MECHANISTIC STUDIES TO ELUCIDATE THE ROLE OF LIPID VEHICLES ON SOLUBILITY, FORMULATION AND BIOA VAILABILITY OF POORLY SOLUBLE COMPOUNDS

Lipids have been utilized to increase the bioavailability of poorly soluble drugs which resulted in with advent of High Throughput Screening (HTS). Although lipids have been used to improve bioavailability of a few drugs for almost twenty years, they are still not well characterized. There are limited publications about the effects of physicochemical properties of the lipid vehicle, such as class of lipid (i.e. glyceride, propylene glycol ester), fatty acid chain length, MW and polarity on lipid solubility, formulation and bioavailability of poorly soluble drugs. Knowledge of such relationship can improve quick screening of these vehicles. The goals of this study are to identify physicochemical properties of the lipids, to investigate the relationship of these properties with solubility of the model drugs, namely nifedipine and griseofulvin, to design lipid-drug formulations with Cremophor EL, to evaluate in vitro dissolution of selected formulations and to compare the bioavailability of nifedipine from the formulations tested. It has been shown that the lipids used improved the solubility of nifedipine and griseofulvin compared to the solubility of drugs in water. Calculated solubility parameter was not sufficient to predict the solubility of the drugs in the lipids. Calculated and measured properties of lipids analyzed with stepwise regression analyses showed that MW, dielectric constant, surface tension and fatty acid chain length are the common factors that govern the lipid solubility. However. the estimates of each factor were different for each drug indicating that the nature of the drug played an important role in lipid solubility. Incorporation of Cremophor EL. a nonionic surfactant into lipid-based nifedipine formulation enhanced the solubility and dissolution of nifedipine. The solubility of ( ( nifedipine showed a linear correlation with increasing surfactant concentration. While solution rate was dependent on the type of lipid used, the nature of the lipid had no affect on the dissolution of nifedipine in presence of the surfactant. Dissolution of nifedipine from the lipids showed that as the fatty acid chain length increases, the dissolution rate increases. This is due to lower solubility of nifedipine in the lipids that have longer chain length. The effect of lipids on dissolution rate and extent of nifedipine showed that even though physicochemical properties of lipids (HLB, interfacial tension, viscosity, density) and solubility of nifedipine in lipids play a role, only partitioning of the drug from lipid to the dissolution medium and the particle size of the formulation in dissolution medium provided a good correlation with dissolution extent of nifedipine. The bioavailability of nifedipine obtained with different lipid formulations in beagle dogs showed that the type lipid and surfactant used in the formulation play important roles. Dissolution is a good predictor for in vivo performance of nifedipine lipid formulation when it was formulated with nondigestible lipids (mostly mono-glycerides). Although, the solubility of nifedipine in lipids, the particle size of the formulation in dissolution medium and partitioning of the drug from the formulation to dissolution medium seem to affect the in vivo performance of the drug, dissolution performance of the formulation and digestibility of the lipid used in the formulation are the major factors in bioavailability of nifedipine from the lipid-based formulations.

relationship can improve quick screening of these vehicles. The goals of this study are to identify physicochemical properties of the lipids, to investigate the relationship of these properties with solubility of the model drugs, namely nifedipine and griseofulvin, to design lipid-drug formulations with Cremophor EL, to evaluate in vitro dissolution of selected formulations and to compare the bioavailability of nifedipine from the formulations tested. It has been shown that the lipids used improved the solubility of nifedipine and griseofulvin compared to the solubility of drugs in water. Calculated solubility parameter was not sufficient to predict the solubility of the drugs in the lipids. Calculated and measured properties of lipids analyzed with stepwise regression analyses showed that MW, dielectric constant, surface tension and fatty acid chain length are the common factors that govern the lipid solubility. However. the estimates of each factor were different for each drug indicating that the nature of the drug played an important role in lipid solubility.
Incorporation of Cremophor EL. a nonionic surfactant into lipid-based nifedipine formulation enhanced the solubility and dissolution of nifedipine. The solubility of ( ( nifedipine showed a linear correlation with increasing surfactant concentration. While solution rate was dependent on the type of lipid used, the nature of the lipid had no affect on the dissolution of nifedipine in presence of the surfactant. Dissolution of nifedipine from the lipids showed that as the fatty acid chain length increases, the dissolution rate increases. This is due to lower solubility of nifedipine in the lipids that have longer chain length. The effect of lipids on dissolution rate and extent of nifedipine showed that even though physicochemical properties of lipids (HLB, interfacial tension, viscosity, density) and solubility of nifedipine in lipids play a role, only partitioning of the drug from lipid to the dissolution medium and the particle size of the formulation in dissolution medium provided a good correlation with dissolution extent of nifedipine. The bioavailability of nifedipine obtained with different lipid formulations in beagle dogs showed that the type lipid and surfactant used in the formulation play important roles. Dissolution is a good predictor for in vivo performance of nifedipine lipid formulation when it was formulated with nondigestible lipids (mostly mono-glycerides). Although, the solubility of nifedipine in lipids, the particle size of the formulation in dissolution medium and partitioning of the drug from the formulation to dissolution medium seem to affect the in vivo   Classification (Redrawn with the permission of publisher of ref. Pouton, 1999) 55 Table 3 (   Table VI. Linear regression analysis with stepwise fit for solubility of griseofulvin 109 Table VII Table II. Jn vivo study experimental design (crossover study in fasted beagle dogs) 172 Table III

INTRODUCTION
During the development of oral pharmaceutical dosage forms of new drug candidates, the formulation scientists are faced with many challenges. One challenge which is very often encountered is poor bioavailability. With the advent of High Throughput Screening (HTS) and Combinatorial Chemistry in drug discovery, modern drug molecules are more lipophilic which often results in their being poorly bioavailable.
The use of lipid vehicles has generated considerable attention in recent years for improving the bioavailability of poorly soluble drugs as one of the approaches to improve their oral bioavailability. The well-known effect of food in improving the bioavailability of many poorly soluble drugs such as griseofulvin, carbamazepine and danazol is evidence that lipids can be beneficial to drug absorption Humberstone and Charman, 1997). Lipids can improve the bioavailability of drugs through several mechanisms such as by enhancing solubility/ dissolution August et al., 1997), by providing uniform gastric emptying and uniform absorption rate (Y amahira et al, 1979a, b, 1980) and by increasing lymphatic absorption August et al., 1993;Porter and Charman, 1997).
Although lipids have been used to improve bioavailability for almost twenty years, they are still not well characterized. In addition, mechanistic studies that explain the effect of physicochemical properties of lipid vehicles on formulation and drug absorption, the use of appropriate techniques to characterize lipid-based systems and in vivo/in vitro correlations have not been adequately addressed and well understood.
In this paper, information regarding these issues is collected and discussed in the

DEFINITION AND CLASSIFICATION OF LIPIDS
Lundberg and Weiner define lipids as a "chemically heterogeneous group of substances, having in common the property of insolubility in water but solubility in non-polar solvents" and the science that deals with lipids is referred to as 'Lipidology' (Lundberg, 1984;Weiner, 1993). However, many lipids have some limited water solubility and are commonly used as vehicles in pharmaceutical formulations. Lipids can be classified based on chemical characteristics such as carbon chain length of fatty acid and ester type rather than on their solubility. Surfactants also fall into this category. Sometimes, there is no distinct difference between a lipid and a surfactant.
Some lipids have surface active properties, for example medium chain monoglycerides can lower interfacial tension and easily dispersed in aqueous media.
Types of lipids include natural and synthetic fatty acids, oils, fats, waxes, phospholipids and some surfactants. Fatty acids can be defined as organic acids with straight hydrocarbon chain and can be classified based on the length of the hydrocarbon chain. Fatty acids with hydrocarbon length C 2 -C 6 are considered to be short~hain, with Cs-C12 medium-chain and with C 14 and higher long-chain fatty acids. Surfactants that are in the lipid category are commonly fatty acid esters of PEGs.
Lipids also include some lipid fractions and synthetic accommodations of glycerides with PEG esters. The fractionated esters are seldom pure and tend to exist as mixtures of various combinations. Commonly used lipid vehicles in oral drug delivery have at least one fatty acid esterified with a polyol such as glycerol/s, propylene glycol or PEG. Tables 1 A-F present lipid vehicles commonly used in the US pharmaceutical industry in oral drug delivery products. Phospholipids are excluded from the scope of this paper since they are mainly used in parenteral formulations.

