Australian public assessment report for Terlipressin

II. Quality findings

Terlipressin is a synthetic vasopressin analog derived from the natural hormone, lysine-vasopressin (LVP). The substance (structure of the monoacetate reproduced below) is comprised of twelve amino acids each with the Lconfiguration (except for glycine, which does not have a chiral centre), and in which the Cterminal glycine is amidated. There is a disulfide bridge between Cys⁴ and Cys⁹.

Figure 1. Chemical structure of terlipressin

H-Gly-Gly-Gly-Cys-Tyr⁵-Phe-Gln-Asn-Cys-Pro¹⁰-Lys-Gly-NH₂

Free Base

Terlipressin Monoacetate

The substance is manufactured by conventional solid phase peptide synthesis and is isolated nominally as the diacetate pentahydrate salt (but containing a variable amount of acetic acid) by lyophilisation. As such, particle size and polymorphism are not relevant. The drug substance is freely soluble in water.

The drug is presented as a white to off white lyophilised powder in clear, USP Type I glass vials fitted with a grey bromobutyl rubber stopper and aluminium seal with a green plastic flip-off cap. Each vial contains terlipressin (as the diacetate) equivalent to 0.85 mg of terlipressin “free base”.¹

The specifications applied to the drug substance are satisfactory. The limit applied to the residual trifluoroacetic acid impurity has been reviewed and accepted by Medicines Toxicology Evaluation Section of the TGA.

Drug product

The product proposed for registration is a sterile, lyophilised powder, presented in clear glass vials for IV administration following reconstitution with 5 mL of sterile 0.9% sodium chloride injection. Each vial of product contains 0.85 mg terlipressin free base and approximately two equivalents (0.084 mg) of acetic acid. The drug substance is present mainly in the form of terlipressin diacetate.

The product (Lucassin) is sterilised by filtration prior to lyophilisation, and contains no antimicrobial preservative.

The finished product specifications were acceptable.

Adequate stability data were provided to support the proposed shelf life of 2 years² at 28°C stored in the original carton to protect from light. The reconstituted injection was shown to be stable when stored at 2° - 8°C such that the following recommendation is supported: “If not administered immediately, the reconstituted solution should be refrigerated (2 - 8°C) up to 24 hrs prior to use”. Terlipressin is incompatible with dextrose solutions. This is reflected in the proposed product information (PI).

Sterility and safety-endotoxins aspects of the submission were evaluated and were acceptable.

Biopharmaceutics

No bioavailability data are required as the product, after reconstitution, is a simple aqueous solution that is given IV.

Quality summary and conclusions

There were no objections in respect of chemistry, manufacturing and controls to registration of the powder for injection.

III. Nonclinical findings

Introduction

Terlipressin (triglycyl-lycine-vasopressin [LVP]) is a synthetic vasopressin analog that acts on vasopressin receptors, both as a pro-drug for lysine vasopressin (LVP) and probably with pharmacological activity on its own, albeit with lower potency. The endogenous vasopressin is arginine vasopressin (AVP), a nonapeptide with arginine in the 8 position. LVP is a porcine vasopressor peptide with lysine in the 8 position. There are 4 known receptors that bind vasopressin with significant affinity: V_1a (vasopressor), V_1b (pituitary), V₂ (renal) and OT (oxytocin; uterine). V_1a receptors are expressed in the liver, vascular smooth muscle, brain and in many other tissues. In the vasculature, V_1a receptors mediate vasoconstriction in response to vasopressin binding. V_1b receptors in the anterior pituitary mediate the ACTH releasing effects of AVP while V₂ receptors in the collecting duct of the kidney mediate the antidiuretic action of AVP. Oxytocin (OT) receptors are expressed in the uterus, the mammary gland, the ovary and several other tissues. OT receptors mediate the uterine contracting effect of OT. Terlipressin is intended to act at the V_1a receptors, mediating vasoconstriction, particularly in the splanchnic area, resulting in an increase in arterial volume, increase in arterial pressure and normalisation of endogenous vasoconstrictor systems resulting in increased renal blood flow and improved renal function.

The nonclinical data consisted of published papers (with terlipressin or LVP) and a small number of sponsor commissioned studies (with terlipressin). A large body of data was submitted to support efficacy in animal models of liver disease but the toxicological dossier was quite limited. While pivotal repeat dose toxicity studies were compliant with Good Laboratory Practice (GLP), the design and conduct of these studies were suboptimal, limiting the value of the findings. Reproductive toxicity studies were restricted to examinations of embryofetal toxicity.

Pharmacology

Primary pharmacology

Pharmacology studies examined receptor binding and vasoconstrictor activity (in vitro and in vivo) of terlipressin and LVP. LVP bound to both the human and rat V_1a receptors with similar affinity (K_i/K_d 2.3−8 nM; 2–8 times the clinical plasma level of LVP). No information was provided on binding to dog vasopressin receptors. Terlipressin had some activity at the human V_1a receptor in a functional assay, albeit with lower potency than LVP (>100 times). The putative metabolite of terlipressin, monoglycyl-LVP, had similar potency to terlipressin (pressor and antidiuretic activities in rats), while another putative metabolite, diglycyl-LVP, had lower potency. All four compounds had 54−100% the efficacy of AVP at the human V_1a receptor as assessed by an in vitro functional assay.

