N0. . 2
`
`519
`
`'
`
`[Chem. Pharm‘ Bull.)
`29(2) 519~531 (1981)
`
`Studies on the Absorption of practically Water—insoluble Drugs following
`Injection.
`1.
`Intramuscular Absorption from Water-
`immisciblc Oil Solutions in Rats
`
`KOICHIRO HIRAN0,* TERUHISA ICHIHASHI, and Home YAMADA
`
`Skioaogi Research Laboratories, Shioaogi and Ca., Ltd, Segz'm 5—13—11,
`Fukaskima-kn,. Osaka, 583, japan
`
`7
`
`(Received Iuly 30, 1980)
`
`The absorption behavior following intramuscular injection of practically water—
`insoluble drugs in Water-immiscible oil solution was investigated with several azo dyes
`and two steroids as model drugs by the iooal clearance method in the m. gastrocnemius
`of the rat. The absorption of the drug component obeyed approximately lst order
`kinetics, while the absorption of the oily solvent was very ‘slow. The injection volume
`(V0) influenced the absorption rate constant (k) and the correlation 1.: cc; V0“ was experi-
`mentally observed (for intact rats, m=—0.14; ior anesthetized rats, m:40.32). This
`was also the case at another injection site,
`the m. ream: femoris in rats. Comparison
`of the absorption rate of a drug in various oil vehicles showed that A: was controlled predo-
`minantly by the oilfwater distribution coefficient (K) and depended little on the viscosity
`of the vehicle. These results suggested that the release process of the drug component
`from the oily depot to the aqueous phase around it was the main route for absorption,
`and that the subsequent transport process in the aqueous phase might be ratevlimiting.
`A plot of log 3; versus log K gave a straight line with a slope close to — 1J but of a slightly
`smaller absolute value. This plot could be applied satisfactorily for estimating the
`absorption rate of other drug-oil systems. A guideline for predicting the absorption rate
`of a. drug in oily suspension from the It Value for its solution is presented.
`
`
`local clearance
`intramuscular injection;
`drug absorption kinetics;
`Keywords
`method ;_ practically water-insoluble drugs; water~immiscible oil vehicles;
`injection
`volume; distribution coefficient;
`intact and anesthetized rats
`
`Parenteral routes of drug administration are as useful and important as oral ones for early
`screening and preclinical testing of drugs in animals. There are many kinds of injections,
`including intramuscular, subcutaneous, intravenous, intradermal, hypodermal, intraarterial,
`intrapleural, intraperitoneal, intraarticular, intracardial, intraspinal and intracerebral. Ab«
`sorption is not involved when a drug is administered parenterally by an intravasular route.
`However, when the drug is administered by any extravascular' route, a depot of some type
`is formed and the drug must leave the depot and reach the blood or lymph Systems by some
`process (es).
`Since the publication of several reviews“ there has been increasing interest in the absorp—
`tion behavior of drugs in aqueous solution from intramuscular“ and Subcutaneous sites.“
`However, compared to water~soluble drugs, little Work has been done on practically water-
`insoluble drugs. These drugs are commonly administered to laboratory animals as parenteral
`preparations in the form of oily solutions (oily suspensions), aqueous suspensions, aqueous
`solutions solubilized with nonionic surfactants, or emulsions. The absorption of such drugs,
`which depends on the dosage form or formulation,
`is in most cases slow enough to be rate—
`determinant in their disposition in the body. This means that even pharmacological responses
`are sometimes governed primarily by the formulation itSelf. Therefore, basic inVestigation
`on drug absorption from such dosage forms by each parenteral route is required in order to
`select the optimal preparation and parenteral route for more complete screening tests in ani—
`mals.
`
`The present study was undertaken to elucidate the absorption behavior of practically
`water—insoluble drugs from water—immiscible oil solutions injected intramuscularly. Early
`
`NI I-Electronic Library Service
`
`AstraZeneca Exhibit 2116 p. 1
`InnoPharma Licensing LLC V. AstraZeneca AB IPR2017-00905
`
`

