An effect of moderate hepatic impairment on the pharmacokinetics and safety of darapladib

Abstract

Aim/Methods

This was a phase 1, open label, non-randomized study designed to assess the pharmacokinetics and safety/tolerability of 10 consecutive once daily 40 mg oral doses of darapladib in subjects with moderate hepatic impairment (n = 12) compared with matched healthy volunteers (n = 12).

Results

For total darapladib, a small increase in total and peak exposure was observed in the subjects with moderate hepatic impairment compared with the subjects with normal hepatic function. The area under the plasma concentration−time curve during a dosing interval of duration τ (AUC(0,τ), geometric mean 223 ng ml⁻¹ h [90% CI 158, 316 ng ml⁻¹h], in moderate hepatic impaired subjects, vs. geometric mean 186 ng ml⁻¹h [90% CI 159, 217 ng ml⁻¹h], in healthy subjects) and maximum concentration (C_max) were 20% and 7% higher, respectively, in the subjects with moderate hepatic impairment than in the healthy control subjects and there was no change in time to maximum concentration (t_max). Protein binding was performed to measure the amount of unbound drug vs. bound. Steady-state was achieved by day 10 for darapladib and its metabolites (M4, M3 and M10). Darapladib was generally well tolerated, with adverse events (AEs) reported by seven subjects in the hepatic impairment group and three subjects in the healthy matched group (five and one of which were drug-related AEs, respectively). The most common AEs were gastrointestinal. These AEs were mostly mild to moderate and there were no deaths, serious AEs or withdrawals due to AEs.

Conclusions

The results of this phase 1 study show that darapladib (40 mg) is well tolerated and its pharmacokinetics remain relatively unchanged in patients with moderate hepatic impairment.

What is Already Known about this Subject

• Darapladib is a novel Lp-PLA₂ inhibitor currently being developed for atherosclerosis.

• The pharmacokinetics of darapladib in hepatic impaired subjects have not been reported previously.

What this Study Adds

• As shown by the results of this study, darapladib can be given to patients with mild or moderate hepatic impairment without dose adjustment.

Introduction

Lipoprotein-associated phospholipase A₂ (Lp-PLA₂), also known as platelet-activating factor acetylhydrolase, is secreted by inflammatory cells involved in the formation of atherosclerosis [1,2]. Through the cleavage of oxidized or polar phospholipids, Lp-PLA₂ generates two key pro-inflammatory mediators, lysophosphatidylcholine and oxidized non-esterified fatty acids, the former of which promotes atherosclerotic plaque development and the formation of a necrotic core [3]. Elevated Lp-PLA₂ concentrations have been shown to be associated with an increased risk of cardiovascular events (e.g. myocardial infarction, stroke, and cardiovascular mortality) in subjects with coronary artery disease [1–4]. Over a 16 year follow-up in the Rancho Bernardo study (n = 1077), baseline Lp-PLA₂ concentrations in the second, third and fourth quartiles predicted an increased risk of coronary heart disease (CHD) compared with the lowest quartile (hazard ratios 1.66, 1.80 and 1.89, respectively; P = 0.05 for each) after adjusting for C-reactive protein and other CHD risk factors [5]. In the Lp-PLA₂ Studies Collaboration meta-analysis of 32 studies (n = 79 036; 474 976 person-years at risk), circulating Lp-PLA₂ mass and activity were associated with log-linear associations risk of CHD and vascular death. The risk ratios, adjusted for conventional risk factors, were 1.10 (95% confidence interval [CI] 1.05, 1.16) with Lp-PLA₂ activity and 1.11 (1.07, 1.16) with Lp-PLA₂ mass for CHD [6]. These findings suggest that Lp-PLA₂ may act as a biomarker, independent of traditional risk factors and other markers of inflammation.

Darapladib (SB-480848) is a novel, selective, orally active inhibitor of Lp-PLA₂ [7] that is currently under clinical investigation for the treatment of atherosclerosis.

