AU2009269958A1

AU2009269958A1 - Opioid agonist compositions for pain management

Info

Publication number: AU2009269958A1
Application number: AU2009269958A
Authority: AU
Inventors: Chin Beng Stephen Lim
Original assignee: IX BIOPHARMA Pte Ltd
Current assignee: IX BIOPHARMA Pte Ltd
Priority date: 2008-07-11
Filing date: 2009-07-13
Publication date: 2010-01-14
Also published as: WO2010005400A1

Description

WO 2010/005400 PCT/SG2009/000248 OPIOID AGONIST COMPOSITIONS FOR PAIN MANAGEMENT Field of the Invention The present invention relates to the administration of analgesics. More particularly, the 5 invention relates to opioid agonist compositions for intranasal administration and to a method of treating or managing pain using such compositions. Background Art The discussion of the background art in this specification is intended to facilitate an understanding of the present invention only. The discussion is not an acknowledgement 10 or admission that any of the material referred to herein is, or was, part of the common general knowledge as at the priority date of the application. A problem encountered using the nasal route for drug administration is that of insufficient drug absorption 15 . There are many different nasal administration devices used in prior studies to administer formulations intranasally, which include syringe-type devices with or 15 without needles attached to deliver fentanyl solution drop-wise, the Baxter-PCIVA device modified to connect to a nasal bottle to be used as nasal spray device, and a nasal applicator bottle manufactured by a local company 6 , and the Mist Actuator Device (MAD*). In the MAD@ applicator, the nasal solution is pushed through the device via a plunger syringe; and an aerosol mist is produced. These spray devices each provide different 20 volumes per spray (from 0.09 mL to 0.5 mL), and the volumes of instillation/spray per administration for most nasal spray devices also vary markedly, typically from 0.05 to 1.08 mL. This may result in inaccurate doses of active compounds being delivered intranasally. In addition, an excessive volume of spray per nostril may also not be retained in the nasal cavity' 4 5 . Any such excesses may run off into the nasopharynx and 25 be swallowed, or out of the nostrils and wiped off, hence the drug in the excess volume being wasted. The stability of active ingredients in nasal solutions when stored in these nasal applicator devices was investigated in this disclosure and the possibility that the plastic components of such devices could adsorb opioid agonist compositions were also investigated. There 30 is also a problem in formulating a nasal solution of a very poorly water-soluble drug into a concentrated aqueous nasal solution so that a small volume can be administered 15 . Permeation enhancers can be used to increase nasal absorption' 5 , but it would be preferable to provide a nasal formulation without such permeation enhancers which would still be effective in relieving pain. It is an object of the present invention to address certain 35 of these shortcomings. Disclosure of the Invention According to one aspect of the invention, there is provided a method of extending the shelf-life of a pharmaceutically acceptable opioid agonist composition, such as a fentanyl or fentanyl citrate composition, said method comprising maintaining the composition at a WO 2010/005400 PCT/SG2009/000248 -2 pH value in a range of between 3.8 and 6.5, between 4.5 and 6.3, between 5.0 and 6.2, or at a pH value of about 6.0. The invention extends, in another aspect thereof, to a pharmaceutically acceptable fentanyl composition having a shelf-life of at least 12 months when stored at room 5 temperature, the composition comprising fentanyl buffered to a pH value in a range of between 3.8 and 6.5, between 4.5 and 6.3, or between 5.0 and 6.2. In one embodiment, the composition has a pH value of about 6.0. The fentanyl may be suspended in an isotonic or hypotonic buffer, such as, for example, a phosphate buffered saline solution. According to another aspect of the invention, there is provided a pharmaceutical 10 composition for intranasal administration, the composition comprising an aqueous solution of an opioid agonist in a concentration sufficient to effectively treat or manage pain in a subject when administered intranasally. The method may include the step of suspending the opioid agonist, fentanyl, or fentanyl citrate in an isotonic or hypotonic buffer, such as a phosphate buffered saline solution. As 15 such, the method may include a composition comprising: fentanyl (or fentanyl citrate); a phosphate buffered saline solution; and sodium chloride, wherein the concentration of fentanyl is between 10 pg/L and 500 pg/L, between 100 pg/mL and 400 pg/mL, or about 300 pg/mL. The method may include the step of sterilising the opioid agonist/fentanyl composition 20 prior to dispensing or aliquoting the composition into a vial, cartridge, applicator, or container. The vial, cartridge, applicator, or container may hermetically sealed following addition of the composition. The vial, cartridge, applicator, or container may be labelled with instructions that the composition is to be used for the treatment of pain and that the composition has a shelf-life of between 12 months and 40 months from the date of 25 manufacture, when stored at room temperature. The vial, cartridge, applicator, or container may be labelled with instructions that the composition has a shelf-life of between 18 months and 30 months from the date of manufacture when stored at room temperature, typically at least 30 months from the date of manufacture when stored at room temperature. The vial, cartridge, applicator, or container may be adapted for the 30 intranasal delivery of the composition in a spray dosage form, typically adapted to deliver a droplet size generated during administration of between 2 pm and 50 pm, between 5 pm and 30 pm, typically about 10 pm. The opioid agonist may be fentanyl (N-(1-phenethyl-4-piperidyl)-N-phenyl-propanamide) or a pharmaceutically acceptable salt or enantiomer thereof. In a preferred embodiment, 35 the opioid agonist is fentanyl citrate. The fentanyl may be included in the composition in a concentration of between 0.5 pg/mL and 1 mg/mL. Preferably, the fentanyl may be included in a concentration of between 10 pg/mL and 500 pg/mL, most preferably about 300 pg/mL. The composition may have a pH in a range of between 3.0 and 8.0, preferably between 40 4.0 and 7.0. Most preferably, the composition has a pH of about 6.0.

WO 2010/005400 PCT/SG2009/000248 -3 In a preferred embodiment, the composition may comprise an aqueous fentanyl citrate solution having a pH of about 6.00, and a fentanyl citrate concentration of between 100 and 400 pg/mL, typically about 300 pg/mL. The composition may be suitable for single dose administration, consisting of about 1.5 pg 5 per dose of fentanyl citrate to 3 mg per dose of fentanyl citrate, preferably about 30 pg per dose of fentanyl citrate to 1.5 mg per dose of fentanyl citrate, most preferably about 50 pg per dose of fentanyl citrate The composition may provide a mean maximum plasma concentration (Cmzx) of fentanyl of about 0.10 to 1.5 ng/ml per 50 pg following nasal administration to a subject, typically 10 about 0.14 to 1.32 ng/ml per 50 pg fentanyl. In a preferred embodiment, the mean plasma concentration of fentanyl or fentanyl citrate provided may be about 0.1 to 2.0 ng/ml per 50 pg fentanyl or fentanyl citrate, following nasal administration. The composition may, in one embodiment, be formulated to be substantially free of preservatives, physiological or mucosal absorption enhancers, or propellants. 15 However, the composition may, in certain embodiments, include an absorption enhancer, present in a concentration of about 0.01 to about 10% by weight of the composition. The absorption enhancer may be a polysaccharide and may be positively charged. Preferably, the absorption enhancer is chitosan. The chitosan may be present in a concentration of about 0.01% to about 10% by weight of the composition, preferably between 0.2% and 20 2%, most preferably about 1%. The composition may include, in addition to fentanyl citrate, an aqueous carrier, such as water, and an organic solvent in an amount sufficient to enhance the solubility of the fentanyl in the water. The organic solvent, if required, may be a polar organic solvent selected from ethanol, 25 propylene glycol, glycerol, polyethylene glycol, and mixtures thereof. When present, the polar organic solvent may be ethanol and may be present in an amount of between 1 and 60%, preferably between 5% and 30%, most preferably about 10%. In a preferred embodiment, the organic solvent comprises about 10% ethanol (v/v) and/or about 5% (v/v) propylene glycol. 30 To aid dissolution of the therapeutic into the aqueous environment, a surfactant might be added. Surfactants may include anionic detergents, such as, for example, sodium lauryl sulfate, dioctyl sodium sulfosuccinate and dioctyl sodium sulfonate. Cationic detergents may be used and may include, for example, benzalkonium chloride or benzethomium chloride. Non-ionic detergents that may be included in the formulation as surfactants are Tween 80, 35 lauromacrogol 400, polyoxyl 40 stearate, polyoxyethylene hydrogenated castor oil 10, 50 and 60, glycerol monostearate, polysorbate 40, 60, 65 and 80, sucrose fatty acid ester, methyl cellulose and carboxymethyl cellulose. These surfactants may be present in the formulation of the compounds either alone or as a mixture in different ratios, and may be between 0.05% and 1%, preferably between 0.1% and 0.5%, preferably about 0.2%, either individually or 40 cumulatively. In one embodiment, the composition includes between 0.05% and 1% Tween WO 2010/005400 PCT/SG2009/000248 -4 80, preferably between 0.1% and 0.5% Tween 80, most preferably about 0.2% Tween 80, or as mixtures of the various surfactants. Formulations or compositions of the invention suitable for use with a spray applicator may also find application in a nebulizer, either jet or ultrasonic, which typically may comprise the 5 active compound suspended in water. The formulation may also include simple or complex carbohydrates and may include polyols, which, in one embodiment, may be mannitol, typically between 0.1% and 3% (v/v) mannitol, more preferably between 0.2% and 1%, most preferably about 0.3% (v/v) mannitol. In addition, the composition may be buffered using a pharmaceutically acceptable buffer, 10 such as a phosphate-based buffer. The composition may be formulated to be an isotonic or hypotonic solution. The composition may have an osmolality of from 200 mOsm/L to 500 mOsm/L, most preferably about 380 mOsm/L. The composition of the invention may, following intranasal administration to a subject, 15 provide a plasma concentration effective to treat, ameliorate or manage pain within a period of no more than 2 hours, preferably within 30 minutes, most preferably within 10 minutes, following intranasal administration. Accordingly, the composition may provide a plasma concentration of at least 0.1 ng/ml of fentanyl within 2 hours, preferably within less than 30 minutes, most preferably within less than 10 minutes. 20 The invention extends, in another aspect thereof, to a method of treating or managing pain by administering intranasally to a subject in need thereof, in an amount to effectively treat, ameliorate or manage pain, a pharmaceutical composition of the invention comprising an aqueous solution of fentanyl citrate or a pharmaceutically acceptable salt or enantiomer thereof. 25 The method may include the step of administering the composition intranasally to the subject at a concentration such that the subject experiences a fentanyl peak plasma concentration of between 10 and 90% of that achieved by intravenous administration, when administered intravenously at an identical fentanyl or fentanyl citrate dose. The method may include the step of administering the composition of the invention 30 intranasally to a subject in need thereof to provide a mean time to maximum plasma concentration (Tmax) of fentanyl of between about 2 minutes and about 180 minutes, preferably between about 2 and about 30 minutes, most preferably between about 2 and about 15 minutes, from the time of administration. Preferably, the method includes administering the composition at a fentanyl or fentanyl 35 - citrate concentration sufficient to provide the subject with a peak plasma concentration of at least 40%, preferably 60%, most preferably more than 70% of such peak plasma concentration. As such, the method may include the step of providing a composition of the invention in a dose sufficient to provide a bioavailability of more than 50% when administered 40 intranasally, when compared to a corresponding intravenous dose. Preferably; the method WO 2010/005400 PCT/SG2009/000248 -5 provides a bioavailability of more than 60%, most preferably more than 70%, when administered intranasally. The method may be used to treat or manage patients experiencing or reasonably expected to experience chronic pain, acute pain, or breakthrough pain. As such, the 5 method may include the step of administering the composition of the invention prophylactically. The method may include administering the composition of the invention to a subject to treat post operative pain; treat fracture pain; and to reduce post-anaesthetic emergency agitation. The method may include the step of providing the composition of the invention in the form 10 of a spray in which a majority of droplets when sprayed may be between 2 pm and 50 pm in diameter, preferably between 5 pm and 30 pm, most preferably about 5 pm to 10 pm in diameter. In one embodiment, more than 80% of the droplets may have a diameter of greater than 10 pm. Preferably, more than 90% of the droplets may have a diameter of greater than 10 pm, most preferably more than 95%. 15 The method may include the step of re-administering the compound of the invention to the subject pre-emptively or when an increase in pain levels is detected in the subject. As such, the compound may be re-administered by the patient him- or herself as a form of patient-controlled analgesia (PCA). The method may include administering the composition in a dosage form adapted for 20 intranasal administration, such as an intranasal spray. Advantageously, administration may be accomplished using a metered dose applicator, such as a pump spray dispensing device or applicator. Preferably, the applicator has a reservoir which is sealed and remains sealed throughout the lifetime of the composition. The method may include the step of administering the composition in a volume not 25 exceeding 300 pl per administration. The method may comprise administering a volume of between 5 pl and 200 pl, preferably between 50 pi and 180 pl per administration. In one embodiment, the method comprises administering a volume of no more than about 150 pl per administration. Accordingly, the invention extends to a dosage form comprising a pharmaceutically 30 effective amount of the pain relieving composition of the invention in a spray applicator suitable for intranasal delivery of the composition to a subject in need thereof. More specifically, the invention extends to an intranasal applicator for intranasal application of fentanyl citrate, the applicator having included therein a fentanyl citrate solution having a pH value within a range of 3.0 and 8.0, preferably between 4.0 and 7.0, 35 most preferably about 6.0, and having a concentration and spray volume to ensure a dose delivery of between about 1.5 pg and 3 mg, preferably between about 30 pg to 1.5 mg, most preferably about 100 pg fentanyl or fentanyl citrate per spray, upon administration. Advantageously, each administration of the dosage form results in an effective dose of.the pharmaceutical composition being delivered to a subject in a volume not exceeding 300 pl 40 per administration, preferably between 5 pl and 200 pi, most preferably between 50 pl and 180 pi per administration. In one embodiment, the dosage form comprises an effective WO 2010/005400 PCT/SG2009/000248 -6 pain relieving dose being delivered to a patient in a volume of no more than about 150 pl per administration, thereby providing adequate pain relief following administration of only a single intranasal application. The composition of the invention, when administered in accordance with the invention, 5 form of a spray in which a majority of droplets when sprayed may be between 2 pm and 50 pm in diameter, preferably between 5 pm and 30 pm, most preferably about 5 pm to 10 pm in diameter. In one embodiment, more than 80% of the droplets may have a diameter of greater than 10 pm. Preferably, more than 90% of the droplets may have a diameter of greater than 10 pm, most preferably more than 95%. 10 The pharmaceutical applicator may be a metered dose applicator, preferably an intranasal applicator. The invention extends to the use of fentanyl or fentanyl citrate in the manufacture of a medicament having pain relieving activity when administered intranasally. Typically, the fentanyl or fentanyl citrate when used in the invention would have a pH 15 value lower than 7.0, lower than 6.5, lower than 6.4, lower than 6.3, lower than 6.2, lower than 6.1, lower than 6.0. Yet another aspect of the invention provides a substance or composition for use in a method of treating or managing pain in a subject, the substance or composition comprising a therapeutically effective amount of fentanyl citrate suitable for intranasal 20 administration. The subject may be an animal or human. The subject, when human, may be an adult or a child. According to a still further aspect of the invention, there is provided a method of minimizing the sorption of fentanyl compounds by polymeric compounds present in 25 pharmaceutical applicators, the method including the step of charging said applicator with a solution comprising a pharmaceutically acceptable concentration of fentanyl, the solution being formulated to remain at a pH value within a range of about of 3.0 to 8.0 for the effective lifetime of the solution. Preferably the solution has a pH value within a range of about 4.0 to 7.0, most preferably 30 the solution has a pH value of about 6.0. In a preferred embodiment of the invention, the solution has a pH of less than, or about, 6.0. A further aspect of the invention provides a method of preparing a sterile intranasal pharmaceutical composition which is stable for at least 30 days without the need for preservatives or stabilizing agents, the method comprising the steps of: 35 mixing fentanyl citrate powder with water to a desired pharmaceutically effective concentration, thereby to provide a pharmaceutical fentanyl composition; sterilizing said composition; WO 2010/005400 PCT/SG2009/000248 -7 charging a reservoir of a suitable intranasal applicator under sterile conditions with the sterilized composition; and sealing the applicator hermetically while under such sterile conditions, the contents of the applicator reservoir remaining hermetically sealed during or following 5 discharging of the composition from the applicator. The composition may be sterilized by way of filter sterilization or autoclaving. The may be a nasal spray applicator manufactured in Western Australia. According to a still further aspect of the invention, there is provided a method of extending the shelf-life or expiry date of a liquid pharmaceutical composition by at least 30 months 10 from the date of manufacture when stored at room temperature, said method comprising maintaining the liquid pharmaceutical composition at a pH of about 6.0, wherein the liquid pharmaceutical composition comprises: fentanyl; a phosphate buffer solution; 15 sodium chloride, wherein the concentration of fentanyl is between 100 pg/mi and 400 pg/ml; the liquid pharmaceutical composition is sterile; the liquid pharmaceutical composition is sealed in a vial, cartridge, 20 applicator, or container; wherein the vial, cartridge, applicator, or container: is labelled with instructions that the composition is used for the treatment of pain; is labelled with instructions that the composition has a shelf-life of at least 30 months from the date of manufacture when stored at room temperature; 25 is adapted for the intranasal delivery of the pharmaceutical composition in a spray dosage form; and is adapted for delivery of a droplet size generated during administration between 2 pm and 50 pm. Further features of the invention will now be described by, way of non-limiting example 30 only, with reference to the accompanying description and figures. Drawings In the drawings: Figure 3.2.0 shows the concentration of sodium chloride g/1OOmL versus milliosmol; WO 2010/005400 PCT/SG2009/000248 -8 Figure 3.5.0.0 shows a chromatogram of fentanyl citrate 2 pg/mL; Figure 3.5.2.0.0 shows the acid hydrolysis of fentanyl (in 4M hydrochloric acid for 24 hours at 500C). Another peak appears at approximately 2.9 and was the breakdown product of fentanyl; 5 Figure 3.5.2.1.0 shows the alkali hydrolysis of fentanyl (in 4 M sodium hydroxide solution for 24 hours at 50*C; Figure 3.5.2.2.0 shows heat degradation of fentanyl (at 900C for 4 days); Figure 3.6.0 shows the sorption of fentanyl into plastic PVC dip-tube and EVA seal as a function of time (concentration of fentanyl nasal solution equals 113 pg/mL in phosphate 10 buffer pH 10 and stored at 370C). Equilibrium condition was reached in 5 hours; Figure 3.6.1 shows sorption of fentanyl into plastic Acetal reservoir as a function of time (concentration of fentanyl nasal solution equals 113 pg/mL in phosphate buffer pH 10 and stored at 370C). Equilibrium condition was reached in 10 days; Figure 3.6.2 shows sorption isotherm for fentanyl in PVC dip-tube, EVA seal and Acetal 15 reservoir [these plastic components were soaked in the 5 mL of fentanyl nasal solution (113 pug/mL), buffered at pH 10 using phosphate buffer and stored at temperature of 370C]. Figure 3.6.3 shows a Langmuir plot for sorption of fentanyl in PVC dip-tube, EVA seal and Acetal reservoir during equilibrium condition [the plastic component was soaked in 5mL of 20 fentanyl nasal solutions (20.9, 39.5, 87.1 and 113 pg/mL) at pH 10 and stored at 370C]; Figure 3.7.0 shows loss of fentanyl (initial concentration of 20.9, 39.5, 87.1 and 113 pg/mL) stored in applicator's amber glass bottle at 37*C, pH 10.0; Figure 3.7.0.0 shows the amount of fentanyl (pg) lost from solution in the presence of PVC dip-tube at pH 10.0 and storage temperature of 370C at initial concentration stated; 25 Figure 3.7.0.1 shows the amount of fentanyl (pg) lost from solution in the presence of EVA seal at pH 10.0 and storage temperature of 370C at the initial concentration stated; Figure 3.7.0.2 shows the amount of fentanyl (pg) lost from solution in the presence of Acetal reservoir at pH 10.0 and storage temperature of 370C at initial concentration stated; Figure 3.7.0.3 shows a reduced sorption curve (short time) for PVC dip-tube stored in 30 fentanyl 20.9, 39.5, 87.1 and 113 pg/mL at 37 0 C, pH 10.0; Figure 3.7.0.4 shows a reduced sorption curve (short time) for EVA seal stored in fentanyl 20.9, 39.5, 87.1 and 113 pg/mL at 37 0 C, pH 10.0; Figure 3.7.0.5 shows a reduced sorption curve (short time) for Acetal reservoir stored in fentanyl 20.9, 39.5, 87.1 and 113 pg/mL at 37*C, pH 10.0; 35 Figure 3.7.1.0: Loss of fentanyl (initial concentrations of 20.9, 39.5, 87.1 and 113 pg/mL) in PVC dip-tube stored at 37"C, pH 10.0; Figure 3.7.1.2: Loss of fentanyl (initial concentrations of 20.9, 39.5, 87.1 and 113 pg/mL) in EVA seal stored at 37C, pH 10.0; WO 2010/005400 PCT/SG2009/000248 -9 Figure 3.7.1.4: Loss of fentanyl (initial concentrations of 20.9, 39.5, 87.1 and 113 pg/mL) in Acetal reservoir stored at 370C, pH 10.0; Figure 3.7.3.0: PVC dip-tube and EVA seal stored in the fentanyl 54.9 pg/mL nasal solution with pH 8.0 (p=0.3 and 0.5), at the storage temperature of 50C; 5 Figure 3.7.3.1: PVC dip-tube stored in the fentanyl 54.9 pg/mL nasal solution with pH 8.0 (p = 0.3 and 0.5), at the storage temperature of 37 and 25*C; Figure 3.7.3.2: EVA seal stored in the fentanyl 54.9 pg/mL nasal solution with pH 8.0 (p = 0.3 and 0.5), at the storage temperature of 25 and 37"C; Figure 3.7.3.3: Acetal reservoir stored in the fentanyl 54.9 pg/mL nasal solution with pH 10 8.0 (p = 0.3 and 0.5) at the storage temperature of 5, 25 and 370C; Figure 3.7.3.4: PVC dip-tube stored in the fentanyl nasal solution (54.9 pg/mL) with pH 6.0 (p = 0.15 and 0.50) at the storage temperature of 5, 25 and 37*C; Figure 3.7.3.5: EVA seal stored in the fentanyl nasal solution (54.9 u/mL) with pH 6.0 (p = 0.15 and 0.50) at the storage temperature of 5, 25 and 37*C; 15 Figure 3.7.3.6: Acetal reservoir stored in the fentanyl nasal solution (54.9 u/mL) with pH 6.0 (p = 0.15 and 0.50) at the storage temperature of 5, 25 and 370C; Figure 3.7.5.0: Arrhenius type plot for the rate constants, a and 8 for the disappearance of fentanyl into the PVC dip-tube (pH 8.0); Figure 3.7.5.1: Arrhenius type plot for the rate constants a and /8 for the disappearance of 20 fentanyl into the EVA seal (pH 8.0); Figure 3.7.5.2: Arrhenius type plot for the intercept B of the bi-exponential equation obtained from the compartment model for the disappearance of fentanyl into the PVC dip tube and EVA seal (pH 8.0); Figure 3.7.7.0: Reduced sorption curve for fentanyl sorbed into PVC dip-tube (PVC dip 25 tube immersed in fentanyl nasal solution 54.9 pg/mL buffered with phosphate buffer at pH 8.0) stored at 5, 25 and 370C; Figure 3.7.7.1: Reduced sorption curve for fentanyl sorbed into EVA seal (EVA seal immersed in fentanyl nasal solution, 54.9 pg/mL, buffered with phosphate buffer at pH 8.0) stored at 5, 25 and 37"C; 30 Figure 3.7.8.0: Arrhenius plots of the natural logarithm of the diffusion coefficients (Ln D), against the reciprocal of the storage temperature (1/T), for the diffusion of fentanyl into PVC dip-tube and EVA seal at pH 8.0; Figure 3.7.9.0: Partition coefficient, LnR (plastic-water at pH 8.0) versus reciprocal temperature in Kevin for PVC dip-tube and EVA seal at the storage temperature of 37, 25 35 and 50C; WO 2010/005400 PCT/SG2009/000248 -10 Figure 3.7.13.0: The loss of fentanyl by sorption into the PVC dip-tube when the PVC dip tube was stored at 370C in the 5 mL fentanyl (50 pg/mL) nasal solutions at the pH 10.0, 8.0 and 6.0; Figure 3.7.13.1: The loss of fentanyl by sorption into the EVA seal when the EVA seal was 5 stored at 37*C in the 5mL fentanyl (50 pg/mL) nasal solutions at pH 6.0, 8.0 and 10.0; Figure 3.7.13.2: The loss of fentanyl by sorption into the Acetal reservoir when Acetal reservoir was stored at 370C in the fentanyl (50 pg/mL) nasal solutions at the pH 6.0, 8.0 and 10.0; Figure 3.7.13.3: The influence of pH on the uptake of fentanyl by PVC, EVA and Acetal 10 plastics when the PVC dip-tube, EVA seal and Acetal reservoir was immersed and stored at 37*C in the fentanyl nasal solution during equilibrium condition (storage time equals to 5 hours for PVC dip-tube and EVA seal respectively and for 10 days for Acetal reservoir). Note the typical S-type or the "titration curve" of this graph. F 0 = fraction of fentanyl remaining in solution during equilibrium condition. Mo, = fraction of fentanyl sorbed into the 15 plastic at equilibrium conditions; Figure 3.7.13.4: Relationship between the reciprocal of the half-time for sorption (1/t 12 ) and the nasal solution's pH value for the sorption of fentanyl (5mL aqueous nasal solution) by PVC dip-tube, EVA seal and Acetal reservoir at the storage temperature of 370C; Figure 3.8.0.0: Sorption number plots for fentanyl when the PVC dip-tube was stored in 20 various fentanyl nasal solutions (initial concentration of 20.9, 39.5, 87.1 and 113.0 pg/mL where the pH value was kept constant at 10.0) and stored at 370C. S, values obtained from the slopes of the graph for the four initial fentanyl concentrations of 20.9, 39.5, 87.1 and 113 pg/mL were 0.08, 0.06, 0.09 and 0.15h 1 respectively; Figure 3.8.0.1: Sorption number plots for fentanyl when the EVA seal was stored in 25 various fentanyl nasal solutions (initial concentrations of 20.9, 39.5, 87.1 and 113.0 pg/mL where the pH value was kept constant at 10.0) and stored at 370C. Sn values obtained from the slopes of the relationship for the four initial fentanyl concentrations of 20.9, 39.5, 87.1 and 113 pg/mL were 0.10, 0.05, 0.13 and 0.14h 1 respectively; Figure 3.8.0.2: Sorption number plots for fentanyl when the Acetal reservoir was stored in 30 various fentanyl nasal solutions (initial concentration of 20.9, 39.5, 87.1 and 113.0 pg/mL where the pH value was kept constant at 10.0) and stored at 370C. Sn values obtained from the slopes of the relationships for the four initial fentanyl concentrations of 20.9, 39.5, 87.1 and 113.0 pg/mL were 0.06, 0.10, 0.09 and 0.09d 1 respectively; Figure 3.8.1.0: Relationship between the sorption number (Sr) and the square of the 35 fraction un-ionised (f% 2 ) for fentanyl (PVC dip-tube and EVA seal stored in fentanyl nasal solution). The regression equations were as follows: Sn = 0.1586fU 2 and S, = 0.0537fu 2 for PVC dip-tube and EVA seal respectively; Figure 4.6.1.0: Chromatograms for fentanyl 0.2 ng/mL (10.66 minutes) and sufentanil 0.2 ng/mL (11.23 minutes); WO 2010/005400 PCT/SG2009/000248 Figure 4.6.1.2: Chromatogram from serum of volunteer number 1 (fentanyl peak at 10.03 minutes and sufentanil peak at 11.35 minutes); Figure 4.6.2.0: Standard Curve for Fentanyl Serum Concentrations using Sufentanil as the Internal Standard; 5 Figure 4.6.6.2.0: An estimate of the absorption slope and the lag time for fentanyl absorption following the intranasal administration of fentanyl nasal spray at the pH value 6.0; Figure 4.6.6.2.1: An estimate of the absorption slope and the lag time for fentanyl absorption following the intranasal administration of fentanyl nasal spray at pH 8.0; and 10 Figure 4.6.8.0: Summed pain intensity differences for movement pain for both cross-over sequences (n=23). Best Mode(s) for Carrying Out the Invention It has now been found that a reduction in the pH of opioid agonist compositions, such as fentanyl citrate compositions, to a range of between 4.0 and about 8.0 improves the shelf 15 life of the composition and decreases the sorption of the fentanyl by plastic components of common applicators, leading to decreased loss of active ingredient without affecting the efficacy of the fentanyl composition. This results in cost savings due to less active ingredient being required for normal storage conditions. This also allows the composition or applicator containing the composition to be stored at room temperature without 20 significant loss, breakdown, or sorption of the fentanyl in such applicators. Throughout the specification and claims, unless the context requires otherwise, the word "comprise" or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated integer or group of integers but not the exclusion of any other integer or group of integers. 25 Other definitions for selected terms used herein may be found within the description of the invention and apply throughout. Unless otherwise defined, all other scientific and technical terms used herein have the same meaning as commonly understood to one of ordinary skill in the art to which the invention belongs. Those skilled in the art will appreciate that the invention described herein is susceptible to 30 variations and modifications other than those specifically described. It is to be understood that the invention includes all such variations and modifications. The invention also includes all of the steps, features, compositions and compounds referred to or indicated in the specification, individually or collectively and any and all combinations or any two or more of the steps or features. 35 The present invention is not to be limited in scope by the specific embodiments described herein, which are intended for the purpose of exemplification only. Functionally equivalent products, compositions and methods are clearly within the scope of the invention as described.

