REVIEW
Clinical studies with oral lipid based formulations of poorly soluble compounds
Dimitrios G Fatouros1 Abstract: This work is an attempt to give an overview of the clinical data available on lipid Ditte M Karpf1 based formulations. Lipid and surfactant based formulations are recognized as a feasible approach Flemming S Nielsen1 to improve bioavailability of poorly soluble compounds. However not many clinical studies Anette Mullertz1 have been published so far. Several drug products intended for oral administration have been marketed utilizing lipid and surfactant based formulations. Sandimmune® and Sandimmune Department of Pharmaceutics and Neoral® (cyclosporin A, Novartis), Norvir® (ritonavir), and Fortovase® (saquinavir) have been Analytical Chemistry, Faculty of Pharmaceutical Sciences, University of formulated in self-emulsifying drug delivery systems (SEDDS). This review summarizes Copenhagen, Copenhagen, Denmark; published pharmacokinetic studies of orally administered lipid based formulations of poorly 1Contributed equally to this work aqueous soluble drugs in human subjects. Special attention has been paid to the physicochemi- cal characteristics of the formulations, when available and the impact of these properties on the in vivo performance of the formulation. Equally important is the effect of concurrent food intake on the bioavailability of poorly soluble compounds. The effect of food on the bioavailability of compounds formulated in lipid and surfactant based formulations is also reviewed. Keywords: lipid formulations, poorly soluble compounds, pharmacokinetics, human clinical studies, lipophilic compounds, SEDDS, emulsions, food effect
Introduction The purpose of this article is to review the clinical trials on lipid and surfactant based formulations of poorly soluble compounds for oral administration. Despite a plethora of articles dealing with orally administered lipid based formulations of poorly soluble compounds, the majority of these articles emphasize issues related to pharmacology and toxicity rather than formulation issues. Therefore special attention has been paid to the differences in the pharmacokinetics upon oral administration of different lipid formulations. The use of lipid and surfactant based formulations is one of several approaches that has been applied in order to improve the oral bioavailability of poorly aqueous soluble compounds intended for oral administration. The approach has proved ef- fi cient and hence has received attention, especially in academia. Lipid and surfactant based systems are physically very different systems including systems like emulsions, microemulsion, self-emulsifying drug delivery systems (SEDDS), micellar solutions, and dry emulsions like lipid solutions and lipid suspensions. The number of excipients which are pharmaceutically acceptable and applicable in the formulation of lipid and surfactant based systems is large. The use of solubilizing Correspondence: Dr Dimitrios G Fatouros, Department of Pharmaceutics and excipients in marketed formulations of poorly soluble compounds has been reviewed Analytical Chemistry, The Faculty of recently (Strickley 2004). Pharmaceutical Sciences, University of Copenhagen, Universitetsparken 2, 2100, The bioavailability enhancing properties of lipid and surfactant based systems Copenhagen, Denmark has most often been attributed to the ability of the vehicles to keep the compound in Tel + 45 35 30 62 88 solution in the gastrointestinal (GI) tract and thereby maintaining a maximal free drug Fax + 45 35 30 60 30 Email [email protected] concentration (Porter and Charman 2001). However the underlying mechanisms of
Therapeutics and Clinical Risk Management 2007:3(4) 591–604 591 © 2007 Dove Medical Press Limited. All rights reserved Fatouros et al absorption from lipid and surfactant based systems is not bioavailability of drug molecules being PGP substrates. fully understood. A number of excipients have shown to infl uence the lym- The majority of the knowledge of the in vivo perfor- phatic transport both in rats, and have an impact to the chy- mance and suggested mechanism of drug absorption from lomicron secretion in caco-2 cells, which is also believed lipid based formulations originates from studies in animals. to be an indicator of lymphatic transport (Rege et al 2002; Several authors have attempted to identify and describe the Seeballuck et al 2003; Karpf et al 2004). important parameters in order to set-up rational strategies These advances in the characterization of the mecha- for development of new lipid based formulations. The nisms underlying the bioavailability enhancing properties parameters that have been emphasized so far are the degree of lipid and surfactant based formulations mark a change in of emulsifi cation (in terms of particle size), and the solubility formulation strategy away from the more empirically based of the drug both in the digested and nondigested formulation formulation testing to a more rational formulation strategy. of the resultant dispersion (Humberstone and Charman 1997; However it also implies that more work is needed in the Pouton 2000; Porter and Charman 2001; Porter et al 2004; fi eld and that a battery of different methods should be used Nielsen et al 2007). to describe the different mechanisms. Furthermore a number of typically employed excipients Among the lipid based formulations used for oral delivery, are susceptible to enzymatic degradation in the GI tract. SEDDS and self-microemulsifying drug delivery systems Excipients susceptible to degradation include natural di- and (SMEDDS) have been characterized more systematically triglycerides as well as some commonly used surfactants like from a physicochemical point of view. They are isotropic Labrasol, Labrafi l, Gelucire, Cremophor, Tween 80 (Khossravi mixtures of oil (pure triglyceride oils, mixed glycerides), et al 2002, Seebaluck et al 2004, Larsen et al 2006). surfactant (hydrophilic or/and lipophilic), water soluble The release of compound from SEDDS based formulation co-solvents and the poorly soluble compound. Development is thought to take place by two major pathways: Interfacial and characterization of SEDDS and SMEDDS has been transfer and degradation of vehicle (de Smidt et al 2004; extensively reviewed (Pouton 