The principal routes of darapladib metabolism are N-de-ethylation to produce M4 (SB-553253), hydroxylation of the cyclopenta pyrimidinone ring to produce M3 (SB-823094) and removal of the 4-fluorophenyl methanethiol to produce M10 (SB-554008). Elimination of darapladib occurs primarily by hepatic metabolism, as unchanged parent compound and renal excretion represents less than 0.5% of the administered dose. Following oral dosing in healthy volunteers, low plasma concentrations (<5% of parent darapladib area under the curve [AUC]) of M4, M3, and M10 are observed. Additionally, the pharmacological activity of M4 and M3 was similar to that of darapladib, while that of M10 was about 100-fold less than that of the parent compound. Given the low exposures relative to parent, darapladib metabolites do not contribute significantly to the pharmacological activity of darapladib in humans.

As darapladib is extensively metabolized in the liver, the primary objective of the present study was to assess the effects of moderate hepatic impairment, compared with normal hepatic function, on the pharmacokinetics and safety/tolerability of oral repeat dosing of darapladib.

Methods

Study design

This was a phase 1, open label, non-randomized study conducted from July to October 2010 at two investigative sites in the United States. The goals of the study were to assess the pharmacokinetics and safety/tolerability of consecutive 40 mg oral doses of enteric-coated darapladib, an Lp-PLA₂ (HSD-PLA₂, also known as serine-dependent phospholipase A₂, PAFAH2, ENSG00000158006) [8] inhibitor, in subjects with moderate hepatic impairment in comparison with matched healthy volunteers. This trial is registered with ClinicalTrials.gov (NCT01154114).

A sample size of up to 12 evaluable subjects per group was selected. Based on FDA Guidance for Industry [9], at least eight evaluable subjects per group should be enrolled in hepatic impairment studies. To allow for possible dropout and provide better variability estimates, 12 subjects per group were enrolled and completed the study. With a typical between subject variability of 39% in darapladib exposure from historical studies, the sample size of 12 per group would provide at least 90% power to detect an effect size of 70% higher exposure in the hepatic impaired group at the α level of 0.05. Subjects were institutionalized from the day before the first dose, assigned to receive darapladib 40 mg enteric-coated tablets daily for 10 consecutive days, and released following collection of the last pharmacokinetic blood sample, which took place 24 h after the last dose of study medication. Subjects were required to return to the unit 10 to 14 days after the last dose of study medication and again 28 to 42 days after the last dose of study medication for clinical assessments.

The investigators and sponsor complied with all regulatory requirements relating to safety reporting to regulatory authorities, institutional review boards, and investigators. This study was conducted in accordance with Good Clinical Practice and all applicable regulatory requirements, as well as the guiding principles of the 2008 Declaration of Helsinki. All subjects gave written informed consent before enrolment, which included compliance with the requirements and restrictions listed in the consent form.

Population studied

Subjects were aged 18 to 65 years, inclusive, at the time of signing the informed consent. Recruitment included male and female subjects in both the moderate hepatic impaired group and the healthy volunteer group.

Hepatic impairment was defined as having a known medical history of liver disease with or without a known history of alcohol abuse, and previous confirmation of liver cirrhosis in moderately hepatically impaired subjects by liver biopsy or other medical imaging technique (including laparoscopy, computed tomography scan or ultrasonography) associated with an unambiguous medical history and a Child-Pugh score of 7 to 9. Other inclusion criteria included a body mass index (BMI) within the range of 19 to 37 kg m⁻² and a QT_cB <480 ms (including subjects with bundle branch block at screening electrocardiogram [ECG]).

Healthy subjects were selected to match as closely as possible with moderate hepatic impairment subjects for gender (1:1), age (age ± 5 years) and BMI (± 15%). Additional inclusion criteria for participants with normal hepatic function included aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase and bilirubin less than or equal to 1.5× the upper limit of normal (ULN; isolated bilirubin >1.5× ULN was acceptable if bilirubin was fractionated and direct bilirubin was <35%).

