British Journal of Clinical DOI:10.1111/bcp.12572 Pharmacology

Correspondence Dr Marion Dahlke, Novartis Pharma AG, Pharmacokinetics of serelaxin CH-4056 Basel, Switzerland. Tel.: +41 79 5590130 Fax: +41 61 3246623 in patients with hepatic E-mail: [email protected] ------impairment: a single-dose, Keywords cardiovascular disease, healthy subjects, hepatic, pharmacokinetics, therapeutics open-label, parallel ------Received 4 September 2014 group study Accepted 9 December 2014 Zhanna Kobalava,1 Svetlana Villevalde,1 Yulia Kotovskaya,1 Accepted Article 2 3 4 Published Online Holger Hinrichsen, Marc Petersen-Sylla, Andreas Zaehringer, 16 December 2014 Yinuo Pang,5 Iris Rajman,4 Jasna Canadi,4 Marion Dahlke,4 Peter Lloyd6 & Atef Halabi3

1Peoples Friendship University of Russia, Moscow, Russia, 2Gastroenterologisch-Hepatologisches Zentrum Kiel, Kiel, 3Clinical Research Services, Kiel, Germany, 4Novartis Institutes for BioMedical Research, Basel, Switzerland, 5Novartis Institutes for BioMedical Research, Cambridge, MA, USA and 6KinDyn Consulting Ltd, Horsham, UK

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT • Data available from completed clinical studies of serelaxin, a recombinant form of AIMS human relaxin-2 currently in clinical Serelaxin is a recombinant form of human relaxin-2 in development for treatment of development for the treatment of acute acute heart failure. This study aimed to evaluate the pharmacokinetics (PK) of heart failure, have supported its clinical serelaxin in patients with hepatic impairment. Secondary objectives included dosing at 30 μgkg−1 day−1 as a 48 h i.v. evaluation of immunogenicity, safety and tolerability of serelaxin. infusion. METHODS • Hepatic impairment, relatively common in a This was an open-label, parallel group study (NCT01433458) comparing the PK of −1 −1 typically aged population of patients with serelaxin following a single 24 h intravenous (i.v.) infusion (30 μgkg day ) between patients with mild, moderate or severe hepatic impairment (Child–Pugh class A, B, C) heart failure, may significantly affect the and healthy matched controls. Blood sampling and standard safety assessments were pharmacokinetics of a drug to the extent conducted. Primary non-compartmental PK parameters [including area under the that dosage adjustment is warranted. serum concentration–time curve AUC(0–48 h) and AUC(0–∞) and serum concentration at 24 h post-dose (C24h)] were compared between each hepatic impairment group and healthy controls. RESULTS WHAT THIS STUDY ADDS A total of 49 subjects (including 25 patients with hepatic impairment) were enrolled, • This study investigated the of which 48 subjects completed the study. In all groups, the serum concentration of pharmacokinetics of serelaxin in patients serelaxin increased over the first few hours of infusion, reached steady-state at 12–24 h and then declined following completion of infusion, with a mean terminal with varying degrees of hepatic impairment, half-life of 7–8 h. All PK parameter estimates were comparable between each group to help inform potential dosing adjustments of patients with hepatic impairment and healthy controls. No serious adverse events, for serelaxin in this patient population. discontinuations due to adverse events or deaths were reported. No serelaxin treatment-related antibodies developed during this study. • This study revealed that the pharmacokinetics of serelaxin is not affected CONCLUSIONS by hepatic impairment, and hence dose The PK and safety profile of serelaxin were not affected by hepatic impairment. No adjustments are unlikely to be required. dose adjustment is needed for serelaxin treatment of 48 h i.v. infusion in patients with hepatic impairment.

© 2014 The British Pharmacological Society Br J Clin Pharmacol / 79:6 / 937–945 / 937 Z. Kobalava et al.