TYPE OF LIPID FORMULATIONS
In addition to simple solutions of drug molecules in lipids, solid dispersions, micelles, liquid crystals, emulsions including self-emulsifying systems and microemulsions and solid lipid nanoparticles (SLN) are also commonly used formulation approaches for lipid based systems.
Because of their unacceptable taste, many lipid formulations are administered in soft or hard gelatin capsules as unit dosage forms.
Lipid formulations were classified by Pouton (1999 and as Types I, II, IIIA and IIIB. This classification is based on composition such as amount(%) of the glyceride, surfactants and hydrophilic co-solvents and physical attributes including particle size ( of the dispersion, impact of aqueous dilution and digestibility (Table 2). This classification can be useful in comparing different studies and can serve as a guide to estimate physicochemical properties of the lipid formulation of interest. The incorporation of a drug in the vehicle can significantly change the physicochemical properties of the system in terms of particle size and dispersibility following aqueous dilution and may result in a shift of the system from one type to another. Therefore, such classification cannot be a significant predictor of the physicochemical properties of the end product and is especially so for surface active and high dose drugs. The classification of lipid systems as solutions, suspensions, solid dispersions, micelles, emulsions including self-emulsifying systems (SEDDS) and microemulsions and solid lipid nanoparticles appears to be universally acceptable. Examples of these formulations that are cited in the literature are summarized in Tables 3-9. As can be seen from these tables, SEDDS are the most commonly used formulations and drugs with molecular weights around 300 are the most commonly used model drugs. By presenting these tables, we wanted give a guideline for the readers. Additional information on selected examples of lipid solutions, suspensions, emulsions and solid dispersions can be found in a review by Humberstone and Charman (1997), SEDDS by  and , and solid lipid nanoparticles by Muller et al. (2000).
Despite the numerous studies shown in these tables, only a relatively few commercial formulations of lipid based systems are available on the market. These are cyclosporine (Sandimmune™ and NeoraFM, Novartis), saquinavir (Fortovase™, ( \, Roche), ritonovir (Norvir™, Abbott) and fat-soluble vitamins. Possible reasons for the rarity of commercially available lipid-based products are the physical complexity of lipids, reliability of supply (quality), lack of knowledge of mechanisms of drug absorption and fate of the lipid in the gastrointestinal tract. An additional factor may be the limited solubility of pharmaceutical compounds in lipid solvents. Overall, the lack of mechanistic understanding of the role oflipids in drug absorption is a key reason for their underutilization for formulation development.

Selection of lipid vehicles
For a lipid to be used in a drug delivery system it must be safe and ,therefore, the first requirement for these excipients is to be classified as GRAS (generally recognized as safe) materials. The majority of lipids used in marketed products do fulfill this requirement. In lipid formulation development, potential toxicity of any surfactants need should be considered carefully. This evaluation includes not only the type of the surfactant but also the concentration/dose (Swenson et al., 1994).
Drug solubility in the lipid is an important factor in the selection of the lipid as a vehicle. Determination of solubility of drugs in lipids is a major challenge since solubility of drugs in lipid based system is a poorly understood concept. There is no published systematic study to show a relationship between solubility in lipid and physicochemical properties of the drug such as molecular weight, logP and melting point. It has been shown that in the case of an investigational anti-HIV agent solubility increases as ester bond concentration of lipid increases .
Determining the solubility parameter of a lipid vehicle and thereby quantifying its cohesive energy can offer a solution for screening of lipid candidates. Calculated solubility parameters can be used for this purpose. The closer the solubility parameter of a drug to that of a lipid vehicle, the more soluble would be that drug in such a vehicle (Dumanli et al., 2000).
The formulation scientist should also take into account the fact that the solubility of a drug may change due to the changes in the vehicle during the products shelf life. For example, partitioning of water from the gelatin capsule shell into the lipid can have a significant effect on the solubility of the drug. Therefore, the hygroscopicity of the lipid used in such formulation should be carefully evaluated (Fig. I). Furthermore, the dissolved drug in a solid lipid may precipitate out during processing when the formulation is cooled down to room temperature. The solubility of a drug in solid lipids should be significantly higher than final concentration in the formulation.
Solubility should be determined under extreme conditions that the product would be subjected to during processing, manufacture and its shelf life. It should also be kept in mind that the release of a drug from the vehicle is the inverse function of its solubility in the lipid solvent. Therefore, chemical potential (concentration/solubility) and partitioning of the drug from lipid to immediate aqueous environment should be considered in addition to drug solubility and concentration .

(
To optimize both the lipid solubility and release of the drug, co-solvent and surfactant/s may be added to the formulation. Cremophor EL a commonly used surfactant has increased the solubility oflipophilic drugs, griseofulvin and nifedipine, Table 10. The ratio of co-solvents should be optimized to avoid potential precipitation of the drug following aqueous dilution. Lipid vehicle, surfactant, co-solvent and drug combinations can be evaluated upon aqueous dilution using phase diagrams Constantinides, 1995 andKim et al. 2000).

2. Effect of HLB
As mentioned above, there is no distinct difference between a surfactant and a lipid that is surface active. In order to achieve good dispersibility and absorption, it may be necessary to use both surfactant/sand lipid vehicle/sin formulations such as SEDDS, emulsions and microemulsions. HLB plays an important role in the selection of lipid vehicles in these types of lipid formulations. Even though most of the lipids have no surface active properties, some of them like medium chain mono-glycerides and PEG esters exhibit surfactant properties. These can have HLB values up to 14, (Table lF).
HLB of a lipid vehicle/surfactant can affect the solubility of a drug. For example, as the HLB increases, the solubility of griseofulvin increases (Fig. 2) (Dumanli et al., 2000). The effect of HLB on drug release rate from a lipid based formulation is shown in Fig.3. HLB value of a lipid vehicle also significantly influences the selfemulsification and dissolution properties of a SEDDS formulation , Kim et al., 2000and Nazzal et al., 2002.

3. Drug Loading
The factors determining the loading capacity of a drug in solid lipid nanoparticles include as (I) drug solubility in molten lipid, (II) miscibility of molten drug in a lipid melt, (III) chemical and physical structure of solid lipid matrix and (IV) polymorphic state of lipid material .
It is not advisable to use a saturated solution in solid lipids because some of the drug may precipitate out at ambient temperature causing structural variations within the system. Resulting crystalliz.ation may lower the bioavailability.
In lipid formulations with high drug concentration, determination of the saturation point of the drug becomes a critical factor in the formulation design. If the drug precipitates out in gastrointestinal tract, its bioavailability may be reduced. However, the use of appropriate lipid vehicle at the optimum level can overcome this problem as it has been shown in the case of lipid formulation ofDMP 323 , a poorly soluble HIV protease inhibitor. A formulation containing Gelucire 44/14 (a mixture of glycerides and PEG esters) increased DMP 323 aqueous solubility and dissolution rate, and enhanced the bioavailability at high doses (August et al. , 1997).
Drug loading can also affect physiochemical properties of the formulation. For example, increasing concentration of L-365 ,260, a poorly soluble benzodiazepine derivative, caused changes in droplet size of emulsions prepared with Labrafil M 2125CS (Craig et al. , 1993. r I A poorly soluble drug substance dissolved in MCM can significantly affect rheological properties and particle size of the formulation as a function of its concentration (Fig. 4). Change in slope of zero shear viscosity vs. concentration shows that the lipid solvent cannot accommodate too many solute molecules so that the drug reaches the saturation point where further addition of the drug may cause entanglement of the molecule (Dumanli and Kislalioglu, 1999).
Concentration and surface active properties of a drug in a lipid vehicle can modify the phase transition of a lipid. However, such an effect is dependent on the properties of the drug. For example, griseofulvin facilitates liquid crystal formation in glyceryl mono-oleate (GMO) at saturated concentration whereas nifedipine does not have the same influence. At the physiological temperature, GMO forms a cubic phase via reverse micellar transformation which is concentration specific. Above a certain concentration, lipophilic drugs such as ibuprofen and propranolol transformed the cubic phase into an inverted hexagonal phase where the system becomes less viscous resulting in modification of the release properties . As the water content of cubic phase is increased, the system may become more viscous  and may sustain the release of the drug. On the other hand this phenomenon may enhance the stability. Enhanced stability of cefazolin and cefuroxime in glyceryl mono-oleate cubic phase gels showed its potential as a chemical stability enhancer, protecting the antibiotics from P-hydrolysis and oxidation as a result ofreduced motility of water, in the gel (Sadhale and Shah, 1998).

