Variability in non-nucleoside reverse transcriptase and protease inhibitors concentrations among HIV-infected adults in routine clinical practice

Abstract

Aims

The objective of this study was to assess interindividual variability in plasma trough concentrations of non-nucleoside reverse transcriptase inhibitors (NNRTI) and protease inhibitors (PI) among HIV-infected adults in an outpatient routine clinical practice setting.

Methods

The study included 117 patients who attended our clinic for routine outpatient blood tests and who were receiving antiretroviral therapy which included NNRTI or PI. Patients were not informed that drug concentrations were going to be assessed until blood sampling. The time of the last antiretroviral treatment intake and blood sampling were recorded. Drug concentrations were considered optimal if they were above the proposed minimum effective concentration. In addition, efavirenz, nevirapine and atazanavir concentrations were considered potentially toxic if they were higher than 4.0 mg l⁻¹, 6.0 mg l⁻¹, and 0.85 mg l⁻¹, respectively.

Results

Overall, interindividual variability in NNRTI and PI plasma concentrations was approximately 50%, and only 68.4% of the patients had drug concentrations within the proposed therapeutic range. Inappropriate adherence only explained 35% of subtherapeutic drug concentrations.

Conclusion

Interindividual variability in trough concentrations of NNRTI and PI among HIV-infected adults is large in routine clinical practice, with drug concentrations being outside the therapeutic window in a significant proportion of patients. Therapeutic drug monitoring may be useful to guide antiretroviral therapy in clinical practice.

Introduction

Current highly active antiretroviral therapy (HAART) can successfully suppress HIV replication, and its widespread use has resulted in a marked decrease in HIV-related morbidity and mortality [1]. However, despite the initial favourable response to HAART, it fails to maintain complete viral suppression in the long term in a significant proportion of HIV-infected subjects [2–4]. Treatment failure may be due to poor adherence on the part of the patients, the development of viral resistance, or pharmacokinetic issues of the antiretroviral agents in use.

There is growing evidence on the relationship between non-nucleoside reverse transcriptase inhibitor (NNRTI) and protease inhibitor (PI) concentrations in plasma and their efficacy and toxicity [5]. In this regard, small decreases in plasma concentrations of antiretroviral drugs can render these concentrations insufficient to maintain complete viral suppression, promoting treatment failure and evolution of viral resistance. On the other hand, excessively high drug concentrations may contribute, at least in part, to the appearance of antiretroviral therapy-related adverse events [6–18].

Marked differences in NNRTI and PI concentrations in plasma have been reported among HIV-infected subjects [5, 13, 15, 19, 20]. Moreover, a large proportion of patients had PI trough concentrations lower than the proposed minimum effective concentration (MEC) in previous studies [19–22]. This interindividual variability in drug concentrations may be explained by differences in drug absorption, distribution, metabolism and elimination among different subjects. In addition, treatment adherence, concomitant diseases, and drug-drug or food-drug interactions may further enhance this variability.

Most of the currently available information regarding interindividual variability in the concentration of antiretroviral agents comes from clinical trials and therapeutic drug monitoring units. However, results obtained in such scenarios may substantially differ from those observed in routine clinical practice with unselected patient populations. Thus, the objective of the present study was to assess the interindividual variability in NNRTI and PI trough concentrations in plasma among HIV-infected adults in an outpatient routine clinical practice setting.

Methods

Patients included in this study were HIV-infected subjects aged 18 years or older who consecutively attended our clinic during a 2 week period for routine outpatient blood tests. To be eligible, patients had to be receiving an NNRTI- or a PI-based HAART for at least 4 weeks. Treatment with more than one PI (ritonavir was not considered as a second PI when given at boosting doses) or with combinations of PI with NNRTI was considered as an exclusion criteria. Complete information about the study was supplied to every subject, and informed consent was obtained from all patients at the time of blood sampling.