Terlipressin had vasoconstrictive activity in isolated systemic and splanchnic vessels from rats and small arteries of the human tubo-ovarian vasculature. The vasoconstrictive activity of terlipressin was less potent than both LVP and AVP. Terlipressin (≥30 nM; ~0.6 times the clinical maximum plasma concentration [C_max] of terlipressin) decreased the coronary blood flow and impaired myocardial performances of an isolated rabbit heart. These cardiac effects were significantly reduced on hearts pre-treated with a selective V_1a receptor antagonist, confirming the vasoconstrictive activity occurred through this receptor.

The chosen models of liver disease were rats with induced portal hypertension or cirrhosis. These are generally considered acceptable models for end stage liver disease. Dysfunction of the splanchnic circulation resembles that in patients with late stage liver disease and changes in handling of sodium and water by the kidney are similar to those observed in HRS patients. In these models, terlipressin decreased portal vein pressure and superior mesenteric arterial blood flow, increased mean arterial pressure and total peripheral resistance. A cardiodepressant effect (decreased heart rate and cardiac output) was also seen. Terlipressin redirected blood flow from the gut and skin (and other organs) to the kidneys. The median effective dose (ED₅₀) was determined to be 12.3 mg/kg based on mean arterial pressure; estimated C_max 28.9 ng/mL³, 0.5 times the clinical Cmax. The onset of pressor responses to terlipressin appeared to be more rapid (immediate) than accounted for by LVP formation, suggesting that terlipressin contributed, at least initially, to the pressor response. Terlipressin was less potent than AVP and LVP in increasing blood pressure. With terlipressin, maximal activity was delayed and the duration of action was longer. The pressor activity of monoglycyl-LVP and diglycyl-LVP showed a similar profile to that of terlipressin. The vasopressor activity was similar in rats and dogs. As the animal studies were acute, normalisation of endogenous vasoconstrictor systems and improvement in renal function were not assessed.

Terlipressin given during haemorrhage was less effective than when given during a stable state in experimental portal hypertension or cirrhosis. This splanchnic hyporesponse was associated with an overexpression of constitutive nitric oxide synthase and cyclooxygenase-1 in the superior mesenteric artery and increased glucagon release due to blood retention in the stomach. While the splanchnic response was diminished during haemorrhage, the systemic responses to terlipressin (increased mean arterial pressure and decreased cardiac output) were retained.

Overall, the animal pharmacology data presented by the sponsor suggested that terlipressin reduces blood flow to the skin, stomach and small intestine, while increasing flow to the liver and kidney. The delay in onset and long duration of action supports the proposed dosage regimen of a bolus IV injection every 6 hours (h), rather than a continuous infusion that would be required for the LVP metabolite. LVP had similar affinity at rat and human V_1a receptors and had similar vasopressor activity in rats and dogs, supporting the use of these animal models in the toxicity studies.

Secondary pharmacodynamics

The binding of terlipressin or LVP to other receptors was not extensively studied. LVP had similar binding affinity at the human V_1b, V₂ and V_1a receptors but less binding affinity at the human oxytocin receptor (K_i 25 nM). The binding affinity of LVP was approximately half that of AVP at all receptors. LVP had similar affinity at the rat V_1a, V_1b, V₂ and oxytocin receptors (K_D 1.7–8 nM). In rats, LVP had similar pressor (V_1a) and antidiuretic (V₂) activity, with 27 times lower oxytocin activity. Terlipressin had similar pressor and antidiuretic activity in one rat study but twice the pressor activity relative to the antidiuretic activity in another study. A dose dependent antidiuretic effect was seen at 0.05–1 µg/kg subcutaneous (SC) terlipressin in rats but marked natriuresis was seen at 5–20 µg/kg SC terlipressin (estimated C_max 12 ng/mL at 5 µg/kg; 0.2 times the clinical C_max). In dogs, the antidiuretic activity of LVP was ~15% that of AVP but similar pressor activity to that seen in rats, suggesting a difference in sensitivity of the V₂ receptor in this species. This is not unusual as species-specific differences in susceptibility at the V₂ receptor are known (Manning et al., 2008).⁴ At the proposed clinical dose, the pressor effects of terlipressin (and/or its metabolite LVP) are likely to predominate over the antidiuretic effect to increase renal perfusion.

