed the observed concentration–

time data adequately. The proportion of nevirapine concentrations

outside the 90% prediction interval was 9.2% in patients

receiving 400 mg of nevirapine once daily, and 11.4% in

patients receiving 200 mg of nevirapine once daily. Individual

predictions of nevirapine trough concentrations were ,3 mg/L

in all six patients who had suboptimal trough concentrations.

Moreover, the final model only predicted suboptimal nevirapine

concentrations in 1 of the 34 patients whose actual nevirapine

concentrations were .3 mg/L (3%).

External validation of the model was performed in a total of

88 observations from 18 patients, which were included in the

Model building n = 40 Model validation n =18

Sex (male) a

31 (77.5) 12 (67.0)

Age (years) 50.8 (23.0) 40.7 (9.4)

Weight (kg) 67.9 (12.0) 69.1 (12.6)

Height (m) 1.67 (0.09) 1.67 (0.07)

Time since HIV diagnosis (years) 9.1 (4.8) 8.4 (4.3)

HCV/HBV co-infection a

3 (7.5) 4 (22.0)

Time on nevirapine therapy (weeks)

NVP dosing regimen

141.3 (122.3) 213.1 (103.2)

a

200 mg bid 14 (35.0) 14 (77.8)

400 mg qd morning 18 (45.0) 2 (11.1)

400 mg qd night

Reverse transcriptase inhibitors

8 (20.0) 2 (11.1)

a

zidovudine 3 (7.5) 0

lamivudine 31 (77.5) 14 (77.8)

stavudine 1 (2.5) 0

didanosine 15 (37.5) 2 (11.1)

abacavir 9 (22.5) 7 (38.9)

tenofovir 15 (37.5) 17 (94.4)

AST (IU/L) 31.3 (13.0) 48.2 (51.6)

ALT (IU/L) 41.3 (21.6) 57.4 (52.3)

Proteins (g/dL) 7.5 (0.5) 6.5 (2.5)

Albumin (g/dL) 4.4 (0.8) 4.1 (1.1)

CD4+ T cell count (cels/mm 3 ) 566 (245) 624 (269)

HIV-1 RNA ,50 copies/mL a

32 (80.0) 18 (100)

AST, aspartate aminotransferase; ALT, alanine aminotransferase; HBV, hepatitis B virus (measured as

the presence of viral surface antigen); HCV, hepatitis C virus; NVP, nevirapine; bid, twice daily; qd,

once daily.

Data are expressed as mean (SD) except when noted.

a n (%).

787

model-validation data set (Table 1). Individual nevirapine predictions

versus actual concentrations in the model-validation data set

are shown in Figure 3. The performance of the individual predictions

was satisfying, and mean (95% CI) bias and precision were

3.4% (20.5 to 6.3) and 8.4% (5.9–10.7), respectively.

Influence of nevirapine dosing regimen and covariates on

nevirapine trough concentrations

In order to elucidate their possible clinical relevance, the influence

of nevirapine dosing regimen and patients’ body weight

on the population prediction of nevirapine concentrations was

further investigated. Figure 4 shows steady-state nevirapine

concentration profiles predicted for typical individuals weighing

50, 70 and 90 kg and undergoing antiretroviral therapy with

nevirapine at a 200 mg twice-daily or at a 400 mg once-daily

dosing regimen.

Overall, the mean (SD) predicted nevirapine trough concentration

was 5.35 (1.97) mg/L in patients receiving 200 mg of

nevirapine twice daily, compared with 4.33 (1.80) mg/L in

patients receiving 400 mg of nevirapine once daily (P , 0.01).

In addition, higher body weight was associated with a lower

predicted nevirapine concentration. Consequently, the proportion

of individuals with predicted nevirapine trough concentrations

Downloaded from

http://jac.oxfordjournals.org/ by guest on March 1, 2013


(a)

NVP concentrations (mg/L)

(b)

NVP concentrations (mg/L)

15

10

5

15

10

5

0

5

Time (h)

0 5 10

Time (h)

15 20 25

Figure 1. Nevirapine (NVP) plasma concentrations in patients receiving

a dose of 200 mg twice daily (a; n = 112) or 400 mg once daily (b; n = 207).

The mean population prediction (continuous line) and the 90% prediction

interval (broken lines) are also shown.