Introduction vided there were no safety or tolerability concerns, as judged by the investigator. The study was initiated on 29 Heart failure (HF) is a growing public health issue world- July 2011 and completed on 16 December 2011. wide due to an ageing population and improvement in The study protocol was reviewed by the Independent myocardial infarction survival [1, 2]. Globally, more than 23 Ethics Committee or Institutional Review Board for each million people suffer from HF [3] and in the United States centre, and the study was conducted according to the alone, each year over 1 million patients with HF are hospi- ethical principles of the Declaration of Helsinki. All subjects talized [4]. Indeed, acute heart failure (AHF) is the most provided written informed consent before any study- common cause of hospitalization in patients over 65 years specific procedures were conducted. of age [5]. Serelaxin is a recombinant form of human relaxin-2, a Inclusion and exclusion criteria naturally occurring peptide hormone [6], and is in clinical Inclusion criteria for all subjects (male or female) included development for the treatment of AHF [7, 8]. In the phase age 18–70 years, weight ≥50 kg and body mass index − III RELAX-AHF study, patients with AHF receiving serelaxin 18–35 kg m 2. For patients with hepatic impairment, inclu- (30 μgkg−1 day−1) on top of standard of care reported sig- sion required sitting vital signs as follows: systolic blood nificantly greater improvement in dyspnoea [as measured pressure (SBP), 100 to 159 mmHg; diastolic blood pressure by the visual analogue scale area under the curve (AUC) (DBP), 60 to 109 mmHg and pulse rate, 45 to 100 beats analysis] and were significantly less likely to experience min–1. Corresponding requirements for healthy controls in-hospital worsening of HF than those receiving conven- were 100 to <140 mmHg, 60 to <95 mmHg and 45 to 100 tional therapy alone. A 37% reduction (P = 0.02) in the 180 beats min–1, respectively. Exclusion criteria for all subjects day cardiovascular and all-cause mortality rates was also included hepatic impairment due to non-liver disease, observed with serelaxin treatment [9]. hypersensitivity to the study drug, significant electrocar- As part of the clinical development of serelaxin, studies diogram abnormalities and any surgical or medical condi- examining the pharmacokinetics (PK) of this agent in tion (other than hepatic impairment) that might have special populations were conducted to help inform clinical significantly altered the distribution or elimination of decision making in a range of patient populations, includ- drugs. ing those with hepatic impairment. Indeed, hepatic impairment is relatively common in patients with HF [10, Sample size calculations 11] and is associated with poor prognosis [11]. Therefore, Sample size calculations were based on the comparison of we report on the findings of a study designed to evaluate the PK profile (in terms of the parameters, AUC from zero the effect of varying degrees of hepatic impairment on the to last measurable concentration [AUC(0–tlast)] and AUC PK of serelaxin administered as a single intravenous (i.v.) from zero to infinity [AUC(0–∞)] within the hepatically continuous infusion. impaired groups vs. the profile of the control group. In previous studies in HF subjects, the coefficient of variation (CV) for AUC in the 30 μgkg−1 day−1 dose group was Methods 21% (unpublished data on file, Novartis; yinuo.pang@ novartis.com). Considering at least eight subjects per Study design and participants group with an observed ratio of 1.4 (equivalent to a 40% This was an open-label, parallel group study increase in drug exposure), the 90% confidence interval (ClinicalTrial.gov identifier, NCT01433458) conducted (CI) for the ratio of PK parameters when CV = 21% would be across two clinical research centres (Clinical Research 1.18 to 1.66. If the CV was increased to 30%, then the 90% Services, Kiel, Germany and ASCENT Clinical Research CI would be 1.10 to 1.78. This was considered as sufficient Solutions, Moscow, Russia) in patients with mild, moderate for the purpose of this study. Under the assumption of a and severe hepatic impairment (Child–Pugh class A, 5–6 10% type I error and a two-sided t-test for the difference of points, class B, 7–9 points and class C, 10–15 points, means on the log normal scale, when the CV = 21%, we respectively), and with demographically matched healthy have a power of 92% and when the CV = 30%, we have a control subjects with normal hepatic function. The criteria power of 70% to test the equality of means (ratio of 1) in for matching included race, age (± 5 years), gender and favour of an alternative of 1.4. weight (± 15%). None of the study participants suffered from HF. All subjects received a single 24 h i.v. infusion of Assessments and analyses – PK assessment serelaxin (30 μgkg−1 day−1). Each subject underwent a Serum samples for PK analysis were collected at baseline, screening visit between 21 and 2 days before dosing and during the 24 h infusion (time points: 15 min, 1, 3, 6, 12 and eligible subjects were admitted to the study site at base- 24 h), and up to 48 h after the end of infusion. The final line (on day –1). Further visits were scheduled on days 4 serum samples were collected on day 15, the last day of and 15 and the end of the study period. Between days 4 the study, to help interpret the immunogenicity assay and 15, subjects were released from the study site pro- results. Serum serelaxin concentrations were determined