1997; Gershanik and Benita Porter et al 2004). Interfacial transfer can be described as a 2000; Gursoy and Benita 2004; Pouton 2006). A number of concentration gradient driven process in which the compound biophysical techniques have been used to characterize these diffuses from the formulation into the bulk or directly over systems (Lawrence and Rees 2000). Scattering techniques the intestinal membrane. The rate and extent of interfacial such as dynamic light scattering (DLS) (Constantinides transfer is thought to be governed by partition coeffi cient and and Scarlet 1997), small angle neutron scattering (SANS) solubility in the donor (formulation) and recipient phase par- (Bergenholtz et al 1995) and small angle x-ray scattering ticle size and hence surface area of formulation (Armstrong (SAXS) (Regev et al 1996) can offer useful information and James 1980). The second pathway is degradation of for the structure of the microemulsions together with freeze the vehicle inducing the release of the compound out of the fracture electron microscopy (Vinson et al 1991). Dielectric vehicle. As mentioned above, for lipid based formulations, viscosity, and conductivity measurements can offer new the most important degradation is the lipolysis catalyzed by insights on the characterization of these systems on a mac- pancreatic lipase. The release rate is thought to be dependent roscopic level (Acosta et al 1996). on the solubility of the compound in the formulation and Equally important are the interactions between drugs and rate and extent of the degradation of the vehicle. Lipolysis food. The impact of food on the bioavailability of drugs in of triacylglycerols (TG) by the pancreatic lipase–colipase clinical studies has recently been reviewed by Schmidt and complex releases monoacylglycerols, diacylglycerols and Dalhoff (2002). Food delays the gastric emptying rate and free fatty acids. These lipolysis products are amphiphiles induces secretion of bile and pancreatic juices, while the pas- that will further assist the solubilization of poorly soluble sage time of the small intestine remains virtually unchanged. compounds in the GI fl uids. Hence food intake can increase the solubilization time and Recently it has been demonstrated that a number of increase the solubility of a poorly soluble drug, which will excipients should be considered as more than just inert sub- affect the pharmacokinetic parameters. stances. It has been shown that Cremophor EL, Tween 80, In the following, data from clinical studies of oral lipid Labrasol, Miglyol polyethoxylated, can inhibit the PGP effl ux based formulations of poorly soluble compounds performed transporter (p-glycoprotein) (Hugger et al 2002; Shono et al are presented. Table 1 summarizes the type of study and the 2004; Cornaire et al 2004) and hereby potentially improve formulations used.
592 Therapeutics and Clinical Risk Management 2007:3(4) Oral lipid based formulations: An overview of clinical studies ) Continued ( al 1994 Authors Year : Neoral > : max max: : Neoral ~ Sandimmune. Neoral ~ Sandimmune. : max : Neoral < Sandimmune Neoral < Sandimmune : max : Oil suspension ~aqueous : max Sandimmune. : Neoral > Sandimmune T Neoral > Sandimmune : : B ~ C > D ~ Sandimmune B ~ C > D Sandimmune : : B < C ~ D Sandimmune : : Mixed with juice ~ Mixed with juice ~ Mixed Mixed : et al Kovarik 1993 max max max lag suspension > tablets Fasted and fed T Fasted and fed exhibit lower Intrasubject and exhibit lower intersubject variability (%CV) Neoral < Sandimunne Neoral Neoral < Sandimunne form form
characteristics and either 150 mg CsA from B ~ C > D Sandimmune AUC: et al Drewe 1992 3 glycol. 180 mg CsA from T glycols. glycols. 3 healthy healthy Sandimmune meal meal fasted four volunteers, The microemulsions C or D. or B, Sandimmune C 2 (n = 10) Aqueous mixture Aqueous amounts of arachis oil A doubling in the dosed > tablets. mixture balanced three-way An oil suspension oil) (fractionated coconut tablets C 1 1 Physicochemical properties and pharmacokinetic data of drugs and formulations used in bioavailability studies with human subjects used in bioavailability and pharmacokinetic data of drugs formulations properties Physicochemical or with a fat- rich with cross-over replication 24 healthy way four volunteers, Sandimmune than Neoral and (A) (Sandimmune) and 300 mg CsA 180 mg CsA (Neoral) T½: Neoral < Sandimmune Fasted C > Neoral. Sandimmune AUC: Fed Neoral > Sandimmune. AUC: Fasted Mueller et al 1994b fasted cross-over, (Sandimmune) C Fed Sandimmune. three volunteers, cross-over way 24 fasted healthy way two volunteers, Sandimmune Neoral and and 300 mg CsA 180 mg CsA (Neoral) Neoral as soft gelatine or as solution capsule with water < Soft gelatine Neoral ~ Sandimmune AUC: with water or juice mixed C All other PK parameters capsule. equal. et al Kovarik 1994 cross-over way (B&D) were Microemulsion B: T weight composed of glycol, polyethylene hydrogenated Solid micellar C: middle castor oil, release. with fast in vitro chain triglycerides molecular and low with Microemulsion D: release. solution with fast in vitro release. in vitro slow (log P = 2.92) Cyclosporine 12 acid mixture suspension contain the corresponded in higher plasma concentrations. arachis oil resulted Clomethiazole (LogP 2.29) study Cross-over Lipid Tablets design (n = 9) crossover suspension contained increasing Lipid mixture Aqueous suspension > lipid Plasma concentrations: Fischler et al 1973 monolaurate propylene The solid micellar solution was composed of sucrose 24 fasted healthy Neoral Atovaquone composition (LogP 3.68) Randomized Aqueous Tablets Oil suspension consisted of miglyol Oil suspension ~aqueous > AUC: Rolan et Table 1 Table Compound Study design Formulation Formulation Comments
Therapeutics and Clinical Risk Management 2007:3(4) 593 Fatouros et al l 1977 Bekerman et al Bekerman 2004
Authors Year
cantly cantly fi Charman et al 1993 (fasted) (fasted) with max > Capsule cant differences were found found were cant differences Macheras 1986 fi (fed) ~emulsion (fed) cant difference in plasma diazepam levels in plasma diazepam levels cant difference Yamahira 1979 fi : Neoral ~ G-SEDDS : : Neoral > Sandimmune Neoral > Sandimmune : et al Llaven : Neoral ~ G-SEDDS : max max max higher than Sandimmune. higher than Sandimmune.