`

`
`
`520 Vol. 29 (1981)
`
`investigators“ noted mainly the vehicle effect on the pharmacologic responses or blood levels
`of drugs but did not refer to the mechanism or kinetics of drug absorption from such prepara-
`tions. Recently, Tanaka et at.“ studied these problems using slightly water—soluble drugs
`in anesthetized and operated rats.
`For our study, several azo dyes and some steroids which are unionized and practically
`water-insoluble under physiological conditions Were used as model drugs, and their intramus-
`cular absorption properties were examined by the local clearance method in the m. gastrm
`enemies of the intact rat. Under various experimental systems and conditions, the intramu—
`scular absorption rates of these model drugs were compared to evaluate the contribution of
`. physicochemical factors and to clarify the kinetic process. On the basis of the results, predic-
`tions of the absorption rates of other drugs in oily solutions were attempted. The relationship
`between absorption rates in oily solution and oily suspension was also investigated.
`
`
`
`Experimental
`
`Azo dyes such as p—aminoazobenzene (PAAB), 1b~hydroxyazobenzene (PHAB), o—amino—
`Materials
`azotoluene (OAAT), 1—phenylazow2~naphthylamine (PANA) and tetrazobenzenerfilinaphthol
`(Sudan III),
`and steroids such as testosterone (TS) and 205,3oc-epithio-Soc-androstam17:3»01 (epitiostanol) Were selected as
`model compounds for practically water-insoluble drugs. These compounds (other than epitiostanol) were
`obtained commercially: PAAB, PANA and TS, Tokyo Kasei Kogyo Co., Ltd. (Tokyo); PHAB, Eastman
`Kodak Co. (N.Y.); OAAT, Ishizu Pharmaceutical Co., Ltd. (Osaka); Sudan III, E. Merck AG (Darmstadt).
`These compounds were of reagent grade and were used without further purification, except for PHAB,
`which was purified by being dissolved in MeOH then recrystallized with water. Epitiostanol was synthesized
`in our laboratory and was of medicinal grade. As water-immiscible oil solvents for the injection preparations,
`sesame oil (SO), medium chain (Cs—C12) triglyceride commercially named Miglyol 812 (Mig), isopropyl myris-
`tate (IPM), diethyl sebacate (DES) and castor oil (CO) were selected in view of relatively wide variety in
`viscosity and dissolving powder. These compounds were also obtained from commercial sources: SO,
`Maruishi Pharmaceutical Co., Ltd. (Osaka); Mig, Chemisohe Werke Witten (Germany); 1PM and DES,
`Nikko Chemicals (30., Ltd. (Tokyo); CO, Kenei Pharmaceutical Co. (Japan). 80 and CO were of JP grade
`and the others were of reagent grade. All were used without further purification. All other chemicals
`
`used in this investigation were of analytical or reagent grade,
`Test Injection Preparations
`An oily solution was prepared by dissolving the desired amount of a
`test compound in an oily solvent and filtering it through an SM 116 membrane filter (Sartorius-Membranfilter
`GmbH, Gottingeu). All test solutions used here were ascertained to be chemically and physically stable
`for at least the experimental period. An oily suspension of PAAB or T8 was formulated in a room held at 37°
`by dispersing the sieved powder (for PAAB, 74—149 pm in diameter; for TS, 37fi74 and 74—149 pm) in the
`oil vehicle presaturated with each drug using a mixer (Micro Thermo Mixer Model I‘M-101, Thermonics
`(30., Ltd. , Tokyo), and the product was stored at 37“ until use. All suspensions were used about 20 hr after
`preparation.
`Animal Experiments—-Male Wistar albino rats Weighing 250i30 g were used in all animal experiments.
`(i) Absorption Experiment Procedure: The injection site for the present study, unless otherwise
`mentioned, was the medial head of the m. gastrocnemius in the m. triceps sum»: of the left hind leg in the
`rat; this site was selected because the location of the needle tip could be easily controlled. The absorption
`time course was followed by the looal clearance method. The rat 'was given light anesthesia with ether,
`then a thin Terumo needle (27G X 314” for oily solutions and 25G x 1” for oily suspensions; Terumo (30., Ltd,
`Tokyo) connected to a volume sca1e~corrected syringe or Terumo micrometer syringe (MS 10, 50 and 100)
`was inserted a. few millimeters above the ankle and the tip of the needle was led to the center of the m. gastro—
`memtus, medial head. Next, the test solution (50 ul, unless otherwise stated) was injected at moderate
`speed.
`Immediately after withdrawal of the needle, the quick adhesive Aron Alpha (Toa Gousei Kagaku
`Kogyo (30., Ltd, Tokyo) was applied to the insertion site in order to prevent leakage of the 0in solution
`injected. During the absorption experiment, the rat was housed in a cage which allowed free movement
`and easy access to water and food. At various intervals after the injection, rats were decapitated and bled,
`and the musele tissues around the injection site, including the oily depot, were excised as completely as
`possible. The removed muscles were mixed with 5 m1 of water and homogenized with a high speed blender
`(Ultra—Turrax, Janka and Kunkel K.G., Ger.) under cooling in an ice-water bath. Another 5 m1 of water
`was used to wash the blender of the homogenizer and this was added to the above homogenate. After
`extraction from the homogenate with 8 ml of ethyl acetate, the residual amount of the drug in the removed
`muscles was analyzed. For absorption studies with sesame oil (SO), the following indirect method was
`adopted. Fifty microliters of 80 solution containing Sudan III (5 mgliml) was administered to the Same
`site in the manner described above in two groups of rats. = At a set time interval, both groups were sacrificed.
`
`NIIeElectronic Library Service
`
`AstraZeneca Exhibit 2116 p.
`
`