938 / 79:6 / Br J Clin Pharmacol Serelaxin pharmacokinetics in patients with hepatic impairment using a commercially available, validated enzyme-linked anti-serelaxin IgG was used as positive controls and a neat immunosorbent assay (Quantikine ELISA kit DRL200, R&D pool of human sera as the negative control. Plates were Systems, Minneapolis, MN, USA), modified so that the cali- incubated for 1.5 h at ambient temperature, washed and bration standard curve and samples were prepared using the bound anti-serelaxin antibodies were detected by the pooled naive male and female human serum (i.e. in matrix), addition of ruthenium-labelled serelaxin. 2 × Read Buffer T as opposed to the kit-supplied assay diluent buffer. The with surfactant (Meso Scale Discovery, Rockville, Massa- method is sandwich ELISA, in which a monoclonal anti- chusetts, USA) was added into the wells to enhance body specific for human relaxin-2 is pre-coated onto wells electrochemiluminescence. of a microtitre plate. Calibration standards, quality control The assay cut-off point was set at a 5% false positive samples and study samples were added to the plate and rate. The assay sensitivity was 118 ng ml−1. Drug tolerance the captured serelaxin was detected by a goat polyclonal of the assay was determined to be at least 1 ng ml−1 (i.e. the anti-serelaxin horseradish peroxidase–conjugated second- assay positive control antibody at 120 ng ml−1 showed tol- ary antibody. Kit standards (in-house produced; Corthera erance for serelaxin, which was ≤1ngml−1). The assay was Inc., San Carlos, CA, USA, a Novartis affiliate company) were considered to have a sufficient drug tolerance to ensure used for the preparation of the calibration standards and the detection of anti-serelaxin antibodies in the presence quality control samples, both prepared from pooled of the very low serelaxin concentrations expected to be human naive male and female sera. Samples were run present at the time point when immunogenicity was undiluted, but if pre-dilution of the study specimens was assessed in this study (i.e. 15 days post-dose, serelaxin con- required, pooled human serum up to 1:100 was used. The centrations in this study were very low or below the limit of chromogen used for detection was tetramethylbenzidine. quantification of 15.6 pg ml−1). The capture assay uses a The working range of ELISA was 15.6 (lower limit of bridging format (serelaxin capture and detection), and the quantitation) to 500 pg ml−1. Serelaxin concentrations assay detects anything that can link the two serelaxin mol- below the lower limit of quantitation were considered to ecules on the plate. As serelaxin is identical to the endog- be zero. The assay was specific for human relaxin-2. There is enous human relaxin-2, the assay does not differentiate no cross reactivity with human relaxin-1 or human antibodies against endogenous relaxin vs. antibodies relaxin-3 at concentrations of up to 50 and 500 ng ml−1, against serelaxin. Samples found to be anti-serelaxin anti- respectively. Quality control samples were stable in serum body positive at screening were, therefore, confirmed by for up to 24 h at room temperature, for up to 1 week at using an immunodepletion assay, in which the excess 2°C–8°C, for up to 3.5 years at −20°C or −80°C and for up to serelaxin (1000 ng ml−1) was added to samples. Specificity six freeze-thaw cycles at −60°C to −90°C. PK samples from for serelaxin was confirmed if the signal in the assay was placebo-treated subjects were not tested. substantially decreased (threshold determined statistically to have a 1% false-positive rate). Confirmed positive Assessments and analyses – safety assessments samples were tested in a titration assay using a two-fold Standard safety assessments included physical examina- dilution series. tions, 12-lead electrocardiography, vital signs, standard Serum samples confirmed to be anti-serelaxin clinical laboratory evaluations (biochemistry, coagulation, antibody-positive were also tested for the ability to neu- haematology and urinalysis) and adverse event (AE)/ tralize the biological activity of serelaxin using a validated serious AE monitoring. bioassay (homogenous time-resolved Fourier resonance Anti-serelaxin antibody levels were evaluated in serum energy transfer). This bioassay measured changes in the samples on day 1 (pre-dose) and day 15 using a validated cyclic adenosine monophosphate (cAMP) production three-tiered assay approach following an initial screening induced by serelaxin binding to the receptors expressed assay. In brief, Multiarray® small spot plates (Meso Scale by human THP-1 (acute monocyte leukaemia) cells. The Discovery, MA, USA) were coated with serelaxin at ability of the sample to inhibit cAMP production in 1.0 μgml−1 (in-house produced; Novartis Pharma AG, response to serelaxin was analyzed to determine the pres- Basel, Switzerland), washed 3 × (phosphate-buffered saline ence of neutralizing antibodies. with 0.05% v/v Tween20®) and blocked with a blocking Control samples (cAMP calibrators and cAMP controls; solution (3% bovine serum albumin v/v in phosphate- CisBio, Codolet, France) were prepared in THP-1 cell buffered saline with 0.05% v/v Tween20®) for 1 h. The medium containing 1 mM of 3 isobutyl-1-methylxanthine serum samples (positive controls, negative controls and (Sigma, St Louis, MO, USA). The pool of human sera spiked study samples) were diluted five-fold in low cross buffer with 3000, 2000 and 1100 ng ml−1 of goat anti-serelaxin (Candor Bioscience, Wangen, Germany) and added in IgG (Corthera Inc.) was used as positive controls and the duplicate on the plate. Goat anti-serelaxin IgG (Corthera neat pool of human sera as negative controls. Anti- Inc., San Carlos, CA, USA, a Novartis affiliated company) serelaxin antibody positive study samples along with posi- served as a positive control for the analysis. The pool of tive and negative controls were diluted 10-fold (MRD) in commercially available naïve human sera spiked with dif- Bioassay diluent (THP-1 cell medium containing forskolin; ferent concentrations (6000, 360 and 120 ng ml−1) of goat Calbiochem, EMD Millipore Corporation, Billerica, MA,

Br J Clin Pharmacol / 79:6 / 939 Z. Kobalava et al.

