Brief Article pubs.acs.org/molecularpharmaceutics
Effect on the Gastrointestinal Absorption of Drugs from Different Classes in the Biopharmaceutics Classification System, When Treating with Liraglutide Monika Malm-Erjefal̈ t,*,†,‡ Marianne Ekblom,†,§ Jan Vouis,∥ Milan Zdravkovic,† and Hans Lennernas̈ ⊥ †
Novo Nordisk A/S, 2860 Søborg, Denmark Phase I Services, Quintiles AB, SE-753 23 Uppsala, Sweden ⊥ Biopharmaceutics, Department of Pharmacy, Uppsala University, SE-751 23 Uppsala, Sweden ∥
S Supporting Information *
ABSTRACT: Like other GLP-1 receptor agonists used for treatment of type 2 diabetes, liraglutide delays gastric emptying. In this clinical absorption study, the primary objective was to investigate the effect of liraglutide (at steady state) on the rate and/or extent of gastrointestinal (GI) absorption of concomitantly orally taken drugs from three classes of the Biopharmaceutics Classification System (BCS). To provide a general prediction on liraglutide drug−drug absorption interaction, single-dose pharmacokinetics of drugs representing BCS classes II (low solubility−high permeability; atorvastatin 40 mg and griseofulvin 500 mg), III (high solubility−low permeability; lisinopril 20 mg), and IV (low solubility−low permeability; digoxin 1 mg) were studied in healthy subjects at steady state of liraglutide 1.8 mg, or placebo, in a two-period crossover design. With liraglutide, the oral drugs atorvastatin, lisinopril, and digoxin showed delayed tmax (by ≤2 h) and did not meet the criterion for bioequivalence for Cmax (reduced Cmax by 27−38%); griseofulvin had similar tmax and 37% increased Cmax. Although the prespecified bioequivalence criterion was not met by all drugs, the overall plasma exposure (AUC) of griseofulvin, atorvastatin, lisinopril, and digoxin only exhibited minor changes and was not considered to be of clinical relevance. KEYWORDS: liraglutide, GLP-1 receptor agonist, gastrointestinal absorption, drug interaction, biopharmaceutical classification system
■
steady state in subjects with T2D.5 However, in a nonclinical study, the delay in GE was diminished after 2 weeks of dosing with liraglutide.6 The present two-period crossover drug−drug interaction trial was conducted in healthy subjects to investigate if liraglutide will alter the GI absorption kinetics of concomitantly taken oral drugs and thereby require dose adjustment. The framework of the Biopharmaceutical Classification System (BCS)7−9 and the single-dose pharmacokinetics of four oral solid drugs, with different rate-limiting biopharmaceutical processes (gastric solubility, intestinal permeability, or both), were used as tools to evaluate the liraglutide drug−drug absorption interaction in vivo. A crossover design and standardized trial setup were applied to compensate for individual differences in metabolism of the oral drugs and potential food effects on GI drug absorption, and thereby highlight the effect of liraglutide on the GI absorption kinetics.
INTRODUCTION The traditional antihyperglycemic agents, such as sulfonylureas, thiazolidinediones, and alpha-glucosidase inhibitors, each have limitations in terms of tolerability and/or efficacy. Consequently, newer therapies such as the incretin-based therapies are increasing in popularity as these antihyperglycemic agents can significantly enhance glycemic control with a low risk of hypoglycemia and without weight gain as reflected in the clinical guideline.1 The glucagon-like peptide-1 receptor agonist (GLP-1RA) liraglutide (molecular weight 3751.2) is approved for once-daily use in adults with type 2 diabetes (T2D) as an adjunct to diet and exercise2 or in combination with oral antidiabetic drugs.2,3 Being a subcutaneously administered peptide, with low potential to influence cytochrome P450 enzymes in vitro, liraglutide is unlikely to cause any metabolic or gastrointestinal (GI) drug transport related drug−drug interactions (data on file, Novo Nordisk). However, like other GLP-1RAs, liraglutide delays gastric emptying (GE). The median time to reach maximum plasma concentration (tmax) of acetaminophen, a model drug often used to study any effects on GE,4 was delayed by up to 15 min when dosed concomitantly with liraglutide at © 2015 American Chemical Society
Received: Revised: Accepted: Published: 4166
April 10, 2015 August 21, 2015 October 1, 2015 October 1, 2015 DOI: 10.1021/acs.molpharmaceut.5b00278 Mol. Pharmaceutics 2015, 12, 4166−4173
Brief Article
Molecular Pharmaceutics
Figure 1. Two groups of subjects went through identical randomized, double-blind placebo-controlled protocols with stepwise dose escalations of liraglutide, or placebo, with subsequent crossover. Oral drugs in single doses were given so their absorption peaks coincided with the expected daily Cmax of liraglutide 1.8 mg at steady state. The upward arrows show start of the in-house period (for group B it includes the pH measurement).