decreased upon 2 g corn oil in the o/w-emulsion upon 2 g corn oil in the o/w-emulsion decreased solution was less variable. solution was less variable. increased with increasing lipid amounts and with increasing increased plateau with 6 g of lipid reached with the higher lipid amounts. compared m T µ m) plasma maximum under the curve, area between et.al (a) µ characteristics cross-over rmatory study), rmatory study), median diameter) (volume fi con ) crossover fed/fasted fed/fasted crossover O/w-emulsion mono-oleate and glycerol Emulsion dried drug-milk,liquid, mg drug particles (49 1 4 fasted washout (n = 4), solution MCT lipid However tablets or lipid solution. between et al (n = 5) Tablets suspension with dispersed drug emulsions Absorption > tablets ~ aqueous suspension. mixture washout (n = 9) days Lipid mixture and beeswax oil, vegetable hydrogenated at 90 min > IR dosage form Lipid mixture Continued 1 1 1 ( Dicumarol Dicumarol (logP of 3.66) fasted Human (n = 4), Redispersed freeze- Caspules containing 300 signi Statistically
Diazepam (logP 2.84) 2 weeks Crossover, study (n = 11 females) Lipid Tablets Lipid solution contained 6 months. stable for was physical No signi (logP = 4.53) crossover Four-way study (latin square O/w-emulsion design) (n = 4) Tablets contained 2, O/w-emulsion 6 or 12 g of corn oil 4, o/w- determined via urinary excretion: Bioavailability Bates et a Absorption was signi ~tablets emulsion
lecithin soy Griseofulvin (logP 0.68) Random crossover Aqueous Lipid mixture contained corn-oil in water Lipid mixture Lipid determined via urinary excretion: Bioavailability Bates et al 1975 in acid Flufenamic Random design 15 (logP 3.98) IR dosage form oil contained vegetable Lipid mixture Mean plasma concentrations: et al Angellucci 1976 and 200 ml milk formulation. concentration with dicumarol-milk Danazol Randomized study open cross-over Danocrine capsule contained Emulsion Capsule AUC: Fasted healthy in (three volunteers (G-SEDDS) and Galactolipid SEDDS Galactolipid 300 mg CsA from Galactolipid SEDDS and Neoral, C Neoral > G-SEDDS AUC: RH 40, Chremophor Span 80, Size Tricarpin, phospholipids, 25–400 nm from with small diameter lipospheres Odeberg et al 2003 study and six prediction Neoral SEDDS particle size of 16–20 Compound Study design Formulation Formulation Comments composition 48 fasted healthy two volunteers, parallel study groups, Neoral and cross-over way four volunteers Sandimmune 23 healthy a fat-rich breakfast fed Neoral and Sandimmune 200 mg to 800 CsA as Neoral or Neoral 7.5 mg/kg CsA from and Sandimmune Sandimmune in the dose Neoral exhibit linear bio availability Neoral ~ Sandimmune AUC: Mueller et al did range 200 – 800 mg but Sandimmune 1994a C Neoral 174% to 239% AUC: not. González- 1999 two-way prior dosing, 30 minutes volunteers 24 healthy open randomized SMEDDS and non- SMEDDS capsules 200 mg single oral dose of each 200 mg of cyclosporin, treatment non- AUC SMEDDS > AUC SMEDDS et al Postolache 2002 volunteers 6 healthy Lipospheres, 80, Tween containing lipospheres and C AUC Higher Table 1 Table
594 Therapeutics and Clinical Risk Management 2007:3(4) Oral lipid based formulations: An overview of clinical studies ) Continued (
et al
.com/LXS-DD.asp
Malingré et al 2001
ritanovir ritanovir Kurowski 2003 + ritanovir ritanovir et al cant Cardiello 2003 : : fi ® + ® max brosis brosis patients and AUC of saquinavir of saquinavir AUC and fi max and AUC: Cremophor Cremophor AUC: and Martin- 2002 : micronized progesterone in LCT > all others progesterone micronized : et al Hargrove 1989 : micronized progesterone in LCT < all others progesterone micronized : max max max
in C > powder. Fed state increased state increased Fed > powder. EL dose dependently increase increase EL dose dependently Facklam et al
200 and 400 mL of milk over the 200 and 400 mL of milk over capsule corresponding C ux solution m urinary estimated by excretion Bioavailability Bruni et al 1970 fl µ m) and the nitrofurantoin-milk et.al (b) µ contain no signi AUC: contain Invirase AUC: ® ® LCT Fortovase Fortovase ® ®
medium chain MG and difference medium chain MG and > Fortovase with 0, with 0, ® ® ® ritanovir ritanovir 2 Invirase × Invirase 2 Fortovase bioavailability. Lipid matrix (LXS) is an organized LXS (lymph-X-Sorb™) obtained Equivalent peak plasma concentrations were http://www. 2003 HIV-patients HIV-patients 6 24) = Invirase brosis) brosis) fi MG and FA lecithin (n = 6) 80 solution cyclosporine to inhibit the P-gp ef milk,liquid, 200 or 400 ml milk with regenerated formulations 1 1 1 1 placebo-controlled, 1000 or 5000 100, double blind controlled TG-based (OLM) CMC = 0.1 mM and consists of cystic each administration solution powder particle size) Oil solution in sesame oil oil solution > suspension (fasted state): IR Plain- Micronized high in long-chained were T 1 1 1 1 ritanovir co-dosed with DG
