.-A-.t~,-ii
`-_._'
`ELSEVIER
`
`International Journal of Pharmaceutics [05 (1994) 189-207
`
`international
`journal of
`_
`pharmaceutics
`
`Invited Review
`
`Release and absorption rates of intramuscularly
`and subcutaneously injected pharmaceuticals (II)
`
`J. Zuidema *, F. Kadir, H.A.C. Titulaer, C. Oussoren
`Department ofBiop.l1armaCc.ultr.‘s, Faculry of Pharmacy. Depa rtmenr of Bfapharmaceuricr PO Box 80032, Utrecht (Jnit=er.t'i.ty,
`3508 T3 Utrecht. The Netherlands
`
`
`
`(Received 22 July 1993; Accepted 22 November 1993)
`
`Abstract
`
`The rate and extent of intramuscular (i.m.) and subcutaneous (s.c.) drug absorption are very erratic and variable.
`The lipophilicity of the compound plays an important role. Aqueous drug solutions and suspensions of the more
`lipophilie compounds are often absorbed incompletely within the therapeutically relevant time. More hydrophilic
`compounds are absorbed completely. Injection depth, drug concentration and vehicle volume, pH-pK,
`relation,
`vehicle, eosolvents and surfactants have strong influences on the absorption profile of lipophilic drugs. Aqueous
`solutions of hydrophilic drugs are less sensitive to these factors. Drug solutions in oil and even suspensions in oil are
`often thought to be sustained release preparations. In fact, rapid absorption has often been observed. Slow release is
`not a property of the oily vehicle but is achieved by a high lipophilicity of the dissolved or suspended compound.
`Liposomal preparations are currently under investigation as i.m. and s.e. injectable sustained release preparations.
`Factors that induce drug release at the injection sites are the proteins and especially lipopruteins in the interstitial
`fluids, originating from serum filtrate and from turnover of inflammatory cells. Phagocytosis by macrophages and fat
`cells may play an important role in the local clearance of liposomal material from the injection site, Sustained
`release of some pharmaceuticals with normal or long half-lives appeared in specific cases preferable to rapid release.
`In addition, high arterial drug concentrations during the absorption phase may result in undesired effects cvcn when
`venous drug concentrations are within the safe range.
`
`Key words: Drug absorption; Intramuscular administration; Subcutaneous administration; Absorption rate
`
`1. Introduction
`
`The intramuscular and subcutaneous routes of
`drug injection are often used when drugs cannot
`be injected intravenously because of their low
`aqueous solubility and / or when high peak con-
`centrations, resulting in local or systemic side
`
`* Corresponding author.
`
`effects, occur with intravenous injection. More-
`over, additional advantages of these routes in-
`clude greater convenience,
`less problems with
`compatibility of the injection components with
`full blood in the ciculation and often less fre-
`quent administration when compared to intra-
`venous administration.
`
`Many variables are known to affect drug re-
`lease after intramuscular or subcutaneous injec-
`tion. Factors such as molecular size, pK,,, drug
`
`0378-5173/94/S07.0tl (E) 1994 Elsevier Science B.V. All rights reserved
`SSDI O3-78—5l73(93)ED34S-T
`
`Astrazeneca Ex. 2114 p. 1
`Mylan Pharms. Inc. V. Astrazeneca AB IPR2016-01326
`
`

`
`190
`
`J. Zm'r.’crrm er mi. _/i'uIt'r'm:.'iamrI .lum'mtl' ofPhrr.v'nurrt1trim‘ 1'05 {I994} .’«W—.?iJ7’
`
`injection
`initial drug concentration,
`solubility.
`depth, body movement. blood supply at the injec-
`tion site. ‘injection technique and properties of
`the vehicle in which the drug is formulated have
`been discussed extensively in a previous review
`(Zuidema et al.. 1988). This article is an update
`with emphasis on factors related to drug trans-
`port
`through the
`tissue.
`the
`role
`of drug
`lipophilicity. recent technology to modulate drug
`absorption from intramuscular and subcutaneous
`in_jection sites by carrier systems such as lipo-
`somes. absorptionby the lymphatic system and
`clinical implications.
`Drugs such as antibiotics, anti-asthmatics, anti-
`convulsics, anxioiyties and analgesics are often
`administered intramuscularly in severe disease
`states. A generally held viewpoint is that the drug
`is rapidly and completely absorbed from the in-
`jcction site. Previously published data have al~
`ready demonstrated that complete absorption
`during a time relevant for therapy is not true in
`every case (Ballard.