WO 2010/005400 PCT/SG2009/000248 - 12 The invention described herein may include one or more ranges of values. A range of values will be understood to include all values within the range, including the values defining the range, and values adjacent to the range which lead to the same or substantially the same outcome as the values immediately adjacent to that value which 5 defines the boundary to the range. The entire disclosures of all publications (including patents, patent applications, journal articles, laboratory manuals, books, or other documents) cited herein are hereby incorporated by reference. Inclusion does not constitute an admission that any of the references constitute prior art or are part of the common general knowledge of those 10 working in the field to which this invention relates. 1.0 Definitions Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although any methods and materials similar or equivalent to those described 15 herein can be used in the practice or testing of the present invention, the preferred methods and materials are now described. All publications mentioned herein are incorporated herein by reference to disclose and describe the methods and/or materials in connection with which the publications are cited. 20 It must be noted that as used herein and in the appended claims, the singular forms "a", "and", and "the" include plural referents unless the context clearly dictates otherwise. As an example, reference to "a composition" includes a plurality of such compositions or formulations, while reference to "the opioid agonist composition" and "the fentanyl composition" would include references to one or more opioid agonist compositions or 25 fentanyl compositions and equivalents thereof known to those skilled in the art, and so forth. As used herein, the terms "active agent", "fentanyl" and "therapeutic agent," used interchangeably herein, are generally meant to refer to fentanyl or a pharmaceutically 30 acceptable fentanyl salt or enantiomer, as well as formulations or compositions including one or more of these compounds. Use of "active agent" or the phrase "fentanyl" or "fentanyl citrate", is not meant to be limited to use of, or formulations or compositions comprising, only one of these selected opioid compounds. Furthermore, reference to fentanyl alone is understood to be only exemplary of the drugs suitable for use in 35 formulations according to the invention, and is not meant to be limiting in any way. The term "subject" is meant any subject, generally a mammal (e.g., human, canine, feline, equine, bovine, etc.), in which management of pain is desired. The term "therapeutically effective amount" is meant an amount of a therapeutic agent, or a rate of delivery of a therapeutic agent, effective to facilitate a desired therapeutic effect. 40 The precise desired therapeutic effect (e.g., the degree of pain relief and source of the pain relieved, and the like) will vary according to the condition to be treated, the formulation to be administered, and a variety of other factors that are appreciated by those of ordinary skill in the art. In general, the method of the invention involves the WO 2010/005400 PCT/SG2009/000248 -13 suppression or mitigation of pain in a subject suffering from pain that may be associated with any of a variety of identifiable or unidentifiable aetiologies. The term "pain management" is used here to generally describe regression, suppression, or mitigation of pain, including acute and chronic pain, so as to make the subject more 5 comfortable as determined by subjective criteria, objective criteria, or both. In general, pain is assessed subjectively by patient report, with the health professional taking into consideration the patient's age, cultural background, environment, and other psychological background factors known to alter a person's subjective reaction to pain. Pain management may also be accomplished by self-administration by the subject. 10 "Applicator" or "device" as used herein is meant to refer to any device suitable for delivering the formulations for pain management to mucosal tissues, particularly nasal mucosal tissues. "Spray applicator" thus encompasses, but is not necessarily limited to, devices capable of delivering a liquid composition contained therein in a spray or droplet form using any mechanism of action, such as, for example, manual activation, manual 15 pumping, assisted pumping, compressed liquids or fluids, pressurisation, or the like. "Treatment" as in "treatment of pain" is used herein to encompass both a decrease in pain severity and/or intensity to provide partial or complete relief of pain and/or pain symptoms. The effect may be prophylactic in terms of completely or partially preventing or reducing the severity of pain. 20 Colourants and flavouring agents may all be included. For example, compounds may be formulated (such as by liposome or microsphere encapsulation) and then further contained within an edible product, such as a refrigerated beverage containing colorants and flavouring agents. To aid dissolution of the therapeutic into the aqueous environment, a surfactant might be 25 added as a wetting agent. Surfactants may include anionic detergents such as sodium lauryl sulfate, dioctyl sodium sulfosuccinate and dioctyl sodium sulfonate. Cationic detergents might be used and could include benzalkonium chloride or benzethomium chloride. The list of potential non-ionic detergents that could be included in the formulation as surfactants are lauromacrogol 400, polyoxyl 40 stearate, polyoxyethylene hydrogenated castor oil 10, 50 and 30 60, glycerol monostearate, polysorbate 40, 60, 65 and 80, sucrose fatty acid ester, methyl cellulose and carboxymethyl cellulose. These surfactants could be present in the formulation of the compounds either alone or as a mixture in different ratios. Additives which potentially enhance uptake of the compounds are for instance the fatty acids oleic acid, linoleic acid and linolenic acid. 35 Formulations suitable for use with a nebulizer, either jet or ultrasonic, will typically comprise the compounds suspended in water. The formulation may also include a buffer and a simple sugar (e.g., for protein stabilization and regulation of osmotic pressure). The nebulizer formulation may also contain a surfactant, to reduce or prevent surface induced aggregation of the compounds caused by atomization of the solution in forming the aerosol.

WO 2010/005400 PCT/SG2009/000248 -14 The formulation may include the use of an excipient, in other words a usually inert substance (such as gum arabic, syrup, lanolin, or starch) that fo-ms a vehicle for an active ingredient of the composition. 2.0 Use of nasal applicators 5 The embodiment demonstrates the effectiveness of intranasal fentanyl as analgesia to treat pain such as, in particular, breakthrough pain. The study is divided into two parts. The first part is an in vitro study which involved the formulation of fentanyl intranasal solution, the testing of its stability and shelf-life when stored in a nasal device which delivered a defined dose at specific volume. The in vivo study involved absorption studies 10 of this intranasal solution when administered in the human nasal cavity of volunteers. This disclosure addresses some of the issues mentioned above. Various concentrations of nasal fentanyl solutions were formulated using pure fentanyl powder to study the initial concentration effect. The nasal solution was buffered to pH values of 6, 8 and 10 using phosphate buffer to study the nasal solution stability in the nasal device at different pH 15 values. The pH values of 6, 8, and 10 were used to study the stability of un-ionised and ionised fentanyl when stored in a nasal device. The nasal solutions were also formulated with different ionic strengths of 0.18, 0.3 and 0.5 to study the potential "salting out" effect on the stability of fentanyl. Finally the nasal solutions at pH 6 and 8 were used to study the bioavailability and bioequivalence of these intranasal sprays. 20 In this study, different fentanyl nasal solutions with different pH values were formulated for the in vitro (pH 6, 8 and 10) and in vivo study (pH 6 and pH 8). At these pH values there are 2 species of fentanyl in the solution depending on the solution's pH value. The pK value for fentanyl is 8.437. Finally, the in vivo study involved the pharmacokinetics study (absorption, distribution and elimination profiles) of this fentanyl nasal solution by 25 comparing a nasal spray with an IV bolus injection in human volunteers. 2.1 Nasal Spray Applicator used in this disclosure: The nasal spray device used in this disclosure was a screw top nasal spray with a 4 minute lockout. This device is available in Western Australia 6 . The device is a hand-held nasal spray bottle with a timed lockout to administer drugs in a non-invasive manner. This 30 nasal device enables the administration of a small volume (0.18 mL per spray) of nasal solution, and it comprises the following parts: . A reservoir where the 1-way valve and the pump mechanism (or piston) are located. The actuator sits on top of this piston device, and when this actuator is activated, a dose of drug is released into the environment. The volume of this 35 reservoir determines the volume of the spray. To enable this unit to function properly the reservoir need to be filled completely with nasal solution before the actuator is activated otherwise the piston system could not function properly and the spray volume and the droplet size would be incorrect (example when activated too early, before the lockout period). This reservoir component of the nasal device 40 sits on top of the glass bottle and is sealed airtight with a plastic seal or flange by a screwing motion or crimping. The seal is made from Ethyl Vinyl Acetate plastic WO 2010/005400 PCT/SG2009/000248 -15 (soft malleable plastic) and usually comes into contact with the nasal solution if the spray bottle is accidentally tip over its side. If the spray device is stored upright, there is no contact of the nasal solution with this seal. The reservoir component is made from Acetal or Polyformaldehyde plastic (a hard rigid opaque type plastic). 5 * A dip-tube (or flow control plastic tube) connects the reservoir to the glass bottle. The length of the dip-tube and the diameter of its bore determined the refilling time of the reservoir. The viscosity of the drug solution also dictates the refilling time of reservoir. This refilling time equals the lockout period for the nasal device. The dip tube is usually made from Polyvinyl Chloride (PVC) of length 42 mm with a bore 10 diameter of 0.067 mm gave a filling time of 3 minutes for most non-viscous aqueous fluid. * A pharmaceutical grade amber coloured glass bottle where the drug solution is stored. The glass material is usually made of silica borate material. * An external nasal piece where the core extension or the stem is located. This stem 15 sits directly above the actuator, which activates the pump device (piston) in the reservoir and the solution is released to the environment via this stem. The stem's opening can be horizontal or vertical position hence directing the spray horizontally or vertically. This external nasal piece is covered by a plastic cap to prevent dust accumulation during storage. 20 Since this nasal device is made from different components with different plastic materials there is a potential for the drug to be sorbed and lost from the nasal solution into the plastic. This in vitro study involved investigating the stability and shelf life of the fentanyl nasal solution in these plastic components of this nasal device. Only the plastic 25 components of the nasal device that are in contact with the nasal solution were studied. Hence the components of the nasal device studied were the reservoir/pump/piston component, which are made of acetal plastic; the seal or flange (made from ethyl vinyl acetate); the glass bottle (made from Silicaborate) and the dip-tube (made from Polyvinyl Chloride). The stability studies involved immersing and storing these plastic components 30 in the fentanyl nasal solution at three different pH values (6, 8 and 10) and kept at different temperatures as well as in different ionic strengths. 2.1.1 Formulation of Fentanyl Nasal Solution for the in vitro study: The formulae for the fentanyl nasal solution (pH values of 6, 8 and 10) for the in vitro study were: 35 2.1.1.0 Isotonic fentanyl nasal solution at pH 6 (ionic strength, p = 0.18): Fentanyl Citrate powder 786 pg (equivalent to 500 pg fentanyl base) Isotonic phosphate buffer at pH 6 sufficient to produce 100 mL 2.1.1.1 Fentanyl nasal solution at pH 6 with ionic strength, p = 0.5: 40 Fentanyl Citrate powder 786 pg (equivalent to 500 pg fentanyl base) Sodium Chloride 0.91g WO 2010/005400 PCT/SG2009/000248 -16 Isotonic phosphate buffer at pH 6 sufficient to produce 100 mL 2.1.1.2 Isotonic fentanyl nasal solution at pH 8 (ionic strength, p = 0.3): Fentanyl Citrate powder 786 pg (equivalent to 500 pg fentanyl base) 5 Isotonic phosphate buffer at pH 8 sufficient to produce 100 mL 2.1.1.3 Fentanyl nasal solution at pH 8 with ionic strength, P = 0.5: Fentanyl Citrate powder 786 pg (equivalent to 500 pg fentanyl base) Sodium Chloride 0.672 g 10 Isotonic phosphate buffer at pH 8 sufficient to produce 100 mL 2.1.1.4 Fentanyl nasal solution at pH 10: Four different concentrations of the fentanyl nasal solution with a pH value of 10 were 15 prepared for a study on the effect of the initial concentration of fentanyl on the stability of fentanyl nasal solution in the three different plastic components. The formulae for preparing these nasal solutions are shown below: * Initial Concentration 20pg/mL fentanyl at pH 10 20 Fentanyl Citrate powder 31440 pg (equivalent to 20000 pg fentanyl base) Isotonic phosphate buffer at pH 10 sufficient to produce 1000 mL * Initial Concentration 40pg/mL fentanyl at pH 10 Fentanyl Citrate powder 62880 pg(equivalent to 40000 pg fentanyl base) Isotonic phosphate buffer at pH 10 sufficient to produce 1000 mL 25 Initial Concentration 80pg/mL fentanyl at pH 10 Fentanyl Citrate powder 125760 pg(equivalent to 80000 pg fentanyl base) Isotonic phosphate buffer at pH 10 sufficient to produce 1000 mL * Initial Concentration 113pg/mL fentanyl at pH 10 Fentanyl Citrate powder 177636 pg(equivalent to 113000 pg fentanyl 30 base) Isotonic phosphate buffer at pH 10 sufficient to produce 1000 mL WO 2010/005400 PCT/SG2009/000248 -17 Embodiments of formulations of the invention may also comprise the following: Example 1: In Example 1, a fentanyl nasal formulation was prepared having a concentration of 0.05 mg/ml. The formulation is as follows: Ingredient Percent [Fentanyl Citrate] % to make 0.05 mg/ml DI Water % (v) QS Example 2 In Example 2, a fentanyl nasal formulation was prepared having a concentration of 0.5 mg/ml. The formulation is listed below: 10 Ingredient Percent [Fentanyl Citrate] % to make 0.05 mg/ml Ethanol % (v) 20 Propylene glycol % (v) 5 DI Water % (v) QS *Contains fentanyl citrate equivalent to 0.05 mg/ml of fentanyl base Example 3 A fentanyl sublingual formulation was prepared having a concentration of 0.5 mg/ml. The formulation is listed below: 15 WO 2010/005400 PCT/SG2009/000248 - 18 Ingredient Percent [Fentanyl Citrate] % to make 0.5 mg/ml Ethanol % (v) 20 Propylene glycol % (v) 5 DI Water % (v) QS Mannitol % (wt) 0.3 Tween 80 % (wt) 0.2 2.1.2 Method to prepare 100mL of isotonic aqueous phosphate buffer solution at pH 6 To prepare 100 mL of isotonic phosphate buffer, mix the following ingredients: 5 0 75 mL of 0.067 M sodium acid phosphate (1.04% NaH 2

PO

4 .2H 2 0 in aqueous solution). * 15 mL of 0.067 M disodium hydrogen phosphate (2.39% t Na 2

HPO

4 .12H 2 0 in aqueous solution). 0 Add 0.5 g sodium chloride to make solution isotonic. 10 0 Adjust to pH 6.0 with 4 M sodium hydroxide and add sufficient water to produce 100 mL. 2.1.3 Method to prepare 100 mL of isotonic aqueous phosphate buffer solution at pH 8: 15 To prepare 100 mL of the isotonic phosphate buffer, mix the following ingredients: 0 5 mL of 0.067 M sodium acid phosphate (1.04% NaH 2

PO

4 .2H 2 0 in aqueous solution). * 90 mL of 0.067 M disodium hydrogen phosphate (2.39% Na 2

HPO

4 .12H 2 0 in aqueous solution). 20 - Add 0.42 g sodium chloride to make solution isotonic. * Adjust the pH of the solution to 8.0 with hydrochloride acid and add sufficient water to produce 100 mL. 2.1.4 Method to prepare 100 mL of fentanyl nasal solution for the in vitro study: 25 To a 100 mL measuring flask add: WO 2010/005400 PCT/SG2009/000248 -19 * 786 mg fentanyl citrate powder (equivalent to 500 mg of fentanyl base) * 0.91 g of sodium chloride (for nasal solution with pH 6 and p = 0.5) or 0.672 g of sodium chloride (for nasal solution with pH 8 and p = 0.5). . Add 90 mL of isotonic phosphate buffer solution (pH 6 or pH 8) 5 0 If necessary, the final pH is adjusted to pH 6 (or pH 8) using 4 M hydrochloric acid (or 4 M sodium hydroxide) as required and add sufficient water to produce 100 mL. * The potency of this nasal solution was confirmed with HPLC analysis. 10 2.1.5 Method to determine the maximum solubility of fentanyl citrate: The method used to determine the solubility (saturated aqueous solution) of fentanyl citrate in aqueous solution was as follow: 1. Accurately weight 100 mg of fentanyl citrate powder using a weighing glass boat. 2. Pour into a 10 mL glass tube with screw cap. 15 3. Add 5 mL of milliQ water into this 1OmL glass tube containing the fentanyl citrate. 4. The Microfilm plastic wrapper was used to seal the mouth of the bottle and the cap was then screwed tight to prevent loss of moisture by evaporation. 5. Make 8 samples by repeating step 1 to 4. Mount these 8 glass tubes with fentanyl citrate mixture onto a shaker (circular shaker) and adjust speed to 60 rpm 20 (revolutions per minute). 6. Shake these solutions by rotary motion for 4 days to ensure maximum solubilisation of the fentanyl citrate into water. 7. Filter the solution to remove the undissolved solute using a paper filter (Whatman paper filter, size 0.2 pm). 25 8. Then 2 mL of the resultant filtrate solution (8 samples) was diluted with the HPLC mobile phase and injected into the HPLC to determine the fentanyl concentration. 9. The pH values of these saturated filtrate solutions at room temperature were also measured using the Hanna Instrument pH meter. 30 2.1.6 Method to measure the tonicity/osmolality of the saturated aqueous solution of fentanyl citrate: The method used for testing the osmolality of the saturated fentanyl citrate solution, using the Osmometer was as followed: 1. The Osmometer (3MOplus) was first calibrated using a 50 mOsm standard 35 solution. 2. A calibration curve was then constructed using different concentration of sodium chloride aqueous solutions. The resultant curve was then use to determine the osmolality of isotonic sodium chloride as well as the osmolality of the saturated fentanyl citrate. 40 2.1.7 Method to measure the droplet size produced by the nasal spray device: 1. Prepare a 50 mL of isotonic fentanyl nasal solution containing 600 pg of fentanyl base per mL using the method as described above.

WO 2010/005400 PCT/SG2009/000248 - 20 2. Fill 5 mL (each) of this nasal solution into the nasal device and prime the device. 3. After priming the nasal device and store the nasal device up right. 4. Repeat step 2 to 3 for 5 more lots of fentanyl nasal spray. 5. Calibrate the spirometer to deliver 60 L/min of air (to simulate the respiratory rate 5 of human subject). 6. Assemble the Multistage Liquid Impinger (MSLI) and inject 10 mL of pure milliQ water into each of the 3 Chambers of the MSLI using a 10 mL plastic syringe. 7. Fix the nose piece onto the Multistage Liquid Impinger. 8. Connect the calibrated spirometer into the outlet (bottom of the MSLI). 10 9. Spray the fentanyl nasal solution into the nose piece 6 times using the 6 different fentanyl nasal spray bottle prepared as above. 10. Remove the spirometer by disconnecting the tubing. 11. Remove the nose piece. Wash the nose piece with milliQ water into a 20 mL volumetric flask. Wash another 2 times to make sure that all fentanyl nasal spray 15 droplets deposited are washed into the volumetric flask. Add milliQ water to produce 20 mL. 12. Suck the solution out of the 3 rd chambers using a plastic pipette and place into a 20 mL volumetric flask. Wash this Chamber twice with milliQ water. Add milliQ water to 20 mL. 20 13. Repeat step 12 for the 2 nd and 1 st chamber. 14. Remove the paper filter using a metal pincer. Wash the paper filter with milliQ water into a 20 mL volumetric flask (using a glass funnel). Wash another 2 times to make sure that all fentanyl nasal spray droplets deposited here are washed into the volumetric flask. Add milliQ water to produce 20 mL. 25 15. Repeat steps 6 to14, five more times to obtain 6 sets of reading. 16. Inject the diluted solution into the HPLC to determine the concentration of the fentanyl solution. The Liquid Impinger (Copley, UK) has three chambers, a nose piece and a 1 pm 30 Polycarbonate Membrane Filter at the bottom end (outlet). The nose piece is connected to the nasal spray and the bottom end has an outlet where a plastic tube is connected to a Compact Spirometer. The spirometer provides one-way airflow by sucking air out of the chambers thus providing a negative pressure so that air will flow from the nose piece to the outlet at the bottom of the chambers. The spirometer is calibrated to deliver airflow of 35 60 L/min (approximately equal to the volume of air exchange from the nostril to the lung). The first Chamber (the upper chamber) retains aerosol particles with size more than 13 pm and allows smaller particles with size less than 13 pm to pass through. The second Chamber (the middle chamber) retains aerosol particles with size from 7 to 13 pm and allows smaller particles with size less than 7 pm to pass through. The third Chamber (the 40 lowest chamber) retains aerosol particles with size 3 to 7 pm and allows smaller particles with size less than 3 pm to pass through. Aerosol particles with size less than 3 pm will flow out of the last Chamber and are caught in the membrane filter (1 pm). Any particles size measuring less than 1 pm will flow out through the outlet port connected to the spirometer.

WO 2010/005400 PCT/SG2009/000248 -21 2.1.8 HPLC method to analyse the concentration of fentanyl in the aqueous solution in the in-vitro study: The HPLC method of analysis was employed to determine the concentration of fentanyl in aqueous solution for all in vitro studies 8 . 5 2.1.8.1 Validation of the HPLC method of analysis: Freshly prepared nasal solution of fentanyl citrate 0.5 and 4 ptg per mL were submitted to HPLC analysis. The within-day coefficient of variation (cv) at 0.5 and 4 tg per mL was determined using eight replicate determinations. 2.1.8.2 Stability indicating HPLC methods: 10 Fentanyl solutions were subjected to heat, acid and alkali degradation to determine whether the degraded products of fentanyl interfered with the peaks of the HPLC of fentanyl. For the heat degradation experiment, 10 mL of the fentanyl solution was placed in the water bath (900C) for 6 hours and left to cool and stored for 24 hours before analysis. For the acid hydrolysis experiment, 5mL of 50 pg/mL fentanyl aqueous solution 15 was added with 5 mL of 4 M hydrochloric acid mixed thoroughly and stored for 24 hours in the water bath (90*C) before analysis. For the alkali hydrolysis experiment, another 5mL of 50 pg/mL fentanyl aqueous solutions was added with 5 mL of 4 M sodium hydroxide solution, mixed thoroughly and stored for 24 hours in the water bath at 900C before analysis. All solutions were stored in glass tubing. 20 2.1.9 Method used to study the stability or disappearance kinetics of fentanyl in different plastic components of the nasal spray device (in vitro study). The fentanyl nasal solutions used in this study for pH, ionic and temperature effect on the sorption profile of fentanyl into various plastic components were: * A 50 pg per mL fentanyl nasal solution buffered with isotonic phosphate buffer at 25 pH 6.0, with ionic strength at p=O.18. * A 50 pg per mL fentanyl nasal solution buffered with phosphate buffer at pH 6.0 with added sodium chloride to made to ionic strength of p=0.5. * Isotonic 50 pg per mL fentanyl nasal solution buffered with isotonic phosphate buffer at pH 8.0 with ionic strength at pt=0.3. 30 0 A 50 pg per mL fentanyl nasal solution buffered with isotonic phosphate buffer at pH 8.0 with added sodium chloride to ionic strength of p=0.5. The four fentanyl nasal solutions used in the study for the initial concentration effect on the sorption profile of fentanyl into various plastic components were: * 20 pg per mL fentanyl nasal solution buffered with isotonic phosphate buffer at pH 35 10. 0 40 pg per mL fentanyl nasal solution buffered with isotonic phosphate buffer at pH 10. * 80 pg per mL fentanyl nasal solution buffered with isotonic phosphate buffer at pH 10.

WO 2010/005400 PCT/SG2009/000248 -22 * 113 pg per mL fentanyl nasal solution buffered with isotonic phosphate buffer at pH 10. The following method was then used for the disappearance kinetics study for fentanyl: * Five mL of the above solutions was measured and placed into a 10 mL glass tube. 5 9 Different components of the nasal spray device (Polyvinyl Chloride dip-tube, Ethyl Vinyl Acetate seal or Acetal reservoir) were weighed and dropped into and soaked in these 5 mL solutions. * These glass tubes were then sealed with Parafilm Film before applying the screw cap to prevent loss of fluid by evaporation during storage. 10 0 These glass tubes were than equilibrated in a water bath which was set at the predetermined temperature (5, 25 or 370C). * At time zero after allowing 5 minutes, previously determined, for temperature equilibration and also at the prescribed time intervals, 2 glass tubes (duplication) were taken out of the water bath. 15 0 2 mL of the fentanyl solutions (from these 2 samples) were withdrawn by pipette added to the 20 mL volumetric flasks and the mobile phase was added to produce 20 mL. * These diluted solutions (nasal solution with the mobile phase) were then injected into the HPLC to measure the concentration of fentanyl. 20 2.2 Storage temperatures used in the disappearance kinetics studies: The temperatures employed were 5, 25 and 37*C. 2.3 Data Analysis: The mechanism and the disappearance kinetics of a penetrator into a plastic material can 25 be described mathematically 9 . Fick's first and second Law of diffusion described the mass transport of molecules across the membrane as similar to the transfer of heat across medium. 2.3.1 Compartment Models: 30 For the compartmental model, a one-compartment or two-compartment model may be used" 2.3.2 Diffusion Model: According to the diffusion model, the rate and extent of sorption is determined by the 35 plastic-water partition coefficient (R) and its diffusion coefficient (D) in the plastic. The diffusion model hence involves a complex mathematical formula1. This model involves data collection at both early times of sorption (hours) and at very long times when uptake into the plastic is approaching equilibrium (may sometimes involve days/months). However, most sorption experiments only involve the initial part of the diffusion model (i.e. 40 a short time).

WO 2010/005400 PCT/SG2009/000248 - 23 In a diffusion model, the aqueous solution is assumed to be well stirred, with the rate of uptake being controlled only by the diffusion of the solute into the plastic material 13 . A typical Fickian absorption experiment involves exposing a polymer sample (by immersing) in a penetrant solution. In the current experiment of this embodiment, the plastic 5 material [cylindrical PVC dip-tube, plane sheet EVA seal and plane sheet Acetal or Polyformaldehyde reservoir] was immersed into 5 mL of the fentanyl aqueous nasal solution (finite solution or limited volume where the fentanyl concentration in the aqueous solution falls as the fentanyl enters the plastic) and stored in the water bath at the predetermined storage temperature of 37, 25 and 5'C. The loss of fentanyl is measured as 10 a function of time. 2.3.3 Langmuir's sorption data (isotherm): Solutes in solution may be adsorbed from the liquid phase onto a solid phase such as charcoal and plastic. The study of the adsorption of solute from liquid phase onto a solid phase is employed in the field of decolourisation of solution, adsorption chromatography, 15 detergent science and wetting agents 14 . Isotherm equations like the Langmuir equation can be used to study this adsorption process. 2.3.4 Sorption Number (Sn): The sorption number provided a method to predict the influence of time, solution volume, plastic surface area and the solution pH on the loss of solute into plastic material 15 . The 20 sorption number was also correlated to the solute's octanol-water partition coefficient. Hence the sorption parameter incorporates the surface area of the plastic, volume of solution, fraction of unionised solute, Rplastc-water and diffusion coefficient in the plastic. The octanol-water partition coefficient for most solutes was most commonly studied and hence could be easily available from the literature. 25 The sorption number method was used to calculate (extrapolate) the expiry time of fentanyl in the plastic components of the nasal device. 2.3.3: Standard Affinity (Ap) Standard Affinity (Ap) is an index to measure the sorptive potential of an agent in a plastic 16 . If the chemical potential of a compound in solution phase is greater than in the 30 solid phase (plastic), there will be a driving force moving the solute from the solution into the solid phase. At equilibrium, both chemical potentials will be equal and the driving force will be reduced to zero. 3.0 Results and Discussions: 3.1 Solubility of fentanyl citrate aqueous solution: 35 The solubility data obtained for the fentanyl citrate is shown in Table 3.1.0 below. The average concentration of fentanyl citrate in the aqueous solution was 10.64 mg/mL. Therefore the saturated solution of fentanyl citrate in water was approximately 10.64 ± 1.03 mg/mL. This solubility is within the range quoted in Martindale 17 which states that the solubility of fentanyl citrate is sparingly soluble to soluble in water (in the range of 10 to 25 WO 2010/005400 PCT/SG2009/000248 -24 mg/mL at the storage temperature of 15 to 250C). The pH value was at 3.64 (250C) of this saturated fentanyl citrate aqueous solution. Table 3.1.0: Mean concentration of fentanyl citrate (mg/mL) in saturated fentanyl citrate aqueous solutions at pH 3.64. Fentanyl Citrate Filtrate Fentanyl Citrate concentration mg/mL 1 9.65 2 9.51 3 12.52 4 10.89 5 11.03 6 9.30 7 10.83 8 11.40 Average 10.64 Std Deviation 1.03 5 3.2 Osmolality of saturated fentanyl citrate aqueous solution. The osmolality of a saturated fentanyl citrate aqueous solution was measured using the 3MOplus Osmometer. Different concentrations of sodium chloride solution were prepared 10 for calibration of the Osmometer. The result was shown below (Figure 3.2.0). From Figure 3.2.0, the osmolality of an isotonic solution of sodium chloride (0.9% or 0.9 g/100 mL or 0.3 M) was calculated to be 287.6 mOsm. The tonicity of the saturated aqueous solution of fentanyl citrate was then measured using the osmometer and the solubility of this saturated fentanyl citrate solution can then be calculated. 15 The osmolality of the saturated solution of fentanyl citrate was found to be 40 mOsm.

WO 2010/005400 PCT/SG2009/000248 -25 Since the osmolality of 0.9% NaCI aqueous solution was 287.6 mOsm hence 40 mOsm of the fentanyl citrate solution was then equivalent to 40/287.6 X 0.9% (0.125% NaCI). Therefore a saturated solution of fentanyl citrate solution is hypotonic. 5 The saturated fentanyl citrate aqueous solution with a measured osmolality of 40 mOsm is equivalent to 10.57 mg/mL of fentanyl citrate. Thus the saturated solution of fentanyl citrate (10.57 mg/mL), as measured by the osmometer, is comparable to that measured by the HPLC method of analysis (10.64 ± 1.03 mg/mL). 10 The tonicity of a nasal solution instilled into the nasal cavity may affect the rate of absorption of a drug. The nasal absorption of secretin was affected by the concentration of sodium chloride in the nasal formulation reaching a maximum when the concentration of sodium chloride was 0.462 M (hypertonic solution). Shrinkage of the nasal epithelia cells was noted as a result of the hypertonicity of this nasal solution. The shrinkage of the 15 nasal epithelia cells opened up the tight junctions between the nasal cells, thereby allowing more drugs to easily cross the cells during this hypertonic condition. Maintaining the constant volume of a cell is an important physiological function not only in defining its intracellular osmolality and its shape, but also in defining other cellular 20 function, such as trans-epithelial transport, cell migration, cell growth, cell death and the regulation of intracellular metabolism. Several transport systems and regulatory signals (proteins in the plasma membrane) have been identified which regulate this process (cell volume maintenance), examples are the swelling-sensitive chloride channels 19 , the Na*/H* exchangers or ion transporters and the C1/HCO 3 - ion exchangers 20 . These transport 25 systems are cations and anions sensitive for example H*, Na*, C, HC0 3 ~ are shown to interfere with these transport systems resulting in the expansion or the shrinkage of the cell volume leading to "holes" formation between the tight junction of the cells. Hence by changing the pH values (H* and HCO 3 ) and the Na* and C contents (tonicity) of the nasal solution, we can manipulate the cell volume in the nasal mucosa leading to an 30 increase or decrease in drug absorption. 3.3 Droplet size measurement of the nasal spray produced by the nasal spray device using the Multistage Liquid Impinger (MSLI): Five fentanyl nasal sprays with fentanyl concentration of 600 pg of fentanyl per mL were used for this study. This higher concentration of fentanyl nasal solution was chosen to 35 ensure that any nasal droplets that were caught in the last chamber (third chamber of the MSLI) or in the filter section (prior to the outlet tubing), which theoretically would be minimal, can be quantifiable. Five sprays (each from the 5 different nasal spray bottles) were sprayed into the nose piece of the MSLI. The actual initial fentanyl concentration of the nasal solution measured by HPLC was 631 pg per mL. This experiment was repeated 40 6 times. Since the spray volume for each spray was 0.18 mL, therefore 5 sprays will give a volume of 0.9 mL or 567.9 pg of fentanyl. The results obtained are tabulated in Table 3.3.0 below. Table 3.3.1 showed the loss of fentanyl (%) during this experiment. The loss of fentanyl may be due to the spraying process where some droplets may escape from the nose piece and loss to the atmosphere.

