REVIEW ARTICLE
`
`Pharmaceutical Applications of Cyclodextrins. 2.
`
`In Vivo Drug Delivery
`
`ROGER A. RAJEWSKix AND VALENTINO J. STELLA
`Received February 9, 1996, from the Higuchi Biosciences Center (or Druf! Delivery Re~earch ~nd the Department of Pharmaceutical
`Fmal rev1sed manuscnpt rece1ved September 6, 1996.
`Chemistry, The University of Kansas, Lawrence, KS 66047.
`Accepted for
`publication September 9, 1996°.
`
`Table 1- General Structures• of Cyclodextrins Referred to in This Review
`and Their Abbreviated Names
`
`oAOC.,:Y~~o ~,OR
`
`
`-::X!
`),:
`
`OR
`
`0
`
`3
`RO 2
`
`OR --------------- --- 0
`
`Abstract o The objective of this Review is to summarize and critique
`recent findings and applications of both unmodified and modified
`cyclodextrins for in vivo drug delivery. This review focuses on the use
`of cyclodextrins for parenteral, oral, ophthalmic, and nasal drug delivery.
`Other routes including dermal, rectal, and pulmonary delivery are also
`briefly addressed. This Review primarily focuses on newer findings
`concerning cyclodextrin derivatives which are likely to receive regulatory
`acceptance due to improved aqueous solubility and safety profiles as
`compared to the unmodified cyclodextrins. Many of the applications
`reviewed involve the use of hydroxypropyl-,8-cycJodextrins (HP-,8-CDs)
`and sulfobutylether-,8-cyclodextrins (SBE-,8-CDs) which show promise of
`greater safety while maintaining the ability to form inclusion complexes.
`The advantages and limitations of HP-,8-CD, SBE-,8-CD, and other
`cyclodextrins are addressed.
`
`Introduction
`The objective of this Review Is to summariZe and critique
`recent findings and applications of the use of unmodified and
`modified cyclodextrlns for in vivo drug delivery. As part of
`this series, Loftssen and Brewster' reviewed the use of
`cyclodextrlns for the solubilization, stabilization, and formula(cid:173)
`tion of drugs through the formation of inclusion complexes
`while Uekama et al.2 will summarize findings on the safety
`profile of cyclodextrlns. Numerous other major reviews on
`the actual and potential pharmaceutical uses of cyclodextrlns
`have been published.3- 11 Recently, as stated by Duchene and
`Wouessldjewe, 3 the availability of new derivatives with better
`safety profiles has renewed Interest In the in vivo uses of
`cyclodextrlns. Also, as more products containing cyclodextrlns
`move toward regulatory approval in the U.S. and elsewhere,
`the general acceptance by researchers and the pharmaceutical
`industry of specific cyclodextrlns as enabling exclplents is
`likely to Increase. Although a number of products containing
`cyclodextrlns have been approved for human use In Japan and
`Europe, no product has yet to be approved In the U.S. The
`approval of specific products by the Food and Drug Admin(cid:173)
`Istration (FDA) In the U.S. will be of paramount Importance
`to the commercial viability of cyclodextrins for worldwide
`pharmaceutical use. It Is generally accepted that the lack of
`an FDA approved product, presumably due to actual or
`perceived safety concerns, continues to Inhibit the universal
`acceptance of these valuable materials.
`Cyclodextrin's Pharmaceutical Nich e-In the early
`1980s the high Incidence of an anaphylactic reaction ac(cid:173)
`companying the parenteral administration of selected formu(cid:173)
`lations led to questioning the use of surfactants which had
`been employed to solubilize or stabilize these formulations.
`This idiosyncratic histamine release was especially evident
`
`e Abstract published in Advance ACS' Abstracts, October 15, 1996.