The primary endpoint of the study was to assess the interindividual variability in NNRTI and PI plasma trough concentrations among HIV-infected adults. In addition, the relationship between drug concentrations and virological outcome or drug-related toxicity was explored as a secondary endpoint.

Demographic and clinical variables including age, sex, time since HIV diagnosis, hepatitis C virus co-infection, concomitant medications, treatment adherence, HIV-1 RNA load evolution and drug-related toxicity were recorded for each subject. Treatment adherence was self-reported by the patients as the number of doses taken among those prescribed within the week before the sample collection. Viral load was recorded at the time of sampling and every 12 weeks thereafter until 48 weeks of follow-up. Virological failure was defined as the presence of HIV-1 RNA load above 50 copies ml⁻¹ in two consecutive determinations after at least 24 weeks on HAART or if previously undetectable. Specific drug-related toxicities were recorded in those patients on treatment with drugs for which an upper limit of the therapeutic window has been proposed. Thus, hypersensitivity reactions or liver enzyme elevations were registered in patients taking nevirapine, central nervous system disorders in those taking efavirenz, or total bilirubin or jaundice in those patients receiving atazanavir.

In an attempt to avoid modifying usual adherence to antiretroviral therapy, subjects were not informed that drug concentrations were going to be measured until blood sampling. The time of the last antiretroviral treatment intake and blood sampling were recorded for each patient. Because trough concentrations have been proposed as the most suitable pharmacokinetic parameter for routine clinical monitoring [5], only blood samples collected from 10 to 13 h or from 21 to 25 h after the last treatment intake were considered in patients taking antiretroviral drugs twice or once daily, respectively. Nevertheless, in the case of patients taking efavirenz once daily at bedtime, samples were collected during the day, at least 8 h after the last drug intake.

Blood samples were collected into 10 ml tubes containing potassium and ethylenediaminetetraaceticacid (EDTA). Plasma was isolated by centrifugation (4000 rev min⁻¹ for 15 min), and stored at −20°C until analysis. Drug concentrations were determined by high-performance liquid chromatography (HPLC) with PDA detector (2996 Waters, Barcelona, Spain). Concentrations of nevirapine were determined using a NovaPak® C18 3.9 × 150 mm analytical column with a NovaPak® C18 guard column (Waters, Barcelona, Spain). The method involved precipitation of proteins with HClO₄. The supernatant was injected and the drug was resolved by isocratic elution with phosphate buffer acetonitrile containing 0.1% triethylammine (pH 6). The method was linear over a concentration range of 0.1–10 mg l⁻¹. Efavirenz concentrations were determined involving solid-phase extraction of the drug, according to the method described by Sarasa-Nacenta et al. [23]. The analytical column was NovaPak® C8, 4.6 × 150 mm, and the guard column was NovaPak® C8 (Waters, Barcelona, Spain). The method was validated over the range of 0.1–10 mg l⁻¹. Concentrations of lopinavir, nelfinavir, saquinavir, amprenavir, atazanavir and indinavir were quantified simultaneously, according to a validated method. The analytical column was NovaPak® C18 3.9 × 150 mm with a guard column of NovaPak® C18 (Waters, Barcelona, Spain). The method involved liquid-liquid extraction of the six drugs from plasma with tert-butyl methyl ether after basification and a second wash with hexane. The mobile phase consisted of a gradient elution with phosphate buffer acetonitrile (pH 6.7). The method was linear over the range of 0.05–20 mg l⁻¹ for lopinavir and amprenavir, 0.042–17 mg l⁻¹ for nelfinavir, 0.044–17.5 mg l⁻¹ for saquinavir and atazanavir, and 0.04–16 mg l⁻¹ for indinavir. The intraday and interday coefficients of variation of each method were lower than 8.2% and 8.7%, respectively. Our laboratory is is a participant in the KKGT quality assurance program organized by Dutch Association for Quality Assessment in Therapeutic Drug Monitoring and Clinical Toxicology of the Radbound University Medical Centre Nijmen, with 36 laboratories involved in 2004 [24]. Plasma HIV-1 RNA was quantified by Amplicor® Ultrasensitive Assay (Roche Amplicor HIV-1 Monitor assay, v1.5), with a limit of detection of 50 copies ml⁻¹, following the manufacturer's instructions.