`

`521
`.
`No. 2
`——ew-—~e—r————-—————m—~—.—,—.—__—___fi__m___—_ww
`
`From one group, the residual amount (W) of Sudan III was determined, and from the other, its concentration
`(C) in SO solution remaining in the muscle was determined. Since Sudan III was pooled and kept homo-
`geneously dissolved in the SO depot for at least 20 hr after injection, the volume (V) of 80 remaining in the
`muscle could be approximately calculated as V=W[C.
`(ii)
`In Vim» Incubation Experiment:
`In order to ascertain whether the clearance of a giVen‘ drug
`from the injection site could be attributed to its transport into the vascular systems or not, an' in vitro incuba—
`tion experiment was undertaken along with the absorption experiment, and the metabolic change at the
`injection site was checked. Fifty microliters of SO solution containing 0.5 mgfml of an azo compound was
`injected into the same site in the manner described abovr: in two groups of rats. For one group, the muscles,
`including the injected solution, were removed immediately after administration, then incubated in 5 ml
`of Ringer-Locke solution under bubbling with air and mild shaking at 37° for 2 or 6 hr (Yamato incubator,
`model BT41; Yamato Scientific Co., Ltd, Tokyo). After incubation, the residual amount of the test com‘
`pound was analyzed and compared with that obtained from the absorption experiment in another group of
`rats. Visible absorption spectra Were also measured before and after the above experiments with a Hitachi
`BPS-3T recording spectrophotometer (Hitachi (20., Ltd., Tokyo) to check for metabolic change of the test
`compounds.
`
`Kinematic viscosities of oily solvents Were determined with an
`Measurement of Absolute Viscosity
`Ubbelohde viscometer (Kaburagi Kagaku Kikai Kogyo 00., Ltd. , Japan) and their densities were measured
`by using a volume scale-corrected Cassia flask. These determinations were performed at 37”, and the absolute
`
`viscosity was calculated from the kinematic viscosity and density.
`Determinafion of the Apparent Distribution Coefficient
`Ten milliliters of 0.9% (w/v) NaCl aqueous
`solution (saline) was added to 4 ml of an oily solution containing the model compound (concentration: for
`PAAB, PHAB and OAAT, 5—40 mgfml; for TS, 2.33—4.65 rug/ml; for epitiostanol, 275 rug/nil) and shaken
`well at 37° with a mechanical shaker until distribution equilibrium was reached. Our preliminary experiment
`had demonstrated that 2 hr was needed for this equilibrium, and that the volume change of each phase after
`equilibrium was small enough to neglect. After equilibrium had been reached, the unstable emulsion was
`centrifuged at 4000 rpm for 10 min to separate the transparent aqueous phase, part of which was withdrawn
`for analysis. The concentration of the model drug in the oily phase was calculated from its concentration
`determined for the aqueous phase and its initial concentration in the oily phase. The apparent distribution
`coefficient (K) was represented as the concentration ratio betWeen the two phases. The value of K as deter-
`mined above depended little on the Concentration and was similar to the solubility ratio at 37° between
`the oil and saline solvents. For PANA, whose concentration in saline was too low to be determined accurate—