,3.0 mg/L was 0.5%, 5.1% and 22.6% when nevirapine was

dosed at 200 mg twice daily to individuals weighing 50, 70 or

90 kg, respectively, compared with 6.6%, 23.5% and 49.2%

10

Moltó et al.

when nevirapine was dosed at 400 mg once daily to individuals

weighing 50, 70 or 90 kg (P , 0.01), respectively.

Discussion

Table 2. Final parameter estimates of the basic and final pharmacokinetic models

Basic model

estimate RSE (%)

CL/F (L/h) 2.85 4.8 2.95 a

V/F (L) 87.5 7.5 95.2 a

A population pharmacokinetic model for nevirapine was developed

and validated in the present study. Our major findings were: (i) the

increase in nevirapine V/F and CL/F as body weight increases; and

(ii) the higher risk of showing nevirapine trough concentrations

,3.0 mg/L with once- rather than twice-daily dosing regimens,

particularly in patients weighing .70 kg.

Our results are consistent with previous descriptions of the

pharmacokinetics of nevirapine. 22–25 The model that best

described nevirapine behaviour was a mono-compartmental

model with first-order absorption and elimination, and estimates

of the pharmacokinetic parameters were in agreement with values

published in the previously cited studies. However, the interindividual

variability in the absorption rate constant (k a)inour

study was greater than that reported previously. 22,24,25 This finding

can be explained by the fact that no blood samples were drawn

until an hour after nevirapine administration, precluding proper

estimation of k a. Body weight was the only covariate that was

significantly related to the pharmacokinetic parameters of nevirapine

in our study. Both CL/F and V/F of nevirapine increased, by

0.4 L/h and 13.7 L, respectively, with body weight increases of

10 kg. In contrast, we did not find any relationship between nevirapine

pharmacokinetic parameters and other covariates previously

described by other authors, such as gender, HCV co-infection or

elevated liver enzyme levels. 22–25 Although women showed a

14% lower nevirapine CL/F than did men in the model developed

by Kappelhoff et al., 23 this relationship was not confirmed by the

present or other studies. 22,24 Similarly, we did not observe a

relationship between HCV co-infection or abnormal liver enzyme

levels and nevirapine CL/F, probably because only three participants

in our study were co-infected with HCV.

As mentioned above, the possible mis-specification of the

absorption phase potentially represented a source of bias in this

study, resulting in an overestimation of the variability associated

Final model

estimate RSE (%)

6.5

ka (h 21 ) 1.67 22 1.8 22

Inter-individual

variability CL/F (%)

29 20 24 22

Inter-individual

variability V/F (%)

36 35 30 42

Inter-individual

variability ka (%)

90 41 96 41

Residual error (%) 8.9 16 8.4 16

CL/F, oral clearance; V/F, apparent volume of distribution; ka, absorption rate constant; BW, body

weight; RSE, relative standard error (as calculated with the COVARIANCE option of NONMEM).

Final pharmacokinetic models: CL/F = u 1 + u 2 .BW; V/F = u 3 + u 4 .BW; k a = u 5.

a Expressed for an individual weighing 70 kg.

788

4.0

Downloaded from

http://jac.oxfordjournals.org/ by guest on March 1, 2013


Table 3. Summary of the models used in NONMEM to examine the influence of patients’ covariates selected by GAM analysis on

nevirapine pharmacokinetics

Hypothesis Model ua ub DOF P

Does AGE influence CL? CL = u a 2 (u b .AGE) 3.97 0.02 23.6 0.05

Does BW influence CL? CL = ua + u .BW b 0 0.04 236.7 ,0.001

(2ub

Does ALT influence CL? CL = u .e

ALT)

a 2.91 3.7 10 210

0 NS

(2ub

Does ALB influence CL? CL = u .e

ALB)

a 6.3 0.17 24.0 0.05

Does BW influence V? V = ua + u .BW b 0 1.36 29.9 0.005

Does ALB influence V? V = ua 2 (u .ALB) b 156 14.9 20.9 NS

Does BW influence ka? ka = u .BW a 0.03 22.1 NS

u a, typical value for the parameter; u b, factor associated with the covariate; DOF, difference in the NONMEM objective function compared with the basic

structural model including no covariates; AGE, age (years); BW, body weight (kg); ALT, alanine transferase (IU/L); ALB, albumin (g/dL); ka, absorption rate

constant; NS, not significant.