(n n = 13 HIV-patients, Fortovase study crossover co-dosed with DG Open-label, washout (n = 8) week randomized, randomized, four double-blind, mg Cremophor 1 phase cross-over. Fed/fasted Q10 (Ubiquinone) randomized way Four (LogP not Oil solution and computable) 1–4 times) repeated amounts Different oil. Oil was soybean (n = 10) cross-over oil solution > other lipid AUC: and Tween Oil, Tween administered. of oil and surfactants were formulations et al Weis 1994 (LogP 2.73) Retinyl palmitate Retinyl randomized, 1-year, (LogP 11.0) Organized lipid matrix OLM has a melting point of 64.6 °C, supplement in TG-based nutritional OLM > AUC: Lepage et al 2002 Micronized milled in LCT unsaturated FA Quingestrone Quingestrone washout (n = 1, 1 week Oil suspension Oil suspension in sesame oil (10 retinamide retinamide (Fenretinide) formulation Capsule lipid matrix that consists of lyso-PC, compound amount. using LXS at 20% of the capsular avantilipids. 4-Hydroxylphenyl Humans 4-Hydroxylphenyl (LogP 4.91) (LogP not computable) washout 2 weeks solution Polysorbate all patients received to administration, EL > Cremophor Polysorbate Progesterone washout (n = 7) 3 days transporter. Plain-milled IR of natural oils that LCT was a mixture C (LogP 3.78) Randomized, Saquinavir trial (n = 78 with cystic supplement nutritional in 1:4:2) FA MG, (lyso-PC, Lym-X-Sorb Paclitaxel Randomized design, EL Cremophor Prior Both solutions contained ethanol. and C AUC (LogP 5.50) Nitrofurantoin (LogP not Human n = 4 fasted Redispersed freeze- containing 100 mg Capsule superiority of Urine data revealed dried drug- drug particles ( 75 Macheras 1986 computable)
Therapeutics and Clinical Risk Management 2007:3(4) 595 Fatouros et al
For practical reasons the studies presented in the current work have been categorized based on the physicochemical characteristics (oil, suspensions, emulsions, SEDDS, and surfactant based systems). Furthermore the studies on the effect of food to oral bioavailability in humans have been reviewed.
Authors Year
Oil solutions Lipid based formulations can enhance the bioavailability of
poorly soluble drug substances by keeping the compound in (fed) solution. Medium chain triaglycerides (MCT) and long chain triaglycerides (LCT) are commonly used for the lipid based formulations. Several major differences exist between MCT
and Holmberg et al 1990 LCT in respect with their in vivo fate, such as lipolytic LCT oil solution
(fasted) products, differences in the modulation on gastric emptying (fed) (fasted) (Hunt and Knox 1968) and the contraction of the gallbladder in humans (Ladas et al 1984). Long chain lipolytic products delay gastric emptying and facilitate the contraction of the gallbladder to a larger extent than the medium chain lipolytic : SEDDS1~SEDDS2>oil solution : max
~ MCT oil solution products (Hunt and Knox 1968). The bioavailability of quingestrone (progesterone 3-cyclopentyl enol ether belonging to the enol ethers of ∆3-kerosteroids) after oral administration to human subjects was evaluated by measuring the urinary excretion of m SEDDS2
µ or an oil solution in 2.5 ml of sesame oil. Bioavailability of quingestrone was determined as the total urinary secretion for characteristics C both metabolites. The urinary excretion was highest when the oil solution (6.43 ± 0.18 mg was administered compared with the oil suspension (2.96 ± 0.27 mg) or and the micronized powder (0.91 ± 0.02 mg). ne) of 10.6 fi The relative bioavailability of clomethiazole (base form)
solution was investigated after oral administration of two arachis oil formulation both in capsules and a tablet formulation to fasting healthy volunteers (Fischler et al 1973). Higher plasma concentrations were obtained from the oil fi lled capsules and the absorption was more rapid, compared with tablets. The two arachis oil formulations consisted of clomethiazole and arachis oil 1:1 (384 mg + 384 mg)
) or 1:2 (384 mg + 768 mg) on a weight basis, respectively. Increasing the arachis oil to clomethiazole ratio from 1:1 to 2:1 resulted in a 1.5 times increase in C suggesting an 1 week study, over Oil solution that consists of pegylated FA-esters solution (n = 6) over SEDDS 2 ( max Oil solution (MCT) 812 MCT was Miglycol 1 1 Continued 1
( increase in bioavailability with increase of co-administered ; 4Bakatselou et al described in the text ; and Neoral are characteristics of Sandimmune 3Composition and formulation et al 1993; Tayar 2el USA); 8.0.3 (Cambridgesoft, Ultra version 1LogP values calculated using ChemDraw® oil amount. Doubling the clomethiazole dose, but keeping -tocopherol -tocopherol 2-sequence cross- Solid dispersions 44/14 Solid dispersion contains Gelucire 44/14 > oil Gelucire AUC: al et Barker 2003 Tocotrienols (LogP 8.56) cross- way Three microscopy washout (n = 6) SEDDS 1 (coarse) diameter a droplet SEDDS 1 produced SEDDS 1~SEDDS2>oil AUC: Confocal DSC, and glycerides. Yuen and Yap 2004 α composition (LogP 7.98) Compound Study design Formulation Formulation Comments Table 1 Table
(logP 7.04) Vitamin E cross-over way Two Soft gelatine\commerci- Span, SEDDS (Tween, study n = 8 SEDDS available ally SEDDS> soft gelatin : AUC palm oil) et al Julianto 2000 Vitamin D n = 24 Fed/fasted, Oil solution (LCT) oil and LCT was peanut LCT oil solution AUC: Oil Notes: of su bjects partici- 6The number of CsA; exposure comparative was selected to provide Neoral and 150 mg CsA from Sandimmune (A) 300 mg CsA from 5See www.emea.eu.int/humandocs/PDFs/EPAR/Fortovase/093398en4.pdf; 1991; pated in this study is not stated. the oil: drug ratio at 1:1 resulted in a 1.7 times increase
