Pharmacokinetics of afatinib in subjects with mild or moderate hepatic impairment

Male or female non-smoker subjects aged 18–75 years and with a body mass index between 18.5 and 34.0 kg/m² with chronic liver disease stable for at least 3 months were enrolled into two pre-specified groups: (a) mild hepatic impairment, Child-Pugh A (scores 5 or 6) [13], and (b) moderate hepatic impairment, Child-Pugh B (scores 7–9). Each individual was matched with a healthy control of the same gender, age (±10 years), weight (±10 %) and, if possible, creatinine clearance (±30 mL/min) according to the Cockcroft–Gault formula (Table 1). Healthy subjects were eligible based on the lack of clinically relevant history or physical examination and electrocardiography (ECG) findings, vital signs and clinical laboratory tests. Overall, creatinine clearance was >70 mL/min for healthy volunteers and >40 mL/min for subjects with hepatic impairment.

Exclusion criteria in hepatic impaired subjects were severe cardiovascular disorders in the preceding 6 months; significant or recent acute gastrointestinal disorders; hepatic encephalopathy >grade 2 by number connection test; changes in chronic medication in the 4 weeks prior to trial drug administration (other than discontinuation of medications known to interact with P-glycoprotein); gastrointestinal bleeding in past 3 months; serum albumin <20 g/L; or haemoglobin <8 g/dL. For both patients and healthy controls, treatment with any medication with a half-life of >24 h in the month before or during the trial, potent P-glycoprotein inhibitors or inducers, or any medication known to interact with afatinib was prohibited. Women who were pregnant, breastfeeding or those of child-bearing potential not using adequate contraception for 3 months before and during the study were excluded.

After an overnight fast (≥10 h), subjects received a single oral dose of afatinib (Boehringer Ingelheim Pharma GmbH & Co. KG, Germany) administered as a film-coated tablet with 240 mL of water in the sitting or standing position. Water was allowed ad libitum except 1 h before and 2 h after dosing, and standardized meals were served at least 2 h after dosing. Subjects were closely observed in the clinic for at least 24 h after dosing until discharge 48 h later and subsequently returned for follow-up blood sampling. Follow-up by telephone was arranged 28 days post-dosing.

Blood samples were collected into potassium-EDTA-anticoagulant tubes before and 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 9, 12, 24, 36, and 48 h after dosing, and then at 24-h intervals up to 10 days after dosing. All samples were centrifuged within 30 min of collection. A 15-mL sample for determination of plasma protein binding was also collected in 3 × 4.9 mL vials (monovettes coated with EDTA) on day 1 before dosing and centrifuged at 4,000 rpm for 10 min at 4 °C. All plasma samples were stored at −20 °C until analysis. Urine was collected in containers before and 0–4, 4–8, 8–12, 12–24, 24–48 and 48–72 h after dosing. To prevent adsorption losses of afatinib, a detergent (Tween 20) was added to the containers before urine collection. For each collection, the urine was weighed and homogenized and aliquots were stored at −20 °C until analysis.

Plasma and urine concentrations of the free base of afatinib (BIBW 2992 BS) were analysed using validated liquid chromatography-mass spectrometry (HPLC–MS/MS) methods (Nuvisan GmbH, Neu-Ulm, Germany) [11, 14]. Plasma concentrations within the validated concentration range were used to calculate the pharmacokinetic parameters. The calibration curves for afatinib covered the range 0.100–50.0 ng/mL for plasma and 5.00–1,000 ng/mL for urine. The lower limit of quantification of afatinib was 0.100 ng/mL in plasma and 5.00 ng/mL in urine. Assay performance was assessed by back calculation of calibration standards, tabulation of the standard curve fit function parameters and measurement of quality control samples. For HPLC–MS/MS bioanalysis of afatinib plasma and urine concentrations, assay accuracy [deviation from nominal value (%)] and precision [coefficient of variation (%)] of quality control samples spiked at three concentrations were between 3.9 and 9.6 % for plasma, and between 4.9 and 8.0 % for urine, respectively.

In vitro plasma protein binding of afatinib was determined in pre-dose plasma samples after spiking of 150 nmol/L (=72.9 ng/mL) [¹⁴C]-radiolabeled afatinib using equilibrium dialysis (Covance Laboratories, Harrogate, UK). Previous in-house studies demonstrated the suitability of this methodology (unpublished data on file, Boehringer Ingelheim, Biberach, Germany).

The safety of afatinib was assessed by 12-lead ECG, vital signs (pulse, blood pressure), routine laboratory assessments, adverse event reporting and assessment of global tolerability by the investigator. Adverse events were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 3.0.

For safety reasons, individuals with hepatic impairment were to be dosed in a 3 + 5 design, with three patients being treated at the initial dose (Fig. 1). The decision to proceed with further afatinib administrations at higher doses was based on a combined assessment of drug-related adverse events and C _max (gMean) of afatinib by interim pharmacokinetic measurements after dosing of three subjects in each hepatic impairment group and their respective controls. Subjects with mild hepatic impairment were initially dosed with afatinib 50 mg; if no subject had a drug-related adverse event of CTCAE grade ≥3, or <2 subjects had a drug-related adverse event of CTCAE grade ≥2 and C _max was ≤30 ng/mL, three patients with moderate hepatic impairment were then dosed with 30 mg afatinib. If no drug-related toxicity was observed, the remaining five subjects with moderate hepatic impairment received 50 mg afatinib or an interim dose of 40 mg if C _max was >30 ng/mL and ≤60 ng/mL. These pharmacokinetic thresholds took into consideration that in a trial of healthy subjects receiving single doses of afatinib (20, 30, 40 or 50 mg), individual C _max values exceeding 60 ng/mL were well tolerated [15]. Further, in a trial of patients with solid tumours who received afatinib 30 and 45 mg once daily for 14 days, no drug-related adverse events of CTCAE grade >2 were reported and individual C _max values after the first dose (day 1) did not exceed 100 ng/mL [14]. To allow for additional safety monitoring, staggered dosing was preferred in each of the subgroups of five subjects (three subjects on 1 day, followed by two subjects on the following day). Afatinib 50 mg was used in this study because it is the maximum tolerated dose and is the highest dose for therapeutic use.