WO 2010/005400 PCT/SG2009/000248 - 26 The MSLI measures the particle size of the spray droplets or aerosol by dividing the particles into 4 size groups. Chamber 1 retains all particles with size greater than 13 pm and letting particles with size less than 13 pm to flow through. Chamber 2 retains particles from size 7 to 13 pm and Chamber 3 retains particles with size 3 to 7 pm. Particles with 5 size less than 3 pm will flow out of the chambers and into the filter. The filter catches these particles with sizes less than 3 pm. Approximately 3.6 ± 3.9% of the aerosol were lost or unaccounted for presumably escaping into the -environment during the spraying process 86.1% of the spray particles were deposited immediately in the nose piece suggesting deposition by inertia since the 10 nose piece has a 900 bend. Of the particles 10.3% were deposited in the 1 st chamber. There were no trace (no detectable) of fentanyl in the second, third chambers or the filter 100% of the measurable aerosol particles (or 96.4% when taking into account the loss of fentanyl into the atmosphere) produced by the nasal spray device. All droplet sizes were greater than 13 pm. This particle size range was found to be ideal for depositing the active 15 composition onto the nasal cavity. The product information for the nasal device states that the nasal device should delivered aerosolised particles with average droplets size of 80 pm. Table 3.3.0: Amount of fentanyl caught in the various compartments of the MSLI after five sprays of fentanyl nasal spray (=567.9 pg): 20 WO 2010/005400 PCT/SG2009/000248 - 27 Fentanyl % 3 rd Chamber 6 ND (13g) Filter I ND Nose piece 1 509.6 89.7 Filter 2 ND Nose piece 2 488.1 85.9 Filter 3 ND Nose piece 3 483.5 85.1 Filter 4 ND Nose piece 4 483 85 Filter 5 ND Nose piece 5 506 89.1 Filter 6 ND Nose piece 6 463.5 81.6 ND = Not detectable. 1st Chambers 1 47.6 8.4 1 st Chamber 2 80.7 14 1 st Chamber 3 25.8 4.6 1 st Chamber 4 69.6 12.3 1 "t Chamber 5 28.6 5 1 st Chamber 6 100.6 17.7 2 nd Chamber I ND 2 nd Chamber 2 ND 2 nd Chamber 3 ND 2 nd Chamber 4 ND 2 nd Chamber 5 ND 2 "d Chamber 6 ND 3 rd Chamber 1 ND 3 rd Chamber 2 ND 3 rd Chamber 3 ND 3rd Chamber 4 ND 3 rd Chamber 5 ND WO 2010/005400 PCT/SG2009/000248 - 28 Table 3.3.1: Percentage loss of fentanyl from the MSLI study: % recovery of fentanyl in the MSLI % loss of fentanyl. (Nose piece + 1 st chamber) 98.1 1.9 99.9 0.1 89.7 10.3 97.3 2.7 94.1 5.9 99.3 0.7 Average = 96.4 Average = 3.6 Std Deviation = 3.86 Standard Deviation = 3.86 The mucous flow rate (mucociliary clearance) is approximately 10 mm per min (ranging from 0.5 to 23.6 mm per min) 21 . The clearance rate is an important factor determining the 5 time a drug remains at its absorption site and hence the half-life of the drug in the nasal cavity. Particles of this size are breathed in and out of the respiratory tract. However, particles with size between 2 to 4 ptm may or may not be deposited in the nasal cavity 22 . Those not deposited in the nasal cavity are carried via the inspired air towards the lungs. 10 Particle size is important during the formulation of fentanyl nasal spray to ensure complete deposition of the fentanyl particles into the intended area of the nasal cavity. For fentanyl to be deposited in the nasal cavity, the spray particle size should be more than 5 pm. Particles less than 2 pm were undesirable because these particles could easily flow in and out of the lung with the inspired and expired air resulting in a loss in potency of the 15 fentanyl nasal spray. Other factors influencing the deposition of particles onto the nasal cavity are shape, density, and electrical charge of the particles, air velocity and resistance 22 There are advantages and disadvantages in using nasal spray and nasal drops. The fate of a particle deposited in the nasal passage is determined by its site of deposition.

WO 2010/005400 PCT/SG2009/000248 -29 Particles deposited anteriorly in the nasal cavity stayed longer by mucociliary movement (0.5 to 3 mm per hour movement), whereas particles distributed posteriorly (the last two thirds of the nose) were removed more rapidly (4 to 10 mm per min) 2 3

-

24 . Hence the distribution of a drug in different areas of the nasal cavity can determines the absorption 5 efficiency of the drug. The olfactory area of the nose covers an area of approximately 10 cm 2 . The olfactory nerves from the olfactory bulb are surrounded by a space containing cerebral spinal fluid (CSF), which is continuous with the cranial subarachnoid space. Hence dopamine and other medication could thus enter into the CSF through this olfactory nerve. 10 The amount of nasally administered drugs absorption is proportional to the retention time of the drug in the mucosal surface and also indirectly proportional to the clearance time of the mucous layer 24 . The nasal mucociliary function could also be affected by air pollutants, chemicals (example some preservatives added to nasal solution) and by certain pathological conditions such as nasal polyposis and also injury as a result of viral 15 infection 25 . Hence changing the mucociliary functions could also increase or decrease the absorption half-life of drug administered nasally. 3.4: Measurement of the volume per spray produced by the nasal device: Six nasal bottles were used in this experiment and were filled with MilliQ water. MilliQ water was used to determine the volume per spray produced by activating the nasal spray 20 device. The specific gravity of MilliQ water is 1 (25'C) which means that each mL of MilliQ water is equal to 1 g. The volume of the spray was obtained indirectly (by subtraction) by weighing the bottle after each activation of the spray bottle. The spray bottles were stored upright after each spray (to prevent malfunctioning in the filling process). The storage time interval after each spray was critical to ensure maximum filling of the MilliQ water into the 25 reservoir, otherwise a suboptimal volume of spray is produced. The pump was primed before the beginning of the experiment. The storage time per spray and the volume per spray data are summarized in Table 3.4.0.

WO 2010/005400 PCT/SG2009/000248 - 30 Table 3.4.0: Volume per spray at 8 to 10 minute intervals (using 5 mL milliQ water as the nasal solution). Spray Max. Min Average Std Max no. of sprays bottle volume/ volumes volume/s deviation achievable for 5mL no. spray pray (mL) pray (mL) (mL) (mL) 1 0.18 0.14 - - 8 2 0.19 0.08 0.18 0.01 26 3 0.18 0.07 0.17 0.01 26 4 0.19 0.13 0.18 0.01 24 5 0.18 0.07 0.18 0.01 26 6 0.18 0.07 0.18 0.01 26 The average volume per spray produced by activating the nasal device was 0.18 mL 2 e 5 The nasal device should be primed twice and allowed to stand upright for 5 minutes between each priming (for the nasal spray device with a 3 minute lockout) before using. It is also advisable to reject the third spray (volume from 0.13 to 0.16 mL). Spray number one was the third spray after two priming sprays. The third spray usually has an inaccurate volume of between 0.13 to 0.16 mL. 10 Spray bottle number one was faulty and could only manage to produce 8 proper sprays. This nasal device could not work after the 8 th spray and no aerosol droplets were produced subsequently. The malfunction could be the result of a "blockage" of the capillary lumen by air bubble in the dip-tube and hence the inability to fill up the reservoir 15 with MilliQ water or could be due to a faulty pump mechanism. The reason for the malfunction was not investigated further. Spray bottle number four (Table 3.4.0) delivered only 24 sprays and maybe due to the inaccuracy in filling the glass bottle with MilliQ water (less than 5 mL). The percentage of 20 faulty spray devices in each batch was not determined. In practise this testing procedure was not used to test for faulty devices because this procedure is a destructive test procedure. This test can only be used to produce a statistically appropriate sample to indicate the rate of faulty nasal devices and then extrapolate to other batches. However the nasal device once working properly can hopefully delivered an accurate and 25 consistent 0.18 mL per spray. The arnount of drug that can then be delivered and WO 2010/005400 PCT/SG2009/000248 -31 deposited in the nasal mucosa can thus be accurate and consistent from each spray applied to the nose using this nasal device. The volume per spray is determined by the size of the reservoir. The reservoir can be set 5 at different volume, in this case, 0.18 mL 2 6 . The pump is connected to the reservoir by a one way valve. On activating the pump (pressing down motion), the fentanyl nasal solution is atomised into aerosolised particles and released into the environment via the outlet. On the other end, the reservoir is connected to the nasal solution (stored in the glass bottle) via a dip-tube. There is a one way valve in the reservoir allowing only the 10 fentanyl nasal solution to flow one way into the reservoir. The diameter of the lumen (hole) in the dip-tube determines the filling time of the reservoir and hence the lockout period and in this case 3 minutes of filling time 2 6 . The lumen hence limits the flow rate of the nasal solution into the reservoir to the desired refilling time. The length of the dip-tube is usually from 30 to 60 mm and the lumen diameter in the range of 0.025 to 0.075 mm. For 15 the filling time of 3 minutes the lumen diameter is 0.025 mm and the length is 10.5 mm. 3.5: Results of the HPLC analytical methods Fentanyl was analysed by the HPLC method of analysis. This method of analysis was chosen because it is a simple, rapid and accurate analytical method 8 . 3.5.0: HPLC chromatogram for fentanyl: 20 The wavelength of detection (210 nm) was used for the detection of fentanyl. A representative chromatograph of fentanyl is shown in Figure 3.5.0.0. The fentanyl eluted at 4.87 minutes. 3.5.1: Standard curve for fentanyl citrate using the HPLC method of analysis: The assay standard curve was found to give a linear relationship between the peak area 25 (AUC) and the concentration over the range of 0.5 to 4 pg/mL of fentanyl citrate with a coefficient of determination of r 2 = 0.9999 (n=8). 3.5.2: Stability indicating HPLC method for the assay of fentanyl Samples of the 50 pg/mL of fentanyl nasal solution (2 mL) were subjected to acid, alkali and heat degradation. 30 3.5.2.0 Acid hydrolysis of fentanyl: A 2 mL dose of 50 ig/mL fentanyl solution was added to 48 mL of 4 M-hydrochloric acid solution (total volume 50 mL) in a conical flask and the solution was left to stand for 24 hours in a 500C water bath after thorough mixing. The pH of the solution was adjusted to 1 with hydrochloric acid. This solution was then injected into the HPLC and the resultant 35 chromatogram is shown in Figure 3.5.2.0.0. There was an unknown breakdown product of fentanyl shown by an extra peak at 2.9 minutes and this peak did not interfere with the peak of the fentanyl. The percent of fentanyl remaining was 82.1%.

WO 2010/005400 PCT/SG2009/000248 - 32 3.5.2.1 Alkali hydrolysis of fentanyl: A 2 mL dose of a 50 1 g/mL fentanyl solution was added to a 48 mL of 4 M sodium hydroxide solution in a 50 mL conical flask, mix thoroughly and allow to stand for 24 hours 5 in a water bath at 50*C. The pH of the solution was adjusted to 12 with 4 M sodium hydroxide. This solution was then injected into the HPLC and the resultant chromatogram is shown in Figure 3.5.2.1.0. The two peaks in the chromatogram have a retention time of 2.7 and 3.6 minutes. These were unknown breakdown products of fentanyl. The percent fentanyl remaining was 77.2% 10 3.5.2.2: Heat degradation of fentanyl: A 2 mL dose of fentanyl 50 tg/mL solution was diluted with milliQ water to 50 mL and immersed in a water bath calibrated at 900C for 4 days. This was then removed and cooled to room temperature. This solution was then injected into the HPLC and the resultant chromatogram is shown in Figure 3.5.2.2.0. There were two extra peaks in the 15 chromatograph at 2.9 and 3.5 minutes. These were unknown breakdown products of fentanyl. The percent fentanyl remaining was 24.3% The stability-indicating property of this USP method is sub-optimal. N-phenyl-N-(4 piperidinyl)propionamide (PPA), which is a known degradation product of acid, hydrogen peroxide, heat and light degradation and also a manufacturing process impurity of 20 fentanyl, was shown to interfere with the active fentanyl peak of the chromatogram and was not adequately resolved 27 . Hence the USP method did not meet the requirements for a stability-indicating assay. The wavelength used (230 nm) for detection was also inadequate for the detection of related products of fentanyl. For a more suitable stability indicating HPLC method of analysis, the wavelength for detection of fentanyl should be 25 change to 206 nm and the mobile phase changed to a perchloric acid/acetonitrile aqueous system 27 . The degradation products identified were phenacyl derivative, acetyl analog, PPA, 4-anilino-1-benzyl-piperidine, benzyl derivative, 4-anilino piperidine and N-phenyl-N 4-piperidinyl propionamide. 3.6 Results of the sorption isotherm and standard affinity, -Ap studies: 30 PVC dip-tube, EVA seal and Acetal reservoir was immersed (and soaked) into four different initial concentrations of fentanyl nasal solutions (20.9, 39.5, 87.1 and 113 pg/mL). These aqueous nasal solutions were buffered in phosphate buffer to kdep the pH value of the nasal solutions constant, at 10. The storage temperature was also kept constant at 370C in a water bath. These experiments were set up to construct the sorption isotherm 35 (Langmuir sorption isotherm) for fentanyl in the 3 plastic components of the nasal device. The percentage of fentanyl loss from the nasal solution during the storage in the plastic components of the nasal device (PVC dip-tube and the EVA seal soaked in the aqueous fentanyl nasal solution of concentration 113 pg/mL buffered to pH 10 in phosphate buffer and stored at 370C) was plotted graphically as shown in the Figure 3.6.0. Equilibrium WO 2010/005400 PCT/SG2009/000248 - 33 conditions were reached within 5 hours as shown in the plateau region of the Figure 3.6.0. When the acetal reservoir was soaked and stored at 37'C in the fentanyl nasal solution (113 pg/mL) buffered with phosphate buffer at pH 10, the equilibrium condition was reached in 10 days as shown in the plateau region of Figure 3.6.1. Sorption isotherms 5 were constructed by plotting the concentration of fentanyl in the nasal solution against the amount of fentanyl loss into the plastic components of the nasal device (see Figure 3.6.2 below). The experimental results have shown that the equilibrium condition was reached within 5 hours of storage for the PVC dip-tube and the EVA seal and 10 days for the acetal 10 reservoir in the fentanyl nasal solution. The concentration of fentanyl remaining in the aqueous nasal solution at this equilibrium condition and the corresponding fentanyl concentration sorbed into the plastics were plotted graphically to obtain the Langmuir sorption isotherms. The Langmuir-type relationship (1/q = 1/kSC] + 1/S, see Equation 2.5.2.0) was constructed for the 3 plastic components (equilibrium condition) as shown in 15 the Figure 3.6.3 below and showed a linear relationship. The intercept at the Y-axis equals to 1/S. The solubility, S (S equals to Langmuir constant for the amount adsorbed at saturation) of fentanyl in the 3 plastic components could then be calculated (see Figure 3.6.0 below) 28 . Hence the solubility, S (or the amount adsorbed at saturation) of fentanyl in the PVC plastic dip-tube, EVA plastic seal and acetal plastic reservoir was calculated to 20 be 4.9, 1.6 and 0.5 mg per gram of plastic respectively. The solubility (S), or Langmuir constant, of fentanyl at 370C in the 3 plastic components of the nasal device was in the following descending order, PVC>EVA >Acetal (pH value of the nasal solution was 10). The "solubilisation" of fentanyl into the acetal or polyformaldehyde plastic to maximum solubility was a slow process taking 10 days to 25 reach equilibrium whereas the "solubilisation" of fentanyl into PVC and EVA plastic to maximum solubility only took up to 5 hours at 370C. The solubility, S, however, is only a theoretical saturation value of fentanyl in the plastic material via adsorption. The Langmuir equilibrium constant, K, is a good measure of the adsorption capacity or potential of the plastic material 2 8 . K for the 3 plastic components is in the following rank 30 order EVA>PVC>Acetal. The adsorption capacity of EVA plastic for fentanyl was thus twice the adsorption capacity of PVC plastic (989 pg compared to 460 pg fentanyl per gram plastic respectively). This however may be due to experimental error or defect. A small difference in q could magnify C value when their reciprocal values were taken. Also, a slight alteration of the plastic structure during the experimental procedure, for example a 35 slight temperature or solvent variation, would lead to more or less available binding sites which in turn could affect the isotherms and the resultant S and K values 28 . Using these two equations the free energy change, AF and the enthalpy change, AH can be calculated for fentanyl adsorption into the 3 plastic components. These results are tabulated in Table 3.6.1 and compared to the results obtained for chlorhexidine on to carbon black.

WO 2010/005400 PCT/SG2009/000248 -34 Table 3.6.1: Thermodynamic functions of adsorption of PVC dip-tube, EVA seal, Acetal reservoir from aqueous nasal solution containing fentanyl (or carbon black from aqueous solution containing chlorhexidine 0 ). Storage temperature = 37 0 C. Materials AF (free energy change) AH (enthalpy change) PVC dip-tube 4.07 kcal/kg -0.17 kcal/kg EVA seal 3.60 kcal/kg -0.15 kcal/kg Acetal reservoir 7.24 kcal/kg -0.31 kcallkg Carbon black 30 -8.46 kcallmol 0.0295 kcallmol 5 The standard enthalpy, AH associated with the sorption process of chlorhexidine digluconate from aqueous solution by poly(2-hydroxyethyl methacrylate) powder was zero between the temperature range of 20 to 500C29. The sorption process was thus associated with ion exchange interaction whose standard enthalpies are general zero or very small. Since the standard enthalpy of the fentanyl sorption into the 3 plastic components falls 10 outside this range (see Table 3.6.1 above), therefore the sorption process was not an ion ion interaction unlike the sorption of chlorhexidine into carbon black which was an ion-ion 30 interaction Standard affinity -Ap, is the difference in the standard chemical potential between the solute (fentanyl) in the liquid phase (aqueous) and the solute in the solid phase (plastic 15 matrix). The standard affinity is thus a measure (an index) of the sorptive potential of fentanyl to the plastic material 28 . The Figure 3.6.2 (isotherm plot) is a plot of Cp (concentration of fentanyl in the plastic) versus C, (concentration of fentanyl in the aqueous nasal solution). The standard affinity, -Ap, for fentanyl in the 3 plastic components can then be calculated from the slope of the data, since -Ap = RT InCp/Cs 20 (Equation 2.5.2.0)28. The results obtained are tabulated in the Table 3.6.2 below. The standard affinity for fentanyl in the 3 plastic components was 2.37, 1.45 and 1.06 kcal per mole for PVC dip-tube, EVA seal and Acetal reservoir respectively. Hence fentanyl has a higher affinity for PVC plastic followed by EVA and acetal plastic.

WO 2010/005400 PCT/SG2009/000248 - 35 Table 3.6.2: Standard affinities (-Ap) at 37*C (pH value of 10) for fentanyl in PVC dip tube, EVA seal and Acetal reservoir. (R = gas constant = 8.3143 JK- 1 mole-1 and 0"C = 273.16 K) Plastic Slope of graph -Ap (Jmole~1) -Ap (kcal mole-) PVC dip-tube 46.43 9897.2 2.37 EVA seal 10.56 6078.4 1.45 Acetal reservoir 5.60 4444.1 1.06 5 3.7 Results of the Disappearance Kinetics of fentanyl into PVC, EVA and Acetal plastics at the pH value of 10 and the storage temperature of 37*C This experiment was conducted to determine whether the four different initial concentrations of fentanyl nasal solution could influence the diffusion coefficient (D) and 10 the disappearance rate constant (k) of fentanyl into the three plastic components (PVC dip-tube, EVA seal and Acetal reservoir) of the nasal device. The PVC dip-tube, EVA seal and Acetal reservoir was immersed (and soaked) in the four different initial concentrations of fentanyl of 20.9, 39.5, 87.1 and 113 pg/mL. These aqueous nasal solutions were buffered in phosphate buffer at pH 10.0. The pH value of 10.0 was chosen for this 15 experiment because at this pH, 97.4% of the fentanyl molecules were calculated to exist as the non ionic species. Non-ionic fentanyl is very lipophilic at the storage temperature of 37C, fentanyl would be expected to sorb rapidly and quickly into the plastic material, hence degradation kinetics data could be obtained within hours. This experiment was performed over 5 hours after a preliminary experimental result found that equilibrium 20 conditions were reached in 5 hours. The plastic components and their specification of the nasal device used in this experiment are listed in Table 3.7.0. Table 3.7.0: Specifications of the three plastic components of the nasal device: *Plastic Thickness (1) Weight (Wp) Specific gravity Volume of components cm g (SG) plastic (Vp) mL PVC dip-tube 0.107 0.2 1.15 0.18 EVA seal 0.1 0.5 1.2 0.42 Acetal 0.058 1.13 0.85 1.33 reservoir These plastic components were in contact with the fentanyl nasal solution when the nasal 25 solution was stored in this nasal device.

WO 2010/005400 PCT/SG2009/000248 -36 None of the fentanyl nasal solutions stored in the amber glass container/reservoir of the nasal device at 370C showed any loss of or change in fentanyl concentration over a period of 12 months (Figure 3.7.0). Fentanyl was found to be stable in the amber silica-borate glass bottle. There was no adsorption of fentanyl onto the glass bottle or any measurable 5 loss of fentanyl by degradation. The loss of fentanyl when the PVC dip-tube, EVA seal and Acetal reservoir were immersed and stored in this fentanyl nasal solution was measured by HPLC analysis. The fraction of fentanyl remaining in solution or the amount of fentanyl in the plastic matrix was plotted against the storage time and the data analysed using the Diffusion Model as well 10 as the Compartment Model. In order to study this as a diffusion process of a drug into the plastic matrix, it is necessary to use a mathematical model that is able to describe or follow the time course (or profile) of the drug loss into the plastic matrix. Two models, the Compartment and the Diffusion Model were use to follow the loss (time course profile) of fentanyl each plastic components of the nasal device. 15 3.7.0: Diffusion Model: The amounts of fentanyl absorbed into the plastic components of the spray applicator device for the various initial fentanyl nasal solution concentrations (at a constant pH 10.0, stored at 370C) were shown in the graphs below (Figure 3.7.0.0, 3.7.0.1, 3.7.0.2). The reduced sorption curves were also plotted for the three plastic components of the 20 nasal device (PVC dip-tube, EVA seal and Acetal reservoir) which was stored in the 4 different initial fentanyl nasal solution concentrations as shown in Figures 3.7.0.3, 3.7.0.4 and 3.7.0.5. The residual equation of best fit for the graphs, the slope of the graph and the calculated diffusion coefficient (D) are tabulated in Table 3.7.0.0.

WO 2010/005400 PCT/SG2009/000248 - 37 Table 3.7.0.0: Diffusion coefficient (D) values for fentanyl when Ethyl vinyl acetate (EVA) seal, Polyvinyl Chloride (PVC) dip-tube and Acetal reservoir was immersed (stored) in various fentanyl nasal solutions buffered at pH of 10.0 (at a constant storage temperature of 37"C). Plastic Materials D (cm 2 /s) EVA seal Mean = 1.15 x 10-7 EVA Seal Std Deviation = 0.06 x 10~7 PVC dip-tube Mean = 1.15 x 10-7 PVC dip-tube Std Deviation = 0.06 x 10 Acetal reservoir Mean = 8.2 x 10-0 Acetal reservoir Std Deviation = 0.33 x 1010 5 3.7.0.0: Diffusion Coefficient (D) derived from the Diffusion Model. Table 3.7.0.3 is a summary of the three diffusion coefficients of fentanyl for the 3 plastic components of the device. The diffusion coefficient, D for fentanyl in the PVC dip-tube was similar to the EVA seal (1.15 ± 0.06 x 10~7 cm 2 s- 1 ). However, the diffusion coefficient 10 of fentanyl in the Acetal reservoir was lower (8.2 ± 0.33 x 10 10 cm 2 s 1 ). The diffusion of fentanyl into the plastic material is a transport process and obeys Fickian Law. Fentanyl is transported from the surface into the inner body of the plastic as a result of the random fentanyl molecular motion due to its thermal energy. This transportation process is a spontaneous process and involves a reduction of the free energy of the system. This 15 experiment also showed that the diffusion of fentanyl (1.15 ± 0.06 x 10- cm 2 S- for PVC and EVA and 8.2 ± 0.33 x 10 1 cm 2 s 1 for Acetal) into the plastic components of the nasal device was between those of the liquid and solid system. The diffusion coefficient (D) for fentanyl in EVA seal, PVC dip-tube and Acetal reservoir (1.15 ± 0.06 x 107, 1.15 ± 0.06 x 10-7 and 8.20 ± 0.33 x 1010 cm 2 /s respectively) were similar for all the 4 initial fentanyl 20 concentrations and hence were independent of the initial fentanyl concentration (Table 3.7.0.3) thus obeying Fick's law. The diffusion coefficient D for fentanyl obtained by using the Diffusion Model may be unreliable if several variables specified in the Diffusion Model Equation are not fulfilled. The assumption that the concentration of fentanyl in aqueous solution at any time is WO 2010/005400 PCT/SG2009/000248 38 independent of its position in the nasal solution from the plastic surface may not be necessarily true. The uptake of non-ionized fentanyl by the plastic components is both rapid and extensive (especially for PVC and EVA plastic). Therefore the concentration of fentanyl in the solution adjacent to the plastic may be significantly lower than elsewhere. 5 In this experiment, the plastic component was soaked or dipped- completely into the 5mL of fentanyl nasal solution at pH value of 10.0. The nasal solution was hence static and non-stirred. Therefore the Diffusion Model Equation may not "hold true" in this situation. Fentanyl is a small molecule and should therefore obey the Fickian characteristics of absorption. For a given polymer, the value of D is determined primarily by the molecular 10 volume of the solute or, for larger drugs, its molecular weight. 3.7.0.1: Distribution or Partition Coefficient/Constant (R) derived from the Diffusion Model. The distribution coefficient or distribution constant, R for fentanyl in PVC-aqueous, EVA 15 aqueous and Acetal-aqueous partition can be calculated 34 where R = (1 - F.)pV/[F.Wp]. The calculated R values are in Table 3.7.0.1.0. The R values were independent of the initial concentration of fentanyl since the amount of fentanyl loss by sorption into the plastic matrix or remaining in solution was independent of the initial concentration. R is dependent only on the amount of fentanyl sorbed into the plastic and the amount of 20 fentanyl remaining in solution at equilibrium since the density of the aqueous nasal solution p, the volume of the nasal solution V and the weight of the plastic component Wp were constant throughout the experiment. The partition or distribution coefficient, R can also be calculated during the equilibrium stage of the experiment. Figures 3.7.0.1.0, 3.7.0.1.1 and 3.7.0.1.2 showed that after 5 hours of storage the PVC dip-tube-aqueous 25 nasal fentanyl solution system and the EVA seal-aqueous nasal fentanyl solution system reached equilibrium conditions (10 days for the Acetal reservoir-aqueous nasal fentanyl solution system). The amount of fentanyl absorbed into the plastic components (pg/mL) divided by the concentration of fentanyl (pg/mL) remaining in solution during equilibrium condition is R. Table 3.7.0.1.1 below tabulates the partition coefficient values obtained by 30 this method. The R values obtained using this method should also be independent of the initial concentration of fentanyl since the amount sorbed into the plastic component and the amount remaining was independent of the initial concentration. The partition coefficient values, R listed in Tables 3.7.0.1.0 and 3.7.0.1.1 were similar even though they were calculated by 2 different methods. The range of R values for the PVC dip-tube .35 aqueous, EVA seal-aqueous and Acetal reservoir-aqueous partition were 16.7 to 28.2, 6.4 to 13.3 and 4.1 to 5.6 respectively. The calculated R values during the equilibrium conditions (Table 3.7.0.1.1) for the PVC dip-tube-aqueous, the EVA seal-aqueous and the Acetal reservoir-aqueous partition were 19.3 to 32.3, 7.6 to 15.6 and 3.2 to 4.8 respectively. These R values were hence similar (similar range of values). The partition 40 coefficients suggest that the order of affinity of fentanyl for the plastic components were PVC dip-tube > EVA seal > Acetal reservoir.

WO 2010/005400 PCT/SG2009/000248 - 39 Table 3.7.0.1.0: Constants for the sorption of fentanyl by PVC dip-tube, EVA seal and Acetal reservoir (pH 10.0, storage temperature 37"C) using Equation 2.5.1.1. Parameter PVC dip-tube EVA seal Acetal reservoir S 20. 39. 87. 113 20. 39. 87. 113 20. 39. 87. 113 99 9 5 1 9 5 1 9 5 1 F,. 0.4 0.6 0.5 0.4 0.4 0.6 0.5 0.4 0.5 0.4 0.4 0.4 9 4 7 3 1 8 2 4 4 5 a [FJ(1 - 0.9 1.5 1.1 0.8 0.7 1.5 1 0.9 1.0 0.7 0.7 0.8 F.)] 6 7 9 5 6 2 8 9 9 2 R 26. 16. 21. 28. 13. 6.4 10 10. 4.1 5.6 5.6 5.4 0 7 3 2 3 8 R mean ± 23.1 ± 5.1 10.1 & 2.1 5.2 t 0.7 SD V (mL) 5 5 5 W, (g) 0.2 0.5 1.13 5 Table 3.7.0.1.1: Calculated partition coefficient, R for fentanyl in PVC dip-tube-, EVA seal- and Acetal reservoir-aqueous fentanyl nasal solution partition during equilibrium condition (pH 10.0, storage temperature 37 0 C). PVC dip-tube Vp (mL) 0.18 Initial conc. ptg/mL 20.9 39.5 87.1 113 Cs pg/mL 10.8 23.6 46.9 53.2 Mp. pg 54.2 79.3 200 299 Cp. pg/mL 312 456 1158 1719 WO 2010/005400 PCT/SG2009/000248 - 40 R = CJ C,. 28.8 19.3 24.6 32.3 EVA seal Vp (mL) 0.42 Initial conc. pg/mL 20.9 39.5 87.1 113 C,. pg/mL 9.1 24.1 43.7 54.3 Mp. pg 59.3 77 217 294 C. jig/mL 141 183 516 699 R = CPJ C, 15.6 7.6 11.8 12.9 Acetal reservoir Vp (mL) 1.33 Initial conc. pg/mL 20.9 39.5 87.1 113 C,. pg/mL 11.3 17.3 38.7 50.9 Mp. 48.1 111 242 311 Cp. pg/mL 36.1 83.3 182 234 R = CPJ C,. 3.2 4.8 4.7 4.6 VP = Volume of plastic; C,. = fentanyl concentration in the aqueous solution during equilibrium; Mp. = Amount of fentanyl sorbed into the plastic during equilibrium; Cp. = concentration of fentanyl in the plastic at equilibrium; R = partition or distribution 5 coefficient for fentanyl in the plastic-aqueous solution partition at equilibrium. Equilibrium was reached after 5 hours of storage in PVC dip-tube-aqueous nasal solution and EVA seal-aqueous fentanyl nasal solution and 10 days for the Acetal reservoir-aqueous fentanyl nasal solution. The storage temperature was 370C and the pH value for the WO 2010/005400 PCT/SG2009/000248 -41 aqueous nasal solution was buffered to 10.0. The rate and amount of sorption of fentanyl into the plastics was therefore dependant on the type of plastic polymer (PVC >:EVA > Acetal plastic). 3.7.1: Compartmental Model 5 The loss of fentanyl (fraction remaining in solution), when the 3 plastic components were immersed in the nasal solution (buffered at the pH value of 10 and stored at 37'C), were plotted against the storage time. The curves obtained are shown in Figures 3.7.1.0, 3.7.1.2, and 3.7.1.4 for the initial fentanyl concentrations. Figures 3.7.1.0, 3.7.1.2 and 3.7.1.4 represent plots of the fraction of fentanyl remaining in 10 solution versus storage time for the initial fentanyl concentrations of 20.9, 39.5, 87.1 and 113 pg/mL, where the data appeared to be superimposed onto each other. Therefore the fraction of fentanyl loss from the nasal solution into the plastic component of the nasal device (PVC dip-tube, EVA seal and Acetal reservoir) was independent of the initial concentration of fentanyl. 15 3.7.1.0: Bi-exponential Equations derived from the compartmental model. The loss of fentanyl into the 3 plastic components follows a bi-exponential decay as shown in Figures 3.7.1.0.0 - 5. The bi-exponential equations obtained were then used to obtained the diffusion constants, A, B, a and 8. These constants are shown in Table 3.7.1.0.3 below.