Safety pharmacology

Specialised safety pharmacology studies were not conducted but effects on the central nervous, cardiovascular, gastrointestinal, renal and respiratory systems were investigated in toxicity and pharmacology studies. Clinical signs generally indicative of central nervous system (CNS) toxicity were seen in mice, rats and dogs. These included lethargy (≥20 mg/kg IV in mice, ≥0.15 mg/kg IV in rats, ≥0.031 mg/kg IV in dogs), piloerection (≥500 mg/kg IV in mice), ataxia and mobility problems (≥1.5 mg/kg IV in rats, ≥0.15 mg/kg IV in dogs). These clinical signs were transient, lasting up to 60 minutes (min) post-dose, consistent with the exposure to terlipressin and are probably related to its vasoconstrictor activity. A No Observable Effect Level (NOEL) was not established but, as lethargy occurred at approximately the clinical Cmax (in dogs), some of these effects may be seen clinically.

In pharmacology studies, terlipressin decreased coronary blood flow, increased mean arterial pressure, increased total peripheral resistance, decreased heart rate and reduced the cardiac index, consistent with its pharmacology on V_1a receptors in the smooth muscle vasculature. In a 28 day repeat dose toxicity study, there was a slight increase in the QT(c) interval in male dogs treated with 0.125 mg/kg IV twice daily (bd) terlipressin. There was no evidence of QT prolongation in treated females. These findings are difficult to interpret as electrocardiogram (ECG) recordings were conducted >16 h post-dose in males and >3 h post-dose in treated females and plasma levels of terlipressin and LVP would have been low to negligible at these times. Therefore no firm conclusions can be drawn from these data. Based on the known pharmacology of the drug, some cardiovascular findings would be expected clinically.

Intraperitoneal (IP) injections of terlipressin (≥0.1 mg/kg; NOEL 0.05 mg/kg, estimated exposure ratio at Cmax [ERCmax] 2) to rats had a gastroparetic effect, reducing the phasic motility of the stomach. This started 3−10 min post-dose and lasted up to 90 min, consistent with the duration of exposure to pharmacologically active material. Vomiting and defaecation were seen in dogs treated with ≥0.031 mg/kg IV bd (ERCmax 1.3; NOEL not established). The clinical signs of gastrointestinal disturbance abated within 1 h post-dose and are consistent with the vasoconstrictive action of terlipressin and LVP on the smooth muscle vasculature. These findings suggest that some gastrointestinal disturbances may be expected in the clinical setting.

SC administration of terlipressin (0.05−1 mg/kg) to rats had a dose dependent antidiuretic effect. The antidiuretic effect of terlipressin was slower and more sustained than LVP, probably due to the conversion of terlipressin to LVP. The antidiuretic activities of terlipressin and LVP were 2.2−3.3 U/mg and 284 U/mg, respectively, in rats. The antidiuretic potency of terlipressin and LVP was approximately equivalent to their pressor potency in this species. In dogs, however, LVP was less potent in the V₂ mediated antidiuretic activity than the V_1a mediated pressor activity. However, at the proposed clinical dose, the pressor effects are likely to predominate over the antidiuretic effect (see Secondary pharmacodynamics).Immediately after terlipressin administration, laboured and heavy breathing were seen in rats (≥0.15 mg/kg IV, estimated ERCmax6; NOEL not established) and dogs (≥0.125 mg/kg IV; ERCmax 3 at the NOEL). These respiratory difficulties lasted for the duration of exposure to pharmacologically active material and are consistent with vasoconstrictive activity on smooth muscle.

Pharmacodynamic drug interactions

The sponsor submitted a number of published papers describing the effects of terlipressin combined with other drugs in portal hypertensive or cirrhotic rats. Octreotide and β-blockers are frequently administered to patients with cirrhosis. The α- and β-adrenoceptor agonist, dobutamine, and the α-adrenoceptor antagonist, DL-028, reduced the systemic pressor effects of terlipressin. The combination of terlipressin and the β-adrenoceptor antagonist, propranolol, produced a greater reduction in portal pressure in portal hypertensive rats, than terlipressin alone. This combinatory effect was not seen in cirrhotic rats. The effect of octreotide in portal hypertensive rats was mixed, depending on the order of administration. Administration of octreotide to portal hypertensive rats 15 min after terlipressin appeared to attenuate the effects of terlipressin, while the reverse order had no significant effect. Alkaloid vasodilators had varying effects on the splanchnic response to terlipressin. Tetramethylpyrazine enhanced the portal hypotensive effects of terlipressin but reduced the systemic pressor and cardiodepressant effects. Whereas tetrandine, when co-administered with terlipressin attenuated both the splanchnic and systemic effects of terlipressin in portal hypertensive rats.