Manufacturing and Stability Considerations
Most of the lipid-based formulations especially the ones that contain semi-solid/solid lipids involve the application of heat during their preparation. Duration of heat application, temperature and cooling rate can have a significant effect on the quality and stability of the final product. Scale-up processes of these formulations are also challenging because duration of heat application and cooling rate can change with increased scale. Therefore, optimal process should be always considered during development to assure scale-up certainty and robustness. DSC and viscosity measurements can be useful tools to monitor the effect of temperature, holding time and cooling rate. Many lipids may exhibit polymorphic transitions upon storage or during processing. It has been shown that cooling rate can affect endotherms of Gelucires upon aging (Sutananta et al., 1994).
Most of liquid and semi-solid lipids contain unsaturated bond that photo-and autooxidize. Peroxide formation with oxidation can damage the drug and induce toxicity.
Lipid peroxides may also form due to autoxidation, which increases with unsaturation level. Hydrolysis of the lipid may be accelerated due to the pH of the solution or from processing energy such as ultrasonic radiation (Weiner, 2002). Including an antioxidant (i.e. cx:-tocopherol, propyl gallate, ascorbate or BHT) can be necessary in lipidbased formulation. Hygroscopicity of the vehicles (Fig. 1) is also important for stability of gelatin capsules. Water migration into capsule fill can cause cracking of the capsule shell.

Lipolysis and systemic absorption
Understanding of lipid digestion process is important in design of lipid-based formulations. Digestion process involves 3 basic steps (Fig. 5): (I) dispersion of lipid into gastric media and emulsification by gastric motility, (II) enzymatic hydrolysis of tri-glycerides into mono-glyceride and fatty acids, (III) the formation of mixed micelles with hydrolysis products and bile salts and ultimately absorption through the intestinal wall . The lipid digestion process has been exclusively described for tri-glycerides. However, there is little information on the fate in Gastrointestinal (GI) lumen of lipids such as propylene glycol esters of fatty acids, polyglycerol and polyethylene glycol esters.
Micelle formation of lipolytic products with bile salts is the crucial step that enhances drug solubilization and absorption. The total solubility of certain drugs is proportional to taurocholate concentration, a major bile salt in the small intestine. A linear relationship was shown between solubilization and logP of steroids, cyclosporine A, griseofulvin (Mithani et al., 1996). A detailed discussion of the lipid digestion process and its impact on drug absorption from the gastro-intestinal tract can be found in other reviews (Eldem and Speiser, 1989;Humberstone and Charman, 1997;.

(
The nature of lipids used in the formulation can affect the lipolysis process. The in vitro lipolysis process can be used to elucidate the effect of lipids on lipolysis. The details of the process are described in section 6.2.2. The nature of oil as an influencing factor on absorption ofSEDDS has been discussed (Craig et al. 2000;. The fatty acid chain length directly influences lipolysis. The long-chain fatty acids are more inhibitory than the mediumchain ones (Bernback et al., 1987;Hutchison, 1994;Zangenberg et al., 1998;Olbrich and Muller, 1999;Lacy et al., 2000). The results in Fig.6 show that the shorter chain length allows more rapid lipolysis among tributyrine, MCT (Miglyol 812) and LCT (soybean oil) (Hutchison, 1994). Lipase enzyme is active only at interface. The greater the interface the higher the lipase activity is. Therefore, the extent of lipolysis is very low for triacetin, this is because it is miscible with aqueous medium so that no oil/water interface can be formed. The comparison study of in vitro lipolysis between medium chain and long chain triglycerides quantified by titrimetric, highpreformance thin-layer chromatography (HPTLC) and ultracentrifugational techniques (Sek et al., 2002) showed that the rate and extent of digestion of the medium chain triglycerides was greater than long chain lipids and independent of bile salt concentration.
The surfactants have also important effect on lipolysis. Hydrophilic surfactants such as Brij 96 (polyoxyethylene 10 oleoyl ether), Tween 80 (polysorbate 80) and Cremophor RH40 (polyoxyl 40 hydrogenated castor oil) that have HLB value more than 10 inhibits lipolysis while lipophilic ones such as Span 20 (sorbitan mono-laureate) and \ Crill 4 (sorbitan mono-oleate) that have HLB less than 10 do not. This inhibitory effect with hydrophilic surfactants can be minimized by adding lipophilic cosurfactants like medium chain mono-/di-glycerides to the formulation of interest. In a recent US Patent (Lacy et al., 2000), it was found that transesterification products of A rapidly digested lipid would promote maximum drug absorption (Hutchison, 1994).
For example, when MCT and LCT have been compared for the absorption of vitamin E, MCT enhanced intestinal absorption of vitamin E .
However, this effect can not be generalized. CyclosporineA absorption, in contrast to vitamin E, was higher in LCT than in MCT (Behrens et al., 1996;Renner et al. l 988a, b ). This difference was interpreted as the result of formation of mixed micelles which occurs during the digestion.
Fluidity and volume of the lipid can affect the gastric retention as it was shown in the absorption ofSL-512, an anti-inflammatory agent, (Yamahira et al., 1978) and Phenytoin (Shinkuma et al., 1985). Medium chain tri-glyceride, medium chain monoglyceride, com oil and N-methylbenzyllinoleamide (MBLA) has been investigated as model lipid vehicles. The solubility of SL-512 was found to be 0. 5, 4.2, 0.3 and 2.2% (w/v) in each vehicle, respectively. Even though, MCM provided the highest solubility ofSL-512, it has not provided the highest serum level and gastric emptying time. It was concluded that serum level and gastric emptying time of SL-512 was correlated with fluidity of the lipids.
Gastric emptying was also affected by the chain length of fatty acids (Hunt and Knox, 1968). Fatty acids with 12-18 carbon atoms slowed down gastric emptying more than those with up to 10 carbon atoms. C 1 4 (myristate) was the most effective one.
The effect of chain length of fatty acid of the lipid vehicle on formulation design and the bioavailability ofHalofantrine (Hf) has been stuided (Khoo et al., 1998). The solubility of Hf was enhanced by increasing the proportion of Captex (MCT) for the medium-chain glyceride formulations based on Captex/Capmul (MCM) of combinations. However, this approach led to less efficient emulsification. A 2: 1 (w/w) ratio of Captex to Capmul provided a good balance between drug loading and efficient emulsification. Hf was less soluble in long-chain glycerides compared to the medium chain ones. Better emulsification was obtained with the formulations that contained medium-chain glycerides. The authors observed a trend towards higher bioavailability of the drug with the long-chain tri-glyceride formulations compared with medium chain glyceride formulations in the beagle dogs. Such anomalous findings apparently based on differences in intra-luminal processing of the medium-and long-chain glycerides where the bile salts and the formation of mixed micelles can influence absorption.
Lipid digestion products pass across an unstirred water layer (UWL) in the small intestine. Within the UWL an acidic microclimate aids micellar dissociation so that monomers pass passively across the brush-border membrane. For long-chain fatty acids passage across the unstirred water layer is rate limiting, whereas passage of short-and medium-chain fatty acids is limited by the brush-border membrane (Thomson et al., 1993).