General exclusion criteria included clinically significant diseases or laboratory abnormalities (except parameters influenced by hepatic impairment), conditions that could potentially alter the absorption, distribution, metabolism and elimination characteristics of the study drug (other than hepatic impairment), history of alcohol or drug abuse within the past 2 months, blood donation or blood loss greater than 500 ml within a 56 day period, current use of oral or injectable strong cytochrome P450 3A4 (CYP3A4) inhibitor(s) or consumption of grapefruit (or grapefruit juice) less than 7 days before initiation of study drug. Women who were breastfeeding, pregnant, planned to become pregnant during the study or who used oral, injected or implanted hormonal methods of contraception were also excluded. Additional exclusion criteria for participants with normal hepatic function included positive prestudy hepatitis B surface antigen, positive hepatitis C antibody or positive hepatitis A IGM antibody result within 3 months of screening, history of cholecystectomy or biliary tract disease or history of liver disease with elevated liver function tests of known or unknown aetiology.

Assessments

Blood samples for the pharmacokinetic analysis of darapladib and its metabolites (M4, M3, and M10) were collected predose on days 2, 4, 6, 8, 9, 10, as well as at the following post-dose time points on day 10: 0.5, 1, 2, 3, 4, 6, 8, 10, 12, 18 and 24 h. Blood samples were taken via an indwelling cannula (or by direct venipuncture), collected into an ethylenediaminetetraacetic acid (EDTA) tube and immediately placed on crushed ice. All samples were centrifuged at 1000 g (∼3000 rev min⁻¹) for approximately 10 to 15 min under refrigeration (4°C) within 30 min of collection. Supernatant plasma was transferred to a polypropylene tube and stored at −20°C before shipping for sample analysis.

SB-480848, SB-553253, SB-554008 and SB-823094 were extracted from human plasma by liquid-liquid extraction using [¹³C₆]-SB-480848, [¹³C₆]-SB-553253, [¹³C₆]-SB-554008 and [¹³C₆]-SB-823094 as internal standards. Extracts were analyzed by liquid chromatography with tandem mass spectrometric detection (LC/MS/MS) using a TurboIonspray™ interface with positive-ion multiple-reaction monitoring. The assay was validated over the SB-480848, SB-553253, SB-554008 and SB-823094 concentration range of 0.10 ng ml⁻¹ to 50 ng ml⁻¹, and the lower limit of quantification (LLoQ) was 0.1 ng ml⁻¹ using a 50 μl aliquot of human plasma.

Quality control (QC) samples, prepared at three different analyte concentrations and stored with study samples, were analyzed with each batch of samples against separately prepared calibration standards. QC samples and calibration standards were prepared using independently prepared stock solutions of darapladib, SB-553253, SB-554008 and SB-823094 reference materials. For the analysis to be acceptable, no more than one-third of the QC results were to deviate from the nominal concentration by more than 15%, and at least 50% of the results from each QC concentration were to be within 15% of nominal.

Plasma protein bound and unbound darapladib were separated by equilibrium dialysis, where the unbound darapladib diffused through a semipermeable membrane into the dialysate until it reached equilibrium. After dialysis, darapladib concentrations in dialysate samples were analyzed using LC/MS/MS with a method over the range of 1 pg ml⁻¹ to 1000 pg ml⁻¹. The LLoQ of this method was 1 pg ml⁻¹ using a 200 μl aliquot of dialysate.

Pharmacokinetic analyses of the concentration−time data for plasma total and unbound darapladib, and M4, M3 and M10, were conducted using non-compartmental Model 200 (for extravascular administration) of WinNonlin Professional Edition version 5.2 (Pharsight Corporation, Mountain View, CA, USA). Actual elapsed time from dosing was used to estimate all individual plasma pharmacokinetic parameters for evaluable subjects. The maximum observed plasma concentration (C_max) and the time at which C_max was observed (t_max) were determined directly from the raw concentration−time data. The area under the plasma concentration−time curve during a dosing interval of duration τ (AUC(0,τ)) and from time zero to the last quantifiable time point (AUC(0,t)) were calculated by a combination of linear and logarithmic trapezoidal methods. The linear trapezoidal method was used for all incremental trapezoids arising from increasing concentrations and the logarithmic trapezoidal method was used for those arising from decreasing concentrations. In addition, the predose (trough) concentration (Cτ) was determined, where τ is the end of dosing interval.

Attempts were made to determine the unbound fraction of darapladib as follows. Human plasma samples were analyzed for total darapladib using a validated analytical method based on liquid-liquid extraction, followed by high-performance liquid chromatography tandem mass spectrometry (HPLC-MS/MS) analysis. The LLoQ was 0.10 ng ml⁻¹ based on a 50 μl aliquot of human plasma with a higher limit of quantification (HLoQ) of 50 ng ml⁻¹.