`
`Cydodextrin
`
`Abbrevia ion
`
`R
`
`n
`a-CO H
`a-Cydodextrin
`4
`{J.CD H
`{3-Cydodextrin
`5
`y-Cydodextrin
`y.CD H
`6
`Carboxymethyl-{3-eyclodextrin
`CM-{J.CD CH2CChHorH
`5
`Carboxymethyl-ethyl-{3-eyclodextrin CME-{3-CD CH2CChH, CH2CH3 or H
`5
`Diethyl-{3-eyclodextrin
`DE-{J.CD CH2CHJorH
`5
`Dimethyi-{J..cydodextrin
`DM-{J.CD CH3or H
`5
`Me hyl-{3-cydodextrin
`M-{J.CD CH3 orH
`5
`Random methyi-{J..cydodextrin
`RM-{3-CD CH3orH
`5
`Glucosyl-{3-cydodextrin
`G,-{J.CD Glucosyl or H
`5
`~{J.CD Maltosyl or H
`Maltosyi-{J..cydodextrin
`5
`Hydroxyethyl-{3-eyclodextrin
`HE-{J.CD CH2CI-bOH or H
`5
`Hydroxypropyi-{J..cyclodextrin
`HP-{3-CD CH2CHOHCH3 or H
`5
`Sulfobutyle her-{3-eyclodextrin
`SBE-{J.CD (C~)4SOJNa or H
`5
`• Derivatives may have differing degrees of substitution of the 2, 3, and 6
`positions.
`with the Increased clinical use of the solubilizers Cremophor(cid:173)
`EL present In cyclosporln A and taxol formulations, and
`Tween 80 which was used In an etoposide product. Other than
`the use ofcosolvents, mlcroemulslon dosage forms, pH adjust(cid:173)
`ment for Ionizable drugs, and experimental dosage forms such
`as mlcropartlculates and liposomes, few new viable formula(cid:173)
`tion options had been made available to address the problems
`associated with administering sparingly water soluble drugs
`In solution form. Although the use of cyclodextrlns as
`potential solubilizing and stabilizing agents had been well
`recognized earlier, safety concerns limited their use for
`parenteral administration.
`Cyclodextrins of pharmaceutical relevance are cyclic oligo(cid:173)
`saccharides composed of dextrose units joined through a 1-4
`bond. Cyclodextrlns with six to eight dextrose units have been
`named a -, {J-, and y-cyclodextrln, respectively (a-, {J-, and
`y-CD; Table 1). As discussed by Loftssen and Brewster.'
`cyclodextrlns are capable of forming Inclusion complexes with
`drugs. These noncovalent, Inclusion complexes can have
`physical, chemical, and biological properties that are dramati(cid:173)
`cally different from those of either the parent drug or cyclo(cid:173)
`dextrln. These complexes can be used to Increase solubility
`and dissolution rate, decrease volatility, alter release rates,
`modify local irritation, and Increase the stability of drugs. The
`driving forces for inclusion complexation, including the re-
`
`1142 I Journal of Pharmaceutical Sciences
`Vol. 85, No. 11, November 1996
`
`S0022-3549(96)00075-5 CCC: $12.00
`
`© 1996, American Chemical Society and
`American Pharmaceutical Association
`SENJU-MITSUBISHI 2025
`
`Hopewell EX1073
`Hopewell v. Merck
`IPR2023-00481
`
`1
`
`

`

`1:1 Complex
`Cyclodextrin
`Drug
`Scheme 1- Scheme illustrating equilibrium binding of drug and cyclodextrin to
`form a 1: 1 complex.
`
`quirement of the drug to "fit" into the cyclodextrin cavity, have
`been discussed elsewhere.1.12-24
`In clus ion
`Mechanisms of Drug Release from
`Complexes- The rational design of formulations which take
`advantage of cyclodextrin inclusion complexation requires an
`understanding of the relationship between intrinsic drug
`solubility, the magnitude of the binding constant for the
`inclusion complex, and dilution effects. Most pharmaceutical
`agents form 1: 1 complexes with cyclodextrins as described by
`Scheme 1. On the basis of the structure and properties of
`the drug as well as the cyclodextrin, higher order complexes
`are also possible.
`The magnitude of the binding constant, K1:1, defmed by eq
`1 for the interaction described by Scheme 1, is generally in
`the range 0- 100 000 M-1, with 0 being the value for a drug
`that is incapable of forming an inclusion complex and 100 000
`M-1 being near the upper value observed experimentally for
`cyclodextrin complexes of drugs. In eq 1, [drug]oomptex repre(cid:173)
`sents the concentration of drug in the complex form, [drug]rree
`represents the free drug concentration, and [cyclodextrinlrree
`represents the concentration of free cyclodextrin.
`
`_
`K
`1·1 -
`·
`
`(druglcomplex
`.