Lymphatic absorption
Once the lipolysis products, fatty acids and mono-glycerides, are taken up by enterocytes, a re-synthesis of the fatty acids and monoglycerides takes place in the endoplasmic reticulum Together with cholesterol, phospholipid and proteins, lipoproteins are synthesized. Lipoproteins are characterized as high density lipoproteins (HDL), low density lipoproteins (LDL), very low density lipoproteins (VLDL) and chylomicrons. Among all these, chylomicrons are the most abundant in the intestine cells. They contain the largest amount of tri-glycerides and they are the largest in size. Since they cannot be taken up by the blood capillaries because of their size, they enter the lymph capillaries (Tso, 1985;Nijs, 1987;Porter and Charman, 1997). Absorption of a drug through lymphatic system escapes first-pass effect.
In order to incorporate a drug in the tri-glyceride core of the chylomicrons, the drug must be lipophilic. Probucol, lipophilic vitamins and halofantrine are the most commonly used model drugs to investigate lymphatic absorption. A lipid used in lymphatic delivery is important to determine the extent of lymphatic drug transport.
The type of lipid, their degree of saturation, chain length and physical state play an ( important role in lymphatic transport of drugs. LCT lipids are more favorable for the lymphatic absorption than MCT. Fatty acids containing 14 or more carbon atoms are considered to be taken up in the lymph, and those containing 8-12 carbon atoms enter the systemic circulation through the portal vein . The effect of different oils on the absorption of probucol has been studied in rat (Palin et al., 1984). The plasma concentration ofprobucol was determined following its administration in arachis oil (LCT), Miglyol 812 (MCT) and liquid paraffm. The total absorption of the drug from lymphatic and portal system was significantly greater for arachis oil formulation than with the other two aforementioned vehicles. The literature also cites the reverse effect: The lymphatic and portal transport ofretinol (vitamin A) in unanesthetized rats (Hollander, 1980) demonstrated that addition of short-chain fatty acids to the infusate did not change the appearance rate of vitamin A in the bile or lymph. Lymphatic uptake rate of vitamin A peaked following the addition of medium chain fatty acids to the infusate. Addition of long-chain unsaturated fatty acids, oleic, linoleic and arachidonic to the infusate inhibited the lymphatic appearance of retinol as the chain length and degree ofunsaturation increased. The absorption rate ofretinol in the bile was increased significantly when long-chain polyunsaturated fatty acids, linoleic and arachidonic acids, were added to the infusate. This indicated that polyunsaturated fatty acids also shifted the exit of vitamin A from the lymphatic to the portal circulation. The preferential transport of the polyunsaturated fatty acids into the portal circulation may also shift the transport of vitamin A into the portal circulation by a co-transport mechanism.
( In another study with rats, three retinoids, namely isotretinoin, etretinate and temarotene, have been used to investigate the effects of solubility and lipophilicity on lymphatic uptake after oral administration in each of three oily vehicles, cottonseed oil, Miglyol 812 and linoleic acid (Nankervis et al., 1996). Lipid solubility increased with increasing partition coefficient of the retinoid. However, the rank order of increasing lymphatic uptake from each of three oils shows an inverse relationship with solubility of the retinoid in each oil. The reason was explained as follows; the retinoid is likely to escape from the dosing vehicle before passing across the diffusional barrier of the unstirred water layer in the gut. It is related to the solubility of the retinoid in the oil. The easier the diffusion of a drug into the intestine, the more thermodynamically favorable it will be for the retinoid to partition out of the oil.
The animal model used is important in studies related to lymphatic absorption. Rat is a commonly used animal model to investigate lymphatic absorption of drugs. The physiological state of the animal is also important. The effect of type of lipid and formulation on lymphatic transport ofHalofantrine (Hf) has been studied in the conscious and the triple-cannulated anesthetized rat model (Porter et al. 1996a,b). It was found that the lymphatic transport ofH£ in conscious rat was independent of the type of lipid (triglyceride or fatty acid) and the type of formulation (micellar or regular solution). On the other hand, the lymphatic transport of Hf in anesthetized rat was dependent on the type of formulation (micellar, emulsion and solution). It was proposed that the lipid vehicle effects in anesthetized rat model reflected the lack of ( l, gastric processing by preduodenal lipase and the shear action of the stomach that presents in the conscious rat model. The degree of unsaturation of lipid was also reported as a factor on the absorption of drug molecules (Craig et al. 2000). The effect of unsaturation on intestinal lymphatic drug absorption has been investigated. The lipoprotein fractions were monitored in mesenteric lymph following intraduodenal administration of arachis oil and fatty acids such as oleic, linoleic and linolenic to rats. It was shown that the greater the degree of unsaturation of the fatty acid, the more rapid the onset of chylomicron synthesis as indicated by more lymphatic absorption of the fatty acid (Renner et al., 1986;. However, absorption of halofantrine was largely independent on triglyceride unsaturation in a conscious rats (Holm et al., 2001 ).
Lymphatic uptake was also affected by isomers oflipid used. The effect of elaidic (9cis) and oleic acids (9-trans) on lymphatic uptake studied showed that oleic acid exhibited higher lymphatic recovery rate compared to elaidic acid (Bernard et al. , 1987).

CHARACTERIZATION OF LIPID BASED FORMULATIONS
Lipids enhance bioavailability for poorly soluble drugs. However, they are challenging to work with because of their physical complexity. It is imperative to use a technique that does not alter any physico-chemical properties of a lipid-based system. Methods used for characterization of SEDDS and SLN were summarized by  and Muller et al. (2000), respectively. In this review, more detailed techniques are cited from the literature as well as from our experience for lipid based delivery systems.

Formulation and Stability Evaluations:
Thermal and particle size analysis are the common methods that can be used to evaluate formulation factors such as HLB, concentration of the components and processing conditions and stability.

Thermal Analysis:
Differential Scanning Calorimeter (DSC) is the most commonly used thermal analysis technique for characterization of lipid based systems. Some of the lipid vehicles are solid or semi-solid, therefore formulations containing such vehicles are processed at temperatures higher than room temperature. Therefore, it is important to characterize the system with thermal analysis before the formulation process. DSC analysis will be useful to determine possible interaction between drug and lipid vehicle especially in solid dispersions. It can be helpful to construct a phase diagram of the drug of interest versus temperature. Solid dispersions ofUC-781, an antiviral agent, with Gelucire 44/14 and PEG 6000 have been characterized using DSC.
Melting points peaks were used to construct the phase diagram for solid dispersions.
Based on these diagrams, no eutectic mixtures were observed. Phase diagrams of solid dispersions and those of physical mixtures were similar suggesting that there was no ( significant chemical change or interaction between the drug and PEG 6000 or Gelucire 44114. It was also showed that based on DSC data, the drug can be dissolved up to a concentration of25% (w/w) in the liquid phase in PEG 6000 and Gelucire 44114 .
DSC can be a useful tool in determination of stability oflipid based systems. The thermal behavior of stored different types Gelucires has been studied using DSC to examine the relationship between preparation condition and stability (Sutananta et al., 1994). Based on DSC results, it was concluded that the thermal traces observed during storage are associated with the segregation or recombination of the Gelucire components into different microscopic regions within the sample, rather than polymorphic changes. They also conducted tensile strength measurements indicated that the strength of Gelucires changed during storage. Modulated DSC has been utilized in softgel capsules to screen the variables such as temperature and humidity influencing the hardness of gelatin capsules (Nazzal and Wang, 2001). The gelling mechanism that occurred in solid lipid nanoparticles that contains Compritol was investigated using DSC . The results showed that the crystallization occur during gelation. However, if the lipid based system is liquid, this kind of stability problem may not be detected by DSC (unpublished data). Therefore, it can be said that DSC analysis is more useful for solid based lipid systems.

Particle size analysis:
( Most lipids are either in emulsion/microemulsion form or they can form an emulsion upon mixing with gastric fluid (self-emulsifying systems). Droplet or particle size will influence performance of the ultimate dosage form. Therefore, particle size analysis is an important analytical tool in characterization of lipid based delivery system. There are number of techniques which are available to measure the particle size. These include optical microscopy, scattering methods such as light scattering and photon correlation spectroscopy and electron microscopy. Scattering methods are mostly commonly used since they cover a wide range particle size (from 2nm to 2 mm).
Particle size measurements provide a quick assessment in comparison of emulsion systems. For example, the efficiency of self-emulsification by measuring the rate of emulsification has been studied by monitoring the relative intensity of light scattered by dispersion and particle size distribution of resulted emulsion using light microscopy and a Coulter Nano-Sizer (Pouton, 1985a). The mixture ofMiglyol 812 (medium chain tri-glyceride) or Miglyol 840 (propylene glycol ester of caprylic/capric acid) with the surfactant Tween 85 (polysorbate 85) provided more efficient selfemulsification system than the mixture liquid paraffin or oleic acid with the surfactant Tween 85. It was concluded that the Coulter Nano-Sizer provided a quick, noninvasive technique for comparing the mean particle sizes of resultant emulsions but was inappropriate for sizing all self-emulsifying systems. It was shown that the determination of particle size with light scattering is more important than comparison of emulsification rate. Light scattering has been used to measure the emulsification rate to explain the mechanism of the spontaneity of self-emulsifiable oils. The ( mechanism was explained by liquid crystal formation and it was concluded that spontaneity was dependent on the nature of the material used in the system . The mechanism of action for self-emulsified systems consisting ofTagat TO (ethoxylated glycerol tri-oleate) and Miglyol 812 (medium chain triglyceride) has been also studied using Low Angle Laser Light Diffraction for the emulsions with droplet distribution above 1 µm (Wakerly et al., 1986). Quasi-elastic light scattering was used for investigations of submicron dispersions. Ternary phase studies showed a specific region of lamellar liquid crystal dispersed in the isotropic phase of solubilized water.
The particle size measurements can be useful to characterize the factors such as HLB of surfactant, type and concentration of co-surfactants affecting the efficacy of selfemulsifying oral delivery system. Such study has been conducted using selfemulsifying formulations containing surfactants (Spans and Tweens) with different HLB values (4.3 to 16.7), co-surfactants (mono-glycerides) with varying fatty acid chain length (Cs to C 1 s) and vegetable oils (peanut, soybean and saffiower oil) with different fatty acid composition (Cs to C1s ). The system was characterized using dissolution test and particle size measurements . It was concluded that a surfactant with an HLB in the range of 11-15 at 5% concentration and a co-emulsifier consisting of a monoglyceride of caprylic/capric acids at 17% concentration were most effective in a formulation having good self-emulsifying properties in terms of particle size and satisfactory dissolution characteristics of an aratinoid model drug used.