WO 2010/005400 PCT/SG2009/000248 -42 Table 3.7.1.0.3: Diffusion constants, A, B, a and f for fentanyl when EVA seal, PVC dip-tube and Acetal reservoir was stored at 37"C in various concentrations of fentanyl nasal solution buffered at pH 10.0 (using the bi-exponential equations). Plastic material pg/mL A B A+B a h-' h PVC diptube 20.9 0.04 0.9 0.94 2.76 0.12 PVC diptube 39.5 0.04 0.94 0.98 6.34 0.10 PVC diptube 87.1 0.14 0.9 1.04 5.84 0.11 PVC diptube 113 0.07 0.91 0.98 1.56 0.14 Average - 0.07 0.91 0.99 4.13 0.12 Std Deviation - 0.05 0.02 0.04 2.33 0.02 EVA seal 20.9 0.13 0.86 0.99 5.36 0.14 EVA seal 39.5 0.06 0.92 0.98 4.73 0.09 EVA seal 87.1 0.16 0.85 1.01 2.36 0.11 EVA seal 113 0.16 0.88 1.04 3.38 0.14 Average - 0.13 0.88 1.00 3.96 0.12 Std Deviation - 0.05 0.03 0.03 1.35 0.02 Acetal Reservoir 20.9 0.26 0.76 1.02 1.07d~1 0.04d Acetal Reservoir 39.5 0.22 0.8 1.02 0.92d- 0.07d Acetal Reservoir 87.1 0.24 0.82 1.06 1.96d- 0.07d Acetal Reservoir 113 0.19 0.84 1.03 0.94d 1 0.07d~1 Average - 0.23 0.81 1.03 1.22d 1 0.06d 1 WO 2010/005400 PCT/SG2009/000248 -43 Std Deviation - 0.03 0.03 0.02 0.50d 1 0.02d 3.7.1.1: Diffusion rate constants a, fl and the intercept A and B for the disappearance of fentanyl into the 3 plastic components. 5 The loss of fentanyl from the aqueous nasal solution into the three plastic matrices can be described using the two compartment model. Hence the loss of fentanyl from the aqueous nasal solution followed two processes. The first part of the curve showed a rapid loss of fentanyl from the aqueous nasal solution. This initial rapid loss was due to the rapid adsorption of fentanyl by the surface of the plastic. The second part of the curve was a 10 slower rate of loss of the fentanyl from the aqueous nasal solution. This second part of the curve was due to the dissolution (or absorption) of fentanyl into the plastic matrix which was a slower process. The adsorbed layer of fentanyl onto the plastic surfaces can be pictured as a few 15 molecular layers thick and this happened instantaneously (due to the lipophilicity of the fentanyl molecule) with a fast rate constant, a (h). The subsequent dissolution or absorption processes was a slower process with a slow rate constant of f (h). The slower rate constants, fl for the 4 initial concentrations of fentanyl were quite similar 20 (0.12 ± 0.02h 1 , 0.12 ± 0.02h 1 and 0.06 ± 0.02d- for the dip-tube, the seal and the reservoir respectively) and hence were independent of the initial concentrations effect (see Table 8.7.2.1.4). It has previously been shown 201 that repetitive determination of the values of B and fl were in close agreement whereas a showed a wide range of values. This is due to the direct consequence of the nature of the biexponential plot and the 25 extreme sensitivity of the curve-fitting procedure. Results showed that for the sorption of acetophenone into a polyethylene plastic container stored at 63 0 C, the repeated values of a yield a standard deviation of 18% for 4 replicate experiments and fl of only 6.5% for the standard deviation. Hence, the a-value should be reported as a range rather than a specific value. 30 In this experiment, the results for the fast rate constants, a were also not closely similar for the four concentrations of fentanyl and could be the result of experimental error. It is quite difficult to quantify this fast rate constant (a) under these experimental conditions and especially at a high temperature of 370C. The first part of the exponential curve was where 35 the instantaneous adsorption process occurred. Fentanyl molecules were attracted to the plastic surfaces by Van der Waals forces or London forces which involved molecular bonding by attraction and these attractions involved only low energies of 1 to 10 kcal/mole which can therefore occur spontaneously 35 . The adsorption process was therefore a rapid process and especially so for lipophilic substances. 40 Adsorption of fentanyl occurred spontaneously when the plastic material is introduced into the fentanyl nasal solution and especially at the higher temperature of 37 0 C. The WO 2010/005400 PCT/SG2009/000248 -44 spontaneous reaction would result in a higher rate of experimental error when determining the rate constant, a.. The adsorption of fentanyl to the Acetal plastic is a much slower process. Therefore the rate constant, a, is subjected to a lesser experimental error and hence the results were more consistent (standard deviation of ± 0.02h- as compared to + 5 2.13 and ± 1.35h-1 for PVC and EVA respectively). The fast rate constant, a, should theoretically be independent of the initial concentrations effect as shown by fentanyl adsorption onto the Acetal reservoir. In a bi-exponential equation for kinetic the values (Equation 2.5.0.1), A + B should add up to one. Table 3.7.1.0.3 indicates that A + B is approximately equal to one, for the three plastic components studied (0.99 ± 0.04, 1.00 10 0.03, and 1.03 ± 0.02 for PVC, EVA and Acetal respectively). The time corresponding to the maximum fraction of fentanyl in the plastic (Fpm) PVC, EVA and Acetal plastic was calculated from Equation 3.7.1.1.0, since a and 8 were known from the biexponential equation. The calculated tm for PVC, EVA and Acetal plastics were 0.88 h, 0.91 h and 2.6 day respectively. Using Equation 3.7.1.1.1, the calculated maximum 15 fraction of solute in the plastic (Fpm) were 0.09, 0.13 and 0.22 respectively for PVC, EVA and Acetal plastic (see Table 3.7.1.1.0 below). These three figures were very close to the experimental values of A as shown in Table 3.7.1.0.3 (0.07, 0.13 and 0.23 for PVC, EVA and Acetal respectively). Table 3.7.1.1.0: Maximum fraction of fentanyl in the plastic (Fpm) as calculated from 20 Equation 3.7.1.1.111; tm the time corresponding to the maximum value of (Fpm) as calculated from Equation 3.7.1.1.111 and intercept A as obtained from the biexponential equation (from Table 3.7.1.0.3). Plastic component tm (Fpm) A PVC dip-tube 0.88 hours 0.09 0.07 EVA seal 0.91 hours 0.13 0.13 Acetal reservoir 2.6 days 0.22 0.23 The quantity of solute adsorbed onto the plastic surface is likely to be small when 25 compared to the-amount of solute migrating (absorbed) into the plastic matrix. However, the rate of adsorption, a, was always much faster than the rate of absorption, /. 3.7.1.2: Estimation of the Clearance (Cis). Permeability (P) and Partition or Distribution Coefficient (R) for fentanyl using the diffusion constants obtained from the Compartment Model. 30 The clearance (Cls), the permeability (P), and the distribution or partition coefficient (R) of fentanyl from the aqueous nasal solution buffered at pH value of 10.0 and stored at 37'C into the 3 plastic materials were calculated respectively and the results were tabulated in Table 3.7.1.2.0 below.

WO 2010/005400 PCT/SG2009/000248 -45 The clearance, Cl,, and the permeability, P, of fentanyl were dependant on the fast rate constant, a. Hence these figures could be subjected to significant experimental errors since a, as mentioned above under section 3.7.1.1, was subjected to a high experimental errors. The clearance for fentanyl from the nasal solution into the PVC dip-tube, EVA seal 5 and Acetal reservoir were 0.64 ± 0.10, 0.68 ± 0.15, and 0.017 ± 0.005 mL/h respectively. The permeability for fentanyl through the PVC dip-tube, EVA seal and Acetal reservoir were 2.15 ± 1.65, 2.90 ± 1.00, and 0.06 ± 0.03 mL/h respectively. However in this experiment fentanyl could not escape or permeate from the plastic matrix into the environment/atmosphere because this was a closed system where the glass tube 10 containing the nasal solution and the plastic component was sealed with Parafilm and capped to prevent escape of moisture. Hence "permeation" of fentanyl could theoretically occur through the plastic matrix and then permeate back into the nasal solution since this was a closed system. The clearance and the permeability of fentanyl were also independent of the initial concentration of fentanyl nasal solutions. At the storage 15 temperature of 370C and at the pH value of 10, the clearance and permeability of fentanyl (non-ionised) into PVC and EVA were similar but were much slower for the Acetal plastic.

WO 2010/005400 PCT/SG2009/000248 -46 Table 3.7.1.2.0: Calculated values for Clearance (Ci,), Permeability (P) and Distribution coefficient (R) of fentanyl, when the three plastic components were immersed in the fentanyl nasal solution buffered to pH 10.0 and stored at 37*C, using the derived diffusion constants of the Compartmental Model Plastic material pg/mL CI, (mL/h) P (mL/h) R PVC diptube 20.9 0.67 1.1 1.23 PVC diptube 39.5 0.53 1.7 1.18 PVC diptube 87.1 0.61 4.6 4.32 PVC diptube 113 0.76 1.2 2.14 Mean - 0.64 2.15 2.22 Std Deviation - 0.10 1.65 1.47 EVA seal 20.9 0.81 4.1 1.80 EVA seal 39.5 0.49 1.85 0.78 EVA seal 87.1 0.64 2.35 2.24 EVA seal 113 0.78 3.3 2.16 Mean - 0.68 2.90 1.75 Std Deviation . 0.15 1.00 0.67 Acetal Reservoir 20.9 0.01 0.05 0.05 Acetal Reservoir 39.5 0.02 0.05 0.04 Acetal Reservoir 87.1 0.02 0.1 0.05 Acetal Reservoir 113 0.02 0.05 0.04 Mean 0.017 0.06 0.045 WO 2010/005400 PCT/SG2009/000248 -47 Std Deviation - 0.005 0.03 0.006 The calculated distribution or partition coefficient, R, for fentanyl in the PVC dip-tube aqueous, EVA seal-aqueous and Acetal reservoir-aqueous system was 2.22 ± 1.47, 1.75 ± 0.67 and 0.045 ± 0.006 respectively. The accuracy of R, obtained using Equation 5 2.5.0.7 depended on the value of Ala and B/fl 10 . The value of A/a must be much smaller than the value of B/fl for Equation 2.5.0.7 to hold true. The distribution or partition coefficient, R, for fentanyl calculated using Equation.2.5.0.7 for the PVC dip-tube aqueous, EVA seal-aqueous and Acetal reservoir-aqueous system was 23.1 5.1, 10.1 ± 2.1 and 5.2 ± 0.7 respectively which was 10 times greater than that obtained from using 10 the Diffusion Model (section 3.7.0.1 above). In the Diffusion Model, R is depended on the value of D/12. This may partly explain the discrepancies in the two estimations of R using the two models if the equilibrium conditions were not reached. However, since equilibrium experimental conditions were obtained in this experiment, the estimate of R would be more accurate using the Diffusion Model because the diffusion equations followed the 15 sorption process from the initial experiment stage through to the equilibrium stage, whereas the Compartment Model only accounts for the initial part of the sorption profile. The open two-compartment model appears to be an adequate empirical model for describing the rate and extent of uptake at the relatively early stages of a given sorption study and has certain deficiencies. The deficiencies arise mainly as a result of the inability 20 of the model to account for the accumulation of solute in the plastic at longer times. Accumulation of solute in the plastic reduces the net rate of solute transfer from the plastic surface into the plastic matrix in comparison to the constant rate of transfer predicted by the open two-compartment model. Consequently, the parameter estimates are not time independent as expected, but vary with the duration of the sorption study. The extent and 25 time course of solute accumulation in the plastic matrix depends on the concentration gradient between the solution and the plastic interface. The concentration in solution declines more slowly as the solution increases in volume and, as a result, the rate constants are found to vary with changes in volume. The apparent partition coefficient, Rapp, calculated using this equation for fentanyl at the 30 pH 10.0 in PVC dip-tube-aqueous, EVA seal-aqueous and Acetal reservoir-aqueous system was 27.8, 11.9 and 0.76 respectively. These apparent partition coefficients for fentanyl in the aqueous nasal solution at pH 10.0 and at the storage temperature of 37 0 C are tabulated in Table 3.7.1.2.1 together with the partition coefficients obtained using the Diffusion Model and the Compartment Model.

WO 2010/005400 PCT/SG2009/000248 - 48 Table 3.7.1.2.1: Partition Coefficients R, obtained from Diffusion and Compartment Model for the three plastic components. Plastics R (Compartment Model) R (Diffusion Rapp using Equation Model) 8.7.2.3.1272 PVC dip-tube 2.22 ± 1.47 23.1 ± 5.1 27.8 EVA seal 1.75 ±0.67 10.1 ±2.1 11.9 Acetal reservoir 0.045 ± 0.006 5.2 ± 0.7 0.76 3.7.1.3: Rate constants k 12 (adsorption), k 23 (absorption) and k 2 1 (desorption) The rate constants, k 12 , k23, and k 2 1 were calculated and the results were tabulated in 5 Table 3.7.1.3.0 below. Since the rate constants were all dependant on the fast rate constant, a, these figures were hence subjected to experimental errors. The k2 or the adsorption rate constant of fentanyl from the nasal solution onto the surface of the plastic (PVC, EVA and Acetal) was 0.43 ± 0.33, 0.58 ± 0.20 and 0.013 ± 0.005 h respectively. The k 2 1 or the desorption rate constant of fentanyl from the PVC, EVA and Acetal plastic 10 back into the nasal solution was 1.23 ± 0.52, 0.84 ± 0.26 and 0.01 ± 0.00 h respectively. The k 23 or the absorption rate constant of fentanyl from the nasal solution into the PVC, EVA and Acetal was 2.58 ± 2.01, 2.66 ± 1.06 and 0.03 ± 0.01 h respectively. The summation of the three rate constants k 23 + k 1 2 - k 2 1 was positive which means that there was a net gain of fentanyl into the plastic matrix from the nasal solution (1.78 h- for PVC 15 dip-tube, 2.4 h for EVA seal and 0.033 h for Acetal reservoir). The experimental results showed that k 12 (adsorption rate constant) for the adsorption of fentanyl from the nasal solution onto the EVA seal (0.58 ± 0.20 h) is faster than the adsorption of fentanyl onto the PVC dip-tube (0.43 ± 0.33 h 1 ). Since k 12 is calculated from the fast rate constant, a, and this is highly subjected to experimental error, hence k 1 2 was 20 also subjected to a wide standard deviation (± 0.33 h for PVC dip-tube and ± 0.20 h- for EVA seal). Taking into account this standard deviation, the adsorption rate constant for fentanyl onto PVC dip-tube could be in fact similar to EVA seal. The adsorption rate of fentanyl onto Acetal reservoir was much slower than the other two plastic components (PVC and EVA), and also has a lower standard deviation (± 0.005 h- 1

).

WO 2010/005400 PCT/SG2009/000248 -49 Table 3.7.1.3.0: Rate constants k 1 2 , k 23 , and k 21 of fentanyl (stored at 37*C, pH 10.0) in the plastic components of the nasal device calculated using the diffusion constants obtained from the Two-Compartment Model. Plastic material pg/mL *k 1 2 (h) k 23 (h 1 ) k 21 (h) PVC diptube 20.9 0.22 1.51 1.15 PVC diptube 39.5 0.35 1.81 4.28 PVC diptube 87.1 0.92 0.7 4.33 PVC diptube 113 0.24 0.91 0.55 Mean - 0.43 1.23 2.58 Std Deviation - 0.33 0.52 2.01 EVA seal 20.9 0.82 0.92 3.76 EVA seal 39.5 0.37 1.15 3.30 EVA seal 87.1 0.47 0.55 1.45 EVA seal 113 0.66 0.72 2.14 Mean 0.58 0.84 2.66 Std Deviation - 0.20 0.26 1.06 Acetal Reservoir 20.9 0.01 0.01 0.03 Acetal Reservoir 39.5 0.01 0.01 0.02 Acetal Reservoir 87.1 0.02 0.01 0.05 Acetal Reservoir 113 0.01 0.01 0.02 Mean 0.013 0.01 0.03 WO 2010/005400 PCT/SG2009/000248 - 50 Std Deviation - 0.005 0 0.01 k 1 2 = adsorption of fentanyl from the nasal solution onto the wall of the plastic components of the nasal device. k 23 = absorption of fentanyl from the nasal solution into the plastic components of the nasal device. k 2 l = release of fentanyl from the plastic components of the nasal device back into the nasal solution. 5 3.7.3: Effects of temperature, pH value and ionic strength (p) differences in the disappearance kinetics of fentanyl in the PVC, EVA and Acetal plastics: The PVC dip-tube, EVA seal and Acetal reservoir was immersed into the fentanyl nasal solutions (strength = 54.9pg/mL of fentanyl), buffered with phosphate buffer to the pH 10 values of 8.0 and 6.0 and in different ionic strengths (p of 0.15, 0.3 and 0.5). These plastic-aqueous solution systems were stored in a water bath, at the preset temperatures of 5, 25 and 370C. There was no loss of fentanyl from the aqueous nasal solution when the 54.9 pg/mL 15 fentanyl aqueous nasal solutions which were buffered with phosphate buffer at pH 6.0 and ionic strengths of p of 0.15 and 0.5, and also at pH 8 (p of 0.3 and 0.5) was stored in the glass amber bottle of the nasal device at the storage temperatures of 5, 25 and 370C after one year's storage. Hence the fentanyl nasal solution was stable for at least a year in this amber silicoborate glass bottle and there was no measurable loss by sorption or by 20 degradation. Table 3.7.3.0 showed the fraction of fentanyl remaining in the aqueous nasal solution after one year storage in this amber glass bottle. However there were losses of fentanyl from the nasal solution when the PVC dip-tube, EVA seal and Acetal reservoir was immersed and stored in the fentanyl nasal solution 25 buffered with the phosphate buffer at pH 8.0 (ionic strength of 0.3 and 0.5) at the storage temperatures of 5, 25 and 37'C. The initial decline in the fentanyl concentration was rapid (especially at the storage temperatures of 25 and 370C in the PVC and EVA plastics) followed by a slower fentanyl loss from solution reaching a plateau (steady state). The fraction of fentanyl remaining in the nasal solution when the PVC dip-tube, EVA seal and 30 Acetal reservoir was immersed in the 54.9 pg/mL fentanyl nasal solutions at pH 8.0 (with p = 0.3 and 0.5) are shown in Figures 3.7.3.0, 3.7.3.1, 3.7.3.2 and 3.7.3.3.

WO 2010/005400 PCT/SG2009/000248 - 51 Table 3.7.3.0: Fraction of fentanyl remaining in the nasal solution buffered with phosphate buffer to pH 6.0 and 8.0 (p = 0.15, 0.3 and 0.5) was stored in the amber silicaborate glass bottle of the nasal device at the storage temperatures of 5, 25, and 37*C. Storage time in Im 3m 5m 7m 9m 12m months 5-C pH6 1.03 1.00 1.08 1.1 1.11 1.09 p=0.15 pH6 0.96 1.08 1.10 1.09 1.08 1.00 p.=0.5 pH8 1.05 1.07 0.99 1.05 1.1 1.06 p=0.3 pH8 1.04 1.05 1.09 1.07 1.1 1.06 p=0.5 25 0 C pH6 0.95 1.01 1.05 1.03 1.07 1.05 p=0.15 pH6 1.01 1.00 1.08 1.03 0.92 1.06 p=0.5 pH8 1.03 0.94 1.01 0.98 1.06 0.92 p=0.3 pH8 0.97 1.05 1.03 1.02 1.01 1.01 p=0.5 370C pH6 1.01 1.05 1.08 0.96 1.03 0.93 pj=0.15 pH6 0.99 1.01 1.1 1.03 1.01 1.07 p=0.5 pH8 1.07 0.98 1.06 0.93 1.03 1.02 p=0.3 pH8 1.08 0.96 1.01 1.02 1.02 1.04 =p0.5 5 Significant losses of fentanyl were observed when the PVC dip-tube and the EVA seal were immersed and stored at the storage temperatures of 25 and 370C and at pH 8.0 (for both the ionic strengths of 0.3 and 0.5). The most rapid and extensive loss of fentanyl occurred when the PVC dip-tube was stored in the fentanyl nasal solution. After 10 approximately 20 hours of storage at the storage temperature of 370C, only 30% of the original fentanyl remained in aqueous nasal solution (see Figure 3.7.3.1). Fentanyl was also rapidly and extensively lost when the EVA seal was stored in the fentanyl nasal solution at 370C (pH 8.0, with p of 0.3 and 0.5). Approximately 40% of the original fentanyl remained in aqueous solution (see Figure 3.7.3.2) after storing for 25 hours. As expected 15 the loss of fentanyl from the aqueous nasal solution (pH 8.0) into the PVC dip-tube and EVA seal at the storage temperature of 50C was much slower. 70% of the original fentanyl remained in the aqueous nasal solution (see Figure 3.7.3.0) after the PVC dip-tube was immersed and stored in the fentanyl nasal solution (buffered with phosphate buffer at pH 8.0, p of 0.3 and 0.5) for two months at the storage temperature of 50C. When the EVA 20 seal was immersed and stored in the fentanyl nasal solution (at pH 8.0 with p of 0.3 and WO 2010/005400 PCT/SG2009/000248 - 52 0.5), 90% of the original fentanyl concentration remained in solution after three months at 50C. Hence, at the lower storage temperature the rate of disappearance of fentanyl by sorption into the plastic matrix was markedly decreased. 5 Figure 3.7.3.3 shows the loss of fentanyl from the aqueous nasal solution into the Acetal plastic, when the Acetal reservoir was immersed and stored in the fentanyl nasal solution buffered with phosphate buffer at a pH 8.0 (p = 0.3 and 0.5). The rate of disappearance of fentanyl by sorption into the Acetal reservoir was significantly slower as compared to the loss by sorption into the PVC dip-tube and EVA seal. 20% and 75% of the original fentanyl 10 remained in the aqueous nasal solution after storing for 100 days at the storage temperatures of 37 and 250C respectively. There was no loss of fentanyl even after storing the Acetal reservoir in the fentanyl nasal solution at pH 8.0 for 350 days at a storage temperature of 50C. Figures 3.7.3.5 and 3.7.3.6 showed that there were no losses of fentanyl from the aqueous nasal solution when the EVA seal and Acetal reservoir were 15 immersed and stored in the fentanyl nasal solution buffered with phosphate buffer to a pH 6.0 (p = 0.15 and 0.50) and stored at the storage temperatures of 5, 25 and 370C. However there was some loss of fentanyl when the PVC dip-tube was immersed and stored in the fentanyl nasal solution buffered with phosphate buffer (p = 0.15 and 0.50) at the pH 6.0 (see Figure 3.7.3.4). Figure 3.7.3.4 showed that at the storage temperatures of 20 370C and 250C, 50% and 80% of the initial fentanyl concentration remained in solution when the PVC dip-tube was immersed and stored for six months. However, there was no loss of fentanyl when the PVC dip-tube was immersed in the fentanyl solution at the pH 6.0 (p = 0.15 and 0.50) for 8 months when the storage temperature was 5*C. Figures 3.7.3.0 to- 3.7.3.6 show that there are no significant differences in the disappearance 25 kinetics of fentanyl into the plastic components of the nasal device whether the ionic strength, p, of the nasal solution was 0.15, 0.3 or 0.50. Hence ionic strength did not influence the disappearance kinetics of fentanyl into the three plastic materials for the solutions tested. The ionic strength did not influence the loss of fentanyl into the plastic components of the nasal device, even when fentanyl is largely ionised at pH 6.0. These 30 data were also found to fit the two-compartmental model (bi-exponential equation fitted the regression line of the curve) for the disappearance kinetics of fentanyl in the three plastic materials. The loss of fentanyl by sorption into the 3 plastic materials hence depended on the type of 35 plastic, the storage temperature as well as the pH value of the nasal solution but was independent of the ionic strength of the solution. 3.7.4 Kinetic parameters obtained using the Compartment Model for fentanyl sorption into the plastic components at the pH 6.0 and 8.0 and at the storage 40 temperatures of 5, 25 and 37*C. The loss of fentanyl into the PVC dip-tube, EVA seal and Acetal reservoir from the 54.9 pg/mL fentanyl nasal solution, buffered with phosphate buffers at pH 8.0 and 6.0 with the ionic strengths p of 0.30, 0.15 and 0.50 and stored at the storage temperatures of 5, 25 WO 2010/005400 PCT/SG2009/000248 - 53 and 370C were analysed using the Compartment Model. The regression lines for the Figures 3.7.3.0, 3.7.3.1, 3.7.3.2, 3.7.3.3 and 3.7.3.4 obeyed the bi-exponential equation. Tables 3.7.4.0, 3.7.4.1, and 3.7.4.1.2 tabulate the kinetics data obtained for the three plastic components using the bi-exponential equations derived from the curve fitting (least 5 square method). Table 3.7.4.0: Disappearance kinetics of fentanyl when the PVC dip-tube was immersed and stored in the fentanyl nasal solution (54.9 pg/mL) buffered to pH 6.0 and 8.0 (p, of 0.15, 0.30 and 0.50) and stored at the storage temperatures of 37, 25 10 and 5 0 C. 370C 250C 50C p 8 8 6 6 8 8 6 6 8 8 6 6 H 0.30 0.50 0.15 0.50 0.30 0.50 0.15 0.50 0.3 0.50 0.1 0.5 0 5 0 A 0.33 0.48 0.12 0.04 0.36 0.39 0.03 0.01 0.0 0.08 - 8 B 0.65 0.54 0.78 0.89 0.61 0.66 0.92 0.97 0.9 .91 - 1 A 0.98 1.02 0.90 0.93 0.97 1.04 0.94 0.98 0.9 .99 - + g B a 0.32 0.40 0.59 0.58 0.07 0.10 0.29 0.1 0.4 1.5 - h 1 h 1 mr mn h 1 hW m- 1 m 1 9 mr 1 p 0.02 0.03 0.08 0.07 0.004 .006 0.03 .02 0.0 .06 - hM h 1 mn m_ W h 1 m m-1 6 m~n m 1 k 12 0.12 0.26 0.12 0.09 0.03 0.04 0.03 0.02 0.1 0.17 - h-W h m 1 m~' h h m 1 m 1 7 m~1 m~ 1 k 21 0.16 0.16 0.17 0.09 0.04 0.05 0.04 0.01 0.8 0.86 - h h" mr 1 m- 1 h~ h1 rm 1 m 1 4 m 1 m- 1 k 23 0.06 0.05 0.35 0.47 0.01 0.01 0.25 0.1 0.5 0.53 - h hW m 1 m h 1 h m" m- 1 4 m~ 1 rn" 1 Cl 0.17 0.23 0.49 0.40 0.03 0.04 0.14 0.1 0.3 0.32 - S mLh~ mLh~ mLm~ mLm~ mLh~ mLh~ mLm mLm~ 3 mLm~ 1 1 1 1 1 1 1 1 mnL 1 r-1 P 0.61 1.29 0.61 0.44 0.13 0.21 0.16 0.1 0.8 .85 - mLh~ mLh~ mLm~ mLm^ mLh~ mLh~ mLm~ mLm~ 4 mLm~ 1 1 1 1 1 1 1 1 mL mL R 14.1 24.7 4.28 1.25 16.4 16.4 0.91 0.29 2.4 2.44 - 4 WO 2010/005400 PCT/SG2009/000248 - 54 Where a = fast rate constant, p = slow rate constant, A and B are the fractional intercept and A + B = 1, k12= adsorption of fentanyl from the solution onto the surface of the plastic, k23 = absorption of fentanyl from the solution into the plastic matrix, k 21 = release of 5 fentanyl from the plastic matrix (desorption) back into the solution, CI the clearance of the fentanyl from the solution, P the permeability coefficient of fentanyl onto the surface of the plastic and R is the distribution or partition coefficient of fentanyl between the plastic matrix and the solution. 10 Table 3.7.4.1: Disappearance kinetics of fentanyl when the EVA seal was immersed and stored in the fentanyl nasal solution (54.9 pg/mL) buffered to pH 6.0 and 8.0 (p, = 0.15, 0.30 and 0.50) at the storage temperatures of 37, 25 and 5 0 C. 370C 250C 50C pH 8 8 6 6 8 8 6 6 8 8 6 6 [t 0.30 0.50 0.1 0.5 0.30 0.50 0.15 0.5 0.30 0.50 0.1 0.5 5 0 0 5 0 A 0.34 0.57 - - 0.25 0.46 - - 0.16 0.2 - B 0.66 0.53 - - 0.67 0.53 - - 0.92 0.9 - A + 1.0 1.1 - - 0.92 1.0 - - 1.08 1.1 - B a 0.43 0.40 - - 1.62 0.56 - - 1.59 1. 1m - h 1 h 1 d 1 d- 1 m __ p 0.02 0.03 - - 0.03 0.01 - - 0.02 0.01m - h- 1 h- 1 d- 1 d- 1 m_ _ I k 12 0.16 0.24 - - 0.43 0.27 - - 0.27m~ 0.23m - hd h-1 d 1 d- 1 1 k21 0.23 0.14 - - 1.1d 0.28 - - 1.24m- 0.83m~ - h1 h I d- 1 1 1 k 23 0.06 0.05 - - 0.12 0.03 - - 0.09m- 0.05m~ - h- h- d 1 d- 1 1 1 Cls 0.18 0.24 - - 0.23 0.12 - - 0.09 0.06 - mLh- mLh- mLd- mLd mLm 1 mLm' 1 1 1 1 P 0.82 1.21 - - 2.15 1.34 - - 1.35 1.15 - mLh~ mLh- mLd~ mLd- mLm 1 mLm 1 1 1 1 1 R 6.13 12.8 - - 4.44 10.3 - - 2.07 2.64 Where a = fast rate constant, 8 = slow rate constant, A and B are the fractional intercept 15 and A + B = 1, k 12 = adsorption of fentanyl from the solution onto the surface of the plastic, k23 = absorption of fentanyl from the solution into the plastic matrix, k 21 = release of fentanyl from the plastic matrix (desorption) back into the solution, Cis the clearance of the fentanyl from the solution, P the permeability coefficient of fentanyl onto the surface of the WO 2010/005400 PCT/SG2009/000248 - 55 plastic and R is the distribution or partition coefficient of fentanyl between the plastic matrix and the solution. Table 3.7.4.2: Disappearance kinetics of fentanyl when the Acetal reservoir was 5 immersed and stored in the fentanyl nasal solution (54.9 pg/mL) buffered to pH 6.0 and 8.0 (p, of 0.15, 0.30 and 0.50) at the storage temperatures of 37, 25 and 5*C. 370C 250C 50C pH 8 8 6 6 8 8 6 6 8 8 6 6 0.30 0.50 0.1 0.5 0.30 0.5 0.1 0.5 0.3 0.5 0.1 0.5 5 0 5 0 0 0 5 0 A 0.33 0.41 - - 0.14 0.15 - - - - - B 0.67 0.67 - - 0.79 0.79 - - - - - A+B 1.0 1.08 - - 0.93 0.94 - - - - - Sd- 1 0.32 0.26 - - 0.05 0.05 - - - - - p d 1 0.02 0.01 - - .0002 0.0002 - - - - - k12d- 0.12 0.11 - - 0.007 0.007 - - - - - k21 d:' 0.18 0.14 - - 0.04 0.04 - - - - - k23 d 0.04 0.02 - - 0.001 0.0001 - - - - - Cis 0.10 0.07 - - 0.001 0.001 - - - - - mLd- 1 P 0.57 0.55 - - 0.04 0.04 - - - - - mLd 1 R 1.85 2.3 - - 0.67 0.71 - - - - - Where a = fast rate constant, # = slow rate constant, A and B are the fractional intercept and A + B = 1, k 1 2 = adsorption of fentanyl from the solution onto the surface of the plastic, 10 k 23 = absorption of fentanyl from the solution into the plastic matrix, k21 = release of fentanyl from the plastic matrix (desorption) back into the solution, Cl, the clearance of the fentanyl from the solution, P the permeability coefficient of fentanyl onto the surface of the plastic and R is the distribution or partition coefficient of fentanyl between the plastic matrix and the solution. 15 Table 3.7.4.3: Equation of best-fit (bi-exponential equation) obtained for the disappearance kinetics of fentanyl when the PVC dip-tube was immersed and stored in the fentanyl nasal solution at the storage temperatures of 37, 25 and 5*C. 0C p pH 8.0 pH 6.0 37 0.3 or Ft = 0.33e~"'+ 0.65e 0 02 t Ft = 0.12e*t+ 0.78e~"08 0.15 37 0.5 Ft = 0.

48 e* 4 t+ 0.54e-0t Ft = 0.04e-0+ 0.

89 e-7t 25 0.3 or Ft = 0.

36 e-07t+ 0.61e-0.004t Ft = 0.03e 9 t + 0.92e-.

03 t 0.15 25 0.5 Ft = 0.39e~t+0.66e 0 00 6 t Ft = 0.01e* +0.97e". 5 0.3 or Ft = 0.08e.