`
`ly, the solubility ratio between the two solvents was used as the apparent distribution coefficient.
`Analytical Method
`(i) Samples from Animal Experiments: Sample ethyl acetate solutions contain-
`ing azo compounds were adequately diluted with ethyl acetate and their optical densities were measured with
`a Hitachi UV—VIS spectrophotometer at 384, 348, 386, 450 and 510 run for PAAB, PHAB, OAAT, PANA
`and Siidan III, respectively. Epitiostanol was determined by GLC after being transformed to olefin (505»
`androst-en—l 75-01)
`in the following manner.” Five milliliters of sample solution (epitiostanol content,
`0-460 pg) 'was shaken with 1 g of A1303 (active, neutral, I, E. Merck, Darmstadt) for 10 min and centrifuged
`at 3000 rpm for 5 min (this procedure was done to minimize interfering materials in the following GLC assay).
`To 1.5 m1 of the supernatant was added 0.5 ml of internal standard solution (Bfi—acetoxy-Ea-androstan in
`ethyl acetate, 0.25 rug/ml), then the mixture was evaporated to dryness in a Vapour Mix instrument (Tokyo
`Rikakikai Co. , Ltd., Tokyo). The residue was dissolved in 5 ml of benzene and refluxed with about 30 mg
`of Spongy active copper (Cu—Zn) for 40 min. The reaction mixture was filtered through Toyo filter paper
`No. 131, and 2 ml of the filtrate was. evaporated to dryness. The residue was redissolved in about 200 yd
`of ethyl acetate and 1W2 ul of this was analyzed with a gas chromatograph, model GLC—4APTF, equipped
`With a flame ionization detector (Shimadzu Seisakusho, Ltd., Kyoto) under the foll0wing conditions. The
`column was a 1.5 mx 5 mm glass tube packed with 3% 813-30 on Gas Chrorn Q (80*100 mesh); carrier gas,‘
`99.999% N, (flow rate, 60 mlj'min); column, injection port and detector temperatures, 250, 268 and 288°,-
`respectively.
`In the case of TS, 3 m1 of sample solution was shaken with 0.6 g of an A'lzoausilica gel” mixture
`(1: 1, w/w) for 10 min then centrifuged at 3000 rpm for 5 min (this procedure was performed for the same
`reason as in the case of epitiostanol). To 1 m1 of the supernatant was added 100 pl of internal standard
`solution (35-acetoxy—5oc-androstan in ethyl acetate, 0.5 rug/ml). Two microliters of this mixture was used
`for GLC analysis'as described above (column, injection port and detector temperatures were modified to
`260, 280 and 293°, respectively).
`‘
`,
`‘
`(ii) Samples from Other Experiments:
`(1) Test Compounds in Oily Solutions: Sample Solutions for
`8.20 compounds were analyzed colorimiytrical‘ly after dilution with chloroform (PAAB, 372 nm; PHAB, 345
`nm; OAAT, 376 nm; PANA, 436 nm; Sudan III, 520 um). Epitiostanol and T8 were assayed by GLC as
`described above after dilution with ethyl acetate.
`(2) Test Compounds in Saline: Sample solutions of azo
`compounds and TS Were analyzed spectrophotometrically as follows: PAAB and PHAB at 376 and 348 nm,
`respectively, after dilution with distilled water; OAA'l‘ and PANA at 379 and 470 nm, respectively, after
`ION-fold dilution with EtOH; TS at 241 nm after 10sfold dilution with EtOH. Epitidstanol was assayed
`by the GLC method described aboveafter extraction with ethyl acetate.
`‘
`
`NI I—Electronic Library Service
`
`AstraZeneca Exhibit 2116 p. 3'
`
`