596 Therapeutics and Clinical Risk Management 2007:3(4) Oral lipid based formulations: An overview of clinical studies compared with a tablet formulation indicating nonlinear The effect of medium chain triglyceride was studied by absorption kinetics. The results emphasizing the impact of Rolan et al in 1994. In this approach, absorption of 500 mg the solubilization capacity of the lipid vehicle to the overall atovaqoune (an antiprotozoal agent) from either a suspen- performance of the formulation sion in 30 ml of Miglyol (fractionated coconut oil, medium A soft capsule contained 200 mg of fl ufenamic acid chain triglyceride), an aqueous suspension using 0.25% (analgesic anti-infl ammatory agent), 100 mg of vegetable oil, w/v methylcellulose as suspending agent, or two 250 mg 40 mg of hydrogenated vegetable oil, 8 mg of beeswax and tablets of atovaquone were used. The AUC values were at 5 mg of soya lecithin and a hard gelatin capsules contained the same levels for the aqueous and oil suspensions (144 ± 200 mg fl ufenamic acid and 20 mg of magnesium stearate 48 µg/ml–1 hr and 144 ± 124 µg/ml–1 hr respectively), the were orally administered to healthy volunteers (Angelluci Cmax increased for both formulations (2.7 ± 1.1 µg/ml for the et al 1976). The plasma levels more than doubled 90 minutes aqueous suspension and 2.0 ± 1.8 µg/ml for the oil) compared after administration of the lipid solution compared with the with the tablets (1.1 ± 0.6 µg/ml). The higher absorption of powder formulation. However after 180 and 360 minutes atovaquone from the aqueous and oil suspensions compared the plasma levels were lower after administration of the soft with tablets was attributed to a better dispersion, allowing a capsules. Due to the low number of plasma samples, it as not faster solubilization of the drug. possible to identify any signifi cant differences. The different Two oil solutions containing long chain and medium plasma curves observed for the two pharmaceuticals forms chain triglycerides with vitamin D were administered in can be attributed to physicochemical factors, triggered by human volunteers (Holmberg et al 1990). Higher AUC values excipients present in the soft gelatin capsule which result in obtained for the LCT solution (peanut oil) compared with the a faster absorption of the drug. MCT solution (Miglyol 812) suggesting that the presence of The absorption of diazepam (benzodiazepine deriva- long chain fatty acids enhance the absorption of vitamin D. tive) in fasted human subjects when dosing 5 mg in either a In four out of seven studies the presence of long chain medium chain triglyceride formulation (2% diazepam-MCT triglycerides (LCT) signifi cantly increased the oral absorption solution) or a commercially available tablet formulation of the drugs (Bruni and Galleti 1970; Fischler et al 1973; (Serenzin®, Sumitomo Chemical Co) in a crossover study was Hargrove et al 1989; Holmberg et al 1990). Furthermore investigated by Yamahira and colleagues (1979). There was LCT formulations demonstrated higher oral absorption no signifi cant difference in the average plasma profi les arising compared with MCT ones (Holmberg et al 1990). In contrast from the two formulations. The authors mainly attributed this no signifi cant difference in the absorption observed among to the large intersubject variation and the limited number of oil formulations containing medium chain triglycerides MCT subjects (n = 4) enrolled in the study. However upon repeated and tablets (Yamahira et al 1979; Rolan et al in 1994). Finally administrations, the MCT solution showed less intrasubject in one case the presence of LCT solutions didn’t enhance the variation. The authors suggested that MCT solution of absorption of the drug compared with the tablets (Angelluci diazepam produced a more uniform drug absorption rate et al 1976). Despite the fact that the results seem to be case compared with the tablets for individual patients. specifi c, there are indications that LCT oil solutions perform The impact of the physical characteristics of a formula- better compared with MCT solutions. However it is rather tion has been demonstrated by Hargrove and colleagues diffi cult to draw conclusions because of the limited number (1989) in a clinical study with progesterone. Progesterone of studies and the fact that there are no clear indications of is indicated for use in the prevention of endometrial hyper- which characteristics of a compound that are determining plasia in nonhysterectomized postmenopausal women. Four whether LCT or MCT oils give better bioavailability. different preparations were made: plain milled progesterone, micronized progesterone, plain-milled progesterone in oil, Emulsions micronized progesterone in oil. Suspending the micronized Another approach to increase the bioavailability of poorly progesterone in oil more than doubled the Cmax, compared water solule drugs is to formulate them in lipid emulsions. with micronized powder, while there was no effect of Bioavailability studies in humans were carried out suspending the nonmicronized progesterone in oil. These using o/w-emulsions, aqueous suspensions and commercial results emphasize the effect of concomitant lipid digestion tablets (Bates and Sequeira 1975) of griseofulvin which in dissolution of drug, at the same time also showing that the is an antifungal antibiotic and it is commonly used in the particle size of the solid drug is important. treatment of dermatophyte infections in humans. Four dosage