Pharmacokinetics

Pharmacokinetic studies with terlipressin were limited. Plasma kinetic profiles were assessed in repeat dose toxicity studies but technical problems limited the usefulness of the acquired information. The severe vasoconstrictive activity of terlipressin impeded blood sampling in the first 2 h post-dose in rats, although adequate data could be obtained. Following IV administration, plasma terlipressin levels fell rapidly in rats, with an elimination half-life of 0.09–0.22 h. The fall in terlipressin levels coincided with an increase in LVP concentration, peaking at 5–10 min and falling below the limit of detection by 1 h post-dose. Results in the dog study are of questionable reliability due to reported sample mislabelling and discrepancies in the time of blood collection. Furthermore, terlipressin levels were unexplainably high in a number of pre-dose samples. Blood samples from the control group were not collected and assayed, as recommended in the TGA-adopted EU guideline, so contamination of blood samples cannot be dismissed and the validity of the results is not assured.⁵ It is noted that, with chronic administration to dogs, terlipressin levels fell below the level of detection 1 h post-dose. The fall in terlipressin levels coincided with an increase in LVP levels, as expected. However, a second peak of terlipressin occurred 2−4 h post-dose, after LVP levels fell below the limit of detection. The second terlipressin peak is difficult to explain. A reformation of terlipressin from LVP is unlikely and terlipressin is unlikely to undergo enterohepatic recirculation intact. Due to the questionable reliability of the plasma concentration data, limited quantitative information can be gained from the dog study. Nonetheless, qualitatively, the plasma kinetic profile in dogs was similar to that in rats and humans; the elimination half-lives for terlipressin and LVP were short and the plasma kinetic profile of terlipressin and LVP in humans support the proposed 4 times daily dosage regimen to achieve sustained vasoconstrictive activity.

Following IV administration of [³H-Tyr]-LVP to rats, radioactivity was widely distributed with the kidney, liver, small intestine, neurohypophysis, adenohypophysis having appreciable levels of radioactivity, 1 h post-dose. The radioactivity in the kidneys, liver and small intestine was attributed to the degradation product ³H-tyrosine, consistent with these organs having high metabolic activity on terlipressin. Radioactivity in the neurohypophysis and adenohypophysis are consistent with the presence of vasopressin receptors in these tissues.

In vitro studies indicated that, in rats, terlipressin is metabolised by various tissues, including the liver, kidney and heart. Significant metabolism of terlipressin was also observed in homogenates of human liver and myometrial tissues. Subcellular fractionation studies indicated the majority of the metabolic activity in these tissues was associated with the cytosolic and mitochondrial fractions and limited activity was associated with the microsomal fraction. Metabolism of terlipressin is likely to involve peptidases, initially to form LVP, then cleavage of the C-terminal glycine amino group, with subsequent degradation. There was no evidence of metabolism of terlipressin in fresh human plasma, erythrocytes or whole blood and serum. The metabolism of terlipressin is likely to be similar in animals and humans.

Only 2−8% of the administered terlipressin was excreted in the urine of cats. In rats that received LVP, only 0.5% of the administered drug was excreted intact. The low level of urinary excretion in animals is consistent with findings in humans where <1% of the injected material is excreted in urine (Forsling et al., 1980).⁶ These data confirm the extensive role of metabolism in the clearance of terlipressin.

Pharmacokinetic drug interactions

In in vitro assays, there was no significant inhibition or induction of the human cytochrome P450 (CYP) isozymes, CYP1A2, 2A6, 2B6, 2C9, 2C19, 2D6, 2E1 and 3A4 or inhibition of CYP2C8 at terlipressin concentrations up to 5000 ng/mL (~80 times the clinical plasma concentration at 5 min [C5 min]). As terlipressin is not appreciably metabolised by CYP450 enzymes, drug interactions involving CYP450 enzymes are unlikely.

Toxicology

General toxicity

Single dose toxicity

Dedicated single dose toxicity studies were not conducted but single dose ranging studies were conducted prior to the mouse micronucleus study and the 7 day repeat dose toxicity studies in rats and dogs. As such, the observation period post-dose was only 3 days, rather than the 14 days recommended in the TGA-adopted EU guideline.⁷ This is not a particular concern as a recovery period was included in the repeat dose toxicity studies. Mortalities in the mouse and rat studies occurred within 60 min of dosing and with perimortem bleeding observed, are likely to be due to severe vasoconstriction. Clinical signs of lethargy, dyspnoea, ataxia and open mouth gasping were observed in all species, starting immediately after dosing and lasting for ~1 h. These clinical signs were most likely a result, either directly or secondary, of the cardiovascular effects of terlipressin (increased peripheral resistance, decreased heart rate, decreased cardiac output, decreased coronary blood flow and decreased skin blood flow). Vomiting and defaecation observed in dogs at ≥0.15 mg/kg IV terlipressin are consistent with the pharmacological action on the gastrointestinal tract. Maximum non-lethal doses were 20 mg/kg IV in mice (60 mg/m2), <2 mg/kg IV in rats (12 mg/m2) and 0.5 mg/kg IV in dogs (10 mg/m2; >7 times the maximum clinical dose of 2 mg [1.3 mg/m2]). Gross pathological analyses were only conducted in rats, where the kidneys and lungs appeared to be target organs. Renal and pulmonary findings were also seen in the repeat dose toxicity studies and therefore are discussed below.