49 +0.91e~ 00 6 WO 2010/005400 PCT/SG2009/000248 - 56 0 .1 5 ._ 5 0.5 Fi=0.08e- +0.91e-t Table 3.7.4.4: Equation of best-fit (bi-exponential equation) obtained for the disappearance kinetics of fentanyl when the EVA seal was immersed and stored in 5 the fentanyl nasal solution at the storage temperatures of 37, 25 and 5"C. 0C P pH 8.0 pH 6.0 37 0.3 (pH8) or 0.15 (pH6) = 0.34e- +0.66e- 37 0.5 = 0.57&o.4t + 0.53e 003t 25 0.3 (pH8) or 0.15 (pH6) = 0.25e 1

.

2t +0.67e 003t 25 0.5 Ft= 0.46e 056t +0.53e 00 t 5 0.3 (pH8) or 0.15 (pH6) F = 0.16e15 '+0.92-et 5 0.5 Ft=0.2e- t +0.9e-O' Table 3.7.4.5: Equation of best-fit (bi-exponential equation) obtained for the disappearance kinetics of fentanyl when the Acetal reservoir was immersed and 10 stored in the fentanyl nasal solution at the storage temperatures of 37, 25 and 5*C. 00 p pH 8.0 pH 6.0 37 0.3 (pH8) or 0.15 (pH6) F= 0.

3 3 e.32t+ 0.

6 7 eo.02t 37 0.5 F~= 0.415e3'+0.67e-olt 25F 0.3 (pH8) or 0.15 (pH6) = 0.14e- T+ 0.79e-oo t 25 0.5 Ft= 0. 1 5e04 - 05t-+0.79e 0002 t 5 0.3F(pH8) or.15 (pH6) 5 0.5 Fi = 0.2e_-"+ 0.9e-"-___ 3.7.5: Temperature dependence of the kinetic constants a, i, A and B 15 The kinetics parameters derived from the Compartment Model were plotted against the inverse of their storage temperatures (lIT) in Kelvin. These graphs were shown in the Figures 3.7.5.0, 3.7.5.1 and 3.7.5.2 (Arrhenius type curve for the fast and slow rate constants a and 0. and also for the intercept constant B). From the Figures 3.7.5.0 and 3.7.5.1, the kinetic rate constants (a, fI) for the disappearance of fentanyl into the PVC 20 and EVA plastics increased proportionately with an increase in the storage temperature whereas the intercept constant B decreased with an increase in the storage temperature. The Figures 3.7.5.0, 3.7.5.1 and 3.7.5.2 showed that there was a linear relationship between the rate constants a, f, and B and the reciprocal of the storage temperature with little difference in slopes for a and f The two linear regression lines for a and g were also 25 near parallel. The slope of the graph can be used to obtain the activation energy needed to move fentanyl molecules into the plastic matrix. The activation energies for sorption of fentanyl into the PVC dip-tube using the fast rate constant a and the slow rate constant WO 2010/005400 PCT/SG2009/000248 -57 from the Figure 3.7.5.0 was 27.6 and 28.4 kJ per mole respectively. The activation energies for sorption of fentanyl into the EVA seal using the fast rate constant a and the slow rate constant f from the Figure 3.7.5.1 was 26.5 and 35.3 kJ per mole respectively. The rate constants for a and f8 can then be obtained by extrapolation of this Arrhenius 5 type graph for any storage temperature. Hence the graphical plot of the logarithm of the hybrid rate constant (and also the intercept B) against the reciprocal of absolute temperature showed linearity according to the Arrhenius type relationship. The regression lines for a and P was also parallel. Thus all the parameters required for the bi-exponential equation are now known for the required storage temperature and hence F, could now be 10 predicted at any time t. 3.7.6: Storage Temperature and Partition coefficient, R calculated using the Compartment Model. In theory the partition coefficient R for fentanyl in the aqueous nasal solution-plastic 15 system should be independent of the storage temperature 12 . However the experimental results (see Tables 3.7.4.0, 3.7.4.1 and 3.7.4.2 above) for R showed some differences in values for the three storage temperatures at pH 8.0. The average partition coefficient, R for fentanyl in the PVC dip-tube-aqueous system at the storage temperatures of 37, 25 and 50C was 19.4, 16.4 and 2.44 respectively. The R for fentanyl in the EVA seal-aqueous 20 system at the storage temperatures of 37, 25 and 5C was 9.47, 7.37 and 2.36 whereas the average R for fentanyl in the Acetal reservoir was 2.08 and 0.69 and there was no result at 50C. 3.7.7 Kinetic parameters obtained using the Diffusion Model for fentanyl at pH 6.0 25 and 8.0 and at the storage temperature of 5, 25 and 37"C: A reduced sorption curve was plotted at different storage temperatures for the pH 6.0 and 8.0. There was no loss of fentanyl into the plastic components when the aqueous nasal solution was buffered at the pH 6.0 except for the PVC dip-tube at the storage 30 temperatures of 25 and 370C for the length of time of this study period. Hence there were only two data sets for plotting the reduced sorption curve for fentanyl sorption into the PVC dip-tube and the EVA seal at the three different storage temperatures. These data were shown in Figures 3.7.7.0 and 3.7.7.1. No reduced sorption curve was plotted for the loss of fentanyl into the Acetal reservoir because of incomplete data (no loss of fentanyl 35 into Acetal reservoir for the study period at the storage temperature of 50C at pH 8.0 and also for the storage temperatures of 5, 25 and 370C at pH 6.0). The diffusion coefficient, D changes with temperature. The increase in diffusion coefficient was proportional to the increase in the storage temperature. Table 3.7.7.0, 3.7.7.1, and 40 3.7.7.2 below showed the diffusion coefficient, D values obtained from the experimental results for the disappearance kinetics of fentanyl into the three plastic components of the nasal device using the Diffusion Model.

WO 2010/005400 PCT/SG2009/000248 -58 Table 3.7.7.0: PVC plastic Dip-tube stored in fentanyl nasal solutions (initial concentration of fentanyl = 54.9 pg/mL, ionic strengths, p, of 0.15, 0.30 and 0.50, pH 8.0 and 6.0) at the storage temperatures of 37, 25 and 5 0 C. Temp. Initial Conc. Ionic pH Diffusion coefficient (D) "C pg/mL strength (p) value cm2/s S 37 54.9 0.30 8.0 1.4 X 10~ 37 54.9 0.50 8.0 1.5X 10 7 37 54.9 0.15 6.0 5.3 X 10 37 54.9 0.50 6.0 4.7 X 10-1 25 54.9 0.30 8.0 2.1 X 10 25 54.9 0.50 8.0 0.9 X 10 25 54.9 0.15 6.0 3.0 X 10-0 25 54.9 0.50 6.0 2.8 X 10-0 5 54.9 0.30 8.0 7.1 X 1010 5 54.9 0.50 8.0 7.1 X 10~'0 5 54.9 0.15 6.0 5 54.9 0.50 6.0 5 Table 3.7.7.1: EVA plastic seal stored in fentanyl nasal solutions (initial concentration of fentanyl = 54.9 pg/mL, ionic strengths, p, of 0.15 and 0.50, pH 8.0 and 6.0) at the storage temperatures of 37, 25 and 5 0 C. Temp. Initial Conc. Ionic pH values Diffusion coefficient (D) *C pg/mL strength (p) cm2/s 37 54.9 0.30 8.0 1.7 X 10-' 37 54.9 0.50 8.0 2.4 X 10-7 37 54.9 0.15 6.0 37 54.9 0.50 6.0 25 54.9 0.30 8.0 1.2 X 10~ 25 54.9 0.50 8.0 2.0 X 10 25 54.9 0.15 6.0 25 54.9 0.50 6.0 5 54.9 0.30 8.0 4.9 X 10-l 5 54.9 0.50 8.0 4.9 X 1010 5 54.9 0.15 6.0 5 54.9 0.50 6.0 10 WO 2010/005400 PCT/SG2009/000248 - 59 Table 3.7.7.2: Acetal reservoir plastic stored in fentanyl nasal solutions (initial concentration of fentanyl = 54.9 pg/mL, ionic strengths, p, of 0.15 and 0.50, pH 8.0 and 6.0) at the storage temperatures of 37, 25 and 5*C. Temp. Initial Conc. Ionic pH Diffusion coefficient (D) "C pg/mL strength (p) values cm 2 /s 37 54.9 0.30 8.0 5.8 X 1010 37 54.9 0.50 8.0 6.3 X 1010 37 54.9 0.15 6.0 37 54.9 0.50 6.0 25 54.9 0.30 8.0 2.5 X 1010 25 54.9 0.50 8.0 2.8 X 1010 25 54.9 0.15 6.0 25 54.9 0.50 6.0 5 54.9 0.30 8.0 5 54.9 0.50 8.0 5 54.9 0.15 6.0 5 54.9 0.50 6.0 5 3.7.8: Temperature dependence of the Diffusion coefficient: Data in Tables 3.7.7.0, 3.7.7.1 and 3.7.7.2 above show that an increase in the storage temperature leads to an increase in the diffusion coefficient, D of fentanyl into the plastic matrix. The natural logarithm of diffusion coefficients (InD) were plotted against the inverse of the storage temperatures and a linear relationship was obtained (Figure 10 3.7.8.0). The activation energy Ed is the energy requirement for fentanyl molecules (at pH value of 8.0, 27.1% of the fentanyl existed as molecules, calculated using Equation 2.5.3.2) to diffuse into the plastic matrix. lonised species contribute only to a very small degree of the total sorption process 28 . The activation energy, Ed for the fentany molecule to diffuse (into holes) in the PVC dip-tube and EVA seal was found to be 26.9 and 29.8 kJ 15 per mole respectively (at the pH value of 8.0). The loss of fentanyl from the aqueous nasal solutions into the plastic component was markedly temperature dependent. As the temperature increases, the diffusion of the fentanyl molecules into the plastic material would be expected to increase due to the additional kinetic/diffusion energy provided by the heat. At the same time, increased temperature may also alter the structure of the 20 plastic polymer 28 . The hydrolysis would produce new binding sites for solutes, increasing the sorption process. An increase in temperature usually increases the diffusion of a solute into the plastic matrix in accordance with the Arrhenius type model. The activation energy is a measure of the energy needed to expand and "break" the polymer strand and against the cohesive forces of the polymer in order to form the gaps through which 25 diffusion can occur. With an increasing molecular size of the solvent molecule there was a decrease in the solvent uptake by the EVA matrix. The uptake of benzene (molecular weight = 78.11) by EVA is maximum, toluene (molecular weight = 92.14) is intermediate, while xylene (molecular weight = 106.2) is minimum. The decrease in uptake with increase in penetrant size might be due to the greater activation energy required for 30 activating the sorption process. Larger penetrant molecules do not fit into the sites or free WO 2010/005400 PCT/SG2009/000248 -60 space already available in the polymer matrix. They therefore require more energy to create additional space within the polymer matrix. The activation energy for nitroglycerin (19.6 kcal/mole) is three times more than the 5 activation energy for fentanyl (26.9 kJ/mole equivalent to 6.43 kcal/mol). Hence nitroglycerin may be less lipophilic than fentanyl for the PVC material. Also, the PVC materials in these two experiments may not be similar 28 , hence a difference in the diffusion properties for the two solutes (nitroglycerin and fentanyl). The mobility of the PVC polymer chains can affect the rate of diffusion and a rise in temperature could affect 10 polymer chain mobility hence resulting in a change in the diffusion coefficient of the diffusion solute. 3.7.9: Storage Temperatures and Partition Coefficient, R, Calculated from the Diffusion Model. 15 The partition or distribution coefficients for fentanyl in the three plastic materials (PVC dip tube, EVA seal and Acetal reservoir) were tabulated in the Table 3.7.9.0 below. The fentanyl nasal aqueous solution was kept at a pH value of 8.0. The partition coefficient, R for fentanyl in the PVC dip-tube-aqueous system at the storage temperatures of 37, 25 and 50C was calculated to be 71.2, 39.1 and 14.1 respectively. The partition coefficient, R 20 for fentanyl in the EVA seal-aqueous system at the storage temperatures of 37, 25 and 50C was calculated to be 48.8, 13.8 and 2.8 whereas the R value for fentanyl in Acetal reservoir-aqueous system, at the storage temperatures of 37, 25 and 50C was 1.8 and 0.7 respectively and there was no effect at 50C. 25 Table 3.7.9.0: Calculated values for R (partition coefficient), for the fentanyl in PVC dip-tube-aqueous, EVA seal-aqueous and Acetal reservoir-aqueous system (the pH value of the fentanyl nasal solution was kept at 8.0) using Equation 2.5.1.1" at 5, 25 and 37*C. 50C 2500 370 C F. Time R F,,. Time R F. Time R toF. toF. toFF. PVC dip-tube 0.64 9.8m 14.1 0.39 90h 39.1 0.26 61h 71.2 EVA seal 0.78 9.8m 2.8 0.42 504h 13.8 0.17 320h 48.8 Acetal reservoir - - - 0.74 8.1m 1.6 0.42 1.2m -6.1 30 The partition coefficient R was shown in some studies to be independent of the storage temperature. The temperature dependence of the sorption process is therefore, a result of the increased diffusivity into the polymer matrix, but the equilibrium fractional uptake being unaffected by changes in temperature. 35 In this experiment, the experimental results for the partition coefficients for fentanyl in the three plastic materials (PVC dip-tube-, EVA seal- and Acetal reservoir-aqueous system) showed some difference in the values for the three storage temperatures (37, 25 and 50C) at the pH 8.0 (Table 3.7.9.0). A plot of Ln partition coefficient versus the reciprocal storage WO 2010/005400 PCT/SG2009/000248 - 61 temperature (K 1 ) showed a linear relationship (Graph 3.7.9.0). AHO obtained from the gradient of the graph for PVC dip-tube = 35.8kJmor- and AH 0 obtained from the gradient of the graph for EVA seal = 62.4kJmol 1 5 3.7.10: Prediction of the partition coefficient of fentanyl into the plastic matrix (RPIastic-water) using its octanol- and hexane-water Partition Coefficient, Ro.w and RH-W. The partition coefficient, R of a solute in an aqueous-plastic system is usually assumed to be independent of temperature changes during the equilibrium condition 12 . R is a useful 10 parameter to use to predict the amount of solute sorbed into a plastic matrix at equilibrium at the stated temperature. The partition coefficient, R can be used to predict the shelf-life of a drug in aqueous solution when stored in a plastic container. 3.7.11: Prediction of the solubility of fentanyl in water (Sw) and in plastic matrix 15 (Spoiymer) using its Partition Coefficient, R: The octanol-water partition coefficient for fentanyl is 8709.6 at a pH 11. The calculated Ro. w using Equation 3.7.10.5 at the pH 10 was 213.4. Therefore REVA-w = 80.3 (partition coefficient for fentanyl in EVA). Hence the solubility of fentanyl in EVA = 854 mg/mL 20 (using Equation 3.7.11.0), then the solubility of fentanyl in the EVA seal equals to 854 mg/mL. Fentanyl is hence 1.2 times more soluble in EVA plastic than in PVC dip-tube. 3.7.12: Ionic Strength, y and temperature differences. 25 This study did not show an effect of ionic strength on the loss of fentanyl from the nasal aqueous solution when the different components of the plastic nasal device were soaked in the nasal solutions with varying ionic strengths. Sodium chloride was added to adjust the ionic strength of the nasal solution to 0.15, 0.30 or 0.50. Figures 3.7.3.0 - 3.7.3.6 in Section 3.7.3 above showed that ionic strength did not interfere with the loss of fentanyl 30 into PVC, EVA or Acetal plastics. Fentanyl in the nasal solutions with different ionic strengths of 0.50, 0.30 and 0.15 showed the same amount of loss in the PVC dip-tube, EVA seal and Acetal reservoir. 3.7.13: Influence of the nasal solution pH on the sorption/disappearance kinetics of 35 fentanyl into the plastic matrixes. A factor which may influence the sorption of acidic or basic drugs into the plastic material is the pH of the solution. The non-ionised species of the drug is the most lipophilic and is the most favourably sorbed by the plastic matrix. The amount of the non-ionised form of 40 the drug present in an aqueous solution is controlled by the solution's pH and the drug's pKa value. Fentanyl was rapidly sorbed into the plastic matrix when the nasal solution has the highest pH value (absorption greatest at pH 10> pH 8 > pH 6). Figures 3.7.13.0, 3.7.13.1, and 3.7.13.2 demonstrate this phenomenon. At the higher pH values, more of the fentanyl molecules existed in the un-ionised form. Using Equation 2.5.3.2 the amount WO 2010/005400 PCT/SG2009/000248 - 62 of un-ionised fentanyl can be calculated. Table 3.7.13.0 below shows the percentage of unionised species (f,) for fentanyl molecule found in the fentanyl nasal solution at the various pH values. The non ionised species of fentanyl is highest when the pH value of the fentanyl nasal solution is 11.0 (99.7%) followed by pH 10.0 (97.4%), 8.0 (27.1%) and 5 least when the pH value is 6.0 (0.4%). Since it is the unionised species of fentanyl that is sorbed into the plastic matrix, therefore fentanyl would be expected to undergo a greater loss by absorption at the higher pH values. Table 3.7.13.0: The amount of unionised species (fu) of fentanyl (%) found in the 10 fentanyl nasal solution at the various pH values of 6, 8, 10 and 11. pH value fu (%) 11 99.7 10 97.4 8 27.1 6 0.4 The amount of uptake of fentanyl from the nasal solution into the plastic matrix during equilibrium condition, plotted against the nasal solution's pH value will result in a typical S 15 shape curve as shown in Figure 3.7.13.3. Both the rate and extent of fentanyl loss increased with increasing pH values because of the increasing unionised fentanyl species at higher pH values (Table 3.7.13.0 above). Therefore the pH value of the solution determines the amount of ionisation of the solutes. lonised species usually do not penetrate the plastic matrix. The uptake of fentanyl into the PVC dip-tube, EVA seal and 20 Acetal reservoir approaches an asymptotic value of 50, 50 and 45 pg/mL for PVC, EVA and Acetal respectively at the pH value of more than 10 and approaches zero at pH value of less than 6. The deviation of sorption data for those solutes that do not conform to matrix-controlled 25 sorption was due to the influence of interfacial or aqueous barriers and not the sorption of ionised species. There was a rapid loss of fentanyl from infusion solution when fluorouracil was added to the infusion solutions containing fentanyl in glucose 5% solution and also in sodium 30 chloride 0.9% solution in the PVC infusion bags 38 . There was a 25% loss of fentanyl within 15 minutes or a 50% loss within 60 minutes and a 70-100% loss within 24 hours in these admixture solutions at the storage temperatures of 4, 23 and 320C respectively. There was no loss of fentanyl when the fentanyl solution (without adding fluorouracil) in glucose 5% or in sodium chloride 0.9% infusion solution was stored in the PVC bags. There was 35 also no loss of fentanyl when fluorouracil was added to the fentanyl solution which was stored in polyethylene container. There was also a rapid loss of fentanyl into the PVC bag when fentanyl was added to glucose 5% infusion when the pH value of the glucose 5% infusion was adjusted to 9, using sodium hydroxide solution. The loss of fentanyl into PVC bags was therefore a pH dependent effect (since there was no loss of fentanyl into the WO 2010/005400 PCT/SG2009/000248 - 63 PVC bags when the solution was in a neutral or an acidic pH). The loss of fentanyl into the plastic matrix was also dependent on the type of plastic used (PVC>>polyethylene). The loss of fentanyl was the result of sorption of fentanyl into the PVC plastic. Sorption occurred as a result of the alkaline pH of the admixture and the loss of fentanyl may occur 5 whenever fentanyl was added to any drug that may sufficiently shift the pH of the solution to the alkaline region. The pKa value for fentanyl was 8.43. lonisation usually decreases its absorption into plastic material. Ions in aqueous solutions have an envelope of solvent around (solubilisation) them and thus have less of a tendency to escape from the solution hence decreasing its movement (sorption) into the plastic matrix. 10 3.7.14: Influence of the nasal solution pH on the diffusion coefficient, D of fentanyl into the plastic matrixes. Table 3.7.14.0 below shows the influence of solution pH value on the diffusion coefficient, 15 D of fentanyl into the three plastic components. As expected the higher the pH value of the nasal solution the greater is the diffusion coefficient of fentanyl. D for fentanyl in the PVC dip-tube is higher at the pH 8.0 and 10.0 compared to the value at pH 6.0. It is the un-ionised fentanyl that is highly absorbed into the plastic matrix. 20 Table 3.7.14.0: The influence of the nasal solution pH values on the diffusion coefficient, D of fentanyl into the 3 plastic components at the storage temperature of 37 0 C. D at pH 10.0 (cm 2 s- 1 ) D at pH 8.0 (cmzs) D at pH 6.0 (cmzs-) PVC dip-tube 1.2 X 10 7 1.45 X 10' 5 X 10_ EVA seal 1.2 X 10' 2.1 X 10- 7 Acetal reservoir 8.2 X 1010 6.1 X 10- 25 3.7.15: Influence of nasal solution pH on the partition coefficient, R of fentanyl into the plastic-water systems. The influence of the solution's pH value on the partition coefficient, R of fentanyl in the plastic-water system (log Rpiastic-water) decreases with decreasing pH. This is because at the 30 lower pH value there is less un-ionised fentanyl in solution. It is the un-ionised fentanyl species that undergoes partition between the aqueous layer and the plastic matrix. The ionised fentanyl species stay in the aqueous solution and do not undergo partitioning into the plastic matrix. Since only the unionised form of the solute reacts with plastic component of the container, the apparent partition coefficient (R,) of a solute at a given 35 pH value is related to the partition coefficient of the unionised portion Ru and the pK of the solute (see Equation 3.7.10.5 above) and can be calculated if R" and the pK value of the drug are known. 3.8: Sorption number (Sn). 40 WO 2010/005400 PCT/SG2009/000248 - 64 The sorption of a solute into a plastic matrix can be described by its sorption number 15 (S,) and has the unit per time (h). The sorption number provides a method to predict the influence of time, solution volume, plastic surface area and the solution pH on the loss of solute into a plastic material. The sorption number may also be correlated to the solute's 5 octanol-water partition coefficient. Hence the sorption parameter incorporates the surface area of the plastic, volume of solution, fraction of unionised solute, Rpiastic-water and diffusion coefficient, D in the plastic. 3.8.0: Sorption number, Sn and concentration effect 10 Sorption numbers can be calculated for fentanyl for the three plastic materials when the PVC dip-tube, EVA seal and Acetal reservoir were stored at 370C in the various initial concentrations of fentanyl nasal solutions (where the pH value was kept constant with phosphate buffer at 10) using the plot of [InFt + 1/(2F 2 ) -0.5] versus t (Equation 2.5.3.5). 15 The sorption number So was obtained from the slope of this linear graph. The sorption number curves are shown in Figures 3.8.0.0, 3.8.0.1 and 3.8.0.2. The sorption numbers S, obtained for fentanyl when the PVC dip-tube, EVA seal and Acetal reservoir were immersed in the fentanyl nasal solution, buffered with phosphate 20 buffer at pH 10.0 and stored at 370C are tabulated in Table 3.8.0.3 below. Table 3.8.0.3 shows that the sorption numbers obtained for the various initial fentanyl nasal solutions were approximately equal. Therefore there was no concentration effect on the sorption number. The average sorption number obtained for fentanyl in the PVC dip-tube, EVA seal and Acetal reservoir was 0.09 ± 0.04, 0.11 ± 0.05, 0.004 ± 0.001 h- respectively. 25 Having obtained the S, values for fentanyl, Equation 2.5.3.5 could then be used to obtain the shelf-life (t 9 o) of the fentanyl nasal solution when the nasal solution was stored at the pH value of 10.0 in PVC, EVA and Acetal plastic materials at the storage temperature of 370C. Therefore at the storage temperature of 370C and at the pH value of 10.0, the shelf 30 life (10% loss of fentanyl or Ft = 0.9) of fentanyl nasal solution stored in PVC, EVA and Acetal plastic was calculated to be 0.09, 0.08 and 2.40 hours respectively. Since the fentanyl nasal solution was in contact with these three plastic components of the nasal device, hence the shelf life of the fentanyl nasal solution stored at 370C and at the pH value of 10.0 is only 4.8 minutes (0.08 x 60 minutes). Hence at this storage temperature of 35 370C, fentanyl nasal solution formulated at pH value of 10.0 when stored in the device would lose its potency within minutes. Table 3.8.0.3: Sorption numbers (Sn) for fentanyl when the PVC dip-tube, EVA seal and Acetal reservoir was stored in fentanyl nasal solution buffered at pH 10.0, and 40 at the storage temperature of 37 0 C. Plastic material Fentanyl pg/mL S, h PVC diptube 20.9 0.08 PVC diptube 39.5 0.06 PVC diptube 87.1 0.09 WO 2010/005400 PCT/SG2009/000248 - 65 PVC diptube 113.0 0.15 Average - 0.09 Standard deviation - 0.04 EVA seal 20.9 0.10 EVA seal 39.5 0.05 EVA seal 87.1 0.13 EVA seal 113.0 0.14 Average - 0.11 Standard deviation - 0.05 Acetal Reservoir 20.9 0.003 Acetal Reservoir 39.5 0.004 Acetal Reservoir 87.1 0.004 Acetal Reservoir 113.0 0.004 Average - 0.004 Standard deviation - 0.001 3.8.1: Sorption number and effect of pH. 5 The shelf lives of the fentanyl solution in the various plastic components at the pH 10.0, 8.0 and 6.0 are tabulated in Table 3.8.1.1 below. Hence the fentanyl nasal solution was more stable at the lower pH value. At room temperature, the shelf lives of the fentanyl nasal solution at the pH 10.0, 8.0 and 6.0, stored in the PVC plastic were 0.36h, 4.5h and 2.4 years respectively. When stored at room temperature in EVA plastic, the shelf lives of 10 the fentanyl nasal solution at the pH 10.0, 8.0 and 6.0, were 1.1h, 13.4h and 7.7 years respectively. Table 3.8.1.0: Storage of PVC dip-tube and EVA seal in the fentanyl nasal solutions buffered at pH 10.0, 8.0 and 6.0 at room temperature. The f" 2 , fraction unionised is 15 obtained using Equation 2.5.3.2 and Sn the sorption number is obtained from the slope of the graph using Equation 2.5.3.5 pH values fU 2 S (h- 1 ) PVC dip-tube EVA seal 10.0 0.95 0.15 0.05 8.0 0.07 0.012 0.004 6.0 0.000016 0.0000025 0.0000008 There were no Sn values for Acetal reservoir because there were no sorption of fentanyl into the Acetal reservoir at pH values of 8.0 and 6.0. 20 Table 3.8.1.1: Shelf-life values of fentanyl nasal solution when PVC dip-tube and EVA seal is immersed in this nasal solution at pH 10.0, 8.0 and 6.0 (room temperature) predicted from Equation 2.5.3.5 WO 2010/005400 PCT/SG2009/000248 - 66 pH values S, (h-) Shelf-life t 9 o h _ PVC dip-tube EVA seal PVC dip-tube EVA seal 10.0 0.150 0.050 0.36 1.07 8.0 0.012 0.004 4.50 13.40 6.0 25.0 x 10' 8.0 x 10 21,450.0 67,031.0 There were no S, values for the Acetal reservoir because there was no sorption of fentanyl into the Acetal reservoir at pH 8.0 and 6.0. 3.8.2: Sorption Number and Temperature Differences: 5 The sorption number Sn for fentanyl at 250C in the PVC plastic at the pH value of 8.0 is 0.012 h. The activation energy for fentanyl in the PVC plastic is 26.9 kJ/mole. The sorption number for fentanyl at the storage temperature of 50C (refrigeration) could then be calculated from Equation 3.8.2.0. The calculated Sn value is 0.0005h 1 . Using the 10 Equation 2.5.3.5 the calculated shelf life for fentanyl nasal solution at a pH value of 8.0 stored in PVC plastic at the storage temperature of 5 0 C is 17.8 hours. Hence keeping this nasal solution in the refrigerator would have increased its shelf life to 17.8 hours as compared to 4.5 hours at room temperature. 15 The sorption number Sn for fentanyl at 25 0 C in the EVA plastic at the pH 8.0 is 0.004 h. The activation energy for fentanyl in the EVA plastic is 29.8kJ/mole (. The sorption number for fentanyl at the storage temperature of 5*C (refrigeration) could then be calculated from Equation 3.8.2.0. The calculated Sn value is 4.00 x 10- h. Using Equation 2.5.3.5 the calculated shelf life for fentanyl nasal solution at the pH value of 8.0 20 stored in the EVA plastic at the storage temperature of 50C is 12.4 days. Hence keeping this nasal solution in the refrigerator would have increased its shelf life to 12.4 days as compared to 13.4 hours (0.6 day) at room temperature. 3.8.3: Sorption number and volume of the container. 25 The uptake of solutes by plastic infusion bags is also dependent on the surface area of plastic-volume of solution ratio. Equation 2.5.3.4 states that Sn is directly related to the square of the reciprocal of the volume. The equation of best fit for the absorption of diazepam into the various sizes of the PVC plastic infusion bags was obtained; Sn = 0.018 30 + 0.0083 V- 2 (where V is the volume of the solution). This relationship only holds true if the plastic acts as an infinite sink. An infinite sink condition is when the concentration of the solute in the solution remains constant. The relation of Sn and V for fentanyl was not studied in this thesis. 35 3.9: Effect of solvents on sorption of solute into plastic matrix. Solvents can have an influence on the sorption of solutes into the plastic matrix. A solvent admixture with an aqueous solution may change the dielectric constant of the aqueous solution hence may decrease the charge effects in both the solutes and the plastic WO 2010/005400 PCT/SG2009/000248 -67 material resulting in an increased (or decreased) sorption 6. Sorption into plastic decreases as the concentration of water content is decreased in alcoholic solution of alcohol, propylene glycol, glycerin or polyethylene glycol 40039. Sorption was optimal when the alcohol concentration in aqueous solution was 10-15% and decreased as the 5 alcoholic content increases. 3.10: Effect of Plasticizer agent on Diffusion. Plasticizers are usually chemicals added to plastics to soften them. Plasticizers lower the 10 glass-transition temperature of the plastic hence have a softening effect on the plastic. Plasticizers may lower the viscosity of the polymer making the plastic material pliable and "mouldable" 4 . Plasticizer molecules are smaller than polymer molecules. In addition, some penetrant molecules can alter the structure of the plastic polymer during the "plasticizing" process. 15 3.11: Differences in the plastic materials and their properties. This study showed that there was a difference in the rate as well as the amount of fentanyl sorbed into the three different plastic materials of the device. There was more absorption 20 of fentanyl into the PVC dip-tube and the EVA seal than the Acetal or Polyformaldehyde reservoir. PVC plastic is by far the most widely used medical plastic materials. The properties of PVC plastic can be changed by adding additives (plasticizer) for examples, their versatility, their hardness and flexibility can be varied widely, their degree of inertness can be altered (example resistance to oil and aqueous fluids), their appearances 25 can be made clear, radio-opaque and their surface properties can be enhanced 4 1 . EVA plastic is a random copolymer of polyethylene and vinyl acetate which also has a broad range of applications (both pharmaceutical and industrial). The EVA copolymer is a clear to translucent rubbery material available in a range of hardness values. EVA polymers are chemical resistant, heat stable, pliable and flexible and have excellent processibility. 30 Hence EVA plastic is used in the nasal device as the seal to prevent leakage of the liquid from the nasal device. They are also used as non-degradable polymers for drug-delivery systems (commonly used as rate-controlling membranes in drug delivery devices). 3.11 Conclusion: in vitro studies 35 These experiments showed the different absorption profiles for the three plastic components of the nasal device. The absorption ability of the plastic material depended on the intrinsic properties of the plastic polymers as well as the penetrators. 40 In summary: . The aqueous solubility of fentanyl citrate at room temperature from the phase solubility method was found to be 10.64 ± 1.03 mg/mL at the pH 3.64. This was confirmed by using the osmometer which gave an osmolality reading of 40 mOsm and an aqueous solubility equivalent to 10.57 mg/mL of fentanyl.