Based on previously published data [18, 25], drug concentrations were considered optimal if they were above the proposed MEC. Thus, target trough concentrations were 3.4 mg l⁻¹ for nevirapine, 1.0 mg l⁻¹ for efavirenz, 0.8 mg l⁻¹ for nelfinavir, 0.1 mg l⁻¹ for indinavir and for saquinavir, 0.15 mg l⁻¹ for atazanavir, 1.0 mg l⁻¹ for lopinavir (4.0 mg l⁻¹ in patients with previous failures to PI therapies), and 0.4 mg l⁻¹ for amprenavir (1.2 mg l⁻¹ in patients with previous failures to PI therapies). In addition, efavirenz, nevirapine and atazanavir concentrations were considered potentially toxic if they were higher than 4.0 mg l⁻¹, 6.0 mg l⁻¹, and 0.85 mg l⁻¹, respectively [18, 25, 26].

Statistical analysis was performed using SPSS Version 11.5 statistical software. Variables with normal distribution were described as mean (standard deviation, SD), whereas median (interquartile range, IQR) was used to describe variables not normally distributed. Interindividual variability in drug concentrations was assessed only for those drugs administered to at least five patients and the coefficient of variation (CV) was calculated as the quotient of the SD divided by the mean. Proportions were compared by the Chi-square or the Fisher exact test, where appropriate.

Results

One hundred and ninety-one patients receiving NNRTI or PI therapy attended our clinic for routine outpatient blood tests. Of these, 117 patients were eligible for the study and were included. Median (IQR) length of time on the current antiretroviral regimen was 20 (8.5–39) months. Overall, compliance with antiretroviral therapy was high, and 82.1% of the patients assured us that they had taken all the prescribed doses within the previous week. Table 1 summarizes patient demographic and clinical characteristics.

Table 1.

Demographic and clinical characteristics of the 117 patients included in the study

	n (%)
Gender
Male	79 (67.5)
Female	38 (32.5)
Age (years)†	41.1 (9.4)
HCV	50 (42.7)
Years since HIV diagnosis†	10.5 (5.2)
Months on ARV therapy‡	20.0 (8.5–39.0)
Current ARV treatment
NNRTI	60 (51.3)
PI	57 (48.7)
Missing doses within the previous week
0	96 (82.1)
1	16 (13.7)
≥2	5 (4.2)

Data are expressed as n (%), except where noted. IVDU intravenous drugs user; HCV hepatitis C virus; ARV antiretroviral; NRTI nucleoside reverse transcriptase inhibitor; NNRTI non-nucleoside reverse transcriptase inhibitor; PI protease inhibitor

^†mean (SD)

^‡Median (interquartile range).

Sixty patients were using one NNRTI and 57 subjects were receiving one PI when blood samples were collected (Table 1). Forty-two patients were receiving efavirenz (600 mg once daily), and 18 were using nevirapine (200 mg twice daily) when the study was performed. Among the PI in current use, 43 subjects were taking lopinavir/ritonavir (400/100 mg twice daily), nelfinavir (1250 mg twice daily) was being used by seven patients, saquinavir/ritonavir (1000/100 mg twice daily) by one patient, amprenavir/ritonavir (600/100 mg twice daily) by two patients, atazanavir (400 mg once daily) by three patients, and indinavir (1200 mg twice daily) by one patient. No patient was receiving ritonavir as single PI when samples were collected.

Drug concentrations

Mean (SD) sampling time was 11.3 (0.8) h after the last drug intake for subjects on a twice daily antiretroviral regimen (n = 72), 22.0 (1.1) h for patients on a once daily regimen at breakfast-time (n = 3), and 10.2 (1.1) h for patients receiving efavirenz once daily at bed-time (n = 42).