`I968: Dundee et al..
`[9742
`Kostenbauder et al.. I975: Tse and Welling, 1980}.
`however. recognition of their significance is lack-
`ing. Such findings may have important clinical
`implications.
`Consequently, this article is aimed at reviewing
`the relevant literature. in order to provide and to
`discuss material for the rational design of intra-
`muscular and subcutaneous drug formulations
`and to examine the clinical aspects of these types
`of injections.
`in contrast
`to the former review
`which was organised in order of the types of
`injection.
`this article is mainly ordered with re-
`spect to elements of the mechanism and further
`subdivided in types of injection.
`
`2. Drugs in conventional systems
`
`Conventional systems are solutions. emulsions
`and suspensions in aqueous or in oily vehicles.
`
`2. .t'. Dm.g.r in rapidly rciea.ring s_v.rtem.r
`
`2.1.1. Aquc'ott.r r'njcc‘.tion.r.‘
`rate
`
`t‘tt.r't'abffit)-' in (tf7.r0rpt‘i0r1
`
`It has frequcntiy been reported that absorp-
`
`0.6
`
`0.4
`
`absorbed
`Fractionnot
`
`.0.2
`
`0.0
`
`O
`
`30
`
`90
`SD
`Time (min)
`I. Fraction remaining to be absorbed after intratr1tIscti|ar'
`Fig,
`injection of In mg/l-tg sodium artelinatc aqueous solution in
`rabbits (H = rm,
`
`120
`
`tion after intramuscular and subcutaneous injec-
`tion is very variable (Gibaldi. 1977). This is first
`illustrated with some aqueous solutions.
`Artelirtic acid is a water-solubic derivative of
`
`artemisinin. an antimalarial drug, of the fast-
`acting schizontocidal
`type. Artemisinin and its
`derivatives are important new drugs- especially
`for the treatment of life~threatcning states of the
`disease. Artclinic acid is even more active than
`artemisinin itself. but
`is very rapidly eliminated
`after intravenous injection. Rapid and complete
`absorption is therefore essential. After intramus-
`cular injeetion in rabbits the absorption rate ap-
`pears to be very variable (Titulaer et al.. 1993}.
`This is depicted in Fig.
`1 where the fractions not
`absorbed are plotted vs time.
`The curves representing the fraction remaining
`to be absorbed suggest an apparent r.cro-order
`absorption. This
`is
`a
`rather unexpected phe-
`nomenon, since diffusion is characterised by a
`first-order mechanism. A possible explanation is
`at solvent
`flow dependent paraeellular transport
`of this highiy hydrophilic solute. This transport
`capacity is very variable. at
`least between sub-
`jects. and it appears that it
`is influenced by sev-
`eral factors including muscle activity.
`inflamma-
`tion and flow of the tissue fluid lZuidema et al..
`1988a). This explanation is supported by the next
`example.
`information on kinetic behaviour of
`Relevant
`i.n't. and s.c. injections has often originated from
`
`Astrazeneca Ex. 2114 p. 2
`
`

`
`J. Zuiclcma er al, /I/irernariomzlJmminf of Pfmrn2ac:mm'c.s‘ 105 ([994) 189-207
`
`19]
`
`veterinary studies. The risk of residual drug at
`injection sites is a considerable problem in meat
`consumption. Fig. 2 shows as a second example
`the large variation in absorption parameters after
`intramusc.ular and subcutaneous injection,
`in a
`fat-rich region. also referred to as intra-adipose
`injection (Kadir et al., 1990a). Carazolol
`is
`a
`,8-blocking agent which is used in veterinary prac-
`tice as a tranquillising agent in cattle and pigs.
`The fraction of earazolol absorbed during the
`first 24 h from an aqueous solution varied from
`24 to 59% after intramuscular and from 25 to
`66% after intra—adipose injection.