USA) and pre-incubated with a fixed concentration of were compared across groups with varying hepatic impair- serelaxin (1.51 ng ml−1) before being mixed with THP-1 ment (mild, moderate and severe), and between the cells. Controls and serum samples were added in triplicate groups with hepatic impairment and healthy matched into a low volume 384-well white microtitre plate (Greiner controls. Bio-One GmbH, Frickenhausen, Germany). THP-1 cells (in- house; provided by Corthera Inc., San Carlos, CA, USA, a Assessments and analyses – statistical analysis Novartis affiliated company) were used directly from liquid Log-transformed PK parameters were analyzed separately nitrogen or nominally stored at −80°C. The cells were using a linear mixed-effect model, with each group of thawed and resuspended in cell medium at a density of hepatically impaired patients (mild, moderate and severe) 12.5 × 105 cells ml−1 (6250 cells per well) and added to set as fixed effect and each matched pair (demographically control and serum samples. Following 30 min incubation matched healthy controls) set as random effect. Least in a 37°C CO2 incubator, cAMP-d2 and labelled anti-cAMP- square means for each group, as well as contrasts between cryptate antibody, both prepared in lysis buffer, were the control and each hepatically impaired group, were cal- added to each well. The plate was then incubated for 1 h in culated on a log scale with corresponding 90% CIs. Back- the CO2 incubator at 37°C to facilitate cell lysis. The labelled transformed ratios and 90% CIs were also generated. In antibody and anti-cAMP-cryptate binds either to the intra- addition, regression analyses of the primary PK parameters cellular, unlabelled cAMP or to the cAMP-d2 molecule. The vs. the Child–Pugh score were performed for all patients amount of anti-cAMP-cryptate bound to cAMP-d2 is with hepatic impairment and illustrated by scatter plots. inversely proportional to the level of intracellular cAMP. To quantitate intracellular cAMP levels, the plate was read Results on a PHERAstar plate reader (BMG LABTECH GmbH, Ortenberg, Germany). A sample was considered positive Study population for neutralizing antibodies, if it generated a signal that was A total of 49 participants were enrolled into the parallel equal to or greater than the assay cut-off point (set at a 1% group study. Twenty-five subjects had varying degrees of μ −1 false-positive rate). The assay sensitivity was 1.45 gml . hepatic impairment (mild, moderate or severe hepatic Drug tolerance of the assay was determined to be at impairment corresponding to Child–Pugh class A, n = 9; −1 23.5 pg ml of serelaxin. class B, n = 8; and class C, n = 8) and 24 subjects were healthy matched controls. The Clinical Research Services Assessments and analyses – PK assessment centre (Kiel, Germany) enrolled 19 subjects (10 subjects The primary non-compartmental PK parameters were the with hepatic impairment and nine healthy subjects) and AUC(0–48 h) AUC(0–∞) and serum concentration at 24 h the ASCENT Clinical Research Solutions centre (Moscow,

(C24h) of serelaxin. To account for the actual delivered doses Russia) enrolled 30 subjects (15 subjects with hepatic in each subject (which sometimes were slightly different impairment and 15 healthy subjects). All were included in from the nominal dose rate of 30 μgkg−1 day−1 due to the safety analysis. Of these, 48 subjects completed the minor protocol deviations), dose-normalised primary PK study and were included for PK and immunogenicity parameters are reported. Additional non-compartmental analyses. One subject from the mild hepatic impairment PK parameters were mean residence time, terminal elimi- group withdrew consent after receiving the infusion for nation half-life (t1/2), systemic clearance from serum and 4 h, subsequently discontinued and was replaced. Overall, volume of distribution at steady-state (Vss). All PK param- demographic characteristics were relatively well balanced eters were determined using WinNonlin Pro Version 5.2 within the hepatic impairment groups and between the (Pharsight Corporation, Mountain View, CA, USA). For ter- hepatic and healthy control groups (Table 1). minal t1/2 estimation, a minimum of three data points were The majority of patients with hepatic impairment used. The AUC(0–48 h), AUC(0–∞) and C24 h of serelaxin had a medical history of liver cirrhosis, and continued with

Table 1 Baseline characteristics (safety population)

Mild hepatic Moderate hepatic Severe hepatic All patients with hepatic Healthy matched impairment (n = 9) impairment (n = 8) impairment (n = 8) impairment (n = 25) subjects (n = 24)

Age (years) mean (SD) 53.8 (5.6) 51.4 (6.4) 50.3 (12.2) 51.9 (8.3) 52.2 (9.5) Men, n (%) 6 (66.7) 5 (62.5) 5 (62.5) 16 (64.0) 15 (62.5) Race, Caucasian (n [%]) 9 (100%) 8 (100%) 8 (100%) 25 (100%) 24 (100%) Body mass index (kg m−2) mean (SD) 28.3 (4.1) 27.9 (5.1) 29.1 (4.4) 28.4 (4.4) 27.1 (4.1)

SD, standard deviation.