The observed plasma concentrations are presented in Figure 1. One sample, which corresponded to lopinavir, was below the lower limit of quantification. Overall, interindividual variability in NNRTI and PI plasma concentrations was approximately 50%. In addition, 12.0% of the patients showed drug concentrations below the MEC, with only 68.4% of NNRTI or PI concentrations being within the proposed therapeutic range (Table 2).

Figure 1.

Observed trough plasma concentrations of NNRTI and PI. The box plot provides a five point summary of the data: minimum, 1st quartile, median, 3rd quartile and maximum. NVP nevirapine; EFV efavirenz; LPV lopinavir; NFV nelfinavir; APV amprenavir; ATV atazanavir; SQV saquinavir; IDV indinavir

Table 2.

Distribution of plasma concentrations of NNRTI and PI

Antiretroviral drug	Number	Concentration (mg l−1)†	CV (%)	Optimal n (%)
Nevirapine	18	5.31 (2.13)	40.1	5 (27.8)
Efavirenz	42	3.62 (1.70)	46.9	26 (61.9)
Lopinavir	43	7.26 (3.91)	53.8	35 (81.4)
Nelfinavir	7	3.02 (1.11)	36.7	7 (100)
Saquinavir/ritonavir	1	0.350		1 (100)
Amprenavir/ritonavir	2	2.06 (1.66–2.45)‡		2 (100)
Atazanavir	3	0.67 (0.62–1.52)‡		3 (100)
Indinavir	1	0.10		1 (100)
Total	117			80 (68.4)

CV: coefficient of variation

^†data are expressed as mean (SD), except where noted.

^‡median (interquartile range).

When drug class was considered, drug concentrations were optimal in only 51.6% of the patients treated with NNRTI, compared with 85.9% of the patients receiving PI (odds ratio 0.17, 95% CI 0.06, 0.46; P < 0.001). In this regard, out of the 18 patients on nevirapine therapy, drug concentrations were subtherapeutic in five (27.8%), potentially toxic in eight (44.4%), and optimal in only five patients (27.8%). Efavirenz concentrations were within the therapeutic range in 61.9% of the patients who were receiving this drug. Although only one patient on efavirenz therapy had drug concentrations below 1.0 mg l⁻¹, concentrations were found to be potentially toxic in 15 (35.7%) of the patients who were being treated with this drug. Regarding PI concentrations, the proportion of patients with drug concentrations above the proposed MEC was approximately 80% in those receiving lopinavir/ritonavir, and 100% in those using nelfinavir, saquinavir, amprenavir atazanavir or indinavir (Table 2).

Out of the 14 patients showing NNRTI or PI plasma concentrations below the MEC, nonadherence was found to be a possible explanation for five cases (35.7%). There was no evident explanation for 64.3% of the patients showing subtherapeutic drug concentrations.

Clinical outcome

Overall, 17 patients developed virological failure within the 48 weeks that followed blood sampling. Drug concentrations were above the proposed MEC in 16 of these patients, and only one subject had subtherapeutic dug concentrations (P = 0.689). Regarding drug-related toxicity, no patient receiving nevirapine or atazanavir therapy developed significant drug-related toxicity during follow-up. On the other hand, nine out of the 42 patients who were receiving efavirenz experienced central nervous system disturbances. However, efavirenz concentrations in plasma were considered optimal in most of these patients, and only two subjects (22.2%) showed efavirenz concentrations above 4.0 mg l⁻¹. Similarly, the proportion of subjects who developed central nervous system symptoms while receiving efavirenz was similar in patients with drug concentrations above or below 4.0 mg l⁻¹ (P = 0.451).

Discussion

Our results show that plasma concentrations of NNRTI and PI may vary widely among HIV-infected adults in an outpatient routine clinical setting. In addition, drug concentrations were outside the recommended therapeutic window in almost one third of our patients.