Repeat dose toxicity

Repeat dose toxicity studies of up to 28 days duration were conducted in rats and dogs. All studies were conducted under GLP conditions. Adequate animal numbers of both sexes were used. The pivotal 28 day studies included a 14 day recovery period to assess the reversibility of toxicity findings. The duration of the repeat dose toxicity studies is not generally considered adequate to support the proposed clinical use of 14 days. Studies of at least 3−6 months duration would normally be required to support the clinical use of up to 1 month for the treatment of life-threatening conditions.^8,9 Therefore, the short duration of studies may not have revealed the full toxicological profile of terlipressin.

Toxicokinetic data were collected in both pivotal studies. Dosing in the rat study was once daily and animals were only exposed to pharmacologically active drug related material for 1 h per day. This does not fully replicate the clinical situation where HRS patients are expected to be continuously exposed to pharmacologically active material (terlipressin and LVP) with the proposed 4 times daily dosing regimen. Daily exposures (based on the area under the plasma concentration time curve [AUC]) were also low compared with the anticipated clinical exposure at the 1 mg/6 h clinical dose (Table 1); AUC data were not available for the maximum recommended human dose (MRHD) (2 mg/6 h) but animal/human exposure ratios would have been even lower at this clinical dose. Higher doses with the single dose regimen in rats would not have been feasible due to the high mortality rate (immediately after dosing) observed at the highest tested dose. However, as mortalities were observed immediately after dosing and were associated with high C_max values, greater exposure (AUC) could have been achieved by more frequent dosing, either twice or three times daily, a regimen more comparable with proposed clinical dosing. Due to the short daily duration of exposure to pharmacologically-active material, and the low overall exposures achieved, the full toxicological profile of terlipressin is unlikely to have been revealed in the submitted rat studies. Relative exposures are shown in Table 1.

Table 1.Relative exposures of terlipressin and LVP achieved in repeat dose toxicity studies

^Species (Strain)	^Study	^Dose (mg/kg/day) IV	^Terlipressin				^LVP
			^AUC0–24h^(ng·h/mL)	^C5 min^(ng/mL)	^ERAUC	^ERC5min	^AUC0–24h^(ng·h/mL)	^Cmax^(ng/mL)	^ERAUC	^ERCmax
^Rat^(SD)	^CB06-5013-R-TX	^0.15	^51	^352	^0.3	^5.7	^1.9	^9.3	^0.2	^8
		^0.5	^246	^1078	^1.5	^17	^16	^22	^1.6	^20
		^1.5	^830	^5385	^5	^87	^54	^75	^5	^68
^Human	^OT-0401	^1 mg every 6 h	^162	^62	^–	^–	^10	^1.1	^–	^–

^{ER, animal/human exposure ratio}

In the study in dogs, twice daily dosing was used and the duration of exposure to pharmacologically active material might be considered adequate to support the proposed clinical dosage regimen. However, some concerns with the conduct of the toxicokinetic portion of the study limit the value of the information gained (see above) and continuous exposure to pharmacologically active material cannot be verified. As exposure comparisons could not be made confidently based on AUC data, dose comparisons based on body surface area (BSA) were made for this species (Table 2). Doses used were generally low compared with clinical dosing at 1 mg every 6 h, and subclinical compared with the MRHD (2 mg every 6 h). Therefore, the full toxicological profile is unlikely to have been revealed in the submitted dog studies.

Table . Relative dose of terlipressin used in repeat dose toxicity studies

^Species (Strain)	^Study	^Dose (mg/kg/day) IV	^Dose (mg/m2/day)^a	^Relative dose based on BSA
^Rat^(SD)	^CB06-5013-R-TX	^0.15	^0.9	^0.4
		^0.5	^3	^1.4
		^1.5	^9	^4
^Dog^(Beagle)	^CB06-5030-D-TX	^0.15 bd	^6	^2.7
	^CB06-5089-D-TX	^0.031 bd	^1.2	^0.5
		^0.0625 bd	^2.5	^1.1
		^0.125 bd	^5	^2.3
^Human	^OT-0401	^1 mg every 6 h^b	^2.2	^–

^aUsing mg/kg to mg/m2 conversion factors of 6, 20 and 33 for rats, dogs and humans, respectively; ^bcorresponds to 0.85 mg terlipressin every 6 h

Despite the flaws in the design and conduct of the submitted toxicity studies, some drug related effects were seen and were largely associated with pharmacological activity. Male rats appeared to be more sensitive to the vasoconstrictive properties of terlipressin, based on mortalities. Deaths occurred at ≥0.5 mg/kg/day in males and 1.5 mg/kg/day in females. Most deaths occurred following the first dose, generally within the first hour. The cause of death was attributed to pulmonary oedema and/or haemorrhage due to reduced perfusion associated with pharmacological action. No mortalities were observed in the dog studies.