medication was not permitted from 14 days prior to study start through to study completion, with the exception of oral contraceptives and medications required for the treatment of adverse events. The protocol, informed consent form, and the subject information sheet were reviewed and approved by the Medical Products Agency, Sweden, and the independent local institutional review board (Regional Ethics Committee in Uppsala, Sweden), according to local regulations. All subjects provided written informed consent before participation. This study was conducted at Quintiles Phase I unit (Uppsala and Luleå, Sweden) in accordance with the International Conference of Harmonization guidelines of good clinical practice and the principles described in the Declaration of Helsinki.16 Study Design. This was a randomized, double-blind, placebo-controlled, two-period crossover trial in two separate groups of subjects (Figure 1). Group A (n = 42) and group B (n = 28) went through identical 35-day protocols with oncedaily subcutaneous administration of liraglutide (Victoza 6 mg/ mL in a 3 mL FlexPen; Novo Nordisk A/S, Bagsvaerd, Denmark) in weekly increasing doses of 0.6 mg, 1.2 mg, and 1.8 mg or corresponding volume of placebo (100 μL, 200 μL, and 300 μL liraglutide vehicle, respectively, from a 3 mL FlexPen device; Novo Nordisk A/S). After approximately 1 week (at steady state) with liraglutide 1.8 mg, the subjects were admitted to the clinic for the drug−drug interaction studies. For group B only, measurements of the intragastric pH over 24 h were conducted at day 20 in each trial study period before the first drug−drug interaction investigation. The wash-out period between the single-dose interactions in the first and second trial periods, respectively, was 9 days (i.e., between atorvastatin and lisinopril in group A, and between griseofulvin and digoxin in group B). The crossover wash-out period (i.e., from liraglutide to placebo or vice versa) was 3−5 weeks. Single doses of atorvastatin 40 mg (Lipitor; Pfizer, New York, NY, USA) and lisinopril 20 mg (Zestril; AstraZeneca, London, U.K.) were investigated in group A while group B received single doses of griseofulvin 500 mg (Essential Generics, Surrey, U.K.) and digoxin 1 mg (Lanoxin; 4 × 250
Interaction was anticipated to be highest with immediaterelease tablets of low gastric solubility and high intestinal permeability: BCS class II drugs. Given in a high dose, a marked increase in the amount of drug in solution and an increased maximum plasma concentration (Cmax) and overall plasma exposure (area under curve; AUC) would be expected. Hence, griseofulvin10 (BCS class II; neutral pKa) was chosen as a model drug to be given in a high dose. As GLP-1 may inhibit gastric acid secretion (by 16− 50%),11,12 the intragastric pH was determined at steady-state concentration of liraglutide in this study. BCS class II weak acid substances would, in particular, increase their dissolution rate and amount of drug in solution at an elevated gastric pH. Consequently, the HMG-CoA reductase inhibitor atorvastatin13 (BCS class II; weak acid) was selected for investigation as statins are frequently used by subjects with T2D. The angiotensin-converting enzyme inhibitor lisinopril14 (BCS class III drug), also commonly used by patients with diabetes, was selected because, although it has high gastric solubility, the low intestinal permeability could be influenced if the intestinal transit time were prolonged by liraglutide treatment. Similarly, BCS class IV drugs with both low intestinal permeability and gastric solubility could be influenced by liraglutide. Hence, the cardiac glycoside and potential comedication digoxin15 (BCS class IV; a low dose drug [with a narrow therapeutic window], requiring a comparatively smaller volume for dissolution) was selected for interaction investigation. The primary objective of this clinical study was to investigate the effect of liraglutide (at steady state) on the rate and/or extent of GI absorption of concomitantly taken oral drugs representing three classes (II, III, and IV) of the BCS. Second, the effect of liraglutide on the intragastric pH was measured for 24 h.