940 / 79:6 / Br J Clin Pharmacol Serelaxin pharmacokinetics in patients with hepatic impairment their regular treatments for concomitant diseases, includ- Overall, both SBP and DBP decreased with the infusion ing propranolol, spironolactone, ursodeoxycholic acid, of the study drug. The study participants showed signs of ornithine aspartate, torsemide, lactulose, vitamin B1 and hypotension or decrease in blood pressure (blood pres- vitamin K. This was not deemed to have affected the study sure <90/50 mmHg) that warranted stopping the study results. drug as specified in the study protocol. However, one subject experienced a transient decrease in SBP to 86 mmHg and in DBP to 51 mmHg (baseline 95/60 mmHg), Effect of hepatic impairment on PK of serelaxin but in the absence of clinical symptoms, infusion was con- Serelaxin was administered to all groups as a single 24 h tinued with no AE reported. Some changes (both increases i.v. infusion at a dose of 30 μgkg−1 day−1. Following the and decreases) in the pulse rate post-infusion were start of the infusion, serum serelaxin concentrations observed, but no clinically relevant trend was identified. increased rapidly over the first few hours and gradually Furthermore, standard clinical laboratory assessments approached a steady-state concentration at approxi- were performed for all subjects at screening, baseline, mately 12–24 h after the start of infusion. Upon comple- study drug infusion period, after the study drug infusion tion of the infusion, serum serelaxin concentrations period and the end of the study visit. Most of the subjects declined rapidly with a mean terminal half-life of 7–8 h had at least one clinical laboratory test result (haematol- (Figure 1). ogy, coagulation, biochemistry and urinalysis) outside of Both the PK profiles and non-compartmental PK the reference range at some time point during the study. parameters of serelaxin were generally comparable There were two subjects reported with mild AEs of between each group of patients with hepatic impairment decreased haemoglobin and hyperbilirubinaemia. These and healthy matched controls (Table 2). In addition, statis- were assessed as not related to the study drug and asymp- tical analysis of the primary non-compartmental PK param- tomatic. Overall, there were no clinically relevant changes eters did not show differences across groups with varying or temporal trends observed in clinical safety laboratory severity of hepatic impairment, and there were no signifi- tests across the study groups. cant differences between the patients with hepatic impair- No serelaxin treatment-related antibodies were ment and matched healthy subjects (Table 3). detected during this study. A total of 18 samples were To analyze further the relationship between serelaxin tested positive in the screening assay, which were con- disposition and hepatic impairment, regression analysis firmed as anti-drug antibody-negative except for one was performed. Data revealed a slight increase in the patient who had confirmed pre-existing antibodies at average value of AUC(0–∞) and C with increasing sever- 24 h baseline and also on day 15, with similarly low titres of ity of liver impairment. However, these effects were not antibodies that were found to be non-neutralizing, indicat- statistically significant (Figure 2). ing that these antibodies were not induced by serelaxin Temporary unplanned infusion interruptions to allow treatment. Overall, the screening assay false-positive rate for bathroom visits occurred in 29 of the 49 study partici- in healthy subjects was approximately 5%, which is con- pants. None of the interruptions occurred within the first sistent with the assay specification. 3 h of the infusion and all lasted maximally for 5–10 min. The impact of unscheduled dose interruptions on the PK of serelaxin was assessed in a sensitivity analysis in which the mean PK profiles and summary statistics of the PK param- Discussion eters for the whole group were compared with the results The degradation of serelaxin, a recombinant form of the for the sub-group of subjects without dose interruptions. naturally occurring human relaxin-2 peptide hormone The comparison indicated that the group mean PK profiles [12], likely occurs in vivo via the same pathways as endog- and PK parameters were similar when including or exclud- enous human relaxin-2. Given that the main elimination ing the subjects with dose interruptions, indicating that pathway for serelaxin is believed to be catabolism by the dose interruptions had minimal impact on the overall peptidases/proteases [13] throughout the body including conclusions of the study. liver, kidneys and other organs and tissues [14], and that there is no evidence to indicate an interaction with Safety and tolerability cytochrome P450 enzymes, it was not anticipated that Overall, five study participants (two patients with moder- hepatic impairment alone would have significant impact ate hepatic impairment and three healthy controls) on the PK of serelaxin. The results of this study confirmed reported AEs. Only one AE, a report of headache in the this expectation, and are consistent with the general healthy control group, was suspected by the investigator understanding of the PK and disposition of therapeutic to be related to the study medication. Two AEs, dyspepsia peptides and proteins. and headache, were reported as moderate AEs, and all This is the first study to investigate the PK of serelaxin in other AEs were reported as mild. No serious AEs were patients with varying degrees of hepatic impairment and reported and all reported AEs resolved without sequelae. healthy matched controls. Overall, there were no apparent

Br J Clin Pharmacol / 79:6 / 941 Z. Kobalava et al.