(
One should be aware of that using laser diffraction requires dilution of the sample that may alter the properties of self-emulsifying systems. Therefore, when the particle size of the sample is above 1000 nm, which is the upper limit of dynamic light scattering, electron microscopy may be used. 6. 1. 3. Other methods X-Ray diffraction can be used to monitor any polymorphic changes and to detect possible lipid-vehicle interactions along with FT-IR (Tandon et al., 2001). Low frequency dielectric spectroscopy (LFDS) has been used to elucidate the mechanism of self-emulsification. It was shown that self-emulsification with Imwitor 742 or Labrafil M2125 CS and Tween 80 occurs via liquid crystal formation depending on surfactant/oil ratio. It was also shown that LFDS is a useful technique to examine the individual components in order to investigate the effects of drug inclusion. Surface tension and particle size measurements were also conducted in order to determine the effect of drug concentration (Craig et al., 1993a andb, 1995). Refractive index, conductance and density measurements were conducted to compare the water-in-oil microemulsions containing long-versus medium chain glycerides (Constantinides and Scalart, 1997). Rheological investigations and NMR can be used to investigate the effect of drug loading in MCM (Dumanli et al., 1999). In this study it has been shown that beyond a certain concentration, the solution reaches a saturation level causing a change in the slope of concentration vs. zero-shear viscosity and particle size (Fig 4).
The effect of drug loading has been further investigated with NMR showing that as the concentration of the drug in MCM increased, the line-widths of drug broadened while line-widths of MCM remained same.
Zeta potential measurements can help to evaluate physical stability of dispersed lipid based systems and charged lipid-based formulations (Washington, 1996;.  studied the gelation mechanisms seen in aqueous dispersions of solid lipid nanoparticles consisting of 10% Compritol and 1.2% Pluronic F68 by measuring zeta potential of the system. The possible mechanisms of gel formation are explained as the structural changes of the lipid phase leading to zeta potential reduction and particle growth. Recently, positively charged self-emulsifying formulations were developed by  to enhance the interaction with mucosal surface and increase cellular uptake. Zeta potential measurements were conducted to characterize such systems, to investigate the effect of progesterone incorporation on the charge of the formulation and to show that the enhanced electrostatic interactions of positively charged droplets with rat intestinal mucosa} surface are responsible for the uptake of cyclosporine A.

2.1 Dissolution studies
Dissolution testing for poorly soluble drugs requires different media than those normally used for water-soluble compounds. The incorporation of surfactant into dissolution media was found useful (Serajuddin, 1988;. The dissolution ofUC-781 , an inhibitor of HIV-1 replication and a poorly soluble drug, has been carried out in aqueous polysorbate 80 solutions . The dissolution of physical mixtures and solid dispersions of the drug compared with PEG 6000 and Gelucire 44/14 showed that the drug release from Gelucire 44/14 was higher than from physical mixture and PEG 6000. It has been also shown that the micelle is sensitive to impurities and electrolytes influencing size, loading capacity, solubility and dissolution rate (Wasan, 2001). The type of surfactant used in dissolution media is important since it may interact with the capsule shells as in the case of sodium lauryl sulfate (SLS), an anionic surfactant that may interact with cationic charges of gelatin at gastric pH . These interactions may retard the disintegration of the shell. Therefore, two-phase dissolution test method has been proposed for the poorly soluble compounds by Grundy et al. (1997a, b). The test system was used as modification of the USP XXII Apparatus 2. The important feature of this system is containing lower phase that has dissolution medium of750 mL SIF and upper phase that has dissolution medium of250 mL n-octanol. Improved in vivo-in vitro correlation was demonstrated with two phase dissolution test by employing Nifedipine gastrointestinal therapeutic system (GITS). The test was used for dissolution studies oflipid-filled capsules employing Nifedipine as a model drug .
Biorelevant dissolution testing to predict the plasma profile of lipophilic drugs after oral administration by RRSBW distribution, known as the Weibull distribution, showed that prediction of plasma profile was possible in seven out of eleven classes (Nicolaides et al., 2001). In this study, it has been also shown that the plasma profile of a lipophilic drug can be predicted with appropriate in vitro dissolution data, provided that the absolute bioavailability of the drug is known and the drug has dissolution limited absorption.

2. 2. In vitro lipolysis
A physiologically representative medium should be used to establish in vitro lipolysis. Lipolysis starts when the enzyme solution is added to tri-glyceride emulsion. The reaction can be followed by monitoring fatty acid generation via continued titration with pH-stat. (Hutchison, 1994;Carriere et al., 1997;Zangenberg et al., 2001 ). Attention should be given to the surfactant that may be present in tri-glyceride emulsion because most of the surfactants used in lipid formulation inhibit lipolysis. There are not many published studies that correlate in vitro lipolysis and in vivo results. Only preliminary results were given for progesterone with good in vivo correlation  whereas no IVIVC was established for cyclosporine. Olive oil as LCT and Miglyol 812 as MCT have been compared for cyclosporine intestinal absorption both in vitro and in vivo (Reymond et al., 1988a, b). An in vitro lipid digestion method established was used to compare LCT and MCT in terms of the drug partitioning into aqueous phase. It was demonstrated that lipid digestion promotes the partition of the drug in the aqueous phase for MCT, whereas for olive oil presence of lipolysis products in the aqueous ( phase after digestion decreases the distribution of Cyclosporine into this phase.
However, their in vivo results showed that drug permeation is higher with LCT.
Digested vehicles promoted the absorption compared to the non-digested ones. It was concluded that in vitro phase quantification could not simulate in vivo absorption events.

Absorption mechanisms
Increased absorption mechanisms with lipids involve increasing drug solubility/dissolution, changing gastric and intestinal transit time, stimulation of bile flow and increased intestinal permeation (Muranishi, S., 1985;Aungst et al., , 1996. Reduced particle/droplet size can be another reason that lipid-based formulation can provide increased absorption. These mechanisms can be elucidated with measuring dissolution test and particle size of the formulation, using cell monolayers and imaging techniques. Zeta potential measurements can also be useful to study the mechanism of charged lipid based systems . Particle size measurements can be used to show the mechanism of increased bioavailability of certain drugs. For example, enhanced the bioavailability of poorly soluble Cyclosporine A has been studied in o/w microemulsions consists of capric/caprylic triglyceride as an oil(Captex 355), polyoxyethylated castor oil (Cremophor EL™) as a surfactant, TranscutolrM as co-surfactant and saline . It was concluded that the enhanced bioavailability of Cyclosporine A loaded in this microemulsion system was due to the reduced droplet size of microemulsion systems.

(
Physicochemical interactions between the self-emulsifying system and intestinal mucosa can also affect drug absorption. The interaction of a self-emulsifying lipid drug delivery system with the everted rat intestinal mucosa has been demonstrated as a function of droplet size and surface charge . A positively and a negatively charged self-emulsifying system was prepared with ethyl-oleate and oleylamine using Cyclosporine A as a lipophilic model drug. Transmission electron microscopy has been used to measure the particle size of the self-emulsifying systems.
They found that the enhanced electrostatic interactions of positively charged droplets with the mucosa} surface are mostly responsible for the preferential uptake of Cyclosporine A from the positively charged droplets as compared to negatively charged droplets. Positively charged self-emulsifying oil formulation containing Tween 80, benzyl alcohoL ethyl oleate and oleylamine also elicited the highest and most satisfactory absorption profile for progesterone .