(a)

2 4 6 8 10 12 14 16

Observations (mg/L)

2 4 6 8 10

Population predictions (mg/L)

with other parameters in the model or in a higher residual error,

which might finally affect the concentration–time profile predicted

by the model. Thus, the potential applicability of our

model in clinical practice made the validation step of particular

importance. The model correctly identified all the participating

patients whose nevirapine trough concentrations were ,3 mg/L,

and only misclassified ,5% of the patients who actually had

nevirapine trough concentrations .3 mg/L. Moreover, it predicted

nevirapine concentrations appropriately in an external set

of patients who were not included during the model-building

step, without evidence of systematic bias in individual model

(b)

Observations (mg/L)

2 4 6 8 10 12 14 16

(c) (d)

Absolute individual

weighted residuals

0.0 0.1 0.2 0.3 0.4 0.5

Population pharmacokinetics of nevirapine

Weighted residuals

–4 –2 0 2 4

2 4 6 8 10 12 14 16

Individual predictions (mg/L)

2 4 6 8 10 12 14 16

0 5 10 15 20 25

Individual predictions (mg/L)

Time post-dose (h)

Figure 2. Goodness-of-fit plots of the final model for nevirapine. Thick grey lines are the smooths of the ordinate values. (a) Observed concentrations versus

population predictions; the thin black line is the line of identity. (b) Observations versus individual predictions; the thin black line is the line of identity. (c) Absolute

individual weighted residuals versus individual predictions. (d) Population weighted residuals versus time post-dose; the thin black line is at ordinate value zero.

789

predictions. However, the abovementioned limitation regarding

the low frequency of HCV co-infected patients in our study as

well as the applicability of our results to other populations who

might differ in genetic variants of the isoenzymes involved in

nevirapine distribution and metabolism should be considered.

Antiretroviral treatment failure is a complex phenomenon in

which several factors may play a role. Treatment adherence is

crucial for attaining sustained viral suppression, and once-daily

administration of antiretroviral agents to improve treatment

convenience is becoming more widely used. Besides, the

importance of maintaining drug concentrations high enough to

Downloaded from

http://jac.oxfordjournals.org/ by guest on March 1, 2013


Observations (mg/L)

2 4 6 8 10 12

2 4 6 8 10 12

Individual predictions (mg/L)

Figure 3. Goodness-of-fit plots of the validation model for nevirapine. The

thick grey line is the smooth of the observed concentrations, and the thin

black line is the line of identity.

NVP concentrations (mg/L)

NVP concentrations (mg/L)

NVP concentrations (mg/L)

15

10

5

0

15

10

5

0

15

10

5

0

NVP 400 mg qd

0 5 10

Time (h)

15 20 25

0 5 10

Time (h)

15 20 25

90 kg

0 5 10

Time (h)

15 20 25

Moltó et al.

50 kg

70 kg

effectively suppress viral replication during the entire dosing

interval has been pointed out in several studies. 14,26 – 28 Although

nevirapine was originally designed for twice-daily administration,

its long plasma half-life has suggested the possibility of

once-daily dosing. However, potential benefits derived from

once- instead of twice-daily administration may be tempered by

pharmacokinetic issues. Consistent with the findings of van

Heeswijk et al., 6 estimates of nevirapine trough concentrations

predicted using a population pharmacokinetic approach in our

study were significantly lower when nevirapine was administered

once instead of twice daily. Moreover, nevirapine trough concentrations

were ,3.0 mg/L in nearly half of the patients when body

weight was 90 kg and nevirapine was given once daily. These

nevirapine concentrations may be considered subtherapeutic, 9

suggesting that twice- instead of once-daily nevirapine dosing

would be more appropriate in this group of patients. Although

the clinical relevance of this cut-off in nevirapine trough concentration

has been questioned, and the sensitivity of this value

was too low to be an adequate predictor for virological failure

NVP concentrations (mg/L)

NVP concentrations (mg/L)

NVP concentrations (mg/L)

15

10

5

0

15

10

5

0

15

10

5

0

NVP 200 mg bid

50 kg

–2 0 2 4 6 8 10 12

Time (h)

70 kg

–2 0 2 4 6 8 10 12

Time (h)

90 kg

–2 0 2 4 6 8 10 12

Time (h)

Figure 4. Steady-state nevirapine (NVP) concentrations predicted after receiving NVP at a 400 mg once-daily (left-hand panels) or at a 200 mg twice-daily

(right-hand panels) dosing regimen for patients weighing 50, 70 or 90 kg. The mean population prediction (continuous line) and the 90% prediction interval

(grey area) are represented for each category. The broken horizontal line is at an ordinate value of 3.0 mg/L.