Linear view Semilogrithmic view A B Infusion Infusion SD) 20 )

± 100 –1 ) (

–1 10

1 10 0.1

Mean serelaxin Mean serelaxin 0.01

0 concentration (ng ml 0.001

concentration (ng ml 0 10203040 50 60 70 80 010203040 50 60 70 80 Time (h) Time (h) C D Infusion Infusion SD) 20 )

± 100 –1 ) (

–1 10

1 10 0.1

Mean serelaxin Mean serelaxin 0.01

0 concentration (ng ml 0.001

concentration (ng ml 0 10203040 50 60 70 80 010203040 50 60 70 80 Time (h) Time (h) E F Infusion Infusion SD) 20 )

± 100 –1 ) (

–1 10

1 10 0.1

Mean serelaxin Mean serelaxin 0.01

0 concentration (ng ml 0.001

concentration (ng ml 0 10203040 50 60 70 80 010203040 50 60 70 80 Time (h) Time (h)

Figure 1 Serelaxin serum concentration–time profiles. Data shown as arithmetic mean with standard deviation in linear and semilogarithmic views for groups of patients with mild (A, B), moderate (C, D) and severe (E, F) hepatic impairment compared with healthy controls. The control groups were demographically matched in terms of ethnicity, gender, age and weight. Subjects received a single 24 h i.v. infusion of serelaxin at 30 μgkg−1 day−1. Serum samples were collected at baseline, during the 24 h infusion and up to 48 h after the end of infusion. Serum serelaxin concentration was analyzed by enzyme-linked immunosorbent assay. , mild hepatic impairment (n = 8); , moderate hepatic impairment (n = 8); , severe hepatic impairment (n = 8); , matched healthy controls (n = 8) differences in the PK profile of serelaxin between the generally well tolerated and no events of hypotension patients with varying degrees of hepatic impairment and were reported in patients with hepatic impairment or healthy subjects matched in terms of ethnicity, gender, healthy controls. age and weight, suggesting comparable distributions of Only one out of the 49 study participants who received the primary PK parameters between the patients with the study drug was found to have detectable levels of hepatic impairment and healthy subjects. Serelaxin was anti-serelaxin antibodies. These were present at similarly

942 / 79:6 / Br J Clin Pharmacol Serelaxin pharmacokinetics in patients with hepatic impairment

Table 2 Primary and secondary non-compartmental pharmacokinetic (PK) parameters

Mild hepatic Moderate hepatic Severe hepatic Healthy matched PK parameter impairment (n = 8) impairment (n = 8) impairment (n = 8) subjects (n = 24)

Dose-normalized AUC(0–48 h) (h.kg.ng ml−1/g) 10.1 (15.7) 10.5 (34.3) 9.9 (27.9) 10.2 (23.1) Dose-normalized AUC(0–∞) (h.kg.ng ml−1/g) 10.2 (15.7) 10.6 (33.9) 10.0 (28.2) 10.3 (23.4) −1 Dose-normalized C24 h (kg.ng ml /g) 0.383 (21.8) 0.347 (70.1) 0.410 (27.0) 0.394 (25.2) MRT (h) 3.55 (12.3) 3.77 (29.5) 4.25 (24.3) 3.29 (38.2)

t1/2 (h) 7.64 (8.5) 7.74 (12.5) 6.87 (24.1) 7.82 (22.0) CL (ml h−1 kg−1) 98.1 (15.7) 94.4 (33.9) 99.6 (28.2) 97.5 (23.4) −1 Vss (ml kg ) 348 (20.7) 355 (49.2) 422 (29.5) 320 (41.1)

Data presented as geometric mean (coefficient of variation, %). AUC(0–48 h), area under the serum concentration–time curve from time 0 to 48 h after administration; AUC(0–∞),

area under the serum concentration–time curve from 0 to infinity; C24 h, serum concentration at 24 h post-dose; CL, systemic clearance from serum; MRT, mean residence time; t1/2,

terminal elimination half-life; Vss, volume of distribution at steady-state.

Table 3 Statistical analysis of primary non-compartmental PK parameters

Geometric means Observed ratio of geometric Degree of hepatic Hepatically impaired Healthy matched means, impaired/healthy Dose-normalized parameter impairment group (n = 8) controls (n = 8) (90% CI)

AUC(0–48 h) (h.kg.ng ml−1/g) Mild 10.1 10.4 0.973 (0.821, 1.15) Moderate 10.5 8.94 1.17 (0.942, 1.45) Severe 9.92 11.3 0.878 (0.667, 1.16) AUC(0–∞) (h.kg.ng ml−1/g) Mild 10.2 10.5 0.975 (0.822, 1.16) Moderate 10.6 9.02 1.18 (0.948, 1.46) Severe 10.0 11.5 0.877 (0.666, 1.15) −1 C24 h (kg.ng ml /g) Mild 0.383 0.41 0.924 (0.689, 1.24) Moderate 0.347 0.34 1.01 (0.665, 1.52) Severe 0.410 0.43 0.955 (0.753, 1.21)