Human plasma phosphate-buffered saline dialysate samples were analyzed for unbound darapladib using a validated analytical method based on liquid-liquid extraction, followed by HPLC-MS/MS analysis. Using a 200 μl aliquot of dialysate, the LLoQ for unbound darapladib was 1 pg ml⁻¹ and the HLoQ was 1000 pg ml⁻¹. The computer systems used to acquire and quantify data included Analyst Version 1.4.2 and SMS2000 Version 2.1.

The darapladib unbound fraction was calculated for each subject by dividing the darapladib concentration in the dialysate by the darapladib concentration in plasma. An average of the three unbound fractions was reported as the unbound fraction for each subject. If the unbound fraction could not be determined for an individual subject, the group (healthy or hepatic impaired) mean unbound fraction was used for that subject.

Complete physical examinations, 12-lead ECGs and clinical laboratory assessments were obtained at screening and on the day 20–24 follow-up visits. Safety evaluations assessed adverse events (AEs) and serious adverse events (SAEs), which were collected from the start of darapladib dosing until the final follow-up visit.

Statistical analysis

This study was designed to estimate the effects of hepatic impairment on the pharmacokinetics of darapladib compared with the effects seen in healthy matched subjects. For the purposes of calculating summary statistics and for statistical analysis, the pharmacokinetic parameters, AUC(0,τ), AUC(0,t), C_max and Cτ, were log_e transformed. Following log_e transformation, darapladib AUC(0,τ) and C_max were separately analyzed using an analysis of variance (anova) model with group (hepatic or healthy) as fixed effect terms in the model. The t_max of darapladib was analyzed with the non-parametric Wilcoxon rank-sum test to compute point estimates and associated 90% CIs for the median differences between the hepatic impaired group and the healthy subjects group. A mixed-effect model was fitted for trough concentration with subject as a random effect and day as a continuous covariate. The coefficient of the slope of the day effect on the log scale was used to determine whether steady-state was achieved on day 10. Using the pooled estimate of variance, the 90% CIs for the slopes were calculated and were back-transformed onto the ordinary scale.

Results

Patient disposition and demographics

A total of 24 subjects (moderate hepatic impairment (n = 12) and healthy (n = 12)) were enrolled to receive darapladib (Table 1). All subjects completed the study as planned and were included in the pharmacokinetic and safety analyses. The demographics were similar between the two groups. The moderate hepatic impairment group had a Child-Pugh total score of 7 to 9 (Table 1).

Table 1.

Disposition, demographics, and Child-Pugh scores

Demographics	Moderate hepatic impairment subjects (n = 12)	Healthy subjects (n = 12)
Age (years), mean (SD)	54.9 (4.85)	54.1 (5.38)
Gender, male, n (%)	10 (83)	10 (83)
BMI (kg m−2), mean (SD)	31.56 (5.115)	28.17 (7.136)
Height (cm), mean (SD)	169.1 (8.10)	167.5 (8.31)
Weight (kg), mean (SD)	90.73 (18.676)	78.35 (19.322)
Ethnicity, n (%)
Hispanic or Latino	3 (25)	4 (33)
Not Hispanic or Latino	9 (75)	8 (67)
Race, n (%)
African American/African Heritage	2 (17)	4 (33)
White, White/Caucasian/European Heritage	10 (83)	8 (67)
Child-Pugh total score	Screening/day 1/follow-up
7	7/6/6	–
8	2/3/4	–
9	3/3/2	–

BMI, body mass index; SD, standard deviation.

Safety

Adverse events were reported by seven subjects in the hepatic impairment group and three subjects in the healthy matched group. There were no deaths or SAEs reported during this study. The AEs were mostly mild to moderate in intensity. The most frequent AEs in the study were gastrointestinal, which were reported in four hepatically impaired subjects and two healthy subjects. Only two subjects, both with hepatic impairment, reported events that were severe in intensity (one with severe constipation, one with severe headache and severe muscle pain). All three events resolved, and the investigator determined that they were related to the study drug. Overall, drug-related AEs were reported in five (42%) subjects in the hepatic impairment group and one (8%) subject in the healthy matched group. Furthermore, there were no clinically significant changes in laboratory values, vital signs and 12-lead ECG parameters throughout the study.