`

`
`
`
`
`522 Vol. 29 (1931)
`
`Results
`
`Comparison of Clearance of Model Compounds in the Injection Site bctWeen in Vitro Incubation
`and in Vivo Absorption Experiments
`
`To date various experimental procedures have been proposed and used to evaluate the
`absorption rate of drugs parenterally administered: (1) observation of changes in the pharmaco-
`logical effect with time; (2) following the blood level and]or urinary excretion, and pharmaco—
`ltinetic analysis of the data; (8) following drug clearance at the injection site (local clearance
`method). We adopted the local clearance method since a sample oily solution injected
`intramuscularly remains localized, allowing almost complete recovery of the remaining drug.
`Apparent elimination of the drug observed by the local clearance method sometimes results not
`only from its transport. to the blood or lymph system but also from its metabolic change at
`the injection site. Hence: in order to check the contribution of the latter factor in the case of
`the model compounds, an m m'im incubation experiment with removed muscle tissues was
`undertaken and the resulting data of % recovery for agiven period were compared with those
`from the in viva absorption experiment. Table I gives this compariSOn for eachlcompound
`in the 80 solution. For all compounds, the recovery was near 100% for the in Miro incubation
`experiment but considerably lower for the in nine absorption experiment. Further, no signifi—
`cant change oi the visible absorption spectrum was observed before and after these two experi-
`ments. These results showed that the metabolic change of these model compounds in the
`muscle was almost negligible and their apparent clearances observed here resulted mainly from
`transport into vascular systems. These findings support the view that the model compounds
`and the absorption experiment procedure adopted here are appropriate for analyzing the true
`absorption phenomena of practically water—insoluble drugs injected intramuscularly.
`
`TABLE I. Comparison of % Recovery between in Wm Incubation with Removed
`Muscle Tissues and in Vivo Absorption Experiment“
`
`
`Compound
`Time (hr)
`
`
`'
`
`% recovery”)
`,__’-._____\
`In Mira
`In vivo
`
`33 l (3.7)
`97 0 (2.9)
`2
`PHAB
`52 9 (5.2)
`98 9 (5.8)
`2
`PAAB
`55 0 (8.5)
`99 6 (4.3)
`6
`OAAT
`91 3 (4.2)
`100 0 (5.7)
`6
`PANA
`
`
`
`6 100 5 (0.9)Sudan III 98 8 (0.8)
`
`a} Vehicle, 80; initial concentration ((3,), 0.5 mglznl; injection volume (Va), 0.051111.
`1:) Each value represents the mean of 3 experiments with the standard deviation in partmthtues.
`
`Clearance of Sesame Oil from the Injection Site
`
`Various vegetable oils such as sesame oil, olive oil, cotton seed oil, peanut oil and castor
`.
`oil are usually used as oily solvents for preparations. Sesame oil (80) is most commonly used
`in japan. Since theSe oils consist not of a single component but of various triglycerides such
`as olein, linolein, stearin, palrnitinand myristin, and other natural products, direct evaluation
`of their absorption characteristics from the injection site seems impossible. Recently, some
`reports have appeared on intramuscular absorption of mineral 0ii,8’ 1'=‘*C—methyl oleate,“ and
`14C—tripalrnitin mixed with sesame oilf“ All these investigations indicated that these oils
`are absorbed very slowly.
`An oil vehicle may be eliminated from the injection site mainly via one or both of two
`routes: (a) absorption after being dissolved in the body fluids or transformed into some hydro—
`philic compounds and (in) direct absorption after division into micro—droplets. Table I shows
`that Sudan III, which could scarcely be released from the SO vehicle into the aqueous phase,
`
`NI I—Electronic Library Service
`
`AstraZeneca Exhibit 2116 p. 41
`
`