The study planned to recruit up to 38 subjects (see Fig. 1) with the aim of entering eight subjects with mild liver impairment (at 50 mg afatinib), eight subjects with moderate liver impairment (at either 30, 40 or 50 mg afatinib) and eight healthy matched controls to each of this two groups (in total 16 healthy subjects). A total of 32 subjects (eight per group) receiving 50 mg afatinib for the primary analysis were judged an adequate sample size, in agreement with regulatory guidance of pharmacokinetic studies in patients with impaired hepatic function [16].

The primary pharmacokinetic endpoints were AUC from time zero extrapolated to infinity (AUC_0–∞) and C _max. AUC from 0 to the time of the last quantifiable data point (AUC_0–tz) was a secondary endpoint. The log-transformed AUC and C _max values for afatinib were compared between groups using an analysis of variance (ANOVA) model with ‘hepatic status’ as fixed effect and ‘matched pair’ as random effect. The least square means and 90 % confidence intervals (CI) based on the t-distribution were calculated and then back-transformed to the original scale to provide gMean and interval estimates for the ratio between response under test (liver impaired subjects) and response under reference (healthy subjects). For all other parameters, descriptive statistics were presented. Non-compartmental analysis of plasma concentration–time data was performed using WinNonlin^® Professional Network version 5.2 software (Pharsight Corporation, Cary, NC, USA). Statistical analyses were performed using SAS^®, version 9.2 (SAS Institute Inc., Cary, NC, USA).

There were no notable differences in the plasma concentration–time profiles between subjects with mild or moderate hepatic impairment and their matched healthy controls (Fig. 2). Afatinib showed a biphasic disposition profile, and a long terminal phase was detected extending out to the last sampling point of 240 h. The extent of exposure as indicated by AUC_0–∞ and C _max was generally comparable between the matched treatment groups (Table 2). For subjects with mild hepatic impairment, the adjusted gMean ratio for AUC_0-∞ was 92.6 % (90 % CI 68.0–126.3 %) and for C _max was 109.5 % (90 % CI 82.7–144.9 %), compared with healthy control subjects. For subjects with moderate hepatic impairment, the corresponding adjusted gMean ratios were 94.9 % (90 % CI 72.3–124.5 %) and 126.9 % (90 % CI 86.0–187.2 %), respectively (Table 3).

For subjects with mild hepatic impairment, median time to peak plasma concentration (t _max) was the same as matched healthy controls (5 h). For subjects with moderate hepatic impairment, t _max occurred earlier than for matched healthy controls (4.0 h for moderate impairment vs. 7.5 h for healthy controls). The range of t _max values was also larger in subjects with hepatic impairment compared with matched healthy controls; for mild impairment, values were between 0.5 to 8 h versus 3 to 7 h for matched controls, and for moderate impairment, values were between 0.5 to 5 h, versus 5 to 9 h for matched controls. The gMean terminal half-life ranged from 60 to 75 h and was comparable for subjects with hepatic impairment and normal hepatic function (Table 2).

There were quantifiable urinary concentrations of afatinib over the entire sampling interval (up to 72 h post-dose) in all subjects. The total cumulative fraction of afatinib excreted in the urine (fe_0–72) in subjects with hepatic impairment was generally low and comparable with matched controls (gMean values between 2.0 and 2.58 %; Table 2). The gMean excretion profiles showed no noteworthy differences between the treatment groups (Fig. 3).

The arithmetic mean ± SD fraction of [¹⁴C] afatinib (target concentration 72.9 ng/mL) bound to plasma proteins in pre-dose plasma samples was 94.6 ± 0.7 % in healthy controls (n = 16), 94.1 ± 1.1 % in subjects with mild hepatic impairment (n = 8) and 93.7 ± 0.7 % in subjects with moderate hepatic impairment (n = 11, three subjects that received afatinib 30 mg and eight subjects that received afatinib 50 mg). The overall mean percentage protein binding was 94.2 ± 0.9 %.

Single-dose afatinib 50 mg was well tolerated with few adverse events. None of the subjects experienced serious adverse events or discontinued the study due to an adverse event. Adverse events were reported in five (26 %) subjects with hepatic impairment (three mild, two moderate) and one (6 %) healthy control subject. Three patients with mild hepatic impairment (50 mg afatinib) had adverse events that were considered treatment-related. One of these subjects had a grade 3 lipase elevation; however, cholecystolithiasis with sludge phenomenon was observed on abdominal ultrasound of this subject, suggesting that this was the most likely cause of the increase. This subject had a similar episode of asymptomatic lipase increase prior to enrolment in the clinical trial. The other two treatment-related events were grade 2 headache and nausea in one subject and grade 1 diarrhoea in one subject. All adverse events had resolved by the end of the trial. There were no other clinically relevant changes in laboratory parameters, vital signs or ECG.