`[druglfree[cyclodextrm]rree
`
`(1)
`
`For illustration of the relationship between drug solubility,
`magnitude of the binding constant, and dilution, a hypotheti(cid:173)
`cal drug (MW 400) with an intrinsic solubility of 10 ,ug·mL - 1
`(2.5 X w-s M) and a K1:1 value of 10 000 M-1 for its interaction
`with a cyclodextrin of unlimited solubility will be examined.
`In the presence of the cyclodextrin, the solubility of the drug
`would be defined by eq 2. In eq 2, [drugltotaJ represents the
`[drugltotal = [drug]intrinsic +
`K 1:1 [drug]intrinsic[cyclodextrin]total
`Kl:l[drug]intrinsic + 1
`
`(2)
`
`total drug solubility, [drugltntrlnstc represents the intrinsic
`solubility of the drug in the absence of the cyclodextrin, and
`[cyclodextrin]totaJ represents the total molar concentration of
`cyclodextrin in the solution. In the presence of 0.1 M
`cyclodextrin, the solubility of the drug would be 8 mg·mL - 1.
`An injectable dosage form containing 40 mg of this drug in 5
`mL of 0.1 M cyclodextrin would contain only 50 ,ug of free or
`noncomplexed drug (0.125%), while the remaining 39.95 mg
`(99.875%) would be in the form of the inclusion complex. The
`ratio of free to complexed drug will change when this 5 mL
`formulation is injected intravenously by bolus dosing. For
`example, when this formulation is injected into a 70 kg subject
`with a plasma volume of about 3.5 L, the total drug and
`cyclodextrin concentrations would be 11.4 ,ug·mL - 1 (2.85 x
`w-s M) and 0.143 mM, respectively. At these concentrations
`the percent free drug would be 43.9% with the balance still
`bound to the cyclodextrin. These values do not reflect drug
`and cyclodextrin binding to plasma proteins, uptake of drug
`into red blood cells or other tissues, and competitive displace(cid:173)
`ment of the drug from the complex by plasma components. If
`
`500
`
`250
`/)·
`'.1/lJt.i
`Ol) {tj
`llJesJ
`
`1000 0
`
`Figure 1- Relationship between fractional percent of a drug in its complex form
`as a function of the strength of the association constant, K,1, and dilution. The
`initial drug and cyclodextrin concentrations were set at 0.1 M, the range of K1:1
`values was 1- 10 000 M-1, and dilutions ranged over 1- 1000 times. Reprinted
`with permission from ref 25. Copyright 1994 Harwood.
`
`the drug has a more extensive volume of distribution (V,J,
`further dilution would occur and a smaller percentage would
`remain bound. Uekama et al. 25 have quantified the free and
`bound fractions of drugs to cyclodextrins as a function of drug
`concentration, binding constant, cyclodextrin concentration,
`and dilution. A graphical representation Is shown in Figure
`1. The kinetics of drug binding to cyclodextrins has been
`studied, and equilibrium binding is usually established with
`half-lives of much less than 1 s.26- 29 Therefore, the kinetics
`of dissociation is generally much faster than most physiologi(cid:173)
`cal processes.
`Frijlink et al.30 studied the effect of dilution with plasma
`on hydroxypropyl-,8-cyclodextrin (HP-,8-CD) complexes of
`naproxen or flurbiprofen. They found experimentally that
`only small fractions of the drugs remained bound to the
`cyclodextrin in plasma in vitro. This effect was due not only
`to dilution but also to the competition between albumin
`binding of the two drugs and cyclodextrin binding. Also
`contributing to the low fraction bound was displacement of
`the drugs from cyclodextrin by a competing agent, plasma
`cholesterol.
`The importance of changes in the ratio of free to complexed
`drug upon dilution of a sparingly water soluble drug in a
`cyclodextrin complex depends on the phase- solubility behav(cid:173)
`Ior of the system. If the cyclodextrin complex of a drug results
`from a 1:1 interaction, there is a linear increase in drug
`solubility with increasing cyclodextrin concentration (Figure
`2A). Therefore, dilution of a true solution of the drug/
`cyclodextrin complex will not result in drug precipitation
`regardless of the extent of dilution. In the example discussed
`above where the 8 mg·mL - 1 solution was diluted to 3.5 L, the
`final drug concentration was 11.4 ,ug·mL - 1, of which the free
`drug concentration was 4.6 ,ug·mL - 1 and the bound concentra(cid:173)
`tion was 6.8 ,ug·mL - 1. The 4.6 ,ug·mL - 1 free drug concentra(cid:173)
`tion is below the intrinsic solubility of the drug, set at 10
`,ug·mL - 1, and precipitation of the drug will not occur.