Therapeutics and Clinical Risk Management 2007:3(4) 597 Fatouros et al forms containing micronized griseofulvin were evaluated a single melting point (64 °C) which is much lower by administration to humans. The corn o/w emulsion and than the melting point of its individual components. It aqueous suspension both contained 300 mg of polysorbate 60 produces an eutectic matrix due to the interaction of in emulsion and 500 mg of suspended griseofulvin per 30 g of monoglycerides with the lyso-phopshatidylcholine. The formulation. Two different commercial tablets contained 500 absorption in cystic fibrosis adolescent patients, who have mg of micronized griseofulvin. The four dosage forms were an impaired pancreatic function, was compared with that administered in a random cross-over fashion to fi ve fasting upon administration of the same dose of retinyl palmitate subjects in the form of 30 g of the corn oil emulsion, 30 g given along with a conventional isocaloric TG based aqueous suspension and the tablets and the drug absorption nutrional supplement. It was concluded that the AUC was monitored from urinary excretion data for the major (over baseline) for the organized lipid matrix was 10-fold metabolite (6-desmethylgriseofulvin). Administration of o/w higher compared with the TG based formula. The orga- emulsion resulted in a 3–4 fold increase approximately, in the nized lipid matrix formula needed no hydrolysis before maximum excretion rate (mg/hr) of total 6-desmethylgriseo- absorption, and could lead to a less variable absorption. fulvin. The bioavailability from the o/w emulsion increased This organized lipid matrix emulsion has also been tested 2 times compared with the other dosage forms. Such behavior in human subjects with the compound 4-Hydroxylphenyl- attributed to the presence of linoleic acid contained in the retinamide (4-HPR) a retinoid compound with antileu- lipolysis products which may inhibit the GI motility and kemic and proapoptotic activity in acute lymphoblastic stimulate the gallbladder evacuation. leukemia. The bioavailability of a 4-HPR oral capsule In a follow up study the effect of different amounts of formulation resulted in poor patient compliance at higher lipid emulsion was investigated (Bates et al 1977). However doses (due to the large number of capsules) in phase I the total 6-desmethylgriseofulvin excretion rate profi les trials of high-dose oral 4-HPR and high inter-patient were quite similar after administration of the ultramicrosize variations. Using a lipid matrix emulsion it was possible griseofulvin tablets and 5, 10, 15, and 30 g of o/w emulsion to reduce the dose to 20% of the traditional formulation used in the that study. and still obtain equivalent peak plasma concentrations Lipid and surfactant based systems have been used (Avanti® Polar Lipids, Inc. 2006). for coenzyme Q10 (ubiquinone) which is a lipid soluble Dry emulsions consisting of redispersed freeze- antioxidant, suggested to have a positive influence in dried drug-milk formulations have been evaluated using congestive cardiac failure. The absorption of coenzyme Q10 reconstitution in vitro to obtain an o/w emulsion before was followed upon administration in different lipid dosage administration for dicumarol (an anticoagulant) (Macheras forms (Weis et al 1994). The tested formulations included and Reppas 1986a) and nitrofurantoin (a bacteriocidal) 100 mg coenzyme Q10 i) suspended in soy bean oil (400 mg) (Macheras and Reppas 1986b) respectively. Capsules administered in a soft gelatine capsule ii) in a mixture of containing the pure drugs were used as controls in both polysorbate 80 (20 mg), phoshatidylcholine (100 mg) and cases. Analysis of urine data revealed superiority of soybean oil (280 mg) and iii) in a mixture of polysorbate the nitrofurantoin-milk formulations regenerated with 80 (20 mg) and soy bean oil (380 mg). The suspension in 200 and 400 ml of milk over the corresponding capsule pure soy bean oil increased the bioavailability of coenzyme formulations in the rates and extents of nitrofurantoin Q10 signifi cantly compared with the lipid mixtures that excretion. Determination of the plasma dicumarol levels included surfactant systems. It was suggested, that possibly indicated superiority of the dicumarol-milk formulation. the formation of mixed micelles containing polysorbate 80 Statistically significant differences were found between phosphatidylcholine and soyabean oil for formulation ii) and area under the curve, maximum plasma concentration, micelles from polysorbate 80 and soyabean oil for formula- and apparent elimination rates. tion iii), in the intestine might induce a decrease of the drug The introduction of the organized lipid matrix emulsion solubilization with the bile salts. which can be used with further hydrolysis is an interesting Recently the effect of an organized lipid matrix emul- approach. The main advantage of this formulation is that no sion containing lyso-phosphatidylcholine, monoglycer- hydrolysis is needed which can result in less variations in ides, and fatty acids (ratio 1:4:2) was given along with absorption. However it is rather diffi cult to draw conclusions a traditional capsule formulation of retinyl palmitate for the applicability of this formulation since the data from (Lepage et al 2002). The lipid matrix emulsion yields the literature are quite limited at the moment. Furthermore