`Many factors which influence the variability in
`the rate and extent of absorption can be postu-
`lated. Firstly. a difference between intra- and
`intermuscular injection is postulated and defined
`as injections within and between the muscle fib-
`rils, respectively (Groothuis et al., 1980). Such a
`supposition must be supported by a bimodal sta-
`tistical distribution in absorption rate. In the lit-
`erature. however, experimental evidence for this
`contention is
`lacking. Secondly. differences in
`drainage and blood flow are possible explana-
`tions. The cause of these differences, however,
`remains unclear. Thirdly, differences in absorp-
`tion rate might also be result from differences in
`osmolality and other formulation factors, how-
`ever, such factors cannot explain variability with
`
`the same preparation and batch. Physiological
`circumstances that vary randomly and physiologi-
`cal reactions to the injection trauma might influ-
`ence absorption.
`A more likely explanation than those men-
`tioned above is a variation in the shape of the
`depot. The shape may vary from merely spherical
`to almost needle-shaped in different subjects.
`These differences depend on the local cohesion
`between the muscle components and the ten-
`dency to be torn open by the injection procedure,
`Differences in shape are accompanied by differ-
`ences in the depot surface (and therefore in the
`effective permeation area), the interface between
`depot and tissue and the absorption rate.
`
`2.J.2. Drug lipophilicity in aqueous systems.‘ extent
`of absorption and absorption rate
`Lipophilic compounds are slowly absorbed
`from intramuscular and subcutaneous injection
`sites (Zuidema et al., 1988). Recent findings show
`that absorption under such conditions often seems
`to be incomplete as well.
`It appeared that the
`apparent half-lives of midazolam in patients after
`intramuscular injection are much longer than af-
`ter intravenous injection, due to rate-limiting sus-
`tained release from the intramuscular injection
`site (Raeder and Nilsen, 1988). In a former study
`by the same group, a reduced apparent bioavail-
`
`lNTFlA-ADIPOSE
`
`INTRAMUSCULAFI
`
`
`
`ii
`
`20
`
`so
`40
`time (ruin)
`
`so
`
`100
`
`500
`
`400
`
`5'‘
`
`Q am
`g
`200
`100
`
`O I
`o
`
`20
`
`
`
`40
`
`so
`
`time (min)
`
`so
`
`in the serum of four pigs following intrarnuseular and subcutaneous
`Fig. 2. Individual concentration-time curves of earazoloi
`administration of0,l)2.5 mg/kg in a fat layer (intra-adipose injection).
`
`Astrazeneca Ex. 2114 p. 3
`
`

`
`192
`
`J. Ziiidermir e: at /lnrcrriaiianal Journal‘ o_fPharm:Jc-c>u:ic.v I 05 (.7904) I30-207
`
`ability of midazolam under these conditions was
`also observed {Raeder and Breivik, 1987).
`Phenobarbital appeared to be completely ab-
`sorbed afte.r
`intramuscular injection in dcltoid
`muscle in children, however,
`the bioavailability
`was only 80% relative to oral administration in
`adults (Viswanathan et al., 1978). Nevertheless,
`the number of investigated subjects was too small
`to permit definite and statistically warranted con-
`clusions. A lack of stability of phenobarbital (the
`amide bonds might be hydrolysed at the injection
`site) might be proposed as an explanation for the
`incomplete bioavailability. This possibility cannot
`excluded, however, it does not explain the differ-
`ence between children and adults nor the find-
`ings in the midazolam study. Further information
`is needed for a better understanding of the fac-
`tors which determine bioavailability in these spe-
`cific cases.
`
`The ,6-blocking agents are an ideal group for
`studying drug lipophilicity and release from intra-
`muscular and subcutaneous injection sites, since
`they have similar molecular weights and pK“
`values but differ markedly in iipophilicity. Studies
`in pigs using crossover experiments with propra-
`nolol, atenoiol, carazoloi, metoprolol and al-
`prenolol have recently been published (Kadir et
`al..
`l990a,b).
`The curves representing the fraction remaining
`to be absorbed of the 3-blocking agents, con-
`structed from intramuscular and subcutaneous
`
`(intra—adipose) plots and using intravenous data
`as references, demonstrate biphasic declines:
`ii
`rapid first phase followed by a very slow second
`phase (Fig. 3). Initial release rates appeared to be
`negatively correlated with drug lipophilicity ex-
`pressed as fat-buffer partition coefficients, espe-
`cially after injection in the subcutaneous fat lay-
`ers, also called intra-adipose layers. Propranolol
`showed greater and faster absorption than ex-
`pected from its lipoph.i1ieity only after intramus-
`cular, but not after intra-adipose, injection. Pro-
`pranolol is known to possess irritating properties
`which may improve blood perfusion in the mus-
`cles and account for the deviation in behaviour
`after intramuscular injection. The subcutaneous
`fat layer or adipose layer is less sensitive to such
`irritating properties and is less perfused.