Pharmacokinetic results

Individual plasma darapladib, M4, M3 and M10 concentrations were determined and selected pharmacokinetic parameters are summarized in Table 2. Following the administration of multiple doses of darapladib 40 mg, darapladib was absorbed with a delay that is consistent with enteric coating (Figure 1). The median t_max values for subjects with moderate hepatic impairment and those with normal liver function were comparable. AUC(0,τ) (geometric mean and 90% CI 223 ng ml⁻¹h, 158, 316 ng ml⁻¹h in moderate hepatic impaired subjects; 186 ng ml⁻¹h, 159, 217 ng ml⁻¹h in healthy subjects) and C_max (geometric mean and 90% CI 14.0 ng ml⁻¹, 9.6, 20.5 ng ml⁻¹ in moderate hepatic impaired subjects; 13.1 ng ml⁻¹, 11.6, 14.7 ng ml⁻¹ in healthy subjects) of total darapladib in moderately hepatic impaired subjects were 20% and 7% higher than those in the healthy control subjects, respectively. For the three metabolites of darapladib (M4, M3 and M10), exposures in the moderate hepatic impaired subjects ranged from 13% lower to 69% higher than the healthy control subjects. In addition, each of these three metabolites circulated at less than 5% of parent darapladib AUC. In healthy subjects, moderate intersubject variability in pharmacokinetic parameters of darapladib was observed, whereas high intersubject variability in its metabolites was observed (coefficient of between subject variation [CV_b] ranged from 19% to 73%). High intersubject variability was observed in moderate hepatic impaired subjects (CV_b ranged from 31% to 158%). Steady-state appears to have been generally achieved by day 10 in both groups for all analytes because the 90% CIs for the back-transformed slope fell within the 0.96 to 1.08 limit.

Table 2.

Summary pharmacokinetic parameters of darapladib and its three metabolites following 10 day once daily dosing of 40 mg of enteric-coated, micronized, free base darapladib in moderate hepatic impaired patients and matched healthy control subjects

Parameter (units)	Geometric		Ratio	90% confidence interval	% CVb
	LS mean (CV%)
	Hepatic	Healthy
Total darapladib
AUC(0,τ) (ng ml−1 h)	223 (76%)	186 (31%)	1.20	(0.83, 1.73)	48.7
Cmax (ng ml−1)	14.0 (85%)	13.1 (23%)	1.07	(0.73, 1.57)	58.8
tmax (h)*	7.00*	7.02*	0†	(–2.00, 2.00)	–
M4
AUC(0,t) (ng ml−1 h)	7.14 (158%)	6.29 (27%)	1.14	(0.62, 2.06)	71.0
Cmax (ng ml−1)	0.69 (117%)	0.79 (59%)	0.87	(0.51, 1.49)	66.4
tmax (h)*	6.00*	6.02*	0†	(–2.00, 2.00)	–
M10
AUC(0,t) (ng ml−1 h)	3.24 (85%)	1.92 (73%)	1.69	(0.84, 3.41)	61.2
Cmax (ng ml−1)	0.17 (31%)	0.13 (19%)	1.38	(1.13, 1.68)	23.3
tmax (h)*	10.0*	8.00*	1.50†	(–6.00, 12.00)	–
M3
AUC(0,t) (ng ml−1 h)	9.87 (62%)	7.11 (46%)	1.39	(0.94, 2.05)	47.0
Cmax (ng ml−1)	0.98 (105%)	1.06 (57%)	0.93	(0.56, 1.56)	62.2
tmax (h)*	6.00*	6.03*	0†	(–2.00, 0.00)	–
Unbound darapladib
AUC(0,τ) (ng ml−1 h)	0.08 (86%)	0.049 (43%)	1.62	(1.06, 2.47)	55.2
Cmax (ng ml−1)	0.005 (102%)	0.003 (41%)	1.45	(0.91, 2.30)	59.5
tmax (h)*	7.00*	7.02*	0†	(–2.00, 2.00)	–

^*t_max values presented are median values.