WO 2010/005400 PCT/SG2009/000248 -68 * The spray particles produced by the nasal spray device, measured using the MSLI, showed that 96.4% had a size greater than 13 pm. The particles size range was ideal for depositing to the nasal cavity. This spray device consistently delivers 0.18 mL per spray. 5 0 The loss of fentanyl from the nasal solution was due to sorption into the plastic components. " Fentanyl nasal solution was stable when stored in amber silicaborate glass bottles. * The shelf life of the fentanyl nasal solution depended on the type of plastic material in contact with the nasal solution, the pH value of the nasal solution and the 10 storage temperature. * Polyvinyl Chloride plastic as well as Ethyl Vinyl Acetate plastic extracted and absorbed high quantities of fentanyl from the nasal solution at alkali pH whereas Acetal or Polyformaldehyde plastic absorbed much less fentanyl. " The loss of fentanyl from the nasal solution increased with an increase in pH value 15 of the nasal solution. This was in accordance to the theory that only the non ionised molecular species were attracted and absorbed into the plastic matrix. * At the pH 10, the shelf-life of the fentanyl nasal solution stored at 370C in the nasal spray device was extremely short (t 9 o = 5 minutes). * At the pH 8, the shelf-life of the fentanyl nasal solution stored at room temperature 20 was 4.5 hours whereas at the pH 6, the shelf-life of the fentanyl nasal solution stored at room temperature was 894 days (30 months). Hence a decrease of 2 pH values (from pH 8 to pH 6), resulted in an increase of the shelf-life of fentanyl nasal solution from 4.5 hours to 30 months. * The ionic strength of the nasal solution did not contribute to or interfere with the 25 sorption process of fentanyl into the plastic components of the nasal device. Since fentanyl was very stable at a pH 6, the fentanyl nasal solution was formulated with phosphate buffer to a pH 6 and used in the in vivo pharmacokinetic and bioavailability study. To avoid nasal irritation, the fentanyl nasal solution was adjusted to isotonicity 30 using sodium chloride. The pH of the nasal solution should be also formulated within the values of 4.5 to 6.5 to avoid nasal irritation. However at pH 6, fentanyl mostly existed as the ionised species. Theoretically, it is the non-ionic fentanyl molecules that are absorbed from the nasal mucosa into the blood stream. The fentanyl nasal solution at pH 8 (27% non-ionised) should have a higher bioavailability compared to a fentanyl solution with a pH 35 6. To test this hypothesis, two fentanyl nasal solutions (InpH6, buffered at pH 6 and inpH8, buffered at pH 8) were formulated for the next phase of this study, viz. the in vivo pharmacokinetics and bioavailability experiments. IN VIVO STUDIES 4.0 Pharmacokinetic Studies: 40 Previous studies have used a commercially available fentanyl injection solution (50 pg/ml). In conscious subjects, fentanyl solution has been instilled intranasally or a spray device used. The latter devices delivered approximately 0.1 mL (fentanyl 5 ptg) 42 . Volumes of WO 2010/005400 PCT/SG2009/000248 - 69 more than 0.15 mL per nostril are likely to drain into the pharynx, be swallowed 4 and detract from patient satisfaction 43 , however, this is not easily achievable when attempting to provide a pharmaceutically effective dose in a single application. A stable aqueous nasal formulation of 300 tg/ml (as fentanyl base) buffered with phosphate buffer to pH 5 values of either 6 or 8 and adjusted with sodium chloride to is tonicity to minimise nasal mucosal irritation was formulated for the in vivo study. This can be administered using a patient-controlled nasal spray bottle delivering an IN volume of approximately 0.18 mL and a dose close to 50 pg of fentanyl base. 10 The lipophilicity of a drug should influence the rate of absorption of drug through a biological membrane. The pKa value of fentanyl is 8.4 and in a nasal solution at pH value of 6, fentanyl citrate should exist essentially in the ionized form. At a pH of 8 a proportion of the fentanyl is in the non-ionized form (fentanyl base). It was postulated that the non ionized (lipophilic) form would rapidly penetrate the nasal mucosal membrane, leading to 15 earlier peak blood fentanyl concentrations and better bioavailability. Consequently, two nasal formulations were evaluated. The aim of this study was to investigate the pharmacokinetic profile of IN fentanyl citrate in two different formulations and to establish the bioavailability of the IN route. A secondary 20 aim was to evaluate the clinical characteristics and acceptability. 4.1 Methodology for Pharmacokinetics study: 4.1.0 Subjects and Methods: 25 4.1.1 Ethical approval: Curtin University Ethics committee approved the study in human and also this trial was registered with the Australian Therapeutic Goods Administration under the Clinical Trial Notification scheme. 30 4.1.2 Subjects: Twenty-four opioid-naive adult female patients requiring gynaecological surgery gave written informed consent. All patients were free of nasopharyngeal disease, of 50-90 kg 35 body weight and consented to general anaesthesia and were allowed to be given intravenous morphine per-operatively as the sole opioid, followed by patient-controlled intravenous morphine on demand, postoperatively for continuous analgesia. IV or IN fentanyl was administered for breakthrough pain. 40 4.1.3 Study Plan: A dedicated intravenous cannula was placed in the forearm intra-operatively for subsequent venous blood sampling. Postoperative analgesia was provided by patient controlled analgesia with intravenous morphine. Parenteral morphine was the only opioid WO 2010/005400 PCT/SG2009/000248 - 70 permitted pre- or post-operatively. Patients were visited after they had recovered from general anaesthesia, within the first 24 hours after surgery. Initiation of the study coincided with a time at which the patient was experiencing pain and was planning to activate the patient-controlled analgesia pump. 5 Each patient received the equivalent of fentanyl citrate 50 tg by both IN and IV routes, with the order of administration randomized according to a Latin cross-over square design and sealed opaque envelope allocation. To prevent a significant "carry over" effect, a period of at least 2 hours was allowed to elapse before fentanyl administration via the 10 second route. This approximated four half-lives. Two formulations of nasal fentanyl spray of different pH values were evaluated, with the first group of 10 patients allocated the nasal solution of the pH 6 (group INpH6) and the second group of 14 patients the solution of the pH 8 (group INpH8). 15 Each IN administration was by means of a patient-controlled spray bottle delivering 0.18 mL per activation (approximately 50 ptg of fentanyl). This bottle was stored securely in an upright position between uses and was primed three times by an investigator using a single press at least 5 minutes apart, (between each priming) before patient administration. Venous samples (5ml) were collected in sterile polypropylene tubes at 2, 20 5, 10, 15, 20 and 25 minutes after fentanyl administration. To guarantee adequate pain relief clinically, the self-administration of further doses of morphine was permitted at any time during the duration of the trial. After collection the venous samples (5 mL each) were allowed to clot and then 25 immediately centrifuged at 2,000 g for 5 minutes and the separated serum stored in the freezer at -70 0 C until analysis. 4.2 Pharmacokinetic (PK) analysis: Pharmacokinetic data were obtained by plotting serum concentration (ng/ml) versus time. 30 The raw pharmacokinetic data were analysed using two methods. The model independent method used the trapezoidal rule to calculate the area under the curve (AUCo-2 5 ) from time 0 to 25 minutes and the remainder (AUC 25 s) from time 25 minutes to infinity using the log trapezoidal rule. 35 Clearance (Cl) was calculated by dividing the IV dose (50,000 ng) with AUCo 0 (CI = Dosei,/AUCo-). The bioavailability (f) of the intranasal formulation was calculated by dividing AUCN with AUCiv (AUCo-a(IN)/AUCo-atiV)), since both the IN and IV doses were equal (50 ptg). 40 For the compartmental model kinetic, variables were calculated by standard equations after the serum concentration/time relationships were analysed by non-linear least-square analysis for a biexponential equation: Ct = Ae-lt + Be t. Volumes of distribution were calculated from these parameters. V, = Dose/(Xl+k2) and Vd = ClI/2. The elimination half- WO 2010/005400 PCT/SG2009/000248 - 71 life tj 1 2 was calculated by dividing X 2 into 0.693 (0.693 2 ) or by multiplying 0.693 with Vd and dividing by clearance (0.693 Vd/CI). For the intranasal study, elimination rate constant kel was obtained from the last 3 points 5 (15, 20 and 25 minute samples) of the serum fentanyl concentration versus time curve (IV results confirmed this elimination phase). A graphic procedure (method of residuals) is use to plot the absorption phase 44 . The observed serum fentanyl concentration for 2, 5 and 10 minutes were subtracted from the corresponding calculated concentrations (using the equation obtained from the log linear plot of the elimination phase) giving the residual 10 serum fentanyl concentrations. A straight-line graph from the semi-logarithmic plot of these residual concentrations versus time denoted the first order absorption process and from the slope of the graph the absorption rate constant (kabs) was determined. The analgesic range for plasma fentanyl was assumed to be 0.2-2 ng/m1 45 . 15 4.3 Statistical Analysis: The maximum serum fentanyl concentration Cmax, the mean bioavailability and patients' age differences between the two intranasal groups were compared for statistical differences using a Mann-Whitney U test. Allowing for a 20% dropout rate, a sample size of at least 20 patients provided 80% power at the 5% level of significance to detect a 25% 20 difference in bioavailability between the pH formulation groups. 4.4 Pain Score Analysis: The measurement of pain intensity included pain scores and summed pain intensity differences (SPID). The latter represent the area under the curve of the pain intensity 25 difference measurements. The normality of the pain measurements, the baseline scores and the SPID were tested using the Smirnov-Kolmogorov test and the SPID scores log transformed to achieve normality. The analysis of baseline pain scores and SPID was carried out using the ANOVA that accounted for the nested crossover design of the study. No transformations to achieve normality were found for side effect data, so a logistic 30 regression model comparing symptom severity at each time point relative to baseline was used. Analysis was performed using SAS and S-Plus software, using two-sided tests and a p-value of 0.05 as statistically significant. 4.5. Experimental 35 4.5.0 Instrumentation (Gas Chromatography with Mass Spectrometer (GC-MS) with Selected Ion Monitoring for analysing serum fentanyl) To measure the serum or plasma concentrations of fentanyl, the method used was gas chromatography-mass spectrometry (GC-MS). This method is sensitive and selective and achieved the sub-nanogram sensitivity required for PK studies. 40 4.5.1 Materials and Reagents: WO 2010/005400 PCT/SG2009/000248 - 72 Fentanyl citrate was purchased from sigma chemicals (analytical grade). Sufentanil citrate injection 50 pg/mL was the commercially available injection. All solvents were analytical (AR) grade (This method was chosen because it involved only a one step extraction. The extraction efficiency was consistently above 50% (see Recovery section below). To 5 minimise loss of fentanyl during the extraction from the 1 mL serum, all glassware used was siliconised. High purity n-butyl chloride was used. A high concentration of sodium hydroxide (4 M) was used to render alkaline the serum samples to ensure fentanyl was in the base form (lipophilic) and also to minimise extraction of organic contaminants for example cholesterol. 10 4.6 Results 4.6.0 Validation for the Bioanalytical Method: The bioanalytical method was validated for specificity (selectivity), linearity, precision and accuracy, sensitivity and the recovery from this method of analysis are given below. 15 4.6.1 Specificity (selectivity): This method of analysing serum fentanyl was specific and could differentiate between other endogenous compounds. The chromatograms for fentanyl and for sufentanil (internal standard) are shown in Figure 4.6.1.0. Figure 4.6.1.2 shows three different sources of serum, spiked with fentanyl and sufentanil from three different volunteers and 20 are free from interference of endogenous peaks at the retention time of fentanyl and sufentanil. 4.6.2 Linearity Figure 4.6.2.0 shows a linear relationship for the standard curve, concentration of serum fentanyl (spiked blank serum with known concentrations of fentanyl) against height of the 25 resultant peaks of fentanyl divided by sufentanil internal standard. The concentrations of fentanyl studied included the low and the high points and covers the entire range of interest. 4.6.3 Precision and Accuracy: The accuracy of this GC-MS analytical method is defined as the closeness of the 30 determined concentration (average values of 6 samples run) to the true concentration of the spiked serum samples (that is the error or percentage error of the observed value from the true value). The mean value for the 4 concentrations analysed was less than 14% of the actual concentration. The precision of an analytical method is the closeness of replicate determinations of an analyte by an assay or the differences among the individual 35 measurements on a set of measurements (same quantity) measured at different times expressed in range or average deviation or the relative standard deviation (coefficient of variation). The precision around the mean for the 4 concentrations analysed was less than 19.9% coefficient of variation (less than 12.9% for the 3 concentrations other than the LOQ). These values are acceptable where the precision of the mean concentration should WO 2010/005400 PCT/SG2009/000248 - 73 not exceed a coefficient of variation of 15% except for the LOQ which should be no more than 20%. 4.6.4 Sensitivity: The minimum detectable quantity or the limit of detection (LOD) is defined as the sample 5 that generates a detector response equal to or greater than three times the background noise of the system or the lowest concentration of an analyte that the analytical process can reliably differentiate from background levels. The LOD for this GC-MS method was 20 pg/mL. Five samples of serum spiked with 20 pg/mL of fentanyl were extracted and analysed for fentanyl concentration. Only three 10 samples gave a detectable peak but no integration results. The signal to background noise was 5:1. The minimum quantifiable limit or limit of quantification (LOQ) is the concentration at which the assay gives an acceptable precision or the lowest concentration of an analyte that can be measured with a stated level of confidence. The LOQ for this GC-MS assay 15 was 50 pg/mL. 4.6.5 Recovery: Recovery can be low provided the method is reproducible and the appropriate standard curve is used for the calculation of the unknown sample 44 . The recovery is consistently 20 above 50% in this study. 4.6.6 Results of serum analysis: 4.6.6.0 Demographic: Twenty-four patients were randomized (10 to receive formulation pH 6 and 14 formulation 25 pH 8) of which 19 had data analysed. Two patients from group INpH 6 and 3 from group INpH 8 were excluded because they requested withdrawal (n=3) or because samples were unsuitable for assay (haemolysis of blood samples, n=2). One sample after IV administration was excluded after it was found that assay samples had been taken from the same intravenous cannula as had been used to inject fentanyl solution. 30 The median age of patients from group INpH 6 was lower (Table 4.6.6.0), but not significantly when compared with group INpH 8 (p=0.12). Table 4.6.6.0: Characteristics of each patient group INpH 6 INpH 8 No. of patients 10 14 WO 2010/005400 PCT/SG2009/000248 - 74 No. of patients with no results* 2 3 n* 8 11 Average age (years) 37 47 Age range (years) 27 to 47 33 to 63 Average weight (kg) 67 67 Weight range (kg) 50 to 90 50 to 89 Fentanyl dose (micrograms) 50 50 Average dose of fentanyl (microgram/kg) 0.75 0.75 * No serum fentanyl concentration results from 5 patients due to haemolysis of the specimen or patient withdrawal from the study. Tables 4.6.6.1.2, 4.6.6.1.3 and tabulate the mean serum fentanyl concentrations for all the 5 patients in the two IN groups and also the combined IV results for the two groups respectively. Table 4.6.6.1.2: Mean serum fentanyl concentrations after 50 pg of nasal fentanyl (1 spray) at pH 6.0 for 8 patients (n=8) from group INpH 6, including the Standard Error 10 of the Mean (SEM) with the 95% confidence intervals (ci): Patients Serum fentanyl concentrations ng/mL 2 mins 5 mins 10 mins 15 mins 20 mins 25 mins Mean 0.20 0.32 0.33 0.3 0.21 0.17 Std Dev 0.1 0.19 0.18 0.19 0.09 0.07 SEM 0.04 0.07 0.07 0.07 0.03 0.2 95% ci 0.07 0.13 0.13 0.13 0.06 0.05 LCL* 0.13 0.19 0.2 0.17 0.15 0.13 UCL** 0.27 0.45 0.45 0.43 0.27 0.22 Two Patients withdrew from the study. *LCL = lower confidence level. **UCL = upper confidence level.

WO 2010/005400 PCT/SG2009/000248 - 75 From the mean serum fentanyl concentration the calculated areas under the curve (AUC) were: AUCo-2 5 = 6.4ngmLimin and AUCo. = 9.4ngmLimin. Table 4.6.6.1.3: Mean serum fentanyl concentrations after 50 pg of fentanyl (1 spray) at pH 8.0 for 11 patients (n=11) from group INpH 8, including the Standard Error of 5 the Mean (SEM) with the 95% confidence intervals (ci): Patients Serum fentanyl concentrations ng/mL 2 mins 5 mins 10 mins 15 mins 20 mins 25 mins Mean 0.31 0.37 0.34 0.31 0.24 0.19 Std Dev 0.27 0.26 0.21 0.17 0.07 0.05 SEM 0.08 0.08 0.06 0.05 0.02 0.02 95% ci 0.16 0.15 0.12 0.1 0.04 0.03 LCL* 0.15 0.22 0.22 0.21 0.2 0.16 UCL** 0.47 0.52 0.46 0.41 0.28 0.22 LCL = lower confidence level. **UCL = upper confidence level. From the mean serum fentanyl concentration the calculated area under the curve (AUC) are: AUCo-2 5 = 7.1ngmLimin and AUCo- = 1lngmLimin. 10 Table 4.6.6.1.4: Mean serum fentanyl concentrations after 50 pg of fentanyl (1 mL IV bolus) for the total of 18 patients (combination of patients groups INpH 6 and INpH 8; n=18), including SEM and 95% confidence intervals: Patients Serum fentanyl concentrations ng/mL 2 mins 5 mins 10 mins 15 mins 20 mins 25 mins Mean 1.65 1.08 0.73 0.45 0.33 0.29 Std Dev 0.76 0.52 0.36 0.19 0.14 0.14 SEM 0.18 0.12 0.08 0.04 0.03 0.03 95% ci 0.35 0.24 0.17 0.09 0.06 0.06 LCL* 1.3 0.84 0.56 0.36 0.27 0.23 UCL** 2 1.32 0.9 0.54 0.39 0.35 LCL = lower confidence evel. **UCL= upper confidence level From the mean serum fentanyl concentration the calculated area under the curve (AUC) 15 are: AUCO-2 5 = 16.9ngmLi'min and AUCo, = 21ngmLlmin. The biexponential curve is C = 0.939e- 0

-

4 t + 0.9e-0 04 5t Co = 1.84ng/mL; Ve = dose/Co = 27.2L; keiim = 0.045 min-; Cl = Dosev/AUCI, = 2.4Lmin 1 ; Vd= Cl/keiim = 53.3 L.

WO 2010/005400 PCT/SG2009/000248 - 76 Two patients withdrew from the study. Blood samples from 1 patient were all haemolysed and therefore could not be analysed. Serum fentanyl concentrations from three subjects were too high. This was because the same cannula was used to give IV bolus injection and to collect blood samples. 5 4.6.6.2 Pharmacokinetic Results: The pharmacokinetic data for fentanyl following the IV bolus (Table 4.6.6.2.0) followed a bi-exponential decay (2-compartment model, bi-exponential model: Ct = Ae-Xt + Be- 2 t). The bi-exponential equation obtained from these data (Table 4.6.6.2.0) is C = 1.22e- .22t + 10 1.1Oe- '- 5 t. Table 4.6.6.2.0: Pharmacokinetic values after intravenous fentanyl (50 pg). n=18* A X B X t1/ 2 e CI Vc Vd AUCO-2 5 AUCo ng/mI /min ng/mI /min min I/min L L ng/ml/min ng/ml/min Mean 1.22 0.22 1.10 0.05 20.8 2.4 27 59.6 17.2 27.4 SD 0.89 0.09 0.71 0.03 17.5 1.2 14.4 45.4 7.3 12.5 CI 0.41 0.04 0.33 0.01 8.1 0.6 6.6 21 3.4 5.8 LCL 0.69 0.18 0.77 0.04 12.7 1.8 20.4 38.6 13.8 21.6 HCL 1.63 0.26 1.43 0.07 28.9 3.0 53.3 80.6 20.6 33.2 SEM 0.21 0.02 0.17 0.01 4.1 0.3 3.4 10.7 1.7 3 A and B = coefficients with units of concentration ng/mL; k, and k 2 = exponential 15 coefficients, min-; t1/2e = elimination half life; CI = clearance; Vc = volume of central compartment; Vd = volume of distribution; AUC = area under the curve; SD = standard deviation, Cl = 95% confidence interval, LCL = lower confidence limit; HCL = higher confidence limit, SEM = standard error of the mean. *One IV result was excluded from the analysis because the serum fentanyl concentrations were unreliable (samples taken from 20 the same cannula through which fentanyl had been injected). The clearance, volume of distribution, elimination rate constant (keiim), half-lives (t 1

/

2 ), C. after IV administration were shown in Table 5.6.6.2.1, Table 5.6.6.2.2. Table 4.6.6.2.1: Clearance (Cl), volume of the central compartment (Ve), volume of distribution (Vd) and the elimination half-Lives (t 1

/

2 ) for the combined groups of 25 InpH 6 and INpH 8 given intravenous bolus of 50 pg (1 mL) of fentanyl (n=18): WO 2010/005400 PCT/SG2009/000248 - 77 Patients Cl Lmin' Vr Litres Vd Litres kenim min t 1

/

2 min Average 2.4 27.0 59.6 0.05 20.8 Std Dev 1.2 14.4 45.4 0.03 17.5 SEM 0.3 3.4 10.7 0.01 4.1 95% ci 0.6 6.6 21 0.01 8.1 LCL 1.8 20.4 38.6 0.04 12.7 UCL 3.0 33.3 80.6 0.07 28.9 Median 2.1 19.8 51.5 0.05 13.9 25th 1.38 15.7 33.6 0.03 10.5 percentile 75 t 3.18 40.2 68.1 0.06 26 percentile Cl = Dose/AUC. Ve = Dose/Serum concentration at time zero. Vd= CI/keim ti/ 2 = 0.6 93 kelim. 5 Table 4.6.6.2.2: Bi-exponential model for the pharmacokinetics for the combined groups of lnpH 6 and INpH 8 given intravenous bolus of 50 pg(1 mL) of fentanyl (n=1 8): Patients

C

1 ng/mL C 2

C

0 ng/mL ng/mL min-' min Average 1.22 1.10 2.33 0.22 0.05 Std Dev 0.89 0.71 1.02 0.09 0.03 SEM 0.21 0.17 0.24 0.02 0.01 95% ci 0.41 0.33 0.47 0.04 0.01 LCL 0.69 0.77 1.86 0.18 0.04 UCL 1.63 1.43 2.80 0.26 0.07 Median 1.12 0.91 2.53 0.21 0.05 2 5 th 0.38 0.63 1.25 0.13 0.03 percentile 7 5 th 1.96 1.28 3.14 0.25 0.06 percentile Co= C1 + C2 10 The mean total area under the curve (AUC..25mins + AUC25mins-) for IV and IN administration of the pH 6.0 and pH 8.0 formulations was 27.4 (IV) and 13.3 and 14.9 (IN) ng/ml/min respectively. The mean bioavailability (f) for the two IN formulations was 0.49 (49% absorbed, pH 6.0) and 0.54 (54% absorbed, pH 8.0) respectively (Tables 4.6.6.2.3, 4.6.6.2.4 and 4.6.6.2.5). 15 WO 2010/005400 PCT/SG2009/000248 - 78 Table 4.6.6.2.3: Area under the curve (AUC), bioavailability (F), maximum serum concentration (Cmax) and time to maximum concentration (tmax) for the 8 patients (n=8) from group INpH 6 given 50 pg(1 spray) of fentanyl intranasally: Patients AUCo.

2 s ngmL~tmin AUCo.. F Cmax tmax min ngmL-min ngmL-1 Mean 6.4 13.3 0.62 0.45 7.1 Std Dev 2.4 6.3 0.38 0.22 4.2 SEM 0.8 2.2 0.14 0.08 1.5 95% ci 1.7 4.3 0.26 0.15 2.9 LCL 4.7 8.9 0.36 0.3 4.2 UCL 8.0 17.6 0.98 0.61 10 5 Table 4.6.6.2.4: Area under the curve (AUC), bioavailability (F), maximum serum concentration (Cmax) and time to maximum serum concentration (tmax) for the 11 patients (n=1 1) from group INpH 8 given 50 pg (1 spray) of fentanyl intranasally: Patients AUCO.

2 5 ngmL~4min AUCo. F Cmax ngmL tmax min ngmL-min1 Mean 6.9 14.9 0.51 0.42 8.1 Std Dev 3.8 6.6 0.25 0.24 5.6 SEM 1.1 2.0 0.08 0.07 1.7 95% ci 2.2 3.9 0.16 0.14 3.3 LCL 4.7 11 0.35 0.28 4.8 UCL 9.1 18.8 0.67 0.57 11.4 10 Table 4.6.6.2.5: Area under the curve (AUC) for the combined groups of inpH 6 and INpH 8 given intravenous bolus of 50 pg (1 mL) of fentanyl (n=18): Patients AUCO.

2 5 ngmL-1min L

AUC

0 .. ngmf' min Mean 17.2 27.4 Std Dev 7.3 12.5 SEM 1.7 3.0 95% ci 3.4 5.8 LCL 13.8 21.6 UCL 20.6 33.2 WO 2010/005400 PCT/SG2009/000248 -79 The absorption of fentanyl (nasal formulation INpH 6) from the nasal mucosa is a first order process confirmed by a linear residual plot of the absorption phase (Figure 4.6.6.2.0). By the method of residual, the absorption begins when the extrapolated curve Y = -0.0568t - 0.3756 (elimination slope = kei = 0.06 min-) and the residual curve Y = 5 0.243t - 0.4026 (absorption slope = kab = 0.243 min-) intersect (that is at -0.145 minutes). Therefore there was no lag time and the nasal absorption of fentanyl was an instantaneous absorption. Also at the nasal formulation of INpH 8, the absorption of fentanyl from the nasal mucosa 10 was also a first-order process confirmed by a linear residual plot of the absorption phase (Figure 4.6.6.2.1). By the method of residual, the absorption begins when the extrapolated curve Y = -0.0405t - 0.5003 (elimination slope = kei = 0.041 min-) and the residual curve Y = -0.211 It - 1.0203 (absorption slope = kab = 0.211 min') intersect (that is at -3 minutes). Therefore there was no lag time and the nasal absorption of fentanyl was an 15 instantaneous absorption. The absorption and elimination rate constants (kab and kelim) and the half-life values after IN administration are shown in Tables 4.6.6.2.6, 4.6.6.2.7, 4.6.6.2.8 and 4.6.6.2.9. 20 Table 4.6.6.2.6: Pharmacokinetic values after intranasal fentanyl (pH 6.0) 50 pg. n=8 kab /min kei /min t 1 /2e] min AUCO.