3.1. Imaging techniques
Imaging techniques such as scintigraphy, ultrasound and magnetic resonance imaging can be used to get detailed visualization of transit absorption oflipids in the body. The effect of oleic acid on human ileal break was determined by measuring the transit of radio-labeled tablets by gamma scintigraphy in volunteers . It was shown that oleic acid activated the ilea} brake that slowed the transit of tablets through small intestine. Readers can refer to a review by Wilson et al. (1997) for more details about imaging techniques.

3. 2. The use of cell monolayers
The Caco-2 cell, derived from a human colorectal carcinoma, has been used extensively to study the permeation of drug compounds. It especially useful to predict the intestinal absorption mechanisms. Mechanisms of lipid uptake and metabolism have been studied in the nutrition area using Caco-2 cells as a model. This information can be applied to study lipid based drug delivery. It has been reported that many of the biochemical and metabolic features of lipid processing in vivo have been shown to exist in Caco-2 cells (O'driscoll, 1998). These features include fatty acid uptake, esterification and triglyceride formation, cholesterol absorption, esterification and synthesis, synthesis and secretion of lipoproteins, expression of P-glycoproteins (P-GP) and CYP3A-like cytochrome P450 activity. However, it is important to be aware of fact that there are also some differences such as no fatty acid binding proteins (F ABP) in Caco-2 cells.
Efficacy and toxicity screening of various enhancers has been studied using Caco-2 cell (Quan et al., 1998). The findings indicated that the effectiveness of absorption enhancers in the Caco-2 monolayer system was similar to an in vivo rat system. The Caco-2 cell model was also used to investigate the charge dependent interactions of positively charged SEOF with human intestinal epithelial cells . It was shown that the positively charged emulsions affected the barrier properties of the cell monolayer at high concentrations and reduce the cell viability.
However, no detectable cytotoxic effect was observed with following dilution.

(
The effects ofCapmul MCM on physiological properties of rabbit ileum and distal colon, including active ion transport, trans-epithelial resistance and passive permeability have been investigated in vitro (Yeh et al., 1994). It was found that Capmul MCM inhibited active ion transport.

CONCLUSIONS
Lipids can play a significant role in improving absorption of poorly soluble drugs.
Their physicochemical and physiological aspects should be well defined.
Physicochemical aspects including characteristics of lipids and its affect on drug solubility and stability should be studied. In physiological aspects, understanding the lipid digestion process will help to design a rational lipid drug delivery systems.
Judicial selection of lipids through knowledge of a lipid's physicochemical and physiological properties can overcome absorption of poorly soluble drugs. This is rapidly evolving area and the future of lipid drug delivery system is quite promising.
Certainly, further research on areas like lipid digestion and absorption is desirable.
More mechanistic studies that will include different variety of lipids are needed. The type of model drugs used in research purpose should be extended to be more representative and rational. The drug/lipid relationship should be carefully investigated to design a cost-effective formulation with enhanced bioavailability together with optimum stability and manufacturability.          Sci., 80 (7), 712-714.
Influence of vehicle on gastrointestinal absorption of Phenitoin in rats. Chem.
Pharm. Bull., 33 (11)              steroid hormone from an emulsion and a tablet fold compare to tablets in fasted 1993 in human in fed and fasted subjects but the differences was state not significant in the fed state.     To investigate the interaction of positively charged self-emulsifying oil formulations with rat everted intestinal mucosa

Outcome of the study
A two-fold increase in the bioavailability was observed compare to tablet formulations Electrostatic interactions were responsible for uptake of the drug from the positively charged droplets Reference Kommuru, et al., 2001 O"I -....)  Serajuddin, et al. 1988      (

INTRODUCTION
The use of lipid vehicles has generated considerable attention in recent years to improve the bioavailability of poorly soluble drugs. The well-known effect of food in improving the bioavailability of many poorly soluble drugs such as griseofulvin, carbamazepine and danazol is the evidence that lipids can be beneficial to drug absorption Humberstone and Charman, 1997).
Lipids can improve the bioavailability of drugs through several mechanisms.
Enhancing the solubility/ dissolution of the drug is one of the commonly observed effects .
Although lipids have the potential for enhancing drug absorption, only few commercial formulations of lipid-based formulations are available on the market.
These are cyclosporine (Sandimmune™ and Neoral™, Novartis), saquinavir (Fortovase™, Roche), ritonavir (Norvir™, Abbott) and fat-soluble vitamins. One of the reasons for rarity of commercially available lipid-based products is the physical complexity of lipids.
Drug solubility in a lipid is the most important factor in the selection of a lipid as a vehicle. However, due to their physical complexity such as existing as mixtures, solubility of drugs in lipids is a poorly understood concept. Formulation scientists many times select an effective lipid that dissolves a given lipophilic drug by trial and error. The lipids that are screened for this purpose as routine are given in Table 1. In hydrogen-bond formation, which is a major mechanism in the solubility of a compound. Polarity of solvents can be defined by dielectric constant (E), which is an important property related to the solubility and hydrophilic-lipophilic balance , Carstensen, J.T., 1971and Rabaron et al. 1993. It has been shown that the solubility ofa solute decreased as the dielectric constant of solvent decreased , Trivedi et al., 1996.
Therefore, dielectric constants of lipids can be evaluated to predict their solvent properties.
An understanding of cohesive energy between the drug and the lipid molecules may help to determine how a lipid will behave as a solvent. Cohesion is result of the London forces, polar interactions and specific ones like hydrogen bonding (Hansen and Alan, 1971, Barton, 197 5). The most commonly used approach in quantifying the cohesion between a solvent and a solute is the solubility parameter,8, which is defined as the square root of the cohesive energy density, expressed as the energy of vaporization.

=(CED) 112 = (LllivNm) 112
(1) ( Where CED is cohesive energy density ~Ev is the energy of vaporization and V m is the molar volume.
This parameter may be useful to predict the solvating ability of a lipid or the lipid mixture. When solubility parameters of lipid and the drug are similar, they are expected to become miscible (Scatchard, 1931, Small, 1953, Krevelen and Hoftyzer, 1972. Determining the solubility parameter of a lipid vehicle and thereby quantifying its cohesive energy may offer a solution for screening solvent properties of the lipid candidates.
There are numbers of indirect and direct methods available that can be used to determine solubility parameter. The practical methods include vapor pressure or boiling point determinations, solubility/miscibility measurement in liquids with known cohesive energy, solution calorimetry, and surface energy measurements.
Theoretically, and more often the group contribution method is used , Samaha et al., 1990, Subrahmanyam et al., 1996, Breitkreutz, 1998, Ohta et al., 1999. According to this calculation, the solubility parameter: Where ~e:The additive atomic group contributions for the energy of vaporization ~v:The additive atomic group contributions for the molar volume In this study, the group contribution method given above (2) (Samaha and Naggar, 1988, Schott, 1984and 1995, Poulain et al., 1997. Similarly, such correlation may exist in lipids. If so this property can be a very useful tool to select a suitable lipid to dissolve a given drug. It has been also shown that surface free energy of solids was correlated well with solubility parameters (Samaha and Naggar, 1990). Therefore, surface tension, defined as the surface free energy change per unit area increase, can be also a useful tool to predict solvent property of lipids.
( Saponification value is another important chemical property to predict the solvent power oflipids. It is defined as the weight of KOH used to saponify the ester and fatty acids. Therefore, it has been used to earlier obtain ester concentration and correlated well with solvent power of lipids ).
Most of the commonly used solvents like water, ethanol and methanol are small molecules. Lipids differ from these solvents in terms of molecular weight and purity.
Lipids like triglycerides are much larger molecules. Therefore, they cannot accommodate large amount of solute. It has been also shown that the oils with larger molar volume like ethyl oleate, Miglyol 812 and soybean oil solubilized testosterone less than the oils with smaller molar volumes (ethyl butyrate and ethyl caprylate) .
The aim of this study is to seek a relationship among the calculated and measured  Table I. lists the lipids used in this study. Their fatty acids distribution is also presented in the same table. The lipids used were limited to the liquid. Selection of the liquid lipid has been made in such a way that the effect of hydrophilic head (glyceryl vs. propylene glycol), number of glyceryl groups and lipophilic tail (fatty acid chain lengths) can be investigated. There was no purification involved; they were used as obtained so that the findings can be helpful for the formulation scientists who use these materials as provided.