`Precipitation of the drug may occur on dilution, however,
`if there is a nonlinear relationship between drug solubility
`and cyclodextrin concentration. Figure 2B is a plot of solubil(cid:173)
`ity versus cyclodextrin concentration for the hypothetical drug
`described earlier where the solubility of 8 mg·mL -1 was
`achieved at 0.1 M cyclodextrin with a 1:1 interaction constant
`of 2000 M- 1 (eq 1) and a 1:21nteraction constant, K1:2. of 50
`M-1 as described by Scheme 2.
`
`Journal of Pharmaceutical Sciences I 1143
`Vol. 85, No. 11, November 1996
`
`2
`
`

`

`+
`
`+
`
`may occur. Precipitation may occur in such systems at any
`dilution where the equilibrium drug solubility is lower than
`the dilution concentration line at a given cyclodextrin con-
`centration. The problem of precipitation on dilution for those
`drugs which display the behavior just illustrated has recently
`been documented in the literature.31 Theoretically, similar
`precipitation could also occur in vivo when such solutions are
`injected intravenously or by other parenteral routes.
`Modified CyclodextrinssPrior to the mid-1970s most
`pharmaceutical research on cyclodextrins focused on the
`unmodified R-, (cid:226)-, and (cid:231)-CDs. The work of Frank et al.32 and
`others33 showing the nephrotoxicity of the unmodified cyclo-
`dextrins limited further studies on the parent cyclodextrins
`to those routes where systemic cyclodextrin absorption was
`limited, namely, nonparenteral routes.
`Most drug molecules tend to interact more favorably with
`(cid:226)-CD than R-CD because the 6 Å cavity diameter for (cid:226)-CD
`accommodates aromatic groups found in many drug molecules.
`In contrast, the cavity diameter for R-CD tends to be too small
`for a favorable fit. Interactions can also be seen between
`many drugs and (cid:231)-CD. However, the cost of (cid:231)-CD has made
`its extensive use economically unfavorable. A severe limita-
`tion of (cid:226)-CD is its limited water solubility of 18.6 mg(cid:226)mL-1 or
`16.4 mM. For example, a hypothetical drug with an aqueous
`solubility of 10 (cid:237)g(cid:226)mL-1 and a 1:1 binding constant with (cid:226)-CD
`of 1 (cid:2) 104 M-1 would have a maximum obtainable solubility
`of 1.3 mg(cid:226)mL-1 in the presence of 16.4 mM (cid:226)-CD. In addition,
`(cid:226)-CD often forms B-type phase-solubility diagrams where the
`complexes themselves have limited aqueous solubility.34
`Because of the solubility limits and the safety concerns with
`(cid:226)-CD, numerous chemical modifications of the cyclodextrins
`have been made. Since the focus of this Review is on the in
`vivo applications of cyclodextrins, only those cyclodextrins
`which have been studied for their pharmaceutical utility will
`be discussed. The cyclodextrins of prime interest to pharma-
`ceutical scientists consist of five general types. The first type
`consists of various methylated and alkylated cyclodextrins,
`especially 2,6-dimethyl-(cid:226)-cyclodextrin (DM-(cid:226)-CD) and ran-
`domly methylated (cid:226)-cyclodextrins (RM-(cid:226)-CD). The second
`type consists of the hydroxypropyl and hydroxyethyl cyclo-
`dextrins, especially 2-hydroxypropyl-(cid:226)-cyclodextrins (HP-(cid:226)-
`CD). Specific products include Encapsin and Molecusol. The
`third type consists of various branched cyclodextrins, espe-
`cially glucosyl, diglucosyl (G2-(cid:226)-CD), maltosyl, and dimaltosyl
`cyclodextrins. The fourth type consists of carboxymethyl
`cyclodextrins, e.g., CM-(cid:226)-CD, and associated derivatives. The
`fifth type consists of the sulfoalkylether cyclodextrins, espe-
`cially sulfobutyl ethers (SBE-(cid:226)-CD) of (cid:226)-CD with degrees of
`substitution of 4 and 7, SBE4-(cid:226)-CD and SBE7-(cid:226)-CD, respec-
`tively. Specific products include Captisol, an SBE7-(cid:226)-CD. The
`cyclodextrin derivatives discussed in this Review may be found
`in Table 1.