`The extent of drug release within the men-
`tioned 24 h also transpired to be dependent on
`the lipophilicity of
`the compound:
`the more
`lipophilic the compound, the lower the bioavail-
`ability at 24 h after injection (the observation
`period) (Fig. 4). The most hydrophilic compound.
`atenolol, was the only one which was completely
`absorbed or bioavailabie within 8 h after intra-
`muscular injeetion and within 24 h after subcuta-
`neous injection (Fig. 3).
`lnjectcd drugs are probably rapidly absorbed,
`provided sufficient vehicle is present to maintain
`the drug in solution or to drive the absorption
`process. After the vehicle has been absorbed the
`
`IN TRA-AD I POSE
`
`|NTFlAMUSf}ULAR
`
`
`
`O
`
`250
`
`500
`
`7'50
`
`IODO
`
`3250
`
`
`
`0
`
`2.50
`
`500
`
`750
`
`It'll)
`
`1150
`
`15m
`
`I500
`
`
`
`fractionnotabsorbed
`
`0.‘
`
`(L2
`
`0.0
`
`time (min)
`time (min)
`Fig. 3. Fraction remaining to be absorbed curves (drug vs time) after intramuscular and inirzi-adipose administration of a series of
`[J—bl0Cl:ing agents. (El) Propranolol. (0) alprenolol. (0) metoprolol, (ll atenolol.
`
`CD
`
`Astrazeneca Ex. 2114 p. 4
`
`

`
`J. Zuidema at al. /International Journal of Pharmaceutics‘ 105 U994} i’c‘i‘9—207
`
`l9}
`
`the rate of
`to the oily system,
`In contrast
`dissolution of arternisinin in the aqueous injec-
`tion is slow and the process appears to cease
`alrnostiizompletely within the first few hours, the
`time during which the aqueous vehicle has been
`absorbed.
`
`In the preceding section, studies have been
`discussed in which the lipophilicity of a drug or
`model compound was the variable in a given
`aqueous medium. Interestingly. a study has ap-
`peared in which the drug was chosen as
`the
`constant and the lipophilicity of the oily vehicle
`was the variable (Table 1) (Al-I-lindawi er al.,
`1986}. The in vivo release of testosterone propi-
`onate in a number of oily vehicles was investi-
`
`u g! I
`400
`
`300
`
`200
`
`ugll
`400
`
`300
`
`Fig. 5. Plots of artemisinin concentrations in serum vs time
`(n=10) after a dose of 400 mg artemisinin to human volun-
`teers: (a) suspension in oil intrarnuscularly; (bl suspension in
`an aqueous vehicle intramuscularly.
`
`Astrazeneca Ex. 2114 p. 5
`
`__s
`
`At
`
`Me
`
`AI
`£
`
`o
`
`1
`
`2
`
`3
`
`Pr
`I
`
`4
`
`55
`E:-14
`
`K-
`
`4 E
`
`3
`
`8
`31
`E
`s:
`'2‘
`o
`
`fat partition coefficient
`Fig. 4. Correlation between the fat-buffer distribution con-
`stants and release rates on intra-adipose administration of
`atenolol (At). metopropo] (Me). alprenolol (Al) and propi-u-
`nolol (Pr).
`
`absorption rate of the drug decreases rapidly.
`This theory explains the midazolam studies but is
`also relevant in the case of the rapid absorption
`of artemisinin from an intramuscularly injected
`suspension in oil and the low and erratic absorp-
`tion of artemisinin from a suspension in water as
`shown in the following (Fig. 5).
`
`2.1.3. Oily injections; influence of the lipophiiiciry
`of the vehicle on the absorption rate
`injection.
`The next
`example
`is
`an oily
`Artemisinin is rather lipophilic and not soluble in
`water, however, it is also sufficiently insoluble in
`oil to allow its preparation as a dissolved injec-
`tion as of a conventional oil system with a suffi-
`ciently high dose.