2 5 AUCo. ng/ml/min ng/ml/min Mean 0.26 0.04 29.8 6.4 13.3 SD 0.14 0.03 25.8 2.4 6.3 SEM 0.05 0.01 9.1 0.8 2.2 Cl 0.1 0.02 17.9 1.7 4.3 LCL 0.16 0.02 11.9 4.7 8.9 HCL 0.36 0.07 47.6 8 17.6 kab /min = absorption rate constant; kei /min = elimination rate constant; tl/2el = elimination half life; AUC = area under the curve; SD = standard deviation, Cl = 95% confidence interval, LCL = lower confidence limit; HCL = upper confidence limit, SEM = standard error of the mean. 25 WO 2010/005400 PCT/SG2009/000248 - 80 Table 4.6.6.2.7: Pharmacokinetic values after intranasal fentanyl (pH 8.0) 50 pg. n=1 I kab /min kei /min tl/2e min AUCO-2 5 AUCa ng/mLmin ng/mLmin Mean 0.22 0.04 31.2 6.9 14.9 SD 0.12 0.04 25.6 3.8 6.6 SEM 0.04 0.01 7.7 1.1 2 CI 0.07 0.02 15.2 2.2 3.9 LCL 0.15 0.02 16.1 4.7 11 HCL 0.29 0.07 46.4 9.1 18.8 kab /min = absorption rate constant; kei /min = elimination rate constant; t1/2e = elimination 5 half life; AUC = area under the curve; SD = standard deviation, Cl = 95% confidence interval, LCL = lower confidence limit; HCL = upper confidence limit, SEM = standard error of the mean Table 4.6.6.2.8: Absorption rate constant (kab), elimination rate constant (kei), 10 elimination half-lives (ti1 2 ei) and time to absorption or lag phase (tiag) for the patients (n=8) from group INpH 6 given 50 pg (1 spray) of fentanyl intranasally. Patients kab min" kei min- tI/2eI min tiag min Average 0.26 0.04 29.8 -0.3 Std Dev 0.14 0.03 25.8 2.1 SEM 0.05 0.01 9.1 0.7 95% ci 0.1 0.02 17.9 1.4 LCL 0.16 0.02 11.9 -1.7 UCL 0.36 0.07 47.6 1.1 WO 2010/005400 PCT/SG2009/000248 - 81 Table 4.6.6.2.9: Absorption rate constant (kab), elimination rate constant (kei), elimination half-lives (t1/ 2 ei) and time to absorption or lag phase (tiag) for the patients (n=11) from group INpH 8 given 50 pg (1 spray) of fentanyl intranasally. Patients kab mi 1 kei min-' tl2eI min tlg min Average 0.22 0.04 31.2 -2.9 Std Dev 0.12 0.04 25.6 5.4 SEM 0.04 0.01 7.7 1.6 95% ci 0.07 0.02 15.2 3.2 LCL 0.15 0.02 16.1 -6.2 UCL 0.29 0.07 46.4 0.3 5 4.6.7 Statistical Analytical Results: Mann-Whitney U test is the non-parametric equivalent of a student's t test for independent groups. The appropriate descriptive statistics to report with this test are medians (mdn) (instead of means) and interquartile ranges (IQR) (instead of standard deviations). 10 There was no significant difference between the bioavailability of these two IN groups (INpH 6: mdn = 55%, IQR = 52.3%; INpH 8: mdn = 71%, IQR = 63%; p = 0.545, Mann Whitney U test). The serum fentanyl concentration versus time curves for the IN formulations are shown in 15 Figure 4.6.7.0. The median peak serum concentrations (Cmax) were 0.36 ng/ml (IQR 0.28 ng/ml) in group INpH 8 and 0.38 ng/ml (IQR 0.41 ng/ml) in group INpH 6 (no significant difference in the 2 peaks, p=0.778, Mann-Whitney U test). There was no carry-over or "order effect" for the pharmacokinetics data whether IN was 20 given first followed by IV administration or IV was given first followed by IN administration. The Mann-Whitney U test showed no statistically significant results (p = 0.486 for AUC INpH 6; p = 0.177 for AUC INpH 8; p = 0.486 Cmax INpH 6; p = 0.537 INpH 8). Tables 4.6.7.0, 5.6.7.1, 4.6.7.2, 5.6.7.3 and 4.6.7.4 report the Mann-Whitney U test data 25 regarding any "order effect" i.e. any carry over effect during the cross-over study for AUC 0 .. .and Cmax and Co when IN spray was given first or when IV bolus was given first. Table 4.6.7.0: AUCo.c , for group INpH 6. Patients Order AUCo.. ngmLmin (IN) AUCo. ngmLAmin (IN) (IV) I IN/IV 10.3 35.2 2 IN/IV 6.7 22.1 3 IN/IV 24 23.2 WO 2010/005400 PCT/SG2009/000248 -82 4 IN/IV 11.2 27.3 AUCO... ngmL-'min AUCo..ngmL-4min (IV) (IN) 5 IV/IN 14.7 13.2 6 IV/IN 16.1 11.9 7 IV/IN 15.9 21.4 8 IV/IN 41.4 7.3 Median 42 32.7 IQR 15.6 25.4 Result of the Mann-Whitney U tests: U = 5, Z = -0.87, p = 0.486 (not statistically significant) Table 4.6.7.1: AUCo.. for group INpH 8. Patients Order AUCo.. ngmL'min AUC 0 -, ngmL'min (IN) (IV) 1 IN/IV 7.2 30.9 2 IN/IV 26 13.5 3 IN/IV 11 25.6 4 IN/IV 16.3 5 IN/IV 5.3 24.5 AUCo., ngmL- min AUCo.. ngmL- 1 min (IN) (IV) 6 IV/IN 8.9 8.7 7 IV/IN 22.4 11.4 8 IV/IN 38.4 19.2 9 IV/IN 44.3 22.2 10 IV/IN 52.3 14.9 11 IV/IN 24.5 21.9 Median 36.6 52 IQR 15.8 36.9 5 Result of the Mann-Whitney U test: U = 7, Z = -1.461, p 0.177 (not statistically significant) Table 4.6.7.2: Cmax for group INpH 6. Patients Order Cmax ngmL- (IN) Co ngmL- 1 (IV) 1 IN/IV 0.64 3.16 2 IN/IV 0.39 3.36 3 IN/IV 0.36 0.8 4 IN/IV 0.16 2.96 WO 2010/005400 PCT/SG2009/000248 - 83 C. ngmL 1 (IV) Cmax ngmL' (IN) 5 IV/IN 1.12 0.72 6 IV/IN 1.5 0.32 7 IV/IN 1.26 0.75 8 IV/IN 2.93 0.28 Median 3.46 1.92 IQR 2.31 1.08 Result of Mann-Whitney U test: U = 5, Z = -0.87, p = 0.486 (not statistically significant). Table 4.6.7.3: Cmax for group INpH 8. Patients Order Cmax ngmLi' (IN) C. ngmL 1 (IV) 1 IN/IV 0.36 3.13 2 IN/IV 0.36 2.01 3 IN/IV 0.65 3.43 4 IN/IV 0.43 5 IN/IV 0.14 3.13 C. ngmL 1 (IV) Cmax ngmL 1 (IN) 6 IV/IN 1.23 0.23 7 IV/IN 2.45 0.37 8 IV/IN 2.98 1.02 9 IV/IN 1.19 0.51 10 IV/N 2.6 0.36 11 IV/IN 1.36 0.22 Median 3.27 2.26 IQC 2.38 1.67 Result of Mann-Whitney U test: U = 11, Z = -0.73, p = 0.537 (not statistically significant). 5 Table 4.6.7.4: Mann-Whitney U test to compare the 2 groups (INpH 6 and INpH 8) for statistical significant. Patients F Cmax ngmLA 1 0.98 0.23 2 0.25 0.36 3 0.52 0.37 4 0.22 0.14 5 0.5 1.02 6 0.5 0.51 7 0.5 0.36 8 0.43 0.65 9 - 0.43 WO 2010/005400 PCT/SG2009/000248 - 84 10 0.3 0.36 11 0.89 0.22 Median 0.71 0.36 IQC 0.63 0.28 Patients F Cmax ngmLA 11 0.9 0.72 12 0.29 0.64 13 0.74 0.32 14 0.3 0.39 15 1.03 0.36 16 1.14 0.75 17 0.1 0.28 18 0.41 0.16 Median 0.55 0.38 IQR 0.52 0.41 Comparing Cmax values between the 2 groups, a Mann-Whitney U test gave U = 40.5, z = 0.29, p = 0.778 (not statistically significant). 5 Comparing the bioavailability (F) between the 2 groups, a Mann-Whitney U test gave U = 36.5, z = -0.62, p = 0.545 (not statistically significant). 4.6.8 Pain Score Results: One patient was recruited who did not require any postoperative opioid analgesia. 10 Twenty-three patients had data analysed, although only 21 had complete data for both routes of administration (Table 4.6.8.0). For both pH formulations of nasal fentanyl the rate of fentanyl absorption (therapeutic serum levels of more than 0.2 ng/ml were achieved within 2 min), the peak serum concentration (0.33 and 0.37 ng/ml for IN fentanyl pH 6.0 and pH 8.0 respectively), the bioavailability and the elimination half-lives were 15 similar 46 . Thus the clinical data for IN administration are represented using data from both formulations combined. Baseline visual analogue pain scores at rest and with movement were similar (mean score with movement 55, 95% CI 45-65 and 57, 95% Cl 47-67 for IV and IN routes respectively, 20 p = 0.78). Rest and movement pain scores were similar 25 minutes after administration (Table 4.6.8.1). ANOVA analysis of log-transformed summed pain intensity differences for both rest and movement pain showed no significant difference between IN and IV administration (for movement pain p = 0.72, Figure 4.6.8.0). The estimated mean summed pain intensity difference was 518, 95% Cl 365-737 versus 563, 95% CI 399-794 for the IN 25 and IV routes respectively. During the 60 minutes after fentanyl administration, 7 patients following IN administration and 6 following IV administration used supplementary patient controlled morphine (dose range 1-3 mg and 3-33 mg respectively) (Table 4.6.8.0).

WO 2010/005400 PCT/SG2009/000248 - 85 Scores for sedation, nausea or pruritus, or for change in oxygen saturation were not significantly different after IN and IV administration and are shown in Table 4.6.8.2. The odds ratio for an increase in sedation from IN administration compared with IV was 1.27 5 (95% Cl 0.57-2.86). The increase in intensity of pruritus did not differ between routes (p = 0.125). There was no order effect for any pain or side effect outcome. Four patients (17%) reported mild stinging in the nose (all received the formulation of pH 6) and 3 (13%) a bitter taste after IN administration. Seven (29%) preferred the IN route, 10 (42%) the IV route and 6 (25%) had no preference. 10 Table 4.6.8.0: Patient Characteristics (n = 23). Age (y) 42(9.0) Weight (kg) 65(12.7) Type of surgery - open abdominal 6 (26%) - major vaginal 17 (74%)* Sequence of fentanyl administration -IN:IV 11 (48%) -IV:IN 12 (52%) Patients using additional morphine during the 60 min after fentanyl administration (n) - IN 7(30%) - IV 6 (26%) IV morphine dose during the 60 min after fentanyl administration (mg) - IN 2 (2,12) IV 1 (1,3) * Data unavailable from one patient. Values are mean and standard deviation except morphine dose (median and interquartile range). Table 4.6.8.1: Pain scores combined for both cross-over sequences (n = 23). Pain IV fentanyl IN fentanyl WO 2010/005400 PCT/SG2009/000248 - 86 At rest: Baseline 48 (25,64) 37 (20,52) 5 min 9(0,20) 18 (6,25) 15 min 7(0,12) 5(0,24) 25 min 5(0,31) 7(0,24) With movement: Baseline 51 (36,80) 64 (40,73) 5 min 20 (9,37) 25(14,41) 15 min 14 (4,35) 15 (8,34) 25 min 16 (7,31) 20 (7,36) Visual analogue scores 0-100. Values are median and interquartile range. Table 4.6.8.2: Scores for sedation, nausea and pruritus; oxygen saturation; and number of patients whose symptoms increased from baseline; combined for both 5 cross-over sequences (n = 23). Symptoms increased (n) IV fentanyl IN fentanyl IV IN Sedation: Baseline 4(3,8) 4(1,8) 15 min 7(3,9) 5(2,8) 25 min 7(4,8) 5(2,9) 14 14 Nausea: Baseline 0 (0,1) 0 (0,2) WO 2010/005400 PCT/SG2009/000248 - 87 15 min 0(0,0) 0(0,1) 25 min 0(0,0) 0(0,1) 3 4 Pruritus: Baseline 0 (0,0) 0 (0,2) 15 min 0 (0,0) 0 (0,0) 25 min 0(0,0) 0 (0,1) 4 0 Oxygen saturation (%): Baseline 97 (96,98) 97 (97,98) 5 min 97 (93,98) 97 (96,98) 10 7 15 min 96 (96,98) 97 (96,97) 25 min 97 (95,98) 97 (96,98) 22 23 Visual analogue scores 0-100. Values are median and interquartile range except number (of 23 patients) for "symptoms increased". There were no significant differences between groups.

WO 2010/005400 PCT/SG2009/000248 - 88 0 0 UJ0 w o 0 0 5 10 15 20 25 30 0 5 10 15 20 25 30 TIME (m in) - INTRAVENOUS FIRST TIME (min) - NASAL SECOND 0 0 0 0 0 0.L o 04 Figure 4.6.8.0: Summed pain intensity differences for movement pain for both cross-over sequences (n=23). 5 4.6.9 Pharmacolkinetic Discussion: 4.6.9.0 Pharmacolkinetics after IV administration of 50 microgram fentanyl: The pharmacokinetic data after giving 50 pig of fentanyl by IV bolus was calculated by two methods: a model Independent method and the Compartmental model. The model 10 independent method uses the Area under the Curve (AUC) of the serum fentanyl concentrations versus time curve. The average AUCO-2 was 17.2 ngmL-min (13.8 to 20.6ngmL~min) and the average AUCo. was 27.4 ngmL~1min (21.6 to 33.2 ngmL min). The calculated average clearance (C) for fentanyl was 2.4 Lmin-1 (1.8 to 3 Lmin~). 15 In this study we found that the pharmacokinetics of fentanyl (50 pg) after IV administration followed a 2-compartment modea, C = 1.22e- mol + 1.10e- . The rate constants obtained were k12 =u0.37 0.19min~1, k21 = 0.03 0.02min and k = 0.1 0.02min-. The volume of distribution values calculated were Ve = 4.4 2L a nd V d = 77 i 15L. The totai body clearance was 428 98 mL/min andt1/2p was 141 gm60 minutes. 20 In this study, after a 50 pg intravenous bolus dose of fentanyl the resultant volumes of distributions (i.e. volume of central compartment Vc and distribution volume Vd) were large and the elimination half-life t 1

/

2 was approximately 21 minutes (13 to 29 minutes). The WO 2010/005400 PCT/SG2009/000248 -89 central compartment (Vc) was estimated to approximately 21.5 litres or 322 mLkg 1 [average 27 litres; range 20.4 to 33.3 litres]. This is larger than the plasma or serum compartment in the body (normally 3 litres) and thus includes tissue compartments that are highly perfused by blood e.g. heart, kidney, liver, lungs and the brain (central nervous 5 system). Hence the onset of the analgesic effect in the central nervous system would be expected to be rapid. The volume of distribution (Vd) for fentanyl was even larger, that is 36.6 litres or 545 mLkg 1 [average 59.6 litres; ranges from 38.6 to 80.6 litres] and thus includes tissues that are less well perfused e.g. muscle and fat. Fentanyl is lipophilic and would be expected to show steady uptake into fatty tissue. 10 4.6.9.1 Pharmacokinetics data after IN administration of 50 pg fentanyl: The bioavailability of a drug following intranasal administration can be influenced by its molecular size (molecular weight), its lipophilicity 5 ' and pKe 47 . In this study the average bioavailability (f) for the two IN formulations was 0.62 (62% absorbed at pH 6) and 0.52 15 (52% absorbed at pH 8) respectively. There was no significant difference between the bioavailability of these two IN groups (INpH6: median (mdn)= 55%, interquartile range; (IQR)=52.3%; INpH8: mdn=71%, IQR=63%; p=0.545, Mann-Whitney U test). 4.6.9.1.0 IN Bioavailability, Drug Size and Molecular weight: 20 A general observation is that compounds whose molecular weight are more than 1000 are likely to result in bioavailabilities of no greater than 10% (averaging 0.5 to 5% via the nasal route of administration) and this holds true in nearly all cases regardless of the physicochemical properties of the compounds 8 . 25 The amount of drug available for nasal absorption will depend on the amount swept away and its permeability through the mucosal surface. A polar compound with low molecular weight may have a slower rate of absorption through the nasal mucosa but nevertheless be fast enough to give a high degree of absorption. Fentanyl (base) has a molecular weight of 336.5. Taking a = 0.001 and b = 1.35, the percent absorption was calculated as 30 being close to 100% without taking into account the polarity of the fentanyl citrate when formulated at the two different pH values of the nasal solution. 4.6.9.1.1 IN bioavailability and pH of the nasal solution: This study found that the serum fentanyl concentrations, including the peak serum 35 concentration (Cmax), were numerically higher (but not significantly) in the nasal formulation with the pH value of 8. The pKa of fentanyl is 8.4. Using the Henderson Hasselbalch equation for a weak base, pk, - pH = logCi/Cu where Ci is the ionic species and Co is the unionised species, and at the pH value of 6, almost all of the fentanyl in the nasal solution would be ionised (99.6%), whereas at the pH value of 8, 40% was 40 unionised. Theoretically it is the unionised fentanyl (base) species that are lipophilic and would be expected to cross the lipophilic nasal mucosal membrane easily. This study however did not support the postulation that the formulation at the higher pH value (with more unionised species) would result in a higher serum concentrations and hence bioavailability. However there was no significant differences between the bioavailability of WO 2010/005400 PCT/SG2009/000248 - 90 these two IN groups (INpH6: mdn = 55%, IQR = 52.3%; INpH8: mdn = 71%, IQR = 63%; p=0.545, Mann-Whitney U test). The buffering capacity of the isotonic fentanyl nasal solutions of the invention (and the 5 volume administered was small, 0.18 mL) may not be enough to maintain the pH values when the nasal solution is in contact with the aqueous mucous. The pH of these nasal solutions may have changed to pH 5.5 to 6.5 (pH value of the mucosa) from their respective pH values of 6 and 8, hence the similarity of the serum fentanyl concentrations (and bioavailability) when the two nasal fentanyl sprays were administered. 10 The average pH value in the anterior of the nose is 6.40 (+0.11, -0.15 standard deviation) when calculated from H* values. The pH in the posterior of the nasal cavity is 6.27 (+0.13, -0.18 standard deviation). The overall range of pH was 5.17 to 8.13 for the anterior and 5.20 to 8.00 for the posterior (a pH value range similar to the fentanyl nasal solution of pH 15 6 and 8). The nasal clearance rate of saccharin was also tested at various pH 5.8 to 7.2 and none of the pH values affected the clearance rate of saccharin (clearance rate of 11.2 ± 6.5 mins with a range of 6 to 25 minutes which was in agreement with most studies) 49 . Hence in this study with fentanyl nasal solutions, (the pH values of 6 and 8), the nasal solutions' pH values should not have affected the clearance rate of the fentanyl from the 20 nasal cavity. 4.6.9.1.3 Comparison with aerosolised and nebulised pulmonary administered fentanyl The fentanyl was formulated into a propellant which was a mixture of 28:72 parts 25 trichlorofluoromethane and dichlorodifluoromethane and 0.05% sorbitan trioleate. This mixture was administered using an oral breadth-actuated metered dose inhaler (SmartMist*) 50 . There were two groups of healthy volunteers given this aerosolised fentanyl. The first group (10 volunteers, seven females and three males, aged between 18 to 50 years old) was given 100, 200 and 300 pg of fentanyl by IV bolus and by 30 pulmonary administration with a washout period of one week. Venous blood samples were taken for analysis of fentanyl. Another group (n=5) was given 100 pg fentanyl by IV bolus and by pulmonary administration after a one-week washout period. Blood samples were taken for analysis via the peripheral artery (artery group). The volunteers were of similar age and weight but included males (our study has only female volunteers). The venous 35 group was set up solely to determine inter and intra-subject variability of the fentanyl pharmacokinetics in IV as well as in aerosolised pulmonary administration but this was not possible due to logistical problems. The results were hence pooled together to obtain the PK data. The results showed that aerosolised fentanyl delivered via the pulmonary system rapidly reached therapeutic plasma levels (0.4 to 3 ng/mL) and the plasma fentanyl 40 concentrations were proportional to the dose administered. The absorption of fentanyl from the lung tissue followed first-order absorption kinetics. In the venous group of volunteers, the overall (95% confidence interval) kinetics data after the administration of IV bolus doses of 100, 200 and 300 pg fentanyl were: Vc = 40 to 63 litres; Vd = 160 to 265 litres; elimination half-life = 162 to 255 minutes and total clearance = 0.76 to 1.05 L/min.

WO 2010/005400 PCT/SG2009/000248 - 91 For their artery group of volunteers who were given 100 micrograms of IV fentanyl the PK data were: mean Vd = 26 Litres (7 to 44L); Vd = 193 Litres (94 to 292L); elimination half life = 131 minutes (89 to 172 mins) and total clearance = 1.19 Lmin 1 (0.93 to 1.46 Lmin 1 ). 5 Comparing the data of the present invention for the IV administration of 50 pg fentanyl (1/2 of the dose for the artery group 5 0 ), our V, was similar but Vd was five times smaller and the clearance was double, hence there was a nine times lower elimination half-life (using t 1

/

2 = 0.693Vd/CI). Both studies used the bi-exponential model to describe the serum fentanyl concentrations time curve. There was no significant difference between 10 the Cmax and tmax for both routes of administration. The estimated mean bioavailability for aerosolised pulmonary administered fentanyl for all subjects/dose (pooled results for venous group) was 56% at 5 minutes, 66% at 20 minutes, 83% at 60 minutes, 93% at 180 minutes, 96% at 360 minutes and 106% at infinite time. The overall estimated mean bioavailability for this venous group was 90% for 100 micrograms, 78% for 200 15 micrograms and 84% for 300 pg dose. The bioavailability of pulmonary administered fentanyl (100 micrograms) for the artery study was 81% (50-113%). In the present fentanyl study, there was no need to add to add polysaccharides to the solution to increase bioavailability. As such, it was decided not to add e.g. chitosan to the IN fentanyl formulation since its bioavailability was already close to 70% without adding an 20 enhancer which was quite satisfactory for its analgesic effect, although certain embodiments may contain chitosan. 5.0 General Discussions The average volume per spray produced by activating the nasal applicator device was 25 0.18 mL. The human nostrils can retain only a limited volume (0.15 mL per nasal cavity) 5 ', and the ideal volume of instillation of nasal solution ranges from 0.05 to 0.15 mL with an upper limit of 0.2 mL 8 . The nasal device gave a spray volume within the upper limit recommended. The droplet sizes produced by this nasal device measured by the Multistage Liquid Impinger were greater than 13 pm. The site of deposition of the aerosol 30 particles in the nasal cavity also determines the amount of absorption of the drug. Particles deposited in the non-ciliated mucosa area of the nasal cavity stayed longer then particles deposited in the ciliated region. The maximum solubility, S (Langmuir constant) of fentanyl in the 3 plastic components, PVC, EVA and Acetal, was 4.9, 1.6 and 0.5mg per gram of plastic respectively. PVC 35 showed the greatest adsorptive capacity followed by EVA and Acetal plastic. The "solubilisation" of fentanyl into the Acetal plastic to maximum solubility was a slow process taking 10 days to reach equilibrium compared to 5 hours for PVC and EVA at the storage temperature of 37 0 C. The diffusion coefficient for fentanyl into PVC was similar to EVA (1.15 ± 0.06 x 10- cm 2 s~ 40 1) but was lower for Acetal plastic (8.2 ± 0.33 x 10 1 cm 2 s 1 ) and obeyed the Fickian Law for diffusion. Fentanyl is transported from the solution into the inner matrix of the plastic by WO 2010/005400 PCT/SG2009/000248 - 92 the random fentanyl molecular motion due to its free thermal energy. The transportation process is a spontaneous process and involves a reduction of the free energy of the system. The diffusion coefficient for fentanyl obtained using the Diffusion Model may be unreliable if several variables were not fulfilled, for example; the assumption that the 5 concentration of fentanyl in aqueous solution at any time is independent of its position in the nasal solution from the plastic surface may not be necessarily true. The uptake of fentanyl by the plastic components is both rapid and extensive (especially for PVC and EVA plastic), therefore the concentration of fentanyl in the solution adjacent to the plastic may be significantly lower than elsewhere. In this experiment, the plastic component was 10 soaked or dipped completely into 5 mL of fentanyl nasal solution. The nasal solution was static and non-stirred and hence the Diffusion Model Equation may not "hold true" in this situation. If the size of the hole next to a polymer strand is large enough, this strand can jump into this hole by cooperative rotation about the single bonds in the polymer molecule resulting 15 in a cascading movement of other polymer strands, hence the rubbery nature of the polymer. The free volume thus moves around at random by the consecutive jumps of different segments of the polymer strands. Fentanyl molecules which are smaller than the free volume can occupy this hole by similar motion (jumps). The distribution coefficient, R for fentanyl in PVC-aqueous, EVA-aqueous and Acetal 20 aqueous partition, calculated using the Diffusion Model were 16.7 - 28.2, 6.4 - 13.3 and 4.1 - 5.6 respectively. R calculated during the equilibrium stage of the experiment for the PVC-aqueous, EVA-aqueous and Acetal -aqueous partition, were 19.3 - 32.3, 7.6 - 15.6 and 3.2 - 4.8 respectively. These R values were hence similar. The R values suggest that the order of affinity of fentanyl for the plastic components were PVC dip-tube > EVA seal > 25 Acetal reservoir. The loss of fentanyl into PVC dip-tube, EVA seal and Acetal reservoir also followed a two Compartment Model (two processes, an initial rapid loss or adsorptive phase, followed by a slower rate of loss or dissolution or absorption phase). The adsorbed layer of fentanyl onto the plastic surfaces can be pictured as a few molecular layers thick and is an 30 instantaneous process due to the lipophilicity of the fentanyl molecule, with a fast rate constant, a. The dissolution or absorption processes was a slower process with a slower rate constant, fl. 8 was independent of the initial concentrations effect, whereas a may showed a wide range of values. This is due to the direct consequence of the nature of the biexponential equation and the extreme sensitivity of the curve-fitting procedure. 35 Adsorption of fentanyl occurred spontaneously when the plastic material is introduced into the fentanyl nasal solution and especially at the higher temperature of 37 0 C. The spontaneous reaction would result in a higher rate of experimental error when determining the rate constant, a. The adsorption of fentanyl to the Acetal plastic is a much slower process and therefore subjected to a lesser experimental error, hence a should also 40 theoretically be independent of the initial concentrations effect.

WO 2010/005400 PCT/SG2009/000248 - 93 The amount of fentanyl adsorbed onto the plastic surface was likely to be smaller than the amount of fentanyl migrating (absorbed) into the plastic matrix. The clearance for fentanyl into the PVC, EVA and Acetal were 0.64 ± 0.10, 0.68 ± 0.15, and 0.017 ± 0.005 mLh 1 respectively (similar for PVC and EVA but much slower for the Acetal plastic). R values for 5 fentanyl in the PVC dip-tube-aqueous, EVA seal-aqueous and Acetal reservoir-aqueous system obtained using the 2-Compartmental Model was 2.22 ± 1.47, 1.75 ± 0.67 and 0.045 ± 0.006 respectively. The accuracy of R depended on the values of a (a is subjected to high experimental error) and hence explained the differences in R values obtained from the Diffusion Model and the Compartmental Model. 10 The loss of fentanyl from the aqueous nasal solutions into the plastic component was markedly temperature dependent. As the temperature increases, the diffusion of fentanyl molecules into the plastic material would be expected to increase due to the additional kinetic/diffusion energy provided by the heat. The activation energy, Ed for the fentanyl 15 molecule to diffuse (into holes) in the PVC and EVA was found to be 26.9 and 29.8 kJ per mole respectively. There was no effect of ionic strength on the loss of fentanyl into the different plastic components. The solute molecules were the removed from the solution by the increasing 20 concentration of other ions and were then deposited onto the plastic matrix. There was no noticeable "salting out" effect when sodium chloride was added to the fentanyl nasal solution. A major factor influencing the sorption of acidic or basic drugs into the plastic matrix is the 25 pH of the solution. The non-ionised species of the drug is the most lipophilic and the most favourably sorbed by the plastic matrix. The amount of the non-ionised form of the drug present in an aqueous solution is controlled by the solution's pH and the drug's pKa value. Fentanyl was rapidly sorbed into the plastic matrix when the nasal solution has the highest pH value (absorption greatest at pH 10> pH 8 > pH 6). The non-ionised species of 30 fentanyl is highest when the pH value of the fentanyl nasal solution is 11.0 (99.7%) followed by pH 10.0 (97.4%), 8.0 (27.1%) and least when the pH value is 6.0 (0.4%). Therefore fentanyl would be expected to undergo a greater loss by absorption at the higher pH values. The amount of uptake of fentanyl from the nasal solution into the plastic matrix during equilibrium condition, plotted against the nasal solution's pH value results in 35 a typical S-shape curve. Both the rate and extent of fentanyl loss increased with increasing pH values because of the increasing unionised fentanyl species. lonised species usually do not penetrate the plastic matrix. The uptake of fentanyl approaches an asymptotic value of 50, 50 and 45 pg/mL for PVC, EVA and Acetal respectively at the pH value of > 10 and approaches zero at pH value of < 6. 40 The pH value of the fentanyl nasal solution also has a great influence on the diffusion coefficient, D of fentanyl into the three plastic components. As expected the higher the pH value of the nasal solution the greater is the diffusion coefficient of fentanyl (D is higher at the pH values of 10.0 > 8.0 > 6.0) confirming that it is the un-ionised fentany that is highly absorbed/diffused into the plastic matrix.

WO 2010/005400 PCT/SG2009/000248 - 94 The average sorption number, Sn obtained for fentanyl in the PVC dip-tube, EVA seal and Acetal reservoir was 0.09 ± 0.04, 0.11 ± 0.05, 0.004 ± 0.001 h- respectively. 5 Sn, values were then used to obtain the shelf-life (t 9 o) of fentanyl nasal. The shelf-lives were 0.09, 0.08 and 2.40 hours when stored in PVC, EVA and Acetal plastic (at 370C, pH value of 10) respectively. The shelf life of the fentanyl nasal solution (pH 10) was only 4.8 minutes (0.08 x 60 minutes). Sn was also used to calculate shelf-lives at different storage temperature and pH values. The calculated shelf-lives, at room temperature, for fentanyl 10 nasal solution at the pH value of 10.0, 8.0 and 6.0, stored in the PVC plastic were 0.36h, 4.5h and 2.4 years respectively (1.1h, 13.4h and 7.7 years respectively, in EVA plastic). There were no Sn values for Acetal reservoir because there were not enough sorption data (during our experimental condition and storage time) for fentanyl when stored in Acetal plastic at the pH 8.0 and 6.0. When extrapolated from PVC and EVA data, the 15 predicted shelf-life should be more than 13.4 h and 7.7 years respectively. Hence the shelf-life at room temperature for fentanyl nasal solution when formulated at the pH 10.0, 8.0 and 6.0 is approximately 22 minutes, 4.5 h and 2.4 years respectively. The calculated shelf life for fentanyl nasal solution at a pH 8.0 stored in PVC plastic at the storage temperature of 50C is 17.8 hours. Hence keeping nasal solution in the refrigerator 20 increased the shelf life. The results presented herein show that there is a difference in the rate and amount of fentanyl sorbed into the three different plastic materials (PVC>EVA>Acetal plastic). PVC plastic is the most widely used medical plastic materials. . EVA plastic is a random 25 copolymer of polyethylene and vinyl acetate which also has a broad range of applications in pharmaceuticals. The EVA copolymer is a clear to translucent rubbery material available in a range of hardness values. Due to its flexibility EVA is used in the nasal device as the seal to prevent leakage of the liquid from the nasal device. Acetal plastic is hard and non-pliable, more inert (less absorptive property) and hence used in device as a 30 reservoir for the nasal solution for maximum stability of the solution. The pharmacokinetic studies showed that the mean total area under the curve for IV and IN administration of the pH 6.0 and pH 8.0 formulations was 27.4 (IV) and 13.3 and 14.9 (IN) ng/ml/min respectively. The mean bioavailability (f) for the two IN formulations was 35 0.49 (49% absorbed, pH 6.0) and 0.54 (54% absorbed, pH 8.0) respectively (no significant differences, INpH 6: mdn = 55%, IQR = 52.3%; INpH 8: mdn = 71%, IQR = 63%; p = 0.545, Mann-Whitney U test). The absorption of fentanyl from the nasal mucosa is a first order process and instantaneous without a lag time. The median peak serum concentrations (Cmex) were 0.36 ng/ml (IQR 0.28 ng/ml) in group INpH 8 and 0.38 ng/ml 40 (IQR 0.41 ng/ml) in group INpH 6 (no significant difference in the 2 peaks, p=0.778, Mann Whitney U test). The pharmacokinetics (PK) of fentanyl after IV administration followed a two-compartment model, C = 1.