`Pharmaceutical scientists have also investigated the anti-
`angiogenic and antiviral properties of sulfated cyclodextrins;
`however, the ability of these derivatives to form inclusion
`complexes is rather minimal.35-44
`In addition, they have
`heparin-like activity resulting in an increase in blood clotting
`times which limits the cyclodextrin dose that can be admin-
`istered to patients. Therefore, the use of sulfated cyclodex-
`trins as potential complexing agents will not be discussed
`further. As noted for the sulfated cyclodextrins, safety is a
`primary concern when the advantages and disadvantages of
`the various cyclodextrins are considered. Uekama et al.2 will
`address in detail the safety status of the various cyclodextrins
`in a Review to be published in a future issue. Questions on
`the safety of the various derivatives will be addressed briefly
`in this paper and then only as they affect the performance of
`the cyclodextrin.
`
`Figure 2s(A) Solubility versus cyclodextrin concentration for a hypothetical drug
`of MW 400 with an intrinsic water solubility of 10 (cid:237)g(cid:226)mL-1 (2.5 · 10-5 M) forming
`a 1:1 cyclodextrin complex with a drug with a K1:1 association constant of 10 000
`M-1. (B) Solubility versus cyclodextrin concentration for a hypothetical drug of
`molecular weight 400 with an intrinsic water solubility of 10 (cid:237)g(cid:226)mL-1 (2.5 · 10-5
`M) forming 1:1 and 1:2 cyclodextrin complexes with the drug wi h K1:1 and K1:2
`constants of 2000 and 50 M-1, respectively.
`
`Scheme 2sScheme illustrating equilibrium binding of a 1:1 complex of drug
`and cyclodextrin with a second molecule of cyclodextrin to form a 1 2
`complex.
`As can be seen from Figure 2B, there is a nonlinear increase
`in solubility with increasing cyclodextrin concentration. If we
`assume that an 8 mg(cid:226)mL-1 solution of the drug in 0.1 M
`cyclodextrin is diluted 1:5 into a minibag, the final drug
`concentration would be 1.6 mg(cid:226)mL-1 while the cyclodextrin
`concentration would be 0.02 M. According to the data in
`Figure 2B, the solubility of the drug in 0.02 M cyclodextrin is
`only 0.71 mg(cid:226)mL-1. Thus, precipitation of the drug on dilution
`
`1144 / Journal of Pharmaceutical Sciences
`Vol. 85, No. 11, November 1996
`
`3
`
`

`

`+
`
`+
`
`Another very important factor in considering the pharma-
`ceutical use of cyclodextrins is cost. Because cyclodextrins
`have molecular weights in the range of 1000-2000, 0.1 M
`solutions require 10-20% (w/v) cyclodextrin. Production of
`dosage forms on a commercial scale would quickly consume
`very large amounts of the cyclodextrin. Therefore the cyclo-
`dextrin must be reasonably inexpensive for the dosage form
`to be economically feasible. (cid:226)-CD itself is quite inexpensive,
`and the costs of R- and (cid:231)-CDs are declining. Any modification
`of the cyclodextrin structure must entail relatively inexpensive
`reagents and purification procedures. However, well-charac-
`terized, pure, single-component materials are rare.
`(cid:226)-CD contains 21 hydroxyl groups, seven primary (the
`6-hydroxy) and 14 secondary (the 2- and 3-hydroxyls). Meth-
`ods have been developed with varying degrees of success either
`to derivatize all 21 of the hydroxyls or to selectively derivatize
`from one to seven of a particular hydroxyl group. Claims of
`selective derivatization are often overstated, and most claimed
`derivatives are mixtures. It may be very difficult to justify
`economically the development of a pure derivative unless it
`can be accomplished with a high yield, using simple and
`inexpensive reagents, and employing a purification procedure
`that can be easily scaled to produce metric ton quantities. The
`use of complex mixtures is not unknown to the pharmaceutical
`community. However, the methods used to characterize such
`mixtures need to be refined so that lot to lot reproducibility
`can be verified.