AUC(0–48 h), area under the serum concentration–time curve from time 0 to 48 h after administration; AUC(0–∞), area under the serum concentration–time curve from zero to

infinity; C24 h, serum concentration at 24 h post-dose; CI, confidence interval. low levels on day 1, before the study drug exposure, as of patients with AHF. RELAX-AHF2 is assessing cardiovas- well as on day 15, and were non-neutralizing, suggesting cular mortality as the primary end point. that they were pre-existing anti-relaxin antibodies and not Overall, the PK profile of serelaxin was not affected by de novo antibodies induced by drug exposure. hepatic impairment and no safety concerns were raised A possible limitation of the current study was the dura- during the study. These results support the finding that tion of infusion. Participants received a 24 h i.v. infusion of adjustment of serelaxin dose will not be required in AHF serelaxin, while the recommended therapeutic dose in subjects with hepatic impairment. patients with AHF is a 48 h i.v. infusion [9]. However, the same dose rate of infusion intended for clinical application (30 μgkg−1 day−1) was used in this study and due to the Competing Interests rapid clearance of serelaxin, steady-state exposure was achieved around 12 to 24 h following the start of infusion All authors have completed the Unified Competing Inter- at a level similar to the intended clinical dose. Therefore, est form at http://www.icmje.org/coi_disclosure.pdf (avail- data from this study are considered relevant and applica- able on request from the corresponding author) and ble to the intended clinical dosing regimen of serelaxin at declare this study was supported by Novartis Pharma AG, 30 μgkg−1 day−1, without a dose adjustment for patients Basel, Switzerland. ZK, HH, AH, MPS, YK and SV received with hepatic impairment. Serelaxin is in clinical develop- remuneration for participating as investigators for the ment for the treatment of AHF. In addition to completed study. AZ, YP, IR, JC and MD are employees of Novartis. PL studies such as RELAX-AHF [9], there is an ongoing was a full time employee of Novartis at the time of the phase III study, RELAX-AHF2 (ClinicalTrial.gov identifier, study and development of this manuscript, and he NCT01870778), in a large population (approximately 6000) resigned from the company at the time of submitting this

Br J Clin Pharmacol / 79:6 / 943 Z. Kobalava et al.

A 20

18 r2 = 0.0013 P = 0.869 16 g)

m 14 / –1 12

10

8 ) (h.kg.ng ml • 6

4

AUC (0– AUC 2

0

45678 9 10 1112 Normal B Child-Pugh score 1.0

0.9 r2 = 0.0231 P = 0.478 0.8

0.7 g) m

/ 0.6 –1 0.5

0.4

(kg.ng ml 0.3 24h C 0.2

0.1

0.0

45678 9 10 1112 Normal Child-Pugh score

Figure 2 Regression analysis of primary pharmacokinetic parameters vs. degree of hepatic impairment. (A) Dose-normalized AUC(0–∞) of serelaxin and (B) dose- normalized C24 h of serelaxin plotted against Child–Pugh score for all subjects with hepatic impairment compared with healthy controls. , mild hepatic impairment (n = 8); , moderate hepatic impairment (n = 8); , severe hepatic impairment (n = 8); , healthy controls (n = 24) manuscript. Medical writing and editorial assistance was liver function of patients and participated in the finaliza- provided by Hannah Fichelson (MediTech Media), funded tion of the manuscript. AZ was the clinical trial leader for by Novartis Pharma AG. Novartis and responsible for the operational aspects of the study design, study planning and conduct (including pro- tocol writing and data analyses plan), and the writing of Contributors the clinical study report. YP was the pharmacokineticist for the project and responsible for the PK and ZK was the principal investigator and responsible for the immunogenicity aspects of the study design (including clinical conduct of the study and review of the final study protocol writing and data analyses plan), the final PK and report. HH participated in patient recruitment, determined IG data analyses and interpretation and the writing of the