^††Estimated median difference presented for t_max.

Figure 1.

Steady-state plasma concentrations of total darapladib (A), unbound darapladib (B), M4 (C), M10 (D) and M3 (E) metabolites. SD, standard deviation. , healthy (mean + SD); , moderately hepatic impaired (mean + SD)

Unbound plasma darapladib pharmacokinetic parameters are summarized in Table 2. There was no concentration-dependent protein binding and, in general, three time points were available in each subject to calculate an average percent unbound. Deriving the unbound darapladib concentration for the other 14 samples per patient resulted in 33 of 288 (11%) of the unbound darapladib concentration data points falling below LLoQ of the assay (1 pg ml⁻¹), giving rise to large variability in unbound darapladib parameters.

Exposures (AUC(0,τ) and C_max) of unbound plasma darapladib were 62% and 45% higher in the moderate hepatic impaired group than in the healthy group. However, the C_max was three-fold and five-fold of the LLoQ in healthy and hepatic impaired subjects, respectively (Table 2). Furthermore, in contrast to total darapladib, which circulated in the ng ml⁻¹ range (Table 2), unbound darapladib concentration−time profiles hovered around the LLoQ (pg ml⁻¹ range, Table 2). Thus, plasma protein binding was calculated at 99.96% and 99.97% in hepatic impaired and healthy subjects, respectively. Similar to what was seen in the total plasma darapladib results, moderate hepatic impaired subjects also seemed to have a variable exposure. The CV_b% of unbound darapladib pharmacokinetic parameters was 86% for AUC and 102% for C_max. The variability of the pharmacokinetic parameters was moderate in healthy subjects. Graphic exploration found no apparent relationship between unbound darapladib AUC(0,τ) or C_max and serum albumin concentrations (Figure 2).

Figure 2.

AUC and C_max of unbound plasma darapladib vs. albumin. AUC, area under the plasma concentration−time curve; C_max, maximum concentration. , healthy; , moderately hepatic impaired

Discussion

This study evaluated the effects of moderate hepatic impairment on the pharmacokinetics of darapladib following once daily dosing of darapladib 40 mg for 10 days. Administration of multiple 40 mg oral doses of darapladib in normal healthy subjects and those with moderate hepatic impairment was well tolerated in this study. The most common AEs, regardless of causality, were headache, flatulence and constipation in hepatically impaired subjects. These AEs were reported as drug-related. Subjects in the hepatically impaired group had drug-related flatulence and constipation, while one healthy patient had diarrhea that was determined to be drug related by the investigator. Most AEs were mild to moderate in intensity and there were no deaths, SAEs or withdrawals due to AEs. Furthermore, the AEs reported in healthy subjects were not different from those previously reported for darapladib in other clinical trials.

While the clinical dose currently employed in the ongoing phase 3 programme is once daily oral doses of 160 mg enteric-coated darapladib, a 40 mg dose was chosen for this hepatic impairment study to guard subject safety in the event of substantial exposure increase in the hepatic impaired group. Darapladib has shown concentration-independent protein binding in humans. Additionally, an earlier non-enteric-coated formulation of darapladib has shown a dose-proportional increase in exposure, indicating linear pharmacokinetics. The less-than-dose-proportional increase in exposure seen with the enteric-coated tablets of darapladib is thought to be absorption related because enteric coating does not alter the elimination of the drug. The impact of enteric coating on absorption is not expected to be affected by hepatic impairment. Therefore, the effects of hepatic impairment observed at 40 mg can be extrapolated to the higher 160 mg dose.

Darapladib has been shown to be a low extraction ratio drug that is extensively bound to plasma proteins. Both total and unbound fractions of darapladib were determined in this study. For total darapladib, small increases in total and peak exposure (20% and 7%, respectively) were observed in the moderately hepatic impaired subjects compared with healthy subjects. Exposures of unnbound darapladib increased in the moderately hepatic impaired subjects with an average 62% and 45% higher AUC(0,24 h) and C_max compared with healthy control subjects, respectively.