■
EXPERIMENTAL SECTION Participants. Healthy male and female subjects between the ages of 18 and 55 years and with a body mass index (BMI) of 18−30 kg/m2 were eligible to participate. Concomitant 4167
DOI: 10.1021/acs.molpharmaceut.5b00278 Mol. Pharmaceutics 2015, 12, 4166−4173
Brief Article
Molecular Pharmaceutics μg, GlaxoSmithKline, Brentford, U.K.). The selected doses of atorvastatin, griseofulvin, or lisinopril were in the recommended dose range for standard therapeutic maintenance treatment while the dose of digoxin represented a standard rapid therapeutic loading dose. To ensure investigation of drug interaction at liraglutide maximum concentration, each oral drug was given so that its expected time to reach maximum plasma concentration (tmax) should coincide with the tmax of 8− 12 h for liraglutide (atorvastatin tmax = 1−2 h; griseofulvin tmax = 4−8 h; lisinopril tmax = 7 h; digoxin tmax = 0.5−1 h). Lisinopril and griseofulvin were given at the same time as administration of liraglutide or placebo while atorvastatin and digoxin were given at +5 and +7 h, respectively. Liraglutide or placebo was administered daily in the morning independently of breakfast. At the days for the drug−drug interaction, liraglutide or placebo was administered in the clinic after an overnight fast. For each individual drug interaction session in the two crossover periods, the oral drug (atorvastatin, griseofulvin, lisinopril, or digoxin) was given with the same standardized volume of water (i.e., one glass of tap water measuring approximately 200 mL). Griseofulvin and lisinopril, respectively, were given at fasting state, concomitantly with the liraglutide/placebo injection. For trial consistency and logistics, it was decided to wait 4 h after oral dosing of lisinopril or griseofulvin before a standardized meal was served. Atorvastatin and digoxin were given after the liraglutide/placebo injection (at +5 h and +7 h, respectively) to ensure Cmax to coincide with that of liraglutide (tmax of 8−12 h). As the actual time points for oral dosing would be between 12:00−14:00 h (atorvastatin) or 14:00−16:00 h (digoxin) after a night of fasting, it was decided, for the convenience of the subjects and for trial consistency and logistics, to serve these subjects a standardized breakfast (2000−2500 kJ) immediately after the liraglutide/placebo injection and two small standardized meals (high in carbohydrates, low in protein and fat) 2 h before and 2 h after the oral dosing. Meal composition and size were the same for each interaction period. Pharmacokinetic Sampling and Bioanalysis. Blood samples anticoagulated with EDTA, 4 mL each, were prepared for plasma analysis of the following: atorvastatin at −30 min, 0.5, 1, 2, 3, 4, 6, 8, 10, 12, 24, 36, 48, and 72 h; lisinopril at −30 min, 2, 4, 6, 8, 10, 12, 14, 16, 24, 36, 48, and 60 h; griseofulvin at −30 min, 1, 2, 3, 4, 6, 8, 10, 24, 36, 48, 60, and 72 h post dose of each respective drug. For digoxin analysis, serum samples were prepared from 4 mL of blood collected at −30 min and 0.5, 0.75, 1, 1.5, 2, 3, 4, 6, 12, 18, 24, 36, 48, 60, and 72 h post dose of digoxin. For each interaction period, a control blood sample was taken before administration of the second oral drug to ensure successful wash-out. In addition, 3 mL samples of EDTAanticoagulated blood were collected at −15 min and 4, 6, 8, 10, 12, 15, 17, and 24 h post dose of liraglutide or placebo treatment for bioanalysis of liraglutide. The drug concentrations were determined using liquid chromatography coupled with tandem mass spectrometry (LC−MS/MS; Shimadzu LC-10ADVP [Shimadzu; Duisburg, Germany], Micromass Quattro II [Waters; Hertfordshire, U.K.]) for atorvastatin and lisinopril. The assay ranges were 0.200−60.0 ng/mL and 0.500−150 ng/mL for atorvastatin and lisinopril, respectively. Griseofulvin was determined using high-performance liquid chromatography (HPLC). The assay range was 0.100−5.00 μg/ mL.