Experimental Solubility Determination
The solubility of the drugs in each lipid vehicle and water was determined by adding an excess amount of the drug into 10 g vehicle and equilibrating the mixture at 25° ± 0.1°C for 72 hr. The mixture was centrifuged for 10 minutes and filtered through a 0.45µm filter. The amount of griseofulvin dissolved was determined by UV spectroscopy at a wavelength of328 nm. The calibration curve used for calculations is given in the appendix, Figure 1. Nifedipine dissolved was quantified by an HPLC method. A Hewlett Packard 1050 HPLC system with a UV detector (A =236 nm) was used. The stationary phase consisted of a micro Bondapak, C 18 reverse phase column (3.9 x 300 mm, Waters Corp., Milford, MA).
The mobile phase used was acetonitrile: methanol: water (2: 3: 3). The samples ( were diluted with methanol before the run. The flow rate was 1.0 mL/min with 30 minutes of total run time per injection. A standard solution for Nifedipine (0.1 mg/mL) and lipid vehicles was prepared in methanol and was run with HPLC. The data acquisition was made using the software Turbochrome 6.1.1.0.0 (Perkin Elmer Corp, Wellesley, MA). Nifedipine standard curve for the method is given in the Appendix, Figure 2. HPLC chromatographs of lipids were also obtained to make sure that they don't interfere with nifedipine peak. All studies were performed in triplicate and coefficient of variance (CV) for precision of the experiments was± 2%.

Surface Tension
A Kruss K 12 Tensiometer (Kruss USA, Charlotte, NC) was used to determine surface tension. The plate method was selected. Lipids were equilibrated and surface tension was measured at 25°C with platinum plate cleaned with flame.

Density Measurements
The density of lipids was measured to use in molecular volume calculations.
The density of a lipid was determined using a Paar Oscillating U-Tube (Anton PAAR, Ashland, VA) at 25°C. The instrument was calibrated with water. The U-Tube was filled slowly with the lipid by making sure that there are no air bubbles.
The display after equilibration was recorded as density (g/cm 3 ).

FT-IR studies
IR spectra for drugs, vehicle and solution were obtained using a Nicolet FT-IR with Attenuated Total Reflection (ATR) accessory (Nicolet Analytical, Madison, WI). The lipid and lipid solution of the drug was added to a sample holder. Solid samples were diluted in mineral oil. First, the spectra of lipids and the model drugs were taken (Appendix, Figures 3-15). The spectra of drugs solution in lipids were also recorded. Omnic software Version 6 (Nicolet Analytical, Madison, WI) was used to obtain spectra and for subtraction of lipid spectrum from solution spectrum.

HLB calculation
The HLB value of each lipid vehicle was estimated as follows: Where S is the saponification number of the ester and A is the acid number of the fatty acid. The HLB ofCapmul MCM (Glyceryl mono-caprylate), for which S=250 and A= 370, is 20 (1-250/370) = 6.5.

Calculation of Solubility Parameters
Solubility parameters U>F ) of lipids and drugs were calculated using the group contribution method devised by Fedors (Eq. 2) Using equation (2), calculation of solubility parameter of oleic acid {CH 3 -(CH 2 ) 6 -(Refer to Table II) (2) In this mode4 the contribution of hydrogen bonding is not included. Therefore, hydrogen bonding contribution was calculated as: where mis the number of hydrogen donor and acceptors, and Vis the molar volume (MW/density) Total solubility parameter (&r) was calculated by adding hydrogen bonding contribution (8H) to the Fedor's solubility parameter (8F): Solubility parameters for griseofulvin and nifedipine were calculated by equation

Constant
The refractive index (N) of each lipid vehicle was measured at 25°C using Abbe Mark II Model 10480 Digital refractometer (Sodium light at 589 wavelengths) (Cambridge Instrument Inc., Buffalo, NY).
Dielectric constant (E) of each vehicle was calculated using the equation (6). .

Statistical Analyses
In order to evaluate the contribution of all the factors together, a stepwise linear regression analysis has been selected for multiple linear regression analyses of factors that affect the solubility of drugs. It is an approach to select a subset of effects for a regression model and facilitates searching and selecting among many models.
Stepwise regression is used when there is little theory to guide the selection of terms for a model and the modeler. It can be done with forward or backward elimination methods. Forward brings in the regressor that most improves the fit, given that term is significant at the level specified by probability to enter. Backward removes the regressor that affects the fit the least, given that term is not significant at the level specified in probability to leave. Backward elimination method has been used in this study. The correlation coefficient (Rad/) and coefficient for each factor was determined. The model used for this analysis is as follows: Where S is the solubility, ai is the coefficient determined by regression analysis to maximize R 2 , Ci is the factor (one of the physicochemical properties of lipid: MW, dielectric constant, saponification number, surface tension and solubility parameter), i is the set of factor and ao is the intercept of the linear regression.
The effect of lipid's lipophilic tail (fatty acid chain length), hydrophilic head (glycerol or propylene glycol and number of glycerol) has been investigated using single factor ANOV A. The model used for this analysis: Sij=µ+fi+Eij (8)  The experimental solubility results for griseofulvin in the tested lipids ranged from 0.318 to 4.031 mg/g and the solubility of nifedipine from 0.322 to 16.983 mg/g (Table   III). Glyceryl mono-caprylate (Capmul MCM) was the best solvent for griseofulvin increasing the solubility 224 times compared to water. For nifedipine, propylene glycol mono-caprylate (Capmul PG-8) was the best solvent and the solubility was increased 1790 times as compared to water solubility.
Lipids used were better solvents for nifedipine than for griseofulvin in increasing solubility because nifedipine has higher lipophilicity with a log P of 4 (Squillante et al., 1997) compared to griseofulvin with log P of2 (Mithani et al, 1996 andNelsen et al. 2001). The solubility values obtained were used to correlate them with solubility parameter, MW, polarity and surface tension oflipids that are given in    previously shown . Therefore, it cannot be generalized. It can be a contributor for more lipophilic drugs like nifedipine as opposed to griseofulvin.
The effect of hydrophilic head (glycerine vs. propylene glycol, number of glyceryl groups) and lipophilic tail (fatty acid chain length) oflipid has been tested with ANOVA for both drugs in accordance with the model described in Section 2.3, ( Equation8. The results in Table VII showed that the effect of fatty acid chain length is significant for both drugs (p< 0.05). The solubility of drugs was increased as the alkyl chain decreased. Also, griseofulvin being more hydrophilic than nifedipine seems to be more sensitive to and affected from hydrophilic head changes.

CONCLUSIONS
In this study, it has been shown that the lipids used improved the solubility of nifedipine and griseofulvin compared to the solubility of drugs in water. Calculated solubility parameter was not sufficient to predict the solubility of drugs in the lipids.      The use of lipid vehicles has generated considerable attention in recent years for improving the bioavailability of poorly soluble drugs as one of the approaches to improve their oral bioavailability. One of the advantages of lipids in delivery of poorly soluble drugs is to improve the bioavailability of drugs by enhancing solubility/dissolution August et al., 1997).
Lipid-based drug development phases can be classified as (I) selection of lipid vehicle/sand design of dosage form, (II) dissolution, product performance and bioavailability testing. Selection of the best lipid vehicle is one of the important chores in lipid formulations.
A suitable solvent for the lipid formulation usually contains a lipid and a surfactant.
Depending on the solubility/concentration of the dug incorporated, saturation can influence the absorption of the drug. A lipid that is suitable for a drug delivery system must be safe and, therefore, the first requirement for these excipients is to be classified as GRAS (generally recognized as safe) materials. The potential toxicity of the surfactant incorporated should be considered carefully especially in chronic use. This evaluation includes not only the type of the surfactant but also its concentration/dose (Swenson et al., 1994).