598 Therapeutics and Clinical Risk Management 2007:3(4) Oral lipid based formulations: An overview of clinical studies the potential of the drug-milk freeze-dried formulations for the market in both hard gelatine capsule (Invirase), containing the enhancement of the bioavailability of sparingly water saquinavir with lactose microcrystalline cellulose, povidone soluble drugs are not yet elucidated. K30, sodium starch glycolate, talc and magnesium stearate and soft gelatin capsules (Fortovase), containing saquinavir Self emulsifying drug delivery in solutions in medium chain mono- and -diglycerides, systems povidone and dl-alpha tocopherol. In a recent study (Roche Self (micro) emulsifying drug delivery systems have attracted Laboratories, Inc 2004), a signifi cant improvement to the a lot of attention lately. Their thermotropic stability and the bioavailability up to 331% of soft gelatine capsules compared high drug loading effi ciency make them a promising system with hard gelatine capsules. Ritonavir (Norvir) is available for poorly water soluble drugs giving particles with small as in soft gelatin capsules with oleic acid, ethanol, polyoxyl size. 35 castor oil titanium dioxide, iron oxide and butylated Julianto and colleagues (2000) demonstrated that a hydroxytoluene (FDA 2006). SEDDS increased the bioavailability of α-tocopherol When healthy subjects were administered with saquinavir/ approximately 210%–410% compared with a commercially ritonavir 1600 mg/100 mg, and no difference in AUC0–24 available soybean oil solution. The antioxidant α-tocopherol, values between hard and soft gelatine capsules was observed, a lipid soluble vitamin (vitamin E), is a poorly aqueous respectively (Cardiello et al 2003). The pharmacokinetics soluble compound. Traditionally α-tocopherol has been and the safety of a boosted saquinavir/ritonavir combination formulated as a capsule containing a simple oil solution. either with hard gelatin capsules or soft gelatin capsules was Both the SEDDS and the commercial α-tocopherol formula- evaluated (Kurowski et al 2003). Comparable plasma expo- tion contained 400 mg α-tocopherol. The SEDDS contained sures with saquinavir achieved when saquinavir boosted with Tween 80: Span 80: Vitamin E (4:2:4) and the commercial ritonavir, accompanied by an improvement in gastrointestinal formulation contained of α-tocopherol dissolved in soy bean system disorders. oil. Both the AUC0–∞ and Cmax values of the SEDDS were The turning point for development of the oral lipid and markedly higher than those of the oil solution, while the Tmax surfactant based formulations of poorly soluble drugs was was shorter, indicating a higher rate and extent of absorption. the introduction to the market of cyclosporine A (CsA) in a No physicochemical characterization was performed and it lipid based formulation. CsA is an immunosuppressant used was not stated if the lipid amount were equal between the in chronic treatment of organ transplant receivers to sup- formulations. press graft rejection and in the treatment of severe rheumatic In a recent study the bioavailabilities of three different arthritis and severe psoriasis. lipid formulations of tocotrienols (Yap et al 2004) were The Sandimmune and Sandimmune Neoral formulations investigated. Two self-emulsifying systems SEDDS, of CsA are perhaps the best known examples of a marketed containing varying amounts of soybean oil, Tween 80 and lipid and surfactant based systems and the pharmaco kinetic Labrasol (C8/C10 polyglycolyzed glycerides from coconut has been studied and reviewed extensively (Holt et al 1994; oil) was analysed according to their self-emulsifying Ritschel 1996). Cyclosporine was introduced in 1981 in properties, droplet sizes and extent of lipolysis. The droplet Europe in a self-emulsifying formulation (Sandimmune) size was 1.5 and 10.6 µm respectively. Formulations con- containing Labrafi l M 1944 CS (polyoxyethylated oleic taining 200 mg mixed tocotrienols administered in healthy glycerides), olive oil and ethanol (Klyashchitsky and Owen adults as SEDDS or simple soybean oil solution stated that 1998). This formulation disperses, when diluted with water, the SEDDS had markedly higher plasma levels and a faster into a polydisperse oil-in-water macroemulsion. In 1994 a new onset of absorption compared with the oily solution. The self-microemulsifying formulation (Sandimmune Neoral, droplet size difference between the two SEDDS did not affect referred to as Neoral in the following) was introduced, the bioavailabilities. However, the total dosed lipid amounts which emulsifi es spontaneously into a microemulsion with a between the three explored formulations ranged from particle size smaller than 100 nm. This formulation contains 31.3–351.3 mg soybean oil thus making the interpretation Cremophor RH40 (polyoxyl hydrogenated castor oil), corn of the bioavailabilities more complex. oil glycerides, propylene glycol and ethanol (Klyashchitsky The potent anti-HIV drug saquinavir has been enrolled and Owen 1998). in studies with human subjects in lipid based formulations. Drewe and colleagues investigated the absorption of It has been formulated into SEDDS. The drug is available in CsA (150 mg) from three experimental formulations using