`

`
`
`
`
`No. 2 523W
`
`was insignificantly absorbed for at least 6 hr after injection. Thus, with Sudan III as a tracing
`material, the absorption of 80 was examined by an indirect method, as described in “Experi—
`mental.” The results appear in Table II. The residual volume (V) 'of SO and the residual
`amount (W). of Sudan III 6 hr after injection had little changed from their initial values, but
`subsequently decreased gradually.
`' This observation showed-that 80 was, scarcely absorbed
`during the first several hours after injection, but that after this it might be very slowly
`absmbed, presumably via both routes mentioned above.
`
`TABLE II. Absorption of Sesame Oil Estimated by the Indirect Method
`
`
`
` Time (hr) . W (03)“) 6 (than) V 0.1)»
`
`
`
`
`
`
`
`
`
`53.6 (0.4)
`5.07 (0.01)
`272.0. ( 2.1)
`0
`53.5 (0.2)
`5.03 (0.02)
`268.7.( 1.0)
`0
`48.6 (1.9)
`5.00 (0.13)
`242.97( 7.3)
`r
`20
`
`
`
`239.4 (24.3) 5 26 (0.28)41 45.5 (5.2)
`
`(1) Residual amount of Sudan III.
`b) Sudan III concentration in the oily depot phase.
`6) Calculated residuai volume of SO. Vz'WIC. Each result is given as the mean of at least 5 experi-
`ments foilowed by the standard deviation in parentheses.
`
`Time Course of Drug Absorption and Efiect of Initial Drug Concentration
`
`Fig. 1 shows a tirne course of intramuscular absorption of PAAB after injection of its
`80 solution. The initial concentration (C0) and injection volume (V0) are given in the legend.
`The upper figure (A) shows the value of the % remaining on a linear scale while the lower one
`(B) uses a logarithmic scale. From the linearity shown in Fig. 1B, this compound was expected
`to be absorbed mono-exponentially, that is, according to a let order rate process.
`In order
`to confirm this, the effect of the initial concentration (C0) of PAAB in the SO solution on absorp- -
`tion was examined.
`Fig. 2 shouts a comparison of time courses of PAAB absorption on
`a semilogarithrnic scale'from three solutions of different C“. All these absorption profiles gave
`
`%remaining
`
`Fig. 2. Effect or Initial Concentration on
`PAAB Absorption following Intramus-
`cular Injection (SO)
`
`
`
`%remaining
`
`Time (hr)
`'
`'
`‘
`_
`I
`Fig. 1. Linear (A) and Semilogarithrmc
`(B) Plots of % Remaining PAAB at the
`Injection Site versus Time (SO)
`Each point represents the mean of 4 experiments.
`The vertical be: about the mean shows the standard
`deviation.
`Ca, 5 rag/ml; V", 0.05 ml.
`
`'
`
`Key: 0, 0.5 rag/ml; A, 40 mglml; solid line, 5
`mgfrnl (shown in Fig. 113). Each point represents
`the mean or at least 3 experiments. V“: 0.05 ml.
`
`NI I—Electronic Library Service
`
`AstraZeneca Exhibit 2116 p. 5:
`
`

`

`
`
`524 Vol. 29 (1981)
`
`straight lines with the same slope, not depending on C0. Other model compounds such as
`PHAB, OAAT and PANA in 80 solution also showad similar behavior. These results were
`analogous to those obtained by Tanaka et at.“ who studied slightly waterwsoiuble drugs and
`oil systems. Others have also reported that intramuScular absorption of neutral drugs in
`. aqueous solution, different from that of acidic or basic drugs, depended little on their initial
`rconcentrationfia’m’ Therefore, it seems reasonable to conclude that intramuscular absorp-
`tion of neutral and practically water~insoluble drugs in water-immiscible oil solutions obeys
`approximately lst order kinetics.
`
`Effect of Injection Volume
`
`Injection volume is one of the important factors which influence drug absorption. Fig. 3
`shows the efiect oi the injection volume on the absorption time profiles following intramuscular
`administration of PAAB in 80. The slope of these profiles increased gradually with decreasing
`injection volume. The quantitative relationship is discussed later.
`
`0
`
`2
`
`Time
`4
`
`6
`
`8
`
`(hr)
`
`as 20
`
`is” 50
`E 40
`5
`t.
`
`100
`80
`
`E”
`.5
`E
`El
`x
`
`Injection Volume on
`Fig_ 3_ 7 Effect of
`Intramuscular Absorption of PAAB from
`50 501%“
`
`Key: JV“ 2'“ '“ll “on “A; _A"10‘u1: _O_' 30
`n1; —.—, 100 311. Each point represents the mean of
`4 or 5 experiments. Co: 5 mgjml.
`
`Intramuscular
`Fig. 4. Comparison of
`Absorption of PAAB from Various Oily
`Solutions
`Key: —1::—. DES.- —a—. Mie; ~v-. co; —<>—, 1PM;
`—O—, SO. Each point. represents the mean of 4 or 5
`experiments. Co, 5 rug/nil; VD, 0.05 ml.
`
`Comparison of Absorption Rates from Various Oily Solvents
`In parenteral preparations for animal experiments, oily solvents other than sesame oil
`are used with or without some adjuvants. Fig. 4 compares the intramuscular absorption
`rates of PAAB in five 0in solvents, DES, Mig, CO, IPM and SO. Each semilogarithmic plot
`was nearly linear with a different slope. Similar results were also obtained for other com—
`pounds, PHAB and OAAT. The major factors which resulted in the remarkable differences in
`the absorption rate from different vehicles will be discussed below.
`
`Discussion
`
`Kinetics of Drug Absorption
`Visual observation throughout all our absorption experiments confirmed that dye solution
`injected intramuscularly was confined to the planes of fascia or connective tissues surrounding
`muscles, did not spread as extensively along the faScial septums as did the aqueous systems,
`and formed a depot which took a flat shape similar to a pod, The geometrical change of the
`oily depot was more abrupt within a short period during and immediately after injection but
`gradually became less with time. Our observations were similar to those of Scaffer.m
`When a practically water~insoluble drug in a water—immiscible oil solution is injected
`intramuscularly, two possible routes for the absorptiOn of the drug component may be con-
`sidered: (1‘) absorption by direct transport of small oil droplets containing the drug and (£13)
`
`NII—Electronic Library Service
`
`AstraZeneca Exhibit 2116 p. 6
`
`