22 e~22t + 1.10e- 005 t. Clearance was found to be 1.53 Lmin 1 , which equated WO 2010/005400 PCT/SG2009/000248 - 95 to the hepatic blood flow (1.5Lmin 1 ) hence the major site of fentanyl elimination in the body was via the liver. Other PK parameters obtained were: V, = 59.7 ± 17.85 L, Vd = 275.7 141.4 L, k 12 = 0.1 ± 0.06 min-, k21 = 0.02 ± 0.01 min-', k 1 o = 0.03 ± 0.01 min-, CI = 1.53 0.25 Lmin- 1 , t 1 /2p = 185.4 ± 70.2 min and the elimination half-life t 1

/

2 was 5 approximately 21 minutes (13 to 29 minutes). The Vc and Vd were large thus indicating that fentanyl was perfused into tissue compartments that are highly perfused by blood e.g. heart, kidney, liver, lungs and the brain (central nervous system). Hence the onset of the analgesic effect in the central nervous system would be expected to be rapid. Since the Vd was even larger, indicating a perfusion into tissues that are less well perfused e.g. muscle 10 and fat. Since fentanyl is lipophilic, it's expected to show steady uptake into fatty tissue. The average bioavailability (f) for the two IN formulations was 0.62 at pH 6, and 0.52 at pH 8. The serum fentanyl concentrations, including the peak serum concentration (Cmax), were numerically higher (but not significantly) in the nasal formulation with the pH 8. At the 15 higher pH there are more unionised fentanyl species (lipophilic, and expected to cross the lipophilic nasal mucosal membrane easily) however these PK studies did not support the postulation that the formulation at the higher pH value would result in a higher bioavailability. This may be due to the inadequate buffering capacity of the isotonic fentanyl nasal solutions (and the volume administered was small, 0.18 mL) and hence 20 may not be enough to maintain the pH values when the nasal solution is in contact with the aqueous mucous. The pH of these nasal solutions may have changed to pH 5.5 to 6.5 (pH value of the mucosa) from their respective pH 6 and 8, hence the similarity of the bioavailability when the two nasal fentanyl sprays were administered. 25 The fentanyl nasal composition of the invention formulated at 300 pg/mL was ideal for nasal administration. The concentrated solution avoided excessive volume administered and prevent "run off" from the nasal cavity. The study showed that the IN formulation at the pH 6 and pH 8 were equal in bioavailability. In a preferred embodiment, the intranasal fentanyl is formulated in a weak acidic buffer (preferably at pH 6) because of (i) its stability 30 in applicators containing plastic components and also (ii) its stimulatory effect on the nasal mucosa to stimulate the reflex bicarbonate production rendering the nasal solution basic and hence increasing the non-ionic fentanyl molecules, increasing penetration/absorption into the nasal mucosa. The therapeutic blood level was obtained within 2 to 5 minutes after a single spray ensuring a rapid analgesic response. Nasal absorption of fentanyl was 35 instantaneous. Fentanyl nasal solution was effective in treating breakthrough pain shown by the therapeutic blood levels as well as the pain score study. In summary, the results from the in vitro study surprisingly showed that: 1. Fentanyl nasal solution is insufficiently stable in the nasal device when the pH of the nasal solution is formulated at pH > 8.0. 40 2. There is a rapid loss of fentanyl via sorption into the plastic components of the nasal device when the pH is greater than 8.0. There is also significant loss of WO 2010/005400 PCT/SG2009/000248 -96 fentanyl via sorption into PVC and ethylene vinyl acetate plastics compare to acetal plastic. 3. The shelf-life fentanyl nasal solution is 4.5 hours at the pH 8.0 at room temperatures. 5 4. Decreasing the pH of fentanyl nasal solution to a pH 6.0 increased the shelf-life of this nasal solution to 30 months at room temperature. 5. Ionic strength of the nasal solution and the initial concentration fentanyl in the nasal solution did not have an effect on the sorption profile of fentanyl into the plastic components of the nasal device. 10 6. Increasing the storage temperature increased the sorption of fentanyl into the plastic components of the nasal device in accordance with the Arrhenius Equation. The results from the in vivo study surprisingly showed that: 1. Instantaneous absorption of fentanyl via the nasal mucosa occurred without a lag period. 15 2. There were no differences in the absorption profile between the two nasal fentanyl formulations of pH 6.0 or 8.0. This may be due to the low buffer capacity of the buffers used in the formulation of the fentanyl nasal spray 3. The bioavailability of fentanyl nasal spray was found to be approximately 70% (compared with IV dose). 20 4. The serum fentanyl levels reached the analgesic range at two minutes following nasal administration 5. Pain score study showed the effectiveness of relieving breakthrough pain using the prototype fentanyl nasal spray The embodiments discussed herein provide evidence of the effectiveness of using 25 intranasal fentanyl for treating breakthrough pain with a single dose application. It provides an alternate route of drug administration especially for patients who are phobic to injection, and by adjusting the pH of the fentanyl solution, the shelf-life of the applicator product can be greatly extended. The following references are included for ease of reference only and are merely included 30 to facilitate an understanding of the present invention. The inclusion of these references is not an acknowledgement or admission that any of the material referred to herein, or the matter contained in these references, is(or was), part of the common general knowledge as at the priority date of the application. REFERENCES 35 1. Agarwal V, Mishra B. Recent trends in drug delivery systems: Intranasal drug delivery. Indian J of Experimental Biology. 1999; 37:6-16 2. Martonen TB, Zhang Z, Yue G, Musante CJ. Fine particles deposition within human nasal airways. Inhalation Toxicology. 2003; 15:283-303 3. O'Rourke. Nasal drug delivery: A nose for a good deal. Scrip's Target World Drug 40 Delivery News. 2003; 20:8-10 WO 2010/005400 PCT/SG2009/000248 - 97 4. Gizurarson S. The relevance of nasal physiology to the design of drugs absorption studies. Adv Drug Del Rev. 1993; 11: 329-347. 5. Chien YW, Chang SF, Su KSE. Intranasal Drug Delivery for Systemic Medications. CRC Critical Reviews in Therapeutic Drug Carrier Systems. 1987, 4: 67-194. 5 6. Quraishi MS, Jones NS, Mason JDT. The nasal delivery of drugs. Clin Otolaryngology. 1997; 22:289-301 7. Product information in the packet insert for Actiq*, Trade Mark for oral transmucosal fentanyl citrate, manufactured by Cephalon Inc, USA, US Patent no. 4671953, June 2004 10 8. Wilson TD, Maloney T, Amsden W. High-performance liquid chromatographic determination of fentanyl citrate in a parenteral dosage form. J of Chromatograph. 1988; 445:299-304 9. Crank J. The mathematics of diffusion. Second Edition, Clarendon Press, Oxford. 1973, page 1. 15 10. Roberts MS, Polack AE, Martin G, Blackburn HD. The storage of selected substances in aqueous solution in polyethylene containers: The effect of some physicochemical factors on the disappearance kinetics of the substances. Int J Pharm. 1979, 2: 295-306. 11. Polack AE, Nunez LJ, Autian J. Transport of solutes into polyethylene bottles from 20 aqueous solutions: Empirical relationships of the data. Int J Pharm. 1979, 3: 157 175. 12. Yuen PH, Denham SL, Sokoloski TD, Burkman AM. Loss of nitroglycerin from aqueous solution into plastic intravenous delivery system. J Pharm Sci 1979, 68: 1163-1166. 25 13. Crank J. The mathematics of Diffusion. 2 nd edition 1975, Clarendon Press, Oxford, Chapter 4, pg 56. 14. Martin A, Swarbrick J, Cammarata A. Physical Pharmacy; Physical Chemical Principles in the Pharmaceutical Sciences. 3 rd Edition, 1983. Lea and Febiger, Philadelphia, USA. Chapter 16, Interfacial Phenomena. Pg 461 30 15. Roberts MS, Kowalik EA, Polack AE. Prediction of solute sorption by Polyvinyl Chloride Plastic Infusion bags. J Pharm Sci 1991, 80: 449-455 16. Martin EW. Dispensing of medication. 7 th Edition, 1971 Mack Publishing Co., USA. Chapter 15, Plastic and medication by Autian J. pg 652-715 17. Martindale. The complete drug reference. 25 th Edition. Pharmaceutical Press, 35 London, UK. page 798 18. Rodell MB, Guess WL, Autian J. Interaction of Sorbic acid with an insoluble polyamide. J Pharm Sci. 1964; 53:873-877 19. Wehner F, Olsen H, Tinel H, Kinne-Saffran E, Kinne RK. Cell volume regulation: osmolytes, osmolyte transport and signal transduction. Reviews of Physiology 40 Biochemistry & Pharmacology. 2003; 148: 1-80 20. Ritter M, Fuerst J, Woll E, Chwatal S, Gschwenttner M, Lang F, Deetjen P, Paulmichi M. Na*/H* exchangers: linking osmotic disequilibrium to modified cell function. Cellular Physiology & Biochemistry. 2001; 11: 1-18 WO 2010/005400 PCT/SG2009/000248 - 98 21. Proctor DF, Anderson I, Lundqvist G. Nasal mucociliary functions in humans: Respiratory defence mechanisms. J Brain, DF Proctor and I Reid. Marcel Dekker, New York 1976 vol1, page 63-93. 22. Philipson K, Falk R, Svartengren M, Jarvis N, Bailey M, Bergmann R, Hofmann W, 5 Camner P. Does lung retention of inhaled particles depend on their geometric diameter? Experimental Lung research. 2000; 26: 437-455 23. Illum L. Nasal drug delivery: new development and strategies. Drug Discovery Today. 2002; 7:1184-1189 24. Fry FA, Black A. Regional deposition and clearance of particles in the human 10 nose. Aerosol Science. 1973, 4 113-124 25. Behl CR, Pimplaskar HK, Sileno AP, deMeireles J and Romeo VD. Effects of physicochemical properties and other factors on systemic nasal drug delivery. Advance Drug Delivery Review. 1998; 29:89-116 26. O'Neil G. Nasal spray resource book. Go Medical Industries, 200 Churchill Ave, 15 Subiaco, Western Australia. 1996. pg 5 27. Lambropoulos J, Spanos GA, Lazaridis NV, Ingallinera TS, Rodriguez VK. Development and validation of an HPLC assay for fentanyl and related substances in fentanyl citrate injection, USP. J. of Pharmaceutical and Biomedical Analysis. 1999; 20: 705-716 20 28. Martin EW. Dispensing of medication. 7 th Edition, 1971 Mack Publishing Co., USA. Chapter 15, Plastic and Medication 29. Plaut BS, Davies DJG, Meakin BJ, Richardson NE. The mechanism of interaction between chlorhexidine digluconate and poly(2-hydroxyethylmethacrylate). J. of Pharmacol. 1981; 33: 82-88 25 30. Akaho E, Fukumori Y. Studies on adsorption characteristics and mechanism of adsorption of chlorhexidine mainly by carbon black. J. of Pharm. Sci. 2001; 90: 1288-1297 31. Kowaluk EA, Roberts MS, Polack AE. Comparison of models describing the sorption of nitroglycerin and diazepam by plastic infusion systems: Diffusion and 30 compartment models. J Pharm Sci. 1985, 74: 625-633 32. Long FA, Thompson LJ. Diffusion of water vapor in polymers. J. of Polymer Science. 1955; 15: 413-426 33. Fornasiero F, Olaya MM, Esprester B, Nguyen V, Frausnitz JM. Distribution coefficients and diffusivities in three polymers for nineteen aqueous non-volatile 35 solutes. J. Apple. Polym. Sci. 2002; 85: 2041-2052. 34. Roberts MS, Cossum PA, Galbraith AJ, Boyd GW. The availability of nitroglycerin from parenteral solutions. J Pharm. Pharmacol. 1980, 32: 237-244. 35. Martin A, Swarbrick J, Cammarata A. Physical Pharmacy; Physical Chemical Principles in the Pharmaceutical Sciences. 3 rd Edition, 1983. Lea and Febiger, 40 Philadelphia, USA. Chapter 3, Atomic and Molecular Structure. Pg 59 36. Wade WH. Adsorption kinetics for CC1 4 on Teflon. J Colloid Interface Sci. 1974; 47: 676-681 37. Martin A, Swarbrick J, Cammarata A. Solutions of electrolytes. In: Physical Pharmacy; 3 d Edition, Philadelphia: Lea and Febiger; 1983:179-181 WO 2010/005400 PCT/SG2009/000248 - 99 38. Xu QA, LA Trissel, JF Martinez. Rapid loss of fentanyl citrate admixed with fluorouracil in PVC containers. The Annals of Pharmacotherapy. 1997, 31:297-301 39. Guess WL, Worrell LF, Autian J. The effect of a quaternary ammonium compound on polyvinyl chloride used in medical practice. Am J Hosp Pharm 1962, 19: 370 5 374 40. Nielsen LE. Mechanical properties of polymers. Reinhold Publishing Corp. (Eds) 1962. New York, USA pp27-28 41. Vincent WW. Plastics in medical tubing applications - manufacturing considerations. Biomaterials. 1981; 2: 194-200. 10 42. Striebel HW, Pommerening J, Rieger A. Intranasal fentanyl titration for postoperative pain management in an unselected population. Anaesthesia 1993; 48:753-757 43. Schwagmeier R, Oelmann T, Dannapel T, Striebel HW. Patientenakzeptanz gegenuber der patientkontrollierten intranasalen analgesie (PCINA). Anaesthetist 15 1996; 45:231-239 44. Rowland M, Tozer TN. Clinical Pharmacokinetics, Concepts and Applications. Second Edition1989, Lea and Febiger, Philadelphia pp. 469-471 45. Fentanyl citrate monograph, AHFS Drug Information. American Society of Health System Pharmacists, 2001 pp.1997-2003 20 46. Lim CB, Paech MJ, Sunderland VB, Roberts MJ, Banks SL, Rucklidge MCM. Pharmacokinetics of nasal fentanyl. J Pharm Pract + Res. 2003; 33(1): 59-64 47. Huang C, Kimura R, Nassar R, Hussein AA. Mechanism of nasal absorption of drugs I: physicochemical parameters influencing the rate of in-situ nasal absorption of drugs in rats, J Pharm Sci 1985; 74:608-611. 25 48. Donovan MD, Huang Y. Large molecule and particulate uptake in the nasal cavity: the effect of size on nasal absorption. Advanced Drug Delivery Reviews. 1998; 29: 147-155 49. Washington N, Steele RJC, Jackson SJ, Bush D, Mason J, Gill DA, Pitt K, Rawlins DA. Determination of baseline human nasal pH and the effect of intranasally 30 administered buffers. International Journal of Pharmaceutics. 2000; 198: 139-146 50. Mather LE, Woodhouse A, Ward ME, Farr SJ, Rubsamen RA, Eltherington LG. Pulmonary administration of aerosolised fentanyl: pharmacokinetics analysis of systemic delivery. Br J Clin Pharmacol 1998; 46:37-43. 35

Claims

1. A method of extending the shelf-life of a pharmaceutically acceptable fentanyl composition, said method comprising maintaining the fentanyl composition at a pH 5 value in a range of between 3.8 and 6.5.

2. The method of claim 1 wherein the composition is maintained at a pH value in a range of between 4.5 and 6.3.

3. The method of claim 2, wherein the composition is maintained at a pH value in a range of between 5.0 and 6.2. 10

4. The method of claim 3, wherein the composition is maintained at a pH value of about 6.0.

5. The method of any one of claims 1 to 4, which includes the step of suspending the fentanyl in an isotonic or hypotonic buffer.

6. The method of claim 5, wherein the fentanyl is suspended in a phosphate buffered 15 saline solution.

7. The method of any one of claims 1 to 6, wherein the composition comprises: fentanyl; a phosphate buffered saline solution; and sodium chloride, 20 wherein the concentration of fentanyl is between 100 pg/mL and 400 pg/mL.

8. The method of claim 7, wherein the concentration of fentanyl is about 300 pg/mL.

9. The method of any one of claims 1 to 8, which includes the step of sterilising the fentanyl composition prior to dispensing or aliquoting the composition into a vial, 25 cartridge, applicator, or container.

10. The method of claim 9, wherein the vial, cartridge, applicator, or container is hermetically sealed following addition of the fentanyl composition.

11. The method of claim 9 or claim 10, wherein the vial, cartridge, applicator, or container is labelled with instructions that the composition is to be used for the 30 treatment of pain and that the composition has a shelf-life of between 12 months and 40 months from the date of manufacture, when stored at room temperature.

12. The method of claim 11, wherein the vial, cartridge, applicator, or container is labelled with instructions that the composition has a shelf-life of between 18 WO 2010/005400 PCT/SG2009/000248 -101 months and 30 months from the date of manufacture when stored at room temperature.

13. The method of claim 12, wherein the vial, cartridge, applicator or container is labelled with instructions that the composition has a shelf-life of about 30 months 5 from the date of manufacture when stored at room temperature.

14. The method of any one of claims 9 to 13, wherein the vial, cartridge, applicator, or container is adapted for the intranasal delivery of the composition in a spray dosage form.

15. The method of any one of claims 9 to 13, wherein the vial, cartridge, applicator, or 10 container is adapted for delivery of a droplet size generated during administration of between 2 pm and 50 pm.

16. The method of claim 15, wherein the vial, cartridge, applicator, or container is adapted for delivery of a droplet size generated during administration of between 5 pm and 30 pm. 15

17. The method of claim 16, wherein the vial, cartridge, applicator, or container is adapted for delivery of a droplet size generated during administration of about 10 pm.

18. The method of any one of claims 1 to 17, wherein the fentanyl is fentanyl citrate.

19. A pharmaceutically acceptable fentanyl composition having a shelf-life of at least 20 12 months when stored at room temperature, the composition comprising fentanyl buffered to a pH value in a range of between 3.8 and 6.5.

20. The composition of claim 19 wherein the composition is buffered to a pH value in a range of between 4.5 and 6.3.

21. The composition of claim 20, wherein the composition is buffered to a pH value in 25 a range of between 5.0 and 6.2.

22. The composition of claim 21, wherein the composition is buffered to or maintained at a pH value of about 6.0.

23. The composition of any one of claims 19 to 22, wherein the fentanyl is suspended in an isotonic or hypotonic buffer. 30

24. The composition of claim 23, wherein the fentanyl is suspended in a phosphate buffered saline solution.

25. A method of minimizing the sorption of fentanyl compounds by polymeric compounds present in pharmaceutical applicators, the method including the step of charging said applicator with a solution comprising a pharmaceutically 35 acceptable concentration of fentanyl, the solution being formulated to remain at a WO 2010/005400 PCT/SG2009/000248 - 102 pH value within a range of between 3.0 and 8.0 for the effective lifetime of the solution.

26. The method of claim 25, wherein the solution is maintained at a pH value within a range of between 4.0 and 7.0. 5

27. The method of claim 26, wherein the solution is maintained at a pH value of about 6.0.

28. A method of preparing a sterile intranasal pharmaceutical composition which is stable for at least 30 days without the need for preservatives or stabilizing agents, the method comprising the steps of: 10 mixing fentanyl powder with water to a desired pharmaceutically effective concentration, thereby to provide a pharmaceutical fentanyl composition; sterilizing said composition; charging a reservoir of a suitable intranasal applicator under sterile conditions with the sterilized composition; and 15 sealing the applicator hermetically while under such sterile conditions, the contents of the applicator reservoir remaining hermetically sealed during or following discharging of the composition from the applicator.

29. The method of claim 28, wherein the composition is sterilized by way of filter sterilization or autoclaving. 20

30. The method of claim 28 or claim 29, wherein the applicator is a timed lock-out intranasal spray applicator.

31. The method of claim 30, wherein the applicator is a nasal spray applicator manufactured in Western Australia.

32. The method of any one of claims 28 to 31, wherein the method includes the step of 25 buffering the fentanyl to a pH value in a range of between 3.8 and 6.5.

33. The method of claim 32, wherein the method includes the step of buffering the fentanyl to a pH value in a range of between 4.0 and 6.3.

34. The method of claim 33, which includes the step of buffering the fentanyl to a pH value in a range of between 5.0 and 6.2. 30

35. The method of claim 34, which includes the step of buffering the fentanyl to a pH value in a range of between 5.5 and about 6.1.

36. The method of claim 35, which includes the step of buffering the fentanyl to a pH value of about 6.0. WO 2010/005400 PCT/SG2009/000248 - 103

37. A pharmaceutical composition for intranasal administration, the composition comprising an aqueous solution of fentanyl in a concentration sufficient to effectively treat or manage pain in a subject when administered intranasally, the composition being buffered to a pH value of between 3.8 and 6.5 and wherein the 5 fentanyl is included in the composition in a concentration of between 0.5 pg/mL and 1 mg/mL.

38. The composition of claim 37, wherein the fentanyl is included in a concentration of between 10 pg/mL and 500 pg/mL.

39. The composition of claim 38, wherein the fentanyl is included in a concentration of 10 between 100 pg/mL and 400 pg/mL.

40. The composition of claim 39, wherein the fentanyl is included in a concentration of about 300 pg/mL.

41. The composition of any one of claims 37 to 40, wherein the composition has a pH value in a range of between 4.0 and 6.3. 15

42. The composition of claim 41, wherein the composition has a pH value in the range of 5.0 to 6.2.

43. The composition of claim 42, wherein the composition has a pH value in the range of 5.5 to 6.1.

44. The composition of claim 43, wherein the composition has a pH of about 6.0. 20

45. The composition of any one of claims 37 to 44, wherein the composition comprises a phosphate buffered saline solution having a pH of about 6.0, and a fentanyl concentration of about 300 pg/mL.

46. The composition of any one of claims 37 to 45, wherein the fentanyl is fentanyl citrate. 25

47. The composition of any one of claims 37 to 46, wherein the composition is suitable for single dose administration, consisting of a dose in a range of between about 1.5 pg per dose of fentanyl to 3 mg per dose of fentanyl.

48. The composition of claim 47, wherein the composition comprises a dose in the range of about 30 pg per dose of fentanyl to 1.5 mg per dose of fentanyl. 30

49. The composition of claim 48, wherein the composition comprises a dose of about 50 pg per dose of fentanyl.

50. The composition of any one of claims 47 to 49, wherein the dose provides a mean maximum plasma concentration (C.x) of fentanyl of about 0.10 to 1.5 ng/ml per 50 pg fentanyl or fentanyl citrate following nasal administration to a subject. WO 2010/005400 PCT/SG2009/000248 - 104

51. The composition of claim 50, wherein the dose provides a mean maximum plasma concentration of fentanyl of about 0.14 to 1.32 ng/ml per 50 pg fentanyl or fentanyl citrate following nasal administration to a subject.

52. The composition of claim 51, wherein the dose provides a mean maximum plasma 5 concentration of fentanyl of about 0.18 ng/ml per 50 pg fentanyl or fentanyl citrate, following nasal administration to a subject.

53. The composition of any one of claims 37 to 52, wherein the composition is formulated to be substantially free of preservatives, physiological or mucosal absorption enhancers, or propellants. 10

54. The composition of any one of claims 37 to 52 which includes a polysaccharide absorption enhancer, present in a concentration of about 0.01 to about 10% by weight of the composition.

55. The composition of claim 54, wherein the absorption enhancer is present in a concentration of between 0.2% and 2%. 15

56. The composition of claim 55, wherein the absorption enhancer is present in a concentration of about 1%.

57. The composition of any one of claims 54 to 56, wherein the absorption enhancer is positively charged.

58. The composition of any one of claims 37 to 52, or claims 54 to 57, which includes 20 one or more organic solvents in an amount sufficient to enhance the solubility of the fentanyl in the water.

59. The composition of claim 58, wherein the organic solvent is a polar organic solvent selected from the group consisting of: ethanol, propylene glycol, glycerol, polyethylene glycol, and mixtures thereof. 25

60. The composition of claim 58 or claim 59, wherein, the polar organic solvent is ethanol in an amount of between 1 and 60% (v/v).

61. The composition of claim 60, wherein the ethanol is present in an amount of between 5% and 30% (v/v).

62. The composition of claim 61 wherein the ethanol is present in an amount of about 30 10% (v/v).

63. The composition of any one of claims 37 to 52, or claims 54 to 62, which includes about 5% (v/v) propylene glycol.

64. The composition of any one of claims 37 to 52, or claims 54 to 62, further comprising one or more anionic, cationic, or non-ionic surfactants selected from the group 35 consisting of: sodium lauryl sulfate, dioctyl sodium sulfosuccinate, dioctyl sodium sulfonate, benzalkonium chloride, benzethomium chloride, Tween 80, lauromacrogol WO 2010/005400 PCT/SG2009/000248 - 105 400, polyoxyl 40 stearate, polyoxyethylene hydrogenated castor oil 10, 50 and 60, glycerol monostearate, polysorbate 40, 60, 65 and 80, sucrose fatty acid ester, methyl cellulose and carboxymethyl cellulose.

65. The composition of claim 64, wherein the one or more surfactants are present in an 5 amount of between. 0.05% and 1%, either individually or cumulatively.

66. The composition of any one of claims 37 to 52, or claims 54 to 62, which includes simple or complex carbohydrates, and/or polyols, including mannitol.

67. The composition of claim 66, comprising between 0.1% and 3% (v/v) mannitol.

68. The composition of claim 67, comprising between 0.2% and 1% (v/v) mannitol. 10

69. The composition of claim 68, comprising about 0.3% (v/v) mannitol.

70. The composition of any one of claim 37 to 69, which is buffered using a phosphate-based buffer.

71. The composition of claim 70, which is formulated to be an isotonic or hypotonic solution. 15

72. The composition of any one of claim 37 to 71, which has an osmolality in the range of between 200 mOsm/L and 500 mOsm/L.

73. The composition of claim 72, which has an osmolality of about 380 mOsm/L.

74. The composition of any one of claims 37 to 73, which provides a plasma concentration of at least 0.1 ng/ml of fentanyl within less than 10 minutes following 20 administration.

75. The composition as claimed in any one of claims 37 to 74, applied using a nebulizer.

76. A method of treating or managing pain by administering intranasally to a subject in need thereof, in an amount to effectively treat, ameliorate or manage pain, a pharmaceutical composition as claimed in any one of claim 37 to 74, which 25 comprises an aqueous solution of fentanyl or fentanyl citrate or pharmaceutically acceptable salts or enantiomers thereof.

77. The method of claim 76, which includes the step of administering the composition intranasally to the subject at a concentration such that the subject experiences a fentanyl peak plasma concentration of between 10 and 90% of that achieved by 30 intravenous administration, when administered intravenously at an identical fentanyl or fentanyl citrate dose.

78. The method of claim 76 or claim 77, which includes the step of administering the composition of the invention intranasally to a subject in need thereof to provide a mean time to maximum plasma concentration (Tmax) of fentanyl of between about 2 35 minutes and about 180 minutes, from the time of administration. WO 2010/005400 PCT/SG2009/000248 - 106

79. The method of claim 78, which includes the step of administering the composition of the invention intranasally to a subject in need thereof to provide a mean time to maximum plasma concentration (Tmax) of fentanyl of between about 2 minutes and about 30 minutes from the time of administration. 5

80. The method of claim 79, which includes the step of administering the composition of the invention intranasally to a subject in need thereof to provide a mean time to maximum plasma concentration (Tmax) of fentanyl of between about 2 minutes and about 15 minutes, from the time of administration.

81. The method of any one of claims 76 to 80, which includes the step of administering 10 the composition at a fentanyl or fentanyl citrate concentration sufficient to provide the subject with a peak plasma concentration of at least 40% of that achieved by intravenous administration, when administered intravenously at an identical fentanyl or fentanyl citrate dose.

82. The method of claim 81, which includes the step of administering the composition 15 at a fentanyl or fentanyl citrate concentration sufficient to provide the subject with a peak plasma concentration of at least 60% of that achieved by intravenous administration, when administered intravenously at an identical fentanyl or fentanyl citrate dose.

83. The method of any claim 82, which includes the step of administering the 20 composition at a fentanyl or fentanyl citrate concentration sufficient to provide the subject with a peak plasma concentration of at least 70% of that achieved by intravenous administration, when administered intravenously at an identical fentanyl or fentanyl citrate dose.

84. The method of any one of claims 76 to 83, which includes the step of providing a 25 composition of the invention in a dose sufficient to provide a bioavailability of more than 50% when administered intranasally, when compared to a corresponding intravenous dose.

85. The method of claim 84, which includes the step of providing a composition of the invention in a dose sufficient to provide a bioavailability of more than 60% when 30 administered intranasally, when compared to a corresponding intravenous dose.

86. The method of claim 85, which includes the step of providing a composition of the invention in a dose sufficient to provide a bioavailability of more than 70% when administered intranasally, when compared to a corresponding intravenous dose.

87. The method of any one of claims 76 to 86, which is used to treat or manage 35 patients experiencing or reasonably expected to experience chronic pain, acute pain, or breakthrough pain.

88. The method of any one of claims 76 to 87, which includes the step of administering the composition prophylactically. WO 2010/005400 PCT/SG2009/000248 -107

89. The method of any one of claims 76 to 88, which includes administering the composition to a subject to treat any one or more of the group selected from: post operative pain; fracture pain; and to reduce post-anaesthetic emergency agitation.

90. The method of any one of claims 76 to 89, which includes the step of providing the 5 composition in the form of a spray in which a majority of droplets when sprayed are between 2 pm and 50 pm in diameter.

91. The method of claim 90, which includes the step of providing the composition in the form of a spray in which a majority of droplets when sprayed are between 5 pm and 30 pm in diameter. 10

92. The method of claim 91, which includes the step of providing the composition in the form of a spray in which a majority of droplets are about 10 pm in diameter.

93. The method of any one of claims 90 to 92, wherein more than 80% of the droplets have a diameter of greater than 10 pm.

94. The method of claim 93, wherein more than 90% of the droplets have a diameter 15 of greater than 10 pm.

95. The method of claim 94, wherein more than 95% of the droplets have a diameter of greater than 10 pm.

96. The method of any one of claims 76 to 95, which includes the step of re administering the compound of the invention to the subject pre-emptively or when 20 an increase in pain levels is detected in the subject.

97. The method of any one of claims 76 to 96, wherein the compound is re administered by the patient him- or herself as a form of patient-controlled analgesia (PCA).

98. The method of any one of claims 76 to 97, which includes administering the 25 composition as an intranasal spray using a metered dose applicator, including a pump spray dispensing device or applicator.

99. The method of any one of claims 76 to 98, which includes the step of administering the composition in a volume not exceeding 300 pl per administration.

100. The method of claim 99, which includes the step of administering the 30 composition in a volume of between 5 pl and 200 pl per administration.

101. The method of claim 100, which includes the step of administering the composition in a volume of between 5 pl and 200 pl per administration.

102. The method of claim 101, which includes the step of administering the composition in a volume of between 50 pl and 180 pl per administration. WO 2010/005400 PCT/SG2009/000248 -108

103. The method of claim 102, which includes the step of achieving a sufficient analgesic dose in a volume of no more than about 150 pi per administration.

104. A dosage form comprising a pharmaceutically effective amount of a fentanyl composition in a spray applicator suitable for intranasal delivery of the 5 composition to a subject in need thereof, the fentanyl composition having a pH value within a range of between 3.8 and 6.5, wherein dosage form has a fentanyl concentration and spray volume sufficient to ensure a dose delivery of between about 1.5 pg and 3 mg of fentanyl per spray, upon administration.

105. The dosage form of claim 104, which has a fentanyl concentration and 10 spray volume sufficient to ensure a dose delivery of between about 30 pg to 1.5 mg fentanyl per spray, upon administration.

106. The dosage form of claim 105, which has a fentanyl concentration and spray volume sufficient to ensure a dose delivery of about 100 pg fentanyl per spray, upon administration. 15

107. The dosage form of any one of claims 104 to 106, wherein the composition has a pH value within a range of between 4.0 and 6.3

108. The dosage form of claim 107, wherein the composition has a pH value within a range of between 5.0 and 6.2.

109. The dosage form of claim 108, wherein the composition has a pH value 20 within a range of between 5.5 and 6.1.

110. The dosage form of claim 109, wherein the composition has a pH value of about 6.0.

111. The dosage form of any one of claims 104 to 110, wherein, advantageously, each administration of the dosage form results in an effective 25 pain-relieving dose of the pharmaceutical composition being delivered to a subject in a volume not exceeding 300 pl per administration.

112. The dosage form of claim 111, wherein the effective dose is delivered in a volume of between 5 pl and 200 pI per administration.

113. The dosage form of claim 112, wherein the effective dose is delivered in a 30 volume of between 50 pl and 180 pi per administration.

114. The dosage form of claim 113, which comprises an effective pain relieving dose being delivered to a patient in a volume of no more than about 150 pl per administration, thereby providing adequate pain relief following administration of only a single intranasal application. 35

115. The dosage form of any one of claims 104 to 115, wherein the pharmaceutical applicator is a metered dose intranasal applicator. WO 2010/005400 PCT/SG2009/000248 - 109

116. A substance or composition for use in a method of treating or managing pain in a subject, the substance or composition comprising a therapeutically effective amount of the fentanyl or fentanyl citrate composition of any one of claims 37 to 75, suitable for intranasal administration. 5

117. The substance or composition, wherein the subject is animal or human.

118. Use of fentanyl buffered to a pH value in a range of between 3.8 and 6.5 in the manufacture of a medicament having pain relieving activity when administered intranasally.

119. A method of extending the shelf-life or expiry date of a liquid 10 pharmaceutical composition by at least 30 months from the date of manufacture when stored at room temperature, said method comprising maintaining the liquid pharmaceutical composition at a pH of about 6.0, wherein the liquid pharmaceutical composition comprises: fentanyl; 15 a phosphate buffer solution; sodium chloride, wherein the concentration of fentanyl is between 100 pg/ml and 400 pg/ml; the liquid pharmaceutical composition is sterile; 20 the liquid pharmaceutical composition is sealed in a vial, cartridge, applicator, or container; wherein the vial, cartridge, applicator, or container: is labelled with instructions that the composition is used for the treatment of pain; is labelled with instructions that the composition has a shelf-life of at least 25 30 months from the date of manufacture when stored at room temperature; is adapted for the intranasal delivery of the pharmaceutical composition in a spray dosage form; and is adapted for delivery of a droplet size generated during administration between 2 pm and 50 pm. 30

120. The method of claim 119, wherein the fentanyl is fentanyl citrate.

121. A method of extending the shelf-life of a pharmaceutically acceptable fentanyl composition as claimed in claim 1 or claim 119, substantially as herein described and illustrated. WO 2010/005400 PCT/SG2009/000248 -110

122. A composition as claimed in claim 19, substantially as herein described and illustrated.

123. A method of minimizing the sorption of fentanyl compounds as claimed in claim 25, substantially as herein described and illustrated. 5

124. A method of preparing a sterile intranasal pharmaceutical composition as claimed in claim 28, substantially as herein described and illustrated.

125. A pharmaceutical composition as claimed in claim 37, substantially as herein described and illustrated.

126. A method of managing or treating pain as claimed in claim 76, substantially 10 as herein described and illustrated.

127. A dosage form as claimed in claim 104, substantially as herein described and illustrated.

128. Use of fentanyl as claimed in claim 118, substantially as herein described an illustrated. 15