790

Downloaded from

http://jac.oxfordjournals.org/ by guest on March 1, 2013


in the study by Leth et al., 10 there does seem to be a higher

risk of virological failure in the presence of low nevirapine concentrations

according to the findings of that study and other

studies. 7,8,14

In conclusion, a population model to describe the pharmacokinetics

of nevirapine was developed and validated in HIVinfected

patients. Both nevirapine CL/F and V/F were found to be

influenced by body weight, which decreased inter-individual

variability in these parameters. Based on these results, twiceinstead

of once-daily administration of nevirapine seems to be

more advisable for patients weighing .70 kg.

Acknowledgements

We thank staff at the clinical sites that collaborated in this study

and the patients who participated. We wish to acknowledge the

contribution of Mary Ellen Kerans who gave her assistance

with English language expression in the final version of the

manuscript.

Funding

The study was supported by a grant from ‘Lluita Contra La

SIDA’ Foundation and by the Spanish AIDS network ‘Red

Temática Cooperativa de Investigación en SIDA’ (RD06/0006).

Transparency declarations

J. Moltó received honoraria for speaking engagements and participating

on advisory boards from Abbott, Bristol-Myers Squibb,

Boehringer-Ingelheim, Gilead Sciences, GlaxoSmithKline, Pfizer

and Roche. M. V. is supported by FIS trough grant CP04/00121

from the Spanish Health Department in collaboration with Institut

de Recerca de l’Hospital de la Santa Creu i Sant Pau, Barcelona,

and is a member of CIBERSAM (supported by the Spanish

Ministry of Health, Carlos III Health Institute). J. R. S. is supported

by the Fundació Lluita contra la SIDA from Hospital

Germans Trias i Pujol, Badalona. E. N. received honoraria for

speaking and participation on advisory boards from Abbott,

Bristol-Myers Squibb, Boehringer-Ingelheim, Gilead Sciences,

GlaxoSmithKline, Pfizer and Roche. B. C. received honoraria

for speaking and participation on advisory boards from Abbott,

Bristol-Myers Squibb, Boehringer-Ingelheim, Gilead Sciences,

GlaxoSmithKline, Pfizer and Roche. The rest of the authors

have had no conflicts to declare during the course of this study.

References

Population pharmacokinetics of nevirapine

1. Montaner JS, Reiss P, Cooper D et al. A randomized, double-blind

trial comparing combinations of nevirapine, didanosine, and zidovudine

for HIV-infected patients: the INCAS Trial. Italy, The Netherlands,

Canada and Australia Study. JAMA 1998; 279: 930–7.

2. Raffi F, Reliquet V, Ferre V et al. The VIRGO study: nevirapine,

didanosine and stavudine combination therapy in antiretroviral-naive

HIV-1-infected adults. Antivir Ther 2000; 5: 267–72.

3. van Leeuwen R, Katlama C, Murphy RL et al. A randomized trial

to study first-line combination therapy with or without a protease inhibitor

791

in HIV-1-infected patients. Acquir Immune Defic Syndr 2003; 17:

987–99.

4. European Medicines Agency. EMEA label [on line]. Accessed

March 2008. www.emea.europa.eu/humandoc/Humans/EPAR/viramune/

riramune.htm (22 May 2008, date last accessed).

5. Murphy R, Montaner J. Nevirapine: a review of its development,

pharmacological profile and potential for clinical use. Expert Opin

Investig Drugs 1996; 5: 1183–99.

6. van Heeswijk RP, Veldkamp AI, Mulder JW et al. The steadystate

pharmacokinetics of nevirapine during once daily and twice daily

dosing in HIV-1-infected individuals. Acquir Immune Defic Syndr 2000;

14: F77–82.

7. Veldkamp AI, Weverling GJ, Lange JM et al. High exposure to

nevirapine in plasma is associated with an improved virological

response in HIV-1-infected individuals. Acquir Immune Defic Syndr

2001; 15: 1089–95.

8. de Vries-Sluijs TE, Dieleman JP, Arts D et al. Low nevirapine

plasma concentrations predict virological failure in an unselected

HIV-1-infected population. Clin Pharmacokinet 2003; 42: 599–605.

9. la Porte C, Back D, Blaschke T et al. Updated guideline to

perform therapeutic drug monitoring for antiretroviral agents. Rev

Antivir Ther 2006; 3: 4–14.

10. Leth FV, Kappelhoff BS, Johnson D et al. Pharmacokinetic

parameters of nevirapine and efavirenz in relation to antiretroviral

efficacy. AIDS Res Hum Retroviruses 2006; 22: 232–9.