`

`
`
`No. 2 525
`
`absorption by the same mechanism as an aqueous Solution after drug release from the 0in
`phase into the aqueous phase. Theabsorption data for the oily solvent (SO) shown in Table
`II imply that for several hours after injection, 80 was scarcely absorbed. Thus, the Second
`route (it) is expected to be the major one for drugs which are more quickly absorbed than the
`oily solvent,
`'
`Thus, assuming the driving force for drug transport to be diffusion, a model was proposed
`to clarify the absorption process kinetically, Fig. 5 shows the absorption process of the drug
`component diagrammatically with a two—dimensional model.
`.The broken lines represent
`the concentration gradient along the distance axis perpendicular to the interface (i) between
`the oily and aqueous phases. First, the drug molecule is transported in the oily phase (I)
`to the interface (i), where it is then distributed into another phase (II).
`Subsequently, the
`drug molecule in the aqueous phase (II) diffuses with or without the aid of body fluid flow
`through the biological matrix, including cell membranes or intercellular spaces of connective,
`muscle and vascular tiSsues, into the blood or lymph stream. The shape of the oily depot
`in the injection site may fluctuate continually in response to the motion of the Surrounding
`muscles and this may cause some convection flow, that is, agitation within the oily depot
`phase (I).
`In the aqueous phaSe (II), the flow of body fluid is expected to be vigorous, espec—
`ially in the region adjacent to vascular systems (III).
`If we assume on the basis of these
`considerations that the rate—limiting process of drug absorption may exist in the diffusion
`in both phases close to the interface (i), the fluxes of the drug molecules per unit time, ]9 and
`j“ are given by the following equations according to Fick's diffusion law:
`6C
`I, = _po ( 3X )Hl
`
`BC
`n = 4):. ( aX)H1
`
`(0<X<Xi)
`
`(X>Xo
`
`(Eq. 1)
`
`(Eq- 2)
`
`where D, and D, represent diffusion coefficients in the oily and aqueous phases, reSpectively,
`and (BCfBX) signifies the concentration gradient of the drug in the oily (O<X(X1) or aqueous
`1(X>Xi) phase. Assuming that the thickness of the diffusion layer may be approximately
`regarded as constant, Eqs. 1 and 2 become:
`'
`
`fa =ko(coil—C,oll)
`
`Ia = kRCC'wa)
`
`'
`
`3)
`
`(EQ- 4)
`
`where Fe, and k, are defined as parameters related to transport in the oily and aqueous phases,
`respectively, and C0“ and Ch repesent concentrations in the oil bulk phase and blood, respec—
`
`
`
`0
`
`.
`X
`X1
`Fig. 5. Model for Intramuscular Drug
`Absorption from Water-immiscible Oil
`Vehicle
`‘
`71 : Oily depot phase,
`JI : Aqueous body fluid phase containing various tis-
`sues adjacent to 0in depot,
`lll: Blood or lymph.
`
`'
`
`1.0
`0.7
`0.5
`
`7,, 03
`:5
`an
`
`0.1
`0.07
`
`9x41
`I
`
`.
`
`A
`
`\
`
`3
`
`5
`
`50 70100
`
`,
`
`30
`7 10
`f
`.
`I” (#1)
`Fig. 6. Relation. between Absorption Rate
`Constant (k) and Injection Volume (V0)
`Key: —Q-, intact, m. gasiracnemius (0,,‘5 rug/ml);
`-<>—, intact, m. Icarus frmorfis (CD, 10 mglml); —A—7,
`anesthetized, m. gastracnmius (Cu, 10 mg/ml).
`Test compound, PAAB; vehicle. 50.
`
`NI IwEJectronic Library Service
`
`AstraZeneca Exhibit 2116 p.
`
`