`Oil systems and suspensions for injection are
`generally considered to be sustained release for-
`mulations. Therefore, the rapid onset of absorp-
`tion shown in Fig. 5a with the artemisinin suspen-
`sion in oil is striking (Titulaer et al., 1990b). The
`oily vehicle is absorbed only very slowly and re-
`mains present at
`the injection site for several
`months. Apparently, artemisinin dissolves rather
`rapidly in the oil phase and the dissolved fraction
`is then depleted. by further absorption.
`In the
`case of arternisinin. this depletion is apparently a
`rapid transit process over the oil
`to the water
`interface to the tissue fluids. This is less favoured
`in the case of highly lipophilic substances.
`
`

`
`N4
`
`J. Zm'zl'mm at u}. / i'mrr.'mn'mmi' Ji)m'rm.l nfPImr:mn'e::ri¢:r H15 {1994} 189-207
`
`I
`Table.‘
`TestoslL'rondui:ano:1te fat/phosp|1atehul‘l‘cr partition and ab-
`sorption rate expressed as half~lil‘e in the muscle
`
`Solvent
`
`Ethyl olezite
`Oelanol
`Isopropy] myrislate
`Light liquid paruffin
`
`Paniiion
`coeffieient
`tx In"-‘I
`raj
`5.3
`4.3
`1.3
`
`1, ,3 in
`muscle
`on
`tll.3
`9.7
`‘Lit
`3.2
`
`gated. Disappearance from the injection site ap-
`peared to be proportionally related to the in vitro
`partition coefficients. This study illustrates and
`emphasises the importance of the vehicle and the
`affinity of the drug to the vehicle for the disap-
`pearance process in the muscles after injection.
`
`2.1.4. hifluencc of infection mlrmre and drug can-
`nsn rm tiem
`Data on the influence of injection volume and
`drug concentration seem to be contradictory. They
`can be classified into the following four groups of
`results on: (i) hydrophilic drugs in aqueous sys-
`tems (concentrations are much lower than drug
`saturation in the vehicle);
`(ii)
`lipophilic drugs
`solved in oily systems;
`(iii)
`lipophilic drugs in
`aqueous systems [concentrations are close to drug
`saturation in the vehicle): and {iv} drugs in sus-
`pension. The latter is discussed in section 2.2.
`since most of the suspension formulations are
`intended to be sustained release formulations.
`The available
`information on the
`first
`three
`groups is summarised below.
`Atropine, sodium chloride, sugars and polyols
`such as mannitol and sorbitol. all hyctrophiifc drug
`solurioiis in aqueoim .‘i‘_!-‘.‘i'.lL’F'l't.S‘, were reported to be
`absorbed more rapidly when the compounds were
`administered in smaller injection volumes (War-
`ner ct al.. 1953: Schriftman and Kondtitzet. 1957:
`Sund and Schou. 1964). The common properties
`of these compounds are that they are very hy-
`drophilie, readily soluble in water and have low
`molecular weights. Examples of compounds with
`higher molecular weights are the clcxtrans, which
`appear to behave similarly but have a slower rate
`of absorption. The molecular weight appears to
`be inversely related to the absorption rate. The
`
`higher absorption rate in smaller volumes can he
`explained by the greater diffusion potential. Ob-
`viously. the absorption of this type oi" compound
`is controlled by passive diffusion or by paracellu-
`lat‘ transport.
`Results with amil-tacin. an antinoglycosidtr, are
`often misinterpreted (Pfeffer and Harl-ten. I981).
`In the available literature, these results are often
`wrongly discussed together with the first group.
`Amikacin is a hydrophilic compound but with a
`high molecular weight. In aqueous systems it is :1
`suspension, which is slowly absorbed from a sin-
`gle large depot. It is preferably given in multiple.
`simultaneously administered separate injections.
`in order to provide sufficiently high concentra-
`tions in serum. The solved umikacin fraction is
`constant and concentration is
`therefore not a
`useful absorption rate-determining variable vari-
`able. This in contrast
`to the situation with at-
`ropinc. sugars and polyols as discussed above.
`The effects of drugs which depress absorption
`rate are illustrated with atropine. At higher doses
`it exerts a sclf-depressive effect on the absorption
`rate by its parasympathicolytic activity. Effects of
`this type have scarcely been investigated and lit-
`erature data are sparse.
`Testosterone and some other model com-
`pounds represent examples of the second group.