Therapeutics and Clinical Risk Management 2007:3(4) 599 Fatouros et al
Sandimmune as reference in fasting healthy volunteers. a statistical signifi cant lower absorption rate. The authors Two of the experimental formulations were microemul- claimed that with this study it has been demonstrated that in sions with respectively fast and slow in vitro release. The vivo the non-SMEDDS capsules are not totally interchange- microemulsions were formulated using polyethylene glycol, able compared with the SMEDDS capsules. In a different hydrogenated castor oil, medium chain triglycerides and low approach cyclosporine lipid nanoparticles (lipospheres) molecular weight glycols. The last experimental formula- were developed (Bekerman et al 2004) and the effect of tion was a solid micellar solution with fast in vitro release composition and particle size investigated. These lipospheres composed of sucrose monolaurate and propylene glycol. consisted of phospholipids, Span 80, Tween 80, Tricaprin, The fast releasing microemulsion and the fast releasing and Cremophor RH 40. Cyclosporin dispersions systems solid micellar solution exhibited signifi cant higher Cmax and resulting in particle size of 25 up to 400 nm prepared. Oral bioavailability (141% and 139% of Sandimmune, respec- bioavailability indicated a correlation between the AUC tively). The slow releasing microemulsion was equivalent and Cmax and the particle size of the dispersion showing the to Sandimmune with respect to Cmax and bioavailability. highest values for particles with 25 nm diameter. The clini-
The Tmax of the tested formulations were in the same range cal formulation of Neoral used as a reference. This study (Drewe et al 1992). demonstrated a correlation between the particle size and the Mueller and colleagues (1994a) investigated the dose oral bioavailability of cyclosporine formulations. Further- linearity of CsA from Sandimmune and Neoral in healthy more the composition and possibly the surface properties volunteers who had been fasting 12 h prior to and 4 h after of the lipid nanoparticles (lipospheres) could affect the oral administration fi nding that Neoral exhibited linear dose bioavailability of cyclosporine. AUC relationship in the range of 200 to 800 mg in contrast The above clearly demonstrate that SEDDS can be a to Sandimmune. The relative bioavailability of Neoral promising formulation approach for poorly water soluble compared with Sandimmune was found to be in the range of drugs. The effi ciency of these systems is rather case-specifi c 174–239% but dependent on the actual dose, with the high- depending on the composition of the formulation used each est dose resulting in the largest difference en bioavailability. time. It must be noted that the bioavailability for Neoral and a number of the generic products of CsA exhibited marked Surfactant-based solutions pharmacokinetic variations in different human subpopula- Cremophor is commonly used excipient in lipid-based tions even though they had been found bioequivalent. This formulations (Strickley 2004). Cremophor EL may inhibit discussion is however beyond the scope of this article and has PGP (p-glycoprotein), a relevant effl ux pump which is been reviewed recently elsewhere (Dunn et al 2001; Pollard expressed in high amount in gut, biliary tract and the blood et al 2001). Studies in fasted healthy volunteers comparing brain barrier, and enzymes like CYP3A which may contribute Sandimmune (300 mg CsA) and Neoral (180 mg CsA) it was to fi rst pass metabolism. found that Neoral exhibit shorter Tmax and higher Cmax and Cremophor EL, is used as a vehicle for the solubilization AUC (Mueller et al 1994a, 1994b; Kovarik et al 1994). of a wide variety of hydrophobic drugs, including anesthetics, Recently an experimental CsA self-emulsifying photosensitizers, sedatives, immunosuppressive agents and galactolipid formulation has been compared with Neoral anticancer drugs including the anticancer agent paclitaxel in fasting healthy volunteers (Odeberg et al 2003). (Taxol). The self-emulsifying galactolipid formulation forms an The role of the surfactants Cremophor EL and Tween 80 emulsion with mean particle size of 16–20 µm. The galac- and the absorption rate of paclitaxel was investigated tolipid formulations exhibited identical Tlag and Tmax but by Malingre and colleagues (2001). In this study the slightly lower Cmax (94%) and AUC (84%) when compared 2 formulations were tested; 6 mg/ml paclitaxel in Cremo- with Neoral (Odeberg et al 2003). In another study the phor EL: ethanol 1:1 v/v or in polysorbate 80: ethanol 1:1 bioavailability of a non-SMEDDS, which is semisolid opaque v/v. The Polysorbate 80 formulation performed better than oily suspension that in aqueous solution forms particles Cremophor EL formulation, both with regard to AUC and with a mean diameter 200 nm, and a SMEDDS formula- Cmax. The results demonstrated that the presence of Cremo- tion of cyclosporine were tested (Postolache et al 2002). phor EL limited the oral absorption of paclitaxel and they The results showed that the non-SMEDDS formulation was highlight the need of designing a better drug formulation in not bioequivalent with the SMEDDS formulation due to order to increase the oral absorption of paclitaxel.
600 Therapeutics and Clinical Risk Management 2007:3(4)