11. Gonzalez de Requena D, Nunez M, Jimenez-Nacher I et al.

Liver toxicity caused by nevirapine. Acquir Immune Defic Syndr 2002; 16:

290–1.

12. Kappelhoff BS, van Leth F, Robinson PA et al. Are adverse

events of nevirapine and efavirenz related to plasma concentrations?

Antivir Ther 2005; 10: 489–98.

13. Molto J, Blanco A, Miranda C et al. Variability in non-nucleoside

reverse transcriptase and protease inhibitor concentrations among

HIV-infected adults in routine clinical practice. Br J Clin Pharmacol

2006; 62: 560–6.

14. Gonzalez de Requena D, Bonora S, Garazzino S et al.

Nevirapine plasma exposure affects both durability of viral suppression

and selection of nevirapine primary resistance mutations in a clinical

setting. Antimicrob Agents Chemother 2005; 49: 3966–9.

15. 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 1995; 171:

537–45.

16. Shapiro RL, Thior I, Gilbert PB et al. Maternal single-dose nevirapine

versus placebo as part of an antiretroviral strategy to prevent

mother-to-child HIV transmission in Botswana. Acquir Immune Defic

Syndr 2006; 20: 1281–8.

17. Droste JA, Aarnoutse RE, Koopmans PP et al. Evaluation of

antiretroviral drug measurements by an interlaboratory quality control

program. J Acquir Immune Defic Syndr 2003; 32: 287–91.

18. Beal S, Sheiner L. NONMEM Users Guides. Ellicott City, MD,

USA: Icon Development solutions, 1989–98.

19. Jonsson EN, Karlsson MO. Xpose—an S-PLUS based population

pharmacokinetic/pharmacodynamic model building aid for NONMEM.

Comput Methods Programs Biomed 1999; 58: 51–64.

20. Pruvost A, Negredo E, Grassi J et al. A pharmacokinetic study

in HIV-infected patient under tenofovir disoproxil fumarate (TDF):

investigation of systemic and intracellular interaction between TDF and

abacavir, or lamivudine or lopinavir/ritonavir. In: Abstracts of the Eighth

International Workshop on Clinical Pharmacology of HIV Therapy,

Budapest, 2007. Abstract 56.

21. Sheiner L, Beal S. Some suggestions for measuring predictive

performance. J Pharmacokin Biopharm 1981; 9: 503–12.

Downloaded from

http://jac.oxfordjournals.org/ by guest on March 1, 2013


22. de Maat MM, Huitema AD, Mulder JW et al. Population pharmacokinetics

of nevirapine in an unselected cohort of HIV-1-infected

individuals. Br J Clin Pharmacol 2002; 54: 378–85.

23. Kappelhoff BS, van Leth F, MacGregor TR et al. Nevirapine and

efavirenz pharmacokinetics and covariate analysis in the 2NN study.

Antivir Ther 2005; 10: 145–55.

24. Sabo JP, Lamson MJ, Leitz G et al. Pharmacokinetics of nevirapine

and lamivudine in patients with HIV-1 infection. AAPS PharmSci

2000; 2: E1.

25. Zhou XJ, Sheiner LB, D’Aquila RT et al. Population

pharmacokinetics of nevirapine, zidovudine, and didanosine in

human immunodeficiency virus-infected patients. The National

Institute of Allergy and Infectious Diseases AIDS Clinical Trials

Moltó et al.

792

Group Protocol 241 Investigators. Antimicrob Agents Chemother

1999; 43: 121–8.

26. Burger DM, Hoetelmans RM, Hugen PW et al. Low plasma

concentrations of indinavir are related to virological treatment failure

in HIV-1-infected patients on indinavir-containing triple therapy. Antivir

Ther 1998; 3: 215–20.

27. Burger DM, Hugen PW, Aarnoutse RE et al. Treatment failure of

nelfinavir-containing triple therapy can largely be explained by low nelfinavir

plasma concentrations. Ther Drug Monit 2003; 25: 73–80.

28. Marzolini C, Telenti A, Decosterd LA et al. Efavirenz plasma

levels can predict treatment failure and central nervous system side

effects in HIV-1-infected patients. Acquir Immune Defic Syndr 2001;

15: 71–5.

Downloaded from

http://jac.oxfordjournals.org/ by guest on March 1, 2013

More magazines by this user
Similar magazines