944 / 79:6 / Br J Clin Pharmacol Serelaxin pharmacokinetics in patients with hepatic impairment clinical study report. YK, MPS, SV and AH were research Johnson MR, Kasper EK, Levy WC, Masoudi FA, McBride PE, physicians participating in the clinical conduct of the McMurray JJ, Mitchell JE, Peterson PN, Riegel B, Sam F, study. IR was responsible for the overall study design, Stevenson LW, Tang WH, Tsai EJ, Wilkoff BL. 2013 ACCF/AHA conduct and approval of the clinical study report and final guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American data interpretation. JC was the bioanalytical laboratory Heart Association Task Force on Practice Guidelines. J Am head for the outsourced sample analysis and acted as the Coll Cardiol 2013; 62: e147–e239. study monitor for PK and immunogenicity sample analysis performed at selected clinical research organizations 6 Teichman SL, Unemori E, Dschietzig T, Conrad K, Voors AA, (including approval of analytical plans for sample analysis, Teerlink JR, Felker GM, Metra M, Cotter G. Relaxin, a pleiotropic vasodilator for the treatment of heart failure. data quality control check and review of bioanalytical data Heart Fail Rev 2009; 14: 321–9. reports). MD was the sponsor’s medical monitor for the study, responsible for the medical and safety aspects of the 7 Teerlink JR, Metra M, Felker GM, Ponikowski P, Voors AA, study design (including the protocol writing and data Weatherley BD, Marmor A, Katz A, Grzybowski J, Unemori E, analyses plan), the safety analyses and interpretation, the Teichman SL, Cotter G. Relaxin for the treatment of patients with acute heart failure (Pre-RELAX-AHF): a multicentre, writing of the clinical study report and acted as the spon- randomised, placebo-controlled, parallel-group, sor’s medical monitor. PL was head of the PK and dose-finding phase IIb study. Lancet 2009; 373: 1429–39. bioanalytical group supporting this study and provided scientific advice in data analysis and interpretation. All 8 Teichman SL, Unemori E, Teerlink JR, Cotter G, Metra M. authors have reviewed and approved the manuscript. Relaxin: review of biology and potential role in treating heart failure. Curr Heart Fail Rep 2010; 7: 75–82. 9 Teerlink JR, Cotter G, Davison BA, Felker GM, Filippatos G, Greenberg BH, Ponikowski P, Unemori E, Voors AA, Adams REFERENCES KF Jr, Dorobantu MI, Grinfeld LR, Jondeau G, Marmor A, Masip J, Pang PS, Werdan K, Teichman SL, Trapani A, Bush 1 Dickstein K, Cohen-Solal A, Filippatos G, McMurray JJ, CA, Saini R, Schumacher C, Severin TM, Metra M. Serelaxin, Ponikowski P, Poole-Wilson PA, Stromberg A, van recombinant human relaxin-2, for treatment of acute heart Veldhuisen DJ, Atar D, Hoes AW, Keren A, Mebazaa A, failure (RELAX-AHF): a randomised, placebo-controlled trial. Nieminen M, Priori SG, Swedberg K. ESC guidelines for the Lancet 2013; 381: 29–39. diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of 10 Giallourakis CC, Rosenberg PM, Friedman LS. The liver in Acute and Chronic Heart Failure 2008 of the European heart failure. Clin Liver Dis 2002; 6: 947–ix. Society of Cardiology. Developed in collaboration with the 11 Allen LA, Felker GM, Pocock S, McMurray JJ, Pfeffer MA, Heart Failure Association of the ESC (HFA) and endorsed by Swedberg K, Wang D, Yusuf S, Michelson EL, Granger CB. the European Society of Intensive Care Medicine (ESICM). Liver function abnormalities and outcome in patients with Eur Heart J 2008; 29: 2388–442. chronic heart failure: data from the Candesartan in Heart 2 Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Failure: assessment of Reduction in Mortality and Morbidity Borden WB, Bravata DM, Dai S, Ford ES, Fox CS, Franco S, (CHARM) program. Eur J Heart Fail 2009; 11: 170–7. Fullerton HJ, Gillespie C, Hailpern SM, Heit JA, Howard VJ, 12 Metra M, Cotter G, Davison BA, Felker GM, Filippatos G, Huffman MD, Kissela BM, Kittner SJ, Lackland DT, Lichtman Greenberg BH, Ponikowski P, Unemori E, Voors AA, Adams JH, Lisabeth LD, Magid D, Marcus GM, Marelli A, Matchar DB, KF Jr, Dorobantu MI, Grinfeld L, Jondeau G, Marmor A, Masip McGuire DK, Mohler ER, Moy CS, Mussolino ME, Nichol G, J, Pang PS, Werdan K, Prescott MF, Edwards C, Teichman SL, Paynter NP, Schreiner PJ, Sorlie PD, Stein J, Turan TN, Virani Trapani A, Bush CA, Saini R, Schumacher C, Severin T, SS, Wong ND, Woo D, Turner MB. Heart disease and stroke Teerlink JR. Effect of serelaxin on cardiac, renal, and hepatic statistics – 2013 update: a report from the American Heart biomarkers in the Relaxin in Acute Heart Failure Association. Circulation 2013; 127: e6–e245. (RELAX-AHF) development program: correlation with 3 Bui AL, Horwich TB, Fonarow GC. Epidemiology and risk outcomes. J Am Coll Cardiol 2013; 61: 196–206. profile of heart failure. Nat Rev Cardiol 2011; 8: 30–41. 13 Bennett RG, Heimann DG, Hamel FG. Degradation of relaxin 4 Fang J, Mensah GA, Croft JB, Keenan NL. Heart family peptides by insulin-degrading enzyme. Ann N Y Acad failure-related hospitalization in the U.S., 1979 to 2004. J Am Sci 2009; 1160: 38–41. Coll Cardiol 2008; 52: 428–34. 14 Cossum PA, Dwyer KA, Roth M, Chen SA, Moffat B, Vandlen 5 Yancy CW, Jessup M, Bozkurt B, Butler J, Casey DE, Jr, R, Ferraiolo BL. The disposition of a human relaxin (hRlx-2) in Drazner MH, Fonarow GC, Geraci SA, Horwich T, Januzzi JL, pregnant and nonpregnant rats. Pharm Res 1992; 9: 419–24.

Br J Clin Pharmacol / 79:6 / 945 Copyright of British Journal of Clinical Pharmacology is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.