The kidney was a target organ in both rats and dogs. Trace to moderate nephritis was seen in rats treated with ≥0.5 mg/kg/day IV terlipressin (NOEL 0.15 mg/kg/day; exposure ratio based on AUC [ERAUC] 0.3) and dogs treated with ≥0.031 mg/kg bd (NOEL not established). Mild to severe renal tubular nephrosis with interstitial inflammation and fibrosis, probably ischaemic in origin, was seen in rats treated with 2 mg/kg/day IV terlipressin for 7 days. Mild lymphocytic infiltration was seen in dogs treated with ≥0.031 mg/kg bd IV terlipressin. These kidney lesions are likely associated with the pharmacological activity and they showed a trend to reversion after a 2 week treatment free recovery period.

Mild to moderate pulmonary inflammation was observed in dogs treated with ≥0.031 mg/kg bd IV (0.5 times the clinical dose based on BSA). This was suggested by the sponsor to be secondary to the pharmacological action of terlipressin – a result of a temporary reduction in blood flow caused by the vasoconstrictor activity of terlipressin followed by a subsequent reperfusion. No inflammation or other pulmonary changes were observed after a 2 week treatment free period, indicating the changes were reversible. While there was no evidence of fibrosis in the study, the duration of the study (1 month) is likely to be too short for fibrosis to develop.

Reduced testicular weights and mild to moderate seminiferous tubular degeneration was observed in male rats treated with ≥0.5 mg/kg/day IV terlipressin (ERAUC 0.3 at the NOEL). This could be due to a direct vasoconstrictor effect on the testes, resulting in reduced blood flow, ischaemia and tubular degeneration and/or a direct effect on V1a receptors in Leydig cells.

Skin pallor was seen in rats treated with ≥0.15 mg/kg/day IV. This is consistent with reduced cutaneous blood flow, a pharmacological effect of terlipressin. Enlarged adrenal glands were also seen in rats treated with ≥0.15 mg/kg/day IV terlipressin but without any corresponding microscopic changes. This could be due to a pharmacological effect, with terlipressin increasing adrenocorticotropic hormone (ACTH) levels by acting on the pituitary V1b receptors. Trace to mild cortical lymphoid depletion was seen in the thymus of female rats treated with ≥0.5 mg/kg/day IV terlipressin. The clinical relevance of this is not known. Due to the short duration of daily exposure, some toxicological findings that may be expected as a result of prolonged vasoconstriction (ischaemic events in cardiac and gastrointestinal tissues and in the skin) were not seen in the toxicity studies. These events are possible during clinical use of terlipressin.

The sponsor noted that the intended clinical treatment is for a maximum period of 2 weeks and that according to the TGA-adopted EU guideline¹⁰, for indications with a treatment duration of up to 2 weeks, 28 day repeat dose toxicity studies are acceptable.

However, the doses, exposures (based on AUC) and daily duration of exposure were all still considered too low to have adequately revealed the toxicological profile of terlipressin. The sponsor acknowledged that the clinical safety profile identified ischaemic events in cardiac, gastrointestinal and skin tissues, which would be associated with prolonged vasoconstriction, as potential risks. As these were not seen in the submitted toxicity studies, it confirms the inadequacy of the repeat dose studies to reveal the full toxicological profile of terlipressin.

Genotoxicity and carcinogenicity

The genotoxic potential of terlipressin was assessed in the standard battery of tests. Appropriate strains were used in the Ames test. Appropriate concentrations were used in the in vitro assays, with high doses used in the mouse micronucleus assay. All assays were appropriately validated. Although toxicokinetic data were not collected in the micronucleus test, clinical signs of toxicity confirmed the animals were adequately exposed. All assays returned negative results, which is to be expected for a peptide like terlipressin. No carcinogenicity studies were submitted. This is considered acceptable given the life threatening condition of the intended patient population, the negative genotoxicity results, and the anticipated short duration of clinical treatment.

Reproductive toxicity

No studies were submitted to assess the effect of terlipressin on male or female fertility. In repeat dose toxicity studies, reduced testicular weights and mild to moderate seminiferous tubular degeneration were seen in rats treated with ≥0.5 mg/kg/day terlipressin (ERAUC 0.3 at the NOEL) but no sperm analysis was conducted. These testicular effects could be the result of decreased blood flow and/or a direct effect on V_1a receptors in Leydig cells (Meidan and Hsueh, 1985).¹¹ After 14 days without treatment, a trend to reversion was seen. Taken together, some effects on male fertility may be expected with terlipressin treatment. No data were provided on the effects of terlipressin on female fertility. In response to a question, the sponsor noted that in the target patient population, both male and female patients with cirrhosis and end stage liver disease are recognised as already having a high baseline incidence of infertility as a result of hypothalamic-pituitary-gonadal dysfunction.