For bioanalysis of serum digoxin a solid phase, two-site chemiluminescent immunometric assay on an Immulite 2000 Analyzer (DPC Scandinavia, Mölndal, Sweden) was used. The assay is a standard commercial assay for diagnostic measurement of serum digoxin further validated for four concentration levels including the lower limit of quantification (LLOQ) and for two times dilution of samples. The assay range was 0.50− 7.8 ng/mL. Liraglutide was analyzed using an enzyme-linked immunosorbent assay (ELISA) with an assay range of 18−4500 pmol/L and a linearity validated up to a 16-fold dilution, as described previously.17 Intragastric pH Measurement. Intragastric pH was measured for 24 h using a glass electrode (Ingold M3; Medical Instruments Corporation, Solothurn, Switzerland) inserted via the nostril into the corpus region of the stomach. During the first 60 min the fasting pH was measured; thereafter the subjects were given the daily liraglutide dose and served standardized meals at predefined time points while pH continued to be measured for 23 h. Data were recorded with a Digitrapper pH 400 (Sierra Scientific Instruments, Los Angeles, CA, USA) at 4-s recording interval and analyzed using software Polygram NET v4.2 (SynMed AB, Spånga, Sweden). Safety Assessments. Adverse events (AEs) were recorded throughout the study. Results from physical examinations, clinical hematology and biochemistry laboratory tests, vital signs, and 12-lead ECGs taken before and as follow-up after each drug interaction period were also evaluated. Statistical Methods. The pharmacokinetic end points were derived from plasma or serum concentrations, and actual sampling times by the standard model-free, noncompartmental method, using model 200 for extravascular administration of WinNonlin Professional, v4.1.b (Pharsight Corp.; Mountain View, CA, USA). Statistical analyses of AUC, Cmax, and intragastric pH parameters, respectively, were performed on log-transformed data using a linear normal model, analysis of variance (ANOVA) with period and treatment as fixed effects and subjects as random effect. The AUC or Cmax of the oral drug was considered equivalent at concurrent liraglutide and placebo treatment if the 90% confidence interval (CI) of the estimated parameter ratio (liraglutide treatment/placebo treatment) was within the predefined interval of 0.80−1.25. The analysis of tmax was performed by use of a nonparametric method for paired samples (Hodges−Lehmann). The sample size was set large enough to provide a power of at least 80%.
■
RESULTS Subject Population and Disposition. A summary of demographics based on all enrolled subjects is presented in Table 1. Of the 70 subjects enrolled, 62 completed the study, 36 in group A and 26 in group B. Four subjects were withdrawn due to AEs, and none of these AEs was considered related to any liraglutide drug−drug absorption interaction effect. Two subjects withdrew after crossover to liraglutide (one due to tonsillitis, unlikely related to treatment; one due to rash, probably related to treatment) and one after crossover to placebo (due to pelvic fracture, unlikely related to treatment). One subject withdrew after crossover to placebo (due to increased aminotransferase activity, possibly related to treatment). One subject withdrew after crossover to placebo due to atorvastatin interaction caused by sensitivity to lisinopril in the previous liraglutide treatment period, and two subjects 4168
DOI: 10.1021/acs.molpharmaceut.5b00278 Mol. Pharmaceutics 2015, 12, 4166−4173
Brief Article
Molecular Pharmaceutics
and showed a 38% lower Cmax compared with placebo treatment (Tables 2 and 3; Figure 2A). The AUC of lisinopril showed a mean reduction of 15% at concurrent liraglutide treatment compared to placebo treatment (mean ratio 0.85, 90% CI: 0.75−0.97; Table 3). Cmax was decreased 27% and tmax appeared 2 h later (Tables 2 and 3; Figure 2B). The plasma exposure of griseofulvin during concurrent liraglutide treatment was equivalent to the exposure with placebo (mean ratio 1.10, 90% CI: 1.01−1.18; Table 3). The tmax of griseofulvin was not changed by liraglutide, although the Cmax of griseofulvin increased by 37% compared with that seen during placebo treatment (Tables 2 and 3 and Figure 2C). The serum exposure of digoxin showed a mean decrease of 16% during liraglutide treatment compared to during placebo treatment (mean ratio 0.84, 90% CI: 0.72−0.98; Table 3). Cmax was decreased by 31% and tmax appeared 1.12 h later (Tables 2 and 3, Figure 2D). The geometric mean steady state plasma exposure (AUCτ) for liraglutide 1.8 mg over 24 h was 789 nmol·h/L (range 483− 1549 nmol·h/L), with a geometric mean maximum concentration of 43.5 nmol/L (range 25.9−78.7 nmol/L) reached after 10 h (range 0−15 h). These data were comparable with results from previous multiple dose pharmacokinetic investigations in healthy subjects.18,19 Gastric pH Assessment. The median intragastric pH in the fasted state did not differ between liraglutide treatment (liraglutide 1.8 mg at steady state, 1.56 (range 1.11−2.81)) and placebo treatment (1.63 (range 1.2−2.01)). There were no clinically significant or statistically significant differences between liraglutide and placebo treatment when studying the entire 24-h measurement period, the 60 min fasting period, the postprandial period, or the supine period (Figures SI1 and SI2). Safety. There were no clinically significant changes from baseline for physical examination, vital signs, or electrocardiography (ECG), or in hematology and clinical chemistry laboratory parameters. Further, no major hypoglycemic episodes were reported during the trial. Minor hypoglycemic episodes of plasma glucose