As stated in the results, only a subset of samples (3 of 17 samples per subject) was directly analyzed for unbound darapladib concentrations. The degree of protein binding (%) in each sample was then calculated from the unbound and corresponding total concentrations. No concentration-dependent protein binding was apparent over the concentration range observed in this study (1.5 to 26.1 ng ml⁻¹). Additionally, protein binding was similarly very high (≥99.9%) at concentrations of 5 μg ml⁻¹ and 10 μg ml⁻¹ in human plasma. Thus, an average of the three protein binding values was taken for each subject. Unbound darapladib concentration in the other 14 samples per subject was then calculated by multiplying the average protein binding value of that subject by the total darapladib concentration. This derivation resulted in 33 of 288 (11%) unbound darapladib concentration data points falling below the LLoQ of the assay (1 pg ml⁻¹). Indeed, darapladib is a very highly plasma protein bound drug with unbound darapladib circulating at low concentrations and the majority of the unbound concentration−time profile hovering around the LLoQ with C_max being three-fold and five-fold of the LLoQ in the healthy and hepatic impaired subjects, respectively. Large variability was observed in the pharmacokinetic parameters of unbound darapladib concentration, especially in the hepatic impaired patients (CV% = 86% for AUC and 102% for C_max). Overall, the difference in unbound darapladib pharmacokinetics between healthy and moderately hepatic impaired patients was observed with considerable ‘noise’ in the data.

Albumin is the major binding component in human plasma and concentrations of serum albumin are reduced in patients with liver disease. Plasma protein binding of darapladib is very high in humans at >99.9%, with serum albumin being the main binding component. In the present study, protein binding of darapladib was not significantly affected by hepatic impairment, with 99.96% (90% CI 99.957, 99.968%) and 99.97% (90% CI 99.964, 99.978%) plasma protein binding in moderately hepatic impaired and healthy matched controls, respectively. Additionally, darapladib has been shown to be a low hepatic clearance drug, with its estimated intrinsic clearance being <20% (GlaxoSmithKline, data on file). Changes in unbound fraction are not important for drugs with low intrinsic clearance. Lastly, a relationship was not apparent between unbound darapladib AUC(0,τ) or C_max and serum albumin concentrations.

To contextualize the difference seen in unbound darapladib exposure, the exposure range in the moderately hepatic impaired subjects observed in this study was compared with the exposure range in CHD patients administered the same 40 mg dose and formulation in the darapladib phase IIB trial. Assuming these CHD patients have similar protein binding characteristics to the healthy control subjects in the hepatic impairment study, post hoc analysis results of unbound darapladib exposure showed significant overlap with that observed from the moderately hepatic impaired patients in this study.

For the three metabolites of darapladib, exposures in the moderately hepatic impaired subjects ranged from 13% lower to 69% higher than the healthy control subjects. However, all three metabolites circulate at very low concentrations compared with parent darapladib (<5% of parent darapladib AUC). Therefore, the slight variations in metabolite levels in moderately hepatic impaired subjects are not considered of concern.

The results of this phase 1 study show that use of darapladib 40 mg appears to be safe and the pharmacokinetics of darapladib remain relatively unchanged in subjects with moderate hepatic impairment. As severe hepatic impaired patients were not included in this study and hepatic metabolism is the primary route of elimination of darapladib, results of this study cannot be extrapolated to severely impaired patients. In conclusion, the difference in darapladib exposure between moderate hepatic impaired patients and healthy subjects is not considered clinically significant. Dosage adjustment is unlikely to be needed in patients with mild or moderate liver disease.

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare the submitted work was sponsored by GlaxoSmithKline, PLC, all authors are current full-time employees of GlaxoSmithKline PLC and there are no other relationships or activities that could appear to have influenced the submitted work.

This study was sponsored by GlaxoSmithKline. The pharmacokinetic analyses were conducted by Covance Laboratories, Inc, under the direction of Clinical Pharmacology Modeling and Simulation, Quantitative Sciences, GlaxoSmithKline. The sponsors and authors thank the subjects who volunteered to participate in this study and acknowledge Ken Wiesen, PhD (Medicus International New York), for writing assistance funded by GlaxoSmithKline. The sponsors and authors also thank the investigators: Dr Thomas C. Marbury from the Orlando Clinical Research Center and Dr Patricia Pardo from Miami Research Associates.