`From the pharmaceutical viewpoint, not only must these
`complex mixtures be reproducible and well characterized lot
`to lot, but they must be free of all potentially reactive and
`toxic components. This is especially important in the case of
`unreacted parent cyclodextrin in products intended for
`parenteral administration. It is critical to have materials free
`of pyrogens and foreign proteins for cyclodextrins which are
`intended for parenteral use and to have the ability to sterilize
`the solution either by heat or filtration. These issues are not
`trivial and must be addressed before a particular cyclodextrin
`can be considered for pharmaceutical use.
`
`Parenteral Applications of Cyclodextrins
`IntroductionsThe major anticipated uses of cyclodextrins
`in parenteral drug delivery include solubilization of drugs,
`allowing for rapid and quantitative delivery of sparingly water
`soluble drugs for intravenous and intramuscular dosing,
`decreasing irritation at the site of administration of parenter-
`ally administered drugs, and stabilization of drugs unstable
`in an aqueous environment.
`To date two types of parenteral studies have been published.
`In the first type a cyclodextrin has been used during animal
`testing to allow for a solution dosage form to be administered.
`Some of these studies have used cyclodextrins of questionable
`safety while other studies have not run controls which might
`have allowed the researcher to determine if the cyclodextrin
`altered the pharmacodynamics of the drug. In these cases
`the cyclodextrin was used as a tool and was not intended as
`a prototypical formulation for ultimate scaling to humans. In
`the second type of study the cyclodextrin was being evaluated
`as a potential tool for improved parenteral delivery of either
`a model drug or an actual clinical candidate. Nevertheless,
`both types of studies provide valuable insight into the
`potential uses of cyclodextrins as enabling excipients for
`parenteral drug delivery.
`Aspects of the use of cyclodextrins to effect parenteral
`delivery of drugs have been reviewed previously.3-11,25,45-60
`The safety record of the cyclodextrin is perhaps more critical
`with parenteral delivery than with other routes of administra-
`tion. Other routes such as oral may result in the systemic
`
`delivery of little or no cyclodextrin. Parenteral delivery, on
`the other hand, guarantees systemic delivery. Although the
`parenteral use of unmodified cyclodextrins and some modified
`cyclodextrins such as DM-(cid:226)-CD has been attempted, their
`systemic safety records are unacceptable. Since the safety
`records of HP-(cid:226)-CDs and SBE-(cid:226)-CDs appear more promising,
`this section will focus primarily on their use.
`The earlier Review in this series by Loftsson and Brewster1
`focused on the solubilization and stabilization of drugs.
`Although most of the examples in the current literature on
`in vivo applications have focused on the use of cyclodextrins
`to solubilize and/or to decrease the irritancy of drugs, future
`uses will also focus on their ability to stabilize drugs in
`aqueous solution, allowing for parenteral delivery. This is
`especially important for those unstable drugs for which long-
`term infusions might be desirable. For example, cyclodextrins
`are currently being studied extensively for their ability to
`stabilize a number of very unstable cytotoxic drugs.
`Parenteral Delivery of Insoluble DrugssEven though
`the safety of a particular cyclodextrin is the critical factor in
`its potential parenteral use, other factors also limit whether
`a cyclodextrin can be used to administer a given drug. One
`such factor is whether the target drug solubility can be
`achieved with the use of an acceptable cyclodextrin concentra-
`tion. In addition, the cyclodextrin safety data must support
`the concentration and total dose of cyclodextrin required to
`solubilize the desired amount of drug. Other factors such as
`the linearity of the relationship between drug solubility and
`cyclodextrin concentration might also affect the acceptability
`of a given concentration or dose. If a nonlinear relationship
`exists and a very high concentration of cyclodextrin is needed
`to solubilize the drug, dilution of that sample either in a large
`volume parenteral (LVP) or on iv or im administration could
`lead to drug precipitation and complications. For example,
`the authors have noticed that some cyclodextrins could be used
`to solubilize the anticancer drug taxol (unpublished data).
`However, the relationships between taxol solubility and
`cyclodextrin concentration were such that any attempt to
`dilute the samples resulted in erratic precipitation of the taxol.