No embryofetal toxicity studies were conducted with terlipressin. The sponsor relied on published papers on embryofetal effects with LVP and AVP. Intramuscular injections of LVP (7 mg/kg) to pregnant rabbits from gestation day (GD)8 to GD11 had no significant effect on pregnancy and parturition. However, a single injection on GD20 resulted in a higher incidence of necrotic young and a greater incidence of abortion and vaginal bleeding. After a single injection of LVP on GD28, approximately 4 days before expected parturition, fetuses were dead within 1 h of the injection. Sites of blood accumulation were grossly visible in the placentas with microscopic evidence of focal dilatation and engorgement of maternal labyrinthine tubules with masses of erythrocytes. These findings were attributed to reduced blood flow to the uterus and placenta, resulting in ischaemic events that cause deterioration of tissues, as well as increased uterine contractions. Oxytocin receptor expression is up-regulated just prior to parturition (starting from GD28 in rabbits; Hinko and Soloff, 1992) and abortions have been reported in pregnant rabbits having elevated plasma oxytocin levels and increased uterine activity (Fuchs and Dawood, 1981).^12,13 LVP has some affinity at the oxytocin receptor and therefore these effects could be attributed to V_1a or oxytocin receptor activity. Intra-amniotic injection of vasopressin (assumed to be AVP) on GD15 in rats resulted in an increased incidence of dysmelia (Love and Vickers, 1973).¹⁴ These malformations of the digits are likely due to local hypoxia possibly associated with reduced blood flow to the fetus. Maternal injection of vasopressin (assumed to be AVP) on GD17 in rats caused transient hypoxia, bradycardia and serum ion changes in fetuses (Chernoff and Grabowski, 1971).¹⁵ Reduced blood flow to the uterus and placenta was seen in guinea pigs following administration of terlipressin (3−10 mg/kg IV). Clinically, terlipressin has also been reported to increase uterine activity and reduce endometrial blood flow (Laudanski and Akerlund, 1980).¹⁶ Therefore, the adverse embryofetal effects reported for LVP and AVP indicate a risk to fetal development during the clinical use of terlipressin. Given the malformations observed in rat studies with vasopressin analogues, Pregnancy Category D, as chosen by the sponsor, was considered appropriate.

No studies examined the excretion of terlipressin or its metabolite, LVP, in milk, or their effects on breastfed young.

Local tolerance

Injection site reactions were monitored as endpoints in repeat dose toxicity studies. Perivascular inflammation, ranging in severity from trace to severe was seen in rats treated with ≥0.15 mg/kg IV terlipressin. The incidence and severity of this inflammatory response did not have a clear relationship with dose. Mild to moderate perivascular inflammation was also noted in dogs at 0.124 mg/kg IV, while some inflammation was seen in single animals at lower doses, as well as in the control group. The incidence of haemorrhage was variable in the dog study but all males treated with ≥0.0625 mg/kg bd IV (similar to the maximum clinical dose on a mg/kg basis) all had moderate to severe haemorrhage at the injection site. Oedema was noted in a single male dog treated with 0.125 mg/kg bd IV. Haemorrhage and oedema are consistent with the pharmacological activity of terlipressin. All injection site reactions reversed after a 2 week treatment free period.

No significant haemolysis of human blood was observed at terlipressin concentrations up to 170 ng/mL (~2.7 times the clinical C_max of a 1 mg dose).

Immunogenicity

The immunogenicity of terlipressin was not assessed in animal studies. No anti-terlipressin antibodies were detected in plasma samples taken from 32 HRS-1 patients who had received terlipressin. Given the small size of terlipressin (12 amino acids) the risk of antibody production is low.

Impurities

The proposed specifications for impurities in the drug substance and degradants in the drug product are either below the ICH qualification thresholds or have been adequately qualified.¹⁷

Nonclinical summary and conclusions

Nonclinical data consisted of published papers and a small number of sponsor commissioned studies. The pharmacology of terlipressin was extensively studied but the overall toxicological component of the dossier was considered inadequate.

Terlipressin is a synthetic vasopressin analogue that acts on vasopressin receptors both as a pro-drug for lysine vasopressin (LVP) and with pharmacological activity on its own. Terlipressin had agonistic activity at the human V_1a receptor, but with >100 times lower potency than LVP. In animal models of liver disease, terlipressin decreased portal vein pressure and superior mesenteric arterial blow flow, increased mean arterial pressure and total peripheral resistance. A cardiodepressant effect was also seen. The delay in onset and long duration of action supports the proposed clinical dosage regimen. A splanchnic hyporesponse, with retention of the systemic effects, was seen during haemorrhage in experimental liver disease models.

LVP had similar binding affinity at the human V_1a (pressor), V_1b (pituitary) and V₂ (antidiuretic) receptors, but less binding affinity at the human oxytocin receptor. During clinical use, some binding to these receptors is possible but the animal studies indicate the pressor effects of terlipressin (and/or its metabolite LVP) are likely to predominate over the antidiuretic effect, to increase renal perfusion.

Specialised safety pharmacology studies were not conducted, but effects on the CNS, cardiovascular, gastrointestinal (GI), renal and respiratory systems were reported in toxicity and pharmacology studies. The majority of findings can be attributed to the pharmacological action of terlipressin or LVP; lethargy, ataxia and mobility problems; decreased coronary blood flow, decreased heart rate and reduced cardiac output; a gastroparetic effect; an antidiuretic effect at low doses but marked natriuresis at higher doses; laboured and heavy breathing. These effects lasted for the duration of exposure to pharmacologically active material. Effects on ECG parameters were not been adequately assessed.