`

`526
`
`Vol. 29 (1981)
`
`tively ; C ’01, and C ’,, mean the concentrations in the oily and aqueous phases at X :X1, respec«
`tively. Under the condition that C ’,,>>Cb, Eq. 4: can be approximated as:
`
`In = kacla
`
`(Eq. 5)
`
`Assuming that the distribution equilibrium of the drug molecule at the interface is reached
`quickly compared withits transport, the distribution coefficient (K,} at the interface is defined
`by
`
`K1 : can/cs
`
`‘
`
`a
`
`‘
`
`(Eq. 6)
`
`After a quasi—steady—State is reached at time t, the following relationship is established:
`
`-dW/dt = [110 = Aft (bf)
`
`(39- 7)
`
`where W represents the amount of drug in the oily depot phase and A is the surface area of
`the oily depot. W is given by the volume of the oily depot, V, and the drug concentration
`in it, Con, according to the following equation:
`'
`
`W 2 CoilV I
`
`From Eqs. 3, 5, 6, 7 and 8, the following equation can be derived:
`
`dW
`
`may" a r"
`
`kg
`
`L V(1+TKI)
`
`7
`
`W
`
`(Eq' 8)
`
`(Es. 9)
`
`Since the absorption of oil itself is very slow compared with that of the drug component, as
`shown in Tables I and II, V can be regarded as equal to the injection volume V0 for several
`hours after administration. Under the condition that the change in the surface area of the
`oily depot, A, is sufficiently small, Eq. 9 shOWs the lst order kinetics and can be integrated.
`When lag time r is sufficiently small compared with the abs0rption time scale, the following
`relations are readily derived:
`
`[MW/Wu} : 4a
`
`k =
`
`k0
`VD(1+”‘"‘K£)
`
`(Eq. 10)
`
`(Eq. 11)
`
`where W0 represents W at $20, that is, the dose, and k is defined as a lst order absorption rate
`constant. For the model drugs used in this study, it was observed that a large portion of the
`amount remaining in the injection site was localized within the oily depot phage. Accordingly,
`the value of W/W0 for these drugs is nearly equal to the residual fraction determined by the
`absorption experiment described above.
`‘
`The experimental results shown in Fig. 2 indicate that the absorption obeys lst order
`kinetics. Therefore, provided that the assumptions and conditions in the above consideration
`based on the model are Valid, the apparent lst order absorption rate constant experimentally
`obtained is governed by A, V0, [2,, k, and K,.
`
`Relationship between Absorption Rate Constant and Injection Volume
`
`Some papers have reported that the intramuscular absorption rate of a drug in aqueousmsm’
`or oily solutionfil'might decrease with increasing injection volume. However, the quantitative
`relationship between absorption rate and injection volume or the influence of anesthesia and
`the injection site on it has not been completely elucidated.
`Fig. 6 shows a plot of the let
`order absorption rate constant (k) against injection volume (V0) on a log-log scale.
`In this
`figure, the data obtained from'Fig. 8 are indicated by open circles. This plot gave a straight
`line with a ,slope of —O.lél (correlation coefficient, r:-0.95). The data from intramuscular
`injection into the other site, the m. reams famoris, in intact rats (indicated as open squares
`in Fig. 6} gave an almost identical straight'line.
`In anesthetized rats, as shown by open.
`
`HIE—Electronic Library Service
`
`AstraZeneca Exhibit 2116 p. 8
`
`

`

`
`
`r ‘No. 2 527
`
`
`
`
`
`i
`l
`
`triangles in Fig. 6, the absorption rate was remarkably sIOWer compared with the corresponding
`data in the unanesthetized (intact) rats. A plot of log k versus log V0 gave a similar straight
`line but with a larger slope, —0.32 (r=——O.97).
`Eq. 11 means that the absorption rate constant k is proportional to the term A [V9, which
`depends on the injection volume (V0). The term A/VD is proportional to 'V0—1/3 when the.
`oily depot retains a similar shape or dimensions for the change in V0, and it is constant when
`the oily depot eXpands two-dimensionally at a constant thickness with increasing V0. The
`relation k 0: Vow-32 observed in anesthetized rats may correspond to the former case, while
`k or. 170—014 observed in unanesthetized rats may correspond to an