`Because nucleation either in an LVP or in vivo is a time
`dependent phenomenon, it might be possible to consider such
`metastable dosage forms even though precipitation can oc-
`cur.31
`Another critical factor would be whether the drug was
`completely released from the cyclodextrin dosage form. Drug
`release from the cyclodextrin dosage form can be determined
`by assessing the pharmacokinetics of the drug from the
`cyclodextrin solution as compared to other dosage forms. The
`pharmacokinetic studies would also indicate if the cyclodextrin
`altered the temporal pattern of the drug in plasma or blood,
`or the drug excretion profile. Either effect could lead to a
`change in pharmacodynamics and perhaps therapeutic ef-
`ficacy and toxicity. Time profiles of the drug in various tissues
`would also be helpful in assessing the role of the dosage form.
`Few well-designed studies have addressed these latter issues
`to date.
`Intravenous AdministrationsMost recent studies on the
`use of cyclodextrins to allow iv administration have utilized
`SBE4-(cid:226)-CD or HP-(cid:226)-CD since safety concerns with other
`cyclodextrins preclude their parenteral use. However, an
`early study by Shirakura et al.61 involving (cid:226)-CD is noted
`because the results demonstrated that iv (cid:226)-CD significantly
`shortened the sleeping time induced by a series of barbituric
`acid derivatives. The amount of (cid:226)-CD administered with the
`barbiturates, however, was about 0.244 mg(cid:226)kg-1. This dose
`will show some significant renal toxicity. The possible mech-
`anisms for the altered sleep times were speculated on by the
`authors, but none were conclusively proven. A similar study
`
`Journal of Pharmaceutical Sciences / 1145
`Vol. 85, No. 11, November 1996
`
`4
`
`

`

`+
`
`+
`
`which utilizes one of the safer cyclodextrins would provide
`valuable insight to the altered barbiturate activity.
`Significant alteration in the pharmacokinetics of a drug was
`also demonstrated in the work of Frijlink et al.30 This study
`evaluated the tissue distribution in rats of naproxen and
`flurbiprofen from HP-(cid:226)-CD containing solutions as compared
`to solutions of the drugs dissolved in rat plasma. The tissue
`distribution in brain, liver, kidney, spleen, muscle, and plasma
`of naproxen 10 and 60 min after iv administration was
`unaffected by coadministration of 7% HP-(cid:226)-CD. For flurbi-
`profen, however, at 10 min postdosing HP-(cid:226)-CD produced
`significantly higher tissue levels in the brain, liver (most
`significant), kidney, and spleen but at 60 min postadminis-
`tration only slightly higher, yet significant, levels in the brain
`tissue alone. Frijlink et al. speculated that the higher levels
`in some tissues reflected a transitory alteration in protein
`binding when HP-(cid:226)-CD was used as the vehicle. In an in vitro
`study, HP-(cid:226)-CD was able to compete with the protein binding
`for both naproxen and flurbiprofen but the effect was more
`pronounced with the flurbiprofen. Another control experiment
`where flurbiprofen was administered in a purely aqueous
`solution might have indicated whether administration in rat
`plasma also affected the relative results.
`LaHann et al.62,63 have studied the pharmacokinetics in
`dogs of p-boronophenylalanine following iv administration of
`4.43 mg(cid:226)kg-1 drug in an aqueous HP-(cid:226)-CD solution (to effect
`greater solubility) versus a more dilute 0.96 mg(cid:226)kg-1 aqueous
`solution. The pharmacokinetics of p-boronophenylalanine
`were found to be significantly different from the two dosage
`forms. The authors intimated that the pharmacokinetics of
`p-boronophenylalanine was dose dependent (nonlinear). There-
`fore, differences observed in the pharmacokinetics including
`longer half-life at the higher dose and a disproportionate
`increase in AUC with dose might be due to the dose differ-
`ences. However, a direct effect of the HP-(cid:226)-CD could not be
`ruled out. This issue could have been clarified by studying
`the pharmacokinetics of the drug at increasing doses in the
`presence of a fixed HP-(cid:226)-CD dose level. Greater toxicity was
`seen with the HP-(cid:226)-CD formulation, but this may have been
`due to the higher dose levels. Some toxicity from the HP-(cid:226)-
`CD vehicle itself was noted in a few dogs. The etiology of this
`response was not assessed.