In summary, the submitted animal pharmacology studies indicated that terlipressin reduced blood flow to the skin, stomach and small intestine, and increased the flow to the liver and kidney, thus supporting the proposed indication. However, as the animal studies were acute, normalisation of endogenous vasoconstrictor systems and improvement in renal function were not assessed.

When used in combination with terlipressin, the α- and β-adrenoceptor agonist, dobutamine, and the α-adrenoceptor antagonist, DL-028, reduced the systemic pressor effects of terlipressin in rat models of liver disease. The combination of terlipressin and the β-adrenoceptor antagonist, propranolol, produced a greater reduction in portal pressure in rats with portal hypertension, but not cirrhosis. Alkaloid vasodilators and octreotide had mixed effects on the splanchnic response to terlipressin.

Pharmacokinetic studies with terlipressin were limited. Qualitatively, the plasma kinetic profile was similar in animals and humans with elimination half-lives for terlipressin and LVP being relatively short. Tissue distribution of radioactivity was widespread following administration of [³H-Tyr]-LVP to rats. Terlipressin was extensively metabolised by peptidases in tissues. Less than 10% of the administered drug was excreted intact in urine. Pharmacokinetic drug interactions involving CYP450 enzymes are unlikely.

The toxicological dossier was limited. Single dose toxicity studies were conducted as dose ranging studies for the mouse micronucleus study and repeat dose toxicity studies. The maximum non-lethal doses were 20 mg/kg IV in mice, <2 mg/kg IV in rats and 0.5 mg/kg IV in dogs; >7 times the 2 mg clinical dose on a body surface area basis. Deaths were attributed to severe vasoconstriction, a pharmacological effect.

Repeat dose toxicity studies of 28 days duration were conducted in rats and dogs. Due to inadequacies in study design (short duration, the relatively low doses used, brief daily exposure [1 h in the rat studies compared with the intended 24 h clinical exposure]), the full toxicological profile of terlipressin is unlikely to have been revealed. Nonetheless, toxicity findings were seen in the kidneys, lungs and testes, which can all be attributed to the pressor activity of terlipressin. All of the findings showed a trend to reversion following a 2 week treatment free recovery period.

Terlipressin was not genotoxic in the standard battery of tests. No carcinogenicity studies were submitted, which is considered acceptable given the short duration of use and the negative genotoxicity findings.

No studies were submitted to assess the effect of terlipressin on male or female fertility. Testicular changes in rats treated with ≥0.5 mg/kg/day IV terlipressin suggest some effects on male fertility may be expected with terlipressin treatment. The exposure (AUC) ratio at the NOEL was 0.3. Assessment of embryofetal toxicity relied on published papers with vasopressins. Gestational administration of LVP or arginine vasopressin caused reduced blood flow to the placenta and increased uterine contractions resulting in abortions in pregnant rabbits and limb malformations in rat fetuses. No studies examined the excretion of terlipressin or LVP in milk, or its effects on breastfed young.

Treatment related reactions were observed at the injection sites of rats and dogs. Trace to severe perivascular inflammation was seen at >0.12 mg/kg IV (3 times the 2 mg clinical dose on a mg/kg basis), with haemorrhage seen in male dogs treated with ≥0.0625 mg/kg IV (similar to the maximum clinical dose on a mg/kg basis). All reactions were reversible, but the data suggest injection site reactions may be seen in the clinical setting. Haemocompatibility was demonstrated in an in vitro assay.

The proposed specifications for impurities in the drug substance and degradants in the drug product are either below the ICH qualification thresholds or have been qualified.

In summary, only limited toxicology data were submitted. Published data with vasopressin analogues provides sufficient evidence of human fetal risk, supporting the proposed Pregnancy Category (D). The repeat dose toxicity studies were of 28 days duration. The short daily exposure to terlipressin, particularly in the studies in rats, indicates toxicities associated with prolonged vasoconstriction would not have been seen. Doses in the studies in dogs were generally quite low, and there were some discrepancies in the conduct of the pivotal study that impacts on the reliability of the findings.

The full toxicological profile of terlipressin acetate is unlikely to have been revealed in the submitted data and the nonclinical studies were considered inadequate for a satisfactory risk assessment, and hence do not offer adequate support for the registration of terlipressin acetate (Lucassin) for the proposed clinical use. It was recognised that terlipressin has been in clinical use for more than 20 years in Europe, and therefore there may be sufficient clinical experience to offset the deficiencies in the nonclinical submission.

The sponsor contended that the nonclinical data should not be interpreted in isolation in the assessment of the risk/safety profile of terlipressin, citing a number of figures relating to clinical usage and citing a number of publications regarding postmarketing experience in other countries. It was, however, noted that it is not the realm of the nonclinical evaluator to comment on the clinical data. The conclusions from the nonclinical data remain as stated above.