Overall, our results are concordant with other previously reported. In a study performed by de Maat et al. [20] which included 1145 samples from 97 subjects, drug concentrations were subtherapeutic in one quarter of the samples analyzed. The lower prevalence of subtherapeutic drug concentrations observed in our study is probably due to the fact that drug determinations in the study by de Maat et al. were all requested for TDM purposes, and were not collected in a systematic way. In addition, many patients in that study had repeated drug determinations over time. As a consequence, the inclusion of patients whose drug concentrations were suspected to be outside the therapeutic range might have been favoured, leading to an overestimation of the real prevalence of subtherapeutic drug concentrations. In an attempt to avoid such concerns, the present study included all the subjects on NNRTI or PI therapy attending our clinic over a representative period of time, and only one sample was collected in each patient. Thus, our results may represent more accurately the global prevalence of drug concentrations outside the therapeutic window among HIV-infected adults in routine clinical practice.

Variability in drug concentrations may have relevant consequences in the field of antiretroviral therapy. Subtherapeutic drug concentrations may preclude sustained viral suppression in HIV-infected patients and may predispose them to the development of viral resistance, limiting future reutilization of antiretroviral agents [27]. On the other hand, inappropriately high drug concentrations have been related to the appearance of treatment-related adverse events, which may have a negative impact on the patients' quality of life and treatment adherence [18, 25, 26]. In our study, the variability in drug concentrations among different subjects receiving the same dose of each NNRTI or PI was near 50%, 14 out of 117 patients included showed drug concentrations below the proposed MEC, and one third of the patients receiving NNRTI therapy had potentially toxic drug concentrations. In addition, inappropriate adherence explained less than half of the subtherapeutic drug concentrations. This finding suggests that individual differences in drug absorption, distribution, metabolism and elimination may be responsible for variability in drug concentrations and that, although other issues such as intraindividual variability in drug concentrations should also be considered [5], monitoring drug concentrations of antiretroviral agents might be of help in clinical practice.

We did not find any relationship between drug exposure and the development of virological failure or drug-related toxicity in this study. In this regard, the proportion of patients who developed virological failure was similar between patients with drug concentrations below or above the proposed MEC. Similarly, we did not observe any relationship between drug exposure and toxicity, although our data were limited to efavirenz. These results are in disagreement with previous observations [11–19], and may suggest that the role of TDM in clinical practice is limited. However, several points should be considered when interpreting these results. First, the main objective of this study was to assess the interindividual variability of drug concentrations in clinical practice, and only few patients developed treatment failure or drug-related toxicity during the follow-up. As a consequence, the study may be underpowered to relate the clinical outcome to drug exposure. In addition, we only measured drug concentrations at an isolated time point, which may not be representative of drug concentrations all through the follow-up. Finally, the clinical usefulness of the limits of the therapeutic window proposed for some antiretroviral agents continues to be debated. In this regard, although Marzolini et al. [15] identified efavirenz concentrations above 4.0 mg l⁻¹ as a predictor for central nervous system disorders, Kappelhoff et al. [28] found no relationship between efavirenz exposure and the development of neuropsychiatric disturbances in the 2NN study. In addition, the target minimum effective concentration may vary widely between different individuals, or even within the same subject over time, due to the accumulation of drug resistance mutations in the viral genome and to the gradual decrease in drug susceptibility.

In conclusion, there is a large interindividual variability in plasma concentrations of PI and NNRTI among HIV-infected adults in routine clinical practice, with drug concentrations being outside the proposed therapeutic window in a significant proportion of patients. These results, in combination with the relatively lower intraindividual variability in PI and NNRTI concentrations previously reported [15, 29, 30], support a potential role for therapeutic drug monitoring of antiretroviral therapy in clinical practice. Prospective, randomized, properly powered trials assessing the clinical usefulness of this strategy are needed.