`In those studies where control experiments could be per-
`formed, i.e., where the drug could be given by means other
`than the cyclodextrins, cyclodextrins have been found to not
`alter the intrinsic pharmacokinetics of the drug. For example,
`Arimori et al. showed that (cid:231)-CD did not alter the pharmaco-
`kinetics of thiopental in rabbits after iv administration.64 The
`activity of the anesthetic propofol (2,6-diisopropylphenol) was
`no different when administered iv as a HP-(cid:226)-CD solution
`versus a commercial oil in water emulsion.65 The responses
`to increasing doses of the anesthetic isoflurane from a HP-(cid:226)-
`CD solution were similar to those after inhalation of the
`agent.66 Estes et al.67 evaluated a nonsurfactant formulation
`that contained HP-(cid:226)-CD for administration of alfaxalone (3R-
`hydroxy-5R-pregnane-11,20-dione, a steroid anesthetic) to rats
`and dogs. This formulation was compared to a veterinary
`product containing the surfactant Cremophor-EL. No differ-
`ence in anesthetic activity was seen between the two formula-
`tions in rats. In dogs, the HP-(cid:226)-CD formulation did not cause
`the significant histamine release and depressed respiratory
`rate and blood pressure drop seen with the surfactant
`formulation. More recently, the same research group provided
`additional data showing the usefulness of HP-(cid:226)-CD in the
`parenteral administration of a series of other steroidal
`anesthetic agents.68
`Doenicke et al.69 compared the iv bolus pharmacokinetics
`and activity of the hypnotic agent etomidate (2 mg(cid:226)mL-1) from
`a 35% propylene glycol aqueous solution compared to a 3%
`
`1146 / Journal of Pharmaceutical Sciences
`Vol. 85, No. 11, November 1996
`
`HP-(cid:226)-CD solution. The pharmacokinetics of etomidate were
`identical from the two solutions, and no differences in hypnotic
`effects were observed. A higher incidence of pain on injection
`was noted in patients receiving the propylene glycol solution
`(58% of the patients) as compared to patients receiving the
`HP-(cid:226)-CD formulation (8% of the patients). Another study70
`did show a high incidence of pain (52%) from the HP-(cid:226)-CD
`formulation when it is was assessed alone. Etomidate itself
`caused irritation and pain on injection.
`Brewster et al.71 were able to develop a parenteral formula-
`tion of the anticonvulsant drug carbamazepine through the
`use of HP-(cid:226)-CD. Carbamazepine solubility increased nonlin-
`early with increasing HP-(cid:226)-CD concentrations, suggesting
`complexes of higher order than 1:1 between carbamazepine
`and HP-(cid:226)-CD. A tolerability and pharmacokinetic study was
`performed in epileptic patients by Lo¨scher et al.72 by compar-
`ing a 10 mg(cid:226)mL-1 carbamazepine solution in 20% HP-(cid:226)-CD
`to a 65% aqueous glycofural (PEG monotetrahydrofurfuryl
`ether) solution. The half-lives and AUC values for carbam-
`azepine from HP-(cid:226)-CD and glycofural were 0.603 ( 0.043 h
`versus 1.0 ( 0.014 h and 3.70 ( 0.72 (cid:237)g(cid:226)h(cid:226)mL-1 versus 7.37
`( 0.72 (cid:237)g(cid:226)h(cid:226)mL-1, respectively. The longer half-life and
`higher AUC value (and therefore slower clearance) of car-
`bamazepine and the delayed appearance of the epoxide
`metabolite of carbamazepine from the glycofural formulation
`strongly suggested that glycofural was an inhibitor of car-
`bamazepine metabolism. Therefore, the nonidentical phar-
`macokinetics of carbamazepine from the HP-(cid:226)-CD solution
`was most probably due to the negative effect of the glycofural
`in the control formulation rather than any catalytic effect of
`the HP-(cid:226)-CD on carbamazepine elimination.
`The pharmacokinetics of the steroids methylprednisolone,
`dexamethasone, and prednisolone have been evaluated from
`solutions of cyclodextrins compared to either cosolvent solu-
`tions or from water soluble prodrugs of the steroids. Arimori
`and Uekama73 studied the iv pharmacokinetics of prednisolone
`in rabbits from (cid:226)-CD and (cid:231)-CD aqueous solutions as compared
`to an aqueous solution of its phosphate ester prodrug. The
`cyclodextrin solutions and the solution of th