(
To optimize the lipid solubility and particularly the release of the drug, addition of a co-solvent and surfactant/scan be useful. Ratios of the co-solvent, lipid and surfactant should be optimized to avoid potential precipitation of the drug in the gastro-intestinal (GI) media. To understand the performance of the formulation in GI media, the lipid, surfactant, co-solvent and drug combinations can be evaluated upon aqueous dilution using the phase diagrams Constantinides, 1995 andKim et al., 2000). Since only few lipids are capable of maintaining the drug in the solution form in the GI fluid, the surfactant addition may be necessary to maintain a good dispersibility of the drug.
Based on these facts, the solubility of a drug in a lipid is an important factor in selection of the lipid as a vehicle. This issue has been discussed earlier by the same group (Manuscript 2). Although, MW, fatty acid chain length and dielectric constant were determined as the common contributors to the solvent property of a given lipid, specific drug -lipid interactions are still important factors.
On the other hand, release of a drug from the vehicle was found to be an inverse function of its solubility in the lipid system . Therefore, chemical potential (concentration/solubility) and partitioning of the drug from the lipid to the immediate aqueous environment should be considered in lipid-based formulation design.
( HLB plays an important role in the selection of the lipid vehicles in such formulations. Even though most of the lipids have no surface-active properties, medium chain mono-glycerides exhibit surfactant properties. HLB provides reliable practical information about the water miscibility of the lipid and the surfactant. HLB of polysorbates varied the dissolution ofaratinoid (Bachynsky et al., 1994).  showed that HLB of the oil surfactant mixtures affected the release property of a freely oil soluble drug.
Dissolution testing for poorly soluble drugs requires modification of the routinely used media. Incorporation of a surfactant into the dissolution media improved dissolution of poorly soluble drugs such as REV5901, UC781 (Serajuddin, 1988;. The mechanism of surfactant facilitated dissolution of poorly soluble drugs has been reviewed by Crison et al. ( 1996) and summarized as mice Ile-facilitated dissolution. Polysorbate 80, sodium lauryl sulfate and polyoxyl-35 castor oil are the most commonly used surfactants added to the medium to improve dissolution of the poorly soluble drugs.
The surfactant concentration selected depends on the critical micellar concentration of the surfactant and micellar solubility of the drug. The type of surfactant used in the dissolution medium is also important because of the possible interactions with the capsule shells as it happens with sodium lauryl sulfate (SLS). This anionic surfactant may interact with cationic charges of gelatin at gastric pH .
( These interactions may retard the disintegration of the capsule shell and thus dissolution of the drug.  demonstrated that partitioning of testosterone from the vegetable oils into a pH 6 buffer was correlated with in vitro release rates. Viscosities of oils have not influenced the release of the drug. However, the viscosities studied were very close to each other.
In the lipid based formulations, the effect of the properties oflipids on dissolution rate and extent of the drugs has been hardly studied. Therefore, in this study, formulations of nifedipine, developed with different classes of lipids and a surfactant combination, were examined to investigate the effect of surfactant on the solubility and dissolution of the drug. The effects ofphysicochemical properties of lipids on dissolution of nifedipine were evaluated. Effect of particle size and partitioning behavior of nifedipine on dissolution properties was also examined.

Materials
Lipids that are used for this study listed in Table I  oleate). By selecting these lipids, the effects of hydrophilic (glyceryl, polyglycerol and propylene glycol) and hydrophobic (fatty acid chain) groups oflipids on the dissolution can be determined. The fatty acid distributions of lipids are also given in the same table. The measured/calculated properties of the lipids (HLB, density, MW, polarity and viscosity) were summarized in Table IL The details of these measurements/calculations are already given in the Manuscript II. Because there is no purification involved in the formulations with lipids, they were used as obtained.
The model drug chosen for this study was nifedipine, a calcium antagonist that is used to treat angina and hypertension. It is 3, 5-pyridinedicarboxylic acid, 1,dihydro-2, 6dimethyl-4-(2-nitrophenyl)-, dimethyl ester. It has a molecular weight of 346.34.
Nifedipine is practically insoluble in water (9.5 µg/mL at 25°C) and less soluble in ethanol. It is also light sensitive and converts to a nitrosophenylpyridine derivative when it is exposed to daylight and UV light. Therefore, all nifedipine related experiments were carried out under yellow light.  ).

Solubility Determination
The mixtures of Cremophor EL and lipid vehicles shown in Table I

Interfacial Tension
A Kruss K 12 Tensiometer (Kruss USA, Charlotte, NC) was used to determine the interfacial tension between the selected lipid and simulated gastric fluid (SGF), USP.
The ring method was selected for the measurements. Both the lipids and SGF were used at 37°C.
All the lipids used except Capmul MCM in Table I were less dense than SGF, USP.
First, surface tension of the lipids was measured in a vessel then the vessel was emptied and cleaned. The heavy phase (SGF) was placed in the vessel, and then the ring was dipped in the liquid. The lipids were slowly added into the vessel by making sure no mixing occurred at the interface. Interfacial tension was determined by pulling the ring away from interface, the "pull" method.

Formulation Development
Among all surfactant/lipid mixtures, the ones containing 0, 10, and 40% w/w Cremophor EL have been selected for formulation development of nifedipine. The drug (10 mg) was added to the lipid/surfactant combination ( 4.5 g) and mixed in a shaker at the room temperature until it was completely dissolved. The final solutions that were examined under a polarizing microscope were crystal-free. The solutions obtained were filled into "000" size hard-gelatin capsules and sealed with a bending solution that contained 20% w/v gelatin, 0.8% w/v polysorbate80 and water to make final volume 100 mL. The capsules were kept at the room temperature and assayed for nifedipine. The assay conducted using the HPLC method showed that the formulations were stable at the room temperature for two months containing 100 ± 2% nifedipine.

Dissolution Test
The USP II apparatus with paddle at a speed of 50 rpm and a temperature of37°± 0.1°C was used to obtain dissolution profiles for the nifedipine lipid capsules.
Simulated gastric fluid (pH =1.2, 900 mL) was used as the dissolution medium. Since the solubility of nifedipine is low, Cremophor EL was added to dissolution medium to increase the solubility of drug. Cremophor EL concentration of 3% (w/v) was selected to increase the solubility of nifedipine (Table III) and can differentiate dissolution of nifedipine from different formulations.
Four capsules that have total of 10-mg drug were placed in capsule cages and dipped into the dissolution medium Samples (3 mL) were withdrawn using a syringe that was attached to a cannula with a 10 µfilter at 3,5,8,10,15,20,40 and 60 minutes. The withdrawn sample was replaced with 3 mL fresh dissolution medium. The amount of dissolved nifedipine was determined by the HPLC method as explained in Section 2.2.1.
( ( The percentage of nifedipine dissolved at 60 minutes was taken as the dissolution extent of the formulation. The dissolution rates were calculated using the Kitalllwa equation : where Ct is the remaining concentration of nifedipine at time t, Co is the initial concentration and k is the dissolution rate. Therefore, the slope of the plot ln (C 1 ) vs.
time in the linear region was taken as the dissolution rate.

Partition Coefficient
Partitioning of nifedipine from a respective formulation to the dissolution medium was obtained by mixing an equal amount of formulation and SGF + 3% Cremophor EL for 1 hr at 25°C. The aqueous and oily phases of the mixture were separated by centrifuging the mixtures for 10 min at 2000 rpm. The oily phase separated was assayed for nifedipine.

Particle Size Measurement
Light scattering was used for the particle size measurements of the formulations that were obtained by dissolving them in SGF. The dissolution conditions were used for the particle size determination. Four capsules that have total of 10-mg drug were placed in capsule cages and dipped into the dissolution medium (900 mL), containing 3% w/v Cremophor EL. 5 mL sample was withdrawn at I 0, 20, 40 and 60 minutes and ( ( the particle sizes were measured using a Dynamic Light Scattering (Brookhaven Instrument Corp. Holtsville, NY). The measurement duration was one minute at the wavelength 488 nm and the detector angle 90°. Polydispersity index monitored was within 0.2% range.

7. Statistical Analyses
The effect of Cremophor EL concentration on the solubility of nifedipine in lipids has been analyzed using linear regression. The model used is expressed as:

Equation 2
Where SN is the solubility of nifedipine (mg/g) in the mixture, SL is the intercept and represents the solubility of nifedipine (mg/g) in the lipid, Cs is the surfactant concentration(% w/w) and a is the slope and represents solubility rate. Correlation coefficient (Rad/) has been obtained. Goodness of fit test has been performed using Shapiro-Wilk hypothesis. For the Shapiro-Wilk test, a test statistic, W, is calculated and is used to test following hypotheses: Ho: The distribution is normal distribution

Ha: The distribution is different from normal distribution
If the probability of Wis higher than 0.05, the conclusion is fail to reject H 0 .
Similarly, the effect of physicochemical properties of lipids, partitioning of nifedipine and particle size of nifedipine lipid-based formulations on dissolution has been analyzed using a simple linear regression analysis and correlation coefficient (Rad/) has been obtained. Goodness of fit test has been performed.

Nifedipine in Lipids
The mixtures of lipid and Cremophor EL were clear except the Migl yo 181 O/Cremophor EL mixtures. The solubility of nifedipine in these mixtures is given in Table IV. As it can be seen from the results, addition 10-50% (w/w) of Cremophor EL increases the solubility of the drug around 