`lipoplziiic‘ .i‘ub.si'a.-i.cc.s' in oil. They have been stud-
`ied in several oily vehicles (Tan-aka ct al.. 1974).
`This type of compound is absorbed more rapidly
`when it
`is dosed in smaller volumes of an oily
`vehicle. Again the diffusion potential is obviously
`the dominating absorption rate-determining lac-
`tor.
`
`injection volume on the
`Thc influence of
`hioavailability of lipophilic drugs in aqtteous sul-
`i"L’fli'S.
`the third group. has been studied in rats
`(Kadir et al.. 1992c}. The aforementioned study
`was performed in order to find an explanation for
`the incomplete absorption of the more lipophilic
`£1’-blocking agents as described in section 2.|.
`Propranolol was used as the model compound.
`Both the rate and the extent of absorption at it
`h appeared to increase with increasing irticction
`volume. When the vehicle volume is increased
`the residence time of the vehicle at the injection
`site increases. maintaining the drug in solution
`
`Astrazeneca Ex. 2114 p. 6
`
`

`
`J‘. Zuidcma rt £1. / lrrrvrwiatfrimrl Jrmrnzai of PImn11acc'u.n'c.t I05 (1 09-1) I89—.?07
`
`195
`
`0.40
`
`050
`
`0.20
`
`0.l0
`
`
`
`releaserateeonsthr‘
`
`one
`
`u
`
`so
`
`too
`
`203
`
`injection volumeflfl)
`Fig. 6. Individual release rate constants after intramuscular
`injection of 3 mg pmpranolol HCl
`in Si), 100 and 2130 iii
`aqueous solution in rats. Lines connect the individual values
`for the same rut (r: = ft).
`
`it
`in other words,
`for a greater length of time;
`remains dissolved over a longer time and absorp-
`tion is faster from the dissolved state (Fig. 6).
`Moreover,
`the increasing volume may increase
`the vehicle flow away from the depot This influ-
`ence is in a certain sense comparable with the
`action of the mobile phase in a chromatographic
`system. in HPLC terms: increasing solvent flow
`diminishes the retention time. This effect may be
`a possible explanation for the greater absorption
`rate during the initial phase.
`
`2.1.5. Influence of pH and /or cosoitlents
`Cosolvents such as ethanol, glycerol, propylene
`glycol and also polyethylene glycol 400 are used
`as solvents together with water, in order to en-
`hanee the solubility of certain drugs for injection
`(Chen-Der and Kent, 1982). They have varying
`influences on the absorption of drugs. Ethanol is
`better discussed separately in the next survey,
`since the mechanism of its influence deviates
`from those of the other cosolvents.
`
`is a liquid of very low viscosity and
`Ethanol
`exerts only a minor influence on the viscosity of
`the vehicle in mixtures with water. When used at
`high concentrations it has a denaturing effect on
`
`proteins at the injection site. It appears to have
`an inhibitory effect on the absorption rate of
`water-soluble ionic as well as non-ionic model
`
`drugs such as isonicotinamide, methylisonicoti-
`nate.
`isonicotinic acid and procaine hydrochlo-
`ride (Kobayashi et al., 1977). This influence could
`be attributed to alterations in the permeability of
`the connective tissue which was significantly de-
`creased. In chromatographic terms: ethanol seems
`to change the properties of the ‘stationary phase‘.
`probably by protein denaturation.
`Other cosolvents such as propylene glycol.
`glycerol and polyethylene glycol 400 have been
`reported eontradictorily to diminish and to en—
`hance the absorption rate of model compounds
`(Kakcrni et al., 1972; Cheng—Der and Kent. 1982).
`The diminishing effect
`is
`ascribed to the
`viscosity-enhancing influence of these compounds
`on the vehicle.
`Effects of eosolvents may partly be explained
`by a change of the properties of the ‘mobile
`phase’. The dielectric constant and consequently
`the hydro/lipophilicity of the solvent is changed
`to that of water. The enhancing effect of cosol—
`vents and the properties of the systems where
`enhancement occurs are discussed below,
`to-
`gether with the effects of pH-pK:, relationships.
`Additives with high molecular weights. known
`as viscosity enhancers, such as cellulose deriva-
`tives, dextrans or long—chain polyethylene glycols
`resulted in a degree of inhibition that was less
`than expected. This suggests that macromolecules
`apparently do not interfere with the absorption
`mechanisms.
`The absorption rate of cefonicid has been de-
`scribed as being pH-dependent (Brumfitt et al.,
`1988). However,
`the authors failed to give an
`explanation nor is it possible to analyse their
`results, since the experimental data were lacking
`in detail. Other studies shed more light on the
`pH dependency in i.m. and s.c. absorption pro-
`cesses.
`Salts with an alkaline or acidic reaction which
`can be neutralised by the tissue fluids have the
`potential to precipitate after injection due to the
`neutralising or buffer capacity of the tissue fluids.
`This has been briefly mentioned in the literature
`for quinidine hydrochloride as well
`(Tse and
`
`Astrazeneca Ex. 2114 p. 7
`
`

`
`E96
`
`J. Zutdetna at at’. /InrcrnatiamtiJrmrnrtiafP}1.m'mrteeuritir I105 {I994} 189-207
`
`is eiearly
`Welling, 1980). The effect, however,
`illustrated in the study using human volunteers by
`Kostenbauder et al. (1975) with intramuscular
`phenytoin (see section ll. Phenytoin is absorbed
`over a period of approx. 5 days. Even after 4U h
`20% of the drug remained unabsorbed. Pheny-
`toin is a drug which is dissolved for injection in
`relatively high concentrations at pH ll or higher,
`and also using cosolvents and/or eornpiexing
`agents.
`The authors reported that precipitation and
`slow redissolution of the drug by tissue fluids at
`the injection site could explain their results and
`developed a mathematical model based on this
`concept. The observed drug concentration curves
`in plasma fitted well with this model. Precipita-
`tion probably has two causes, i.e., the pH neutral-
`ising effect of the tissue components and the
`rapid absorption of some of the solvents.
`The influence of eosolvents on the absorption
`of salts is illustrated by a study of the effect of
`propylene glycol on the absorption of benzimida-
`zole hydrochloride (Cheng—Der and Kent. 1982).
`It appeared that propylene glycol, which appar-
`ently is absorbed more slowly than water, may
`prevent in certain circumstances, at least partly,
`the precipitation of the free base (or free- acid).
`In this manner,
`it might enhance the absorption
`of the drugs in question.
`
`2.1.6. Influence of surfactants
`The influence of surfactants and injection
`depth on the kinetics of absorption has been
`discussed previously (Zuidema et al., 1988}. Non-
`ionic surfactants at low concentrations probably
`exert a retaining effect on well-absorbed water-
`soluble drugs. however, the promotion of absorp-
`tion has been reported for a poorly absorbed high
`molecular weight polypeptide. The mechanisms
`underlying this phenomenon are still not well
`understood. Again the model of reversed-phase
`chromatography may be helpful.
`Surfactants in the mobile phase might have a
`coating effect on the lipophiiie stationary phase.
`making the stationary phase more hydrophilic.
`This might explain the retaining effects on hy-
`drophilic drugs. Such a model predicts the re-
`verse effect on lipophilic drugs.
`
`2. I. 7. Infhmtrsc of injection dept)‘:
`'
`in the articles discussed above. the difference
`in absorption rate between deep intramuscular
`injections and shallow subcutaneous or
`intra-
`adipose injections is shown in Fig. 3 (Kadir cl al..
`1990b). Even the hydrophilic atenolol
`is already
`completely absorbed 8 h after intramuscular in-
`jection whereas this lasts about 24 h after subcu-
`taneous or intra-adipose injection. The fatty sub-
`cutaneous connective tissues and adipose layers
`are more lipophilic and perfused less than muscu-
`lar
`tissues. These phenomena have been dis-
`cussed extensively previously (Zuidema cl al..
`1988).
`
`2.2. Long-acting .v_v.ttem.r
`
`Long-acting systems consist either of lipophilic
`drugs in aqueous solvents as suspensions or of
`highly lipophilic drugs dissolved or suspended in
`oil. In the first case, the release or absorption is
`dissolution rate controlled. while in the latter. the
`compounds with lower molecular weights, which
`are soluble in oil, are ‘phase transfer controlled‘
`released from the system (Zuidema et al.. I988}:
`the compounds with high molecular weights.
`which are not soluble in the oil. are released by
`dissolution and/or phase transfer control, these
`factors alone or in combination.
`
`absorption mechanism is
`Another possible
`phagocytosis. The participation of direct absorp-
`tion of fine particles by phagocytosis appears to
`be possible in certain cases but the contribution
`to the overall absorption process can mostly be
`neglected. Absorption via the mechanisms of lym-
`phatic transport and inflammatiommediated ap-
`pearance of phagocytosing macrophages (24-48 h
`after injection) have been demonstrated for iron
`complexes (Beresford et al., 1957}.
`Long-lasting residues after intramuscular in-
`jection or subcutaneous injection. of water-in-
`soluble penicillin derivatives and dihydrostrcpto-
`myein in aqueous vehicle in cattle have been
`detected 30-45 days after administration (Mercer
`ct al.. 197i). Examples of long-acting systems in
`oil are the depot neuroleptics as discussed previ-
`ously (Zuidema et al., 1988).
`Ober et al. (1958) stated that aggregation in
`
`Astrazeneca Ex. 2114 p. 8
`
`

`
`J’, Zuidema er al. /International Journal of Phannaceufics 105 (1994) 189-207
`
`197
`
`concentrated systems often gives rise to increased
`viscosities, specific rheological features and a di-
`minished ra.te of dissolution after injection. It is
`known that viscosities and specific rheological
`features such as (pseudo)plastic behaviour in sus-
`pensions and emulsions increase with increasing '
`concentration and with decreasing _particle size.
`In fact,
`this process is associated with the
`so-called flocculation phenomenon. This is be-
`cause gel formation consists of aggregation to a
`large viscous aggregate in which vehicle might be
`included (structured water or,
`in oil, structured
`oil), which is a special case of flocculation of
`lyophilic dispersions. These systems display mostly
`(pseudo)plastic properties. This theory is largely
`expounded in the context of colloidal systems but
`is it also valid for systems of small non-colloidal
`particles.
`When either lyophilic or lyophobic systems are
`concentrated by precipitation under gravitational
`forces or by other compression forces, they might
`develop a more compact
`form of aggregation,
`comparable with the phenomenon referred to as
`calcing.
`This theory, applied to suspension prepara-
`tions for injection, was provided with a solid and
`mathematical foundation by the work of Hirano
`et al.
`(1981) and Hirano and Yamada (1982,
`l983a,b). It
`is relevant to review this work in a
`little more detail, since it is also pertinent to the
`behaviour of liposomes, as discussed in section 3.
`The authors studied the local clearance of
`suspensions of practically water-insoluble drugs
`injected intramuscularly or subcutaneously in the
`rat. They used these results to develop or to
`check their mathematical model. This work is of
`great importance for understanding the influence
`of the phenomenon of aggregation on the release
`kinetics. The theory is mainly applicable to those
`drugs for which the absorption into blood or
`lymph is controlled by dissolution. Phase transi-
`tion of oily systems and blood supply are not
`thought
`to be relevant. As is clear from the
`former study, this is a simplification and not true
`for every system (Zuidema et al., 1988).
`Hirano and co-workers expressed the influ-
`ence of aggregation as a factor e with two limits:
`no aggregation and complete aggregation. In the
`
`first case, the suspension particles behave inde-
`pendently, while in the latter, the aggregate be-
`haves as a single clot with the clot surface acting
`as the effective area for release or in vivo dissolu-
`tion. The parameter e
`is governed by particle
`size, concentration, volume, hydrodynamic factors
`such as injection speed and pressure and histolog-
`ical and physiological states at the injection site.
`This means that particle size has two opposite
`influences on the dissolution rate: on the one
`hand, smaller sizes without aggregation lead to a
`greater effective dissolution surface and therefore
`to higher dissolution rates; on the other, smaller
`particle sizes may lead to stronger aggregation
`effects with more compact aggregates and subse-
`quently to a smaller effective dissolution surface.
`A deviation of this pattern occurs for particles
`smaller than about 2-3 ,u.m, especially after sub-
`cutaneous injection. These particles appear
`to
`pass more easily through the fibrous networks
`accompanying the spreading of
`the dispersion
`medium during injection and seem to form conse-
`quently looser agglomerate. Differences between
`muscle or subcutaneous networks, however, were
`not investigated and therefore not discussed.
`Dose adjustment can be effected as a change
`in drug concentration for a constant volume or as
`a change in vehicle volume at a constant drug
`concentration. In both cases, the absorption rate
`appears to decrease with increasing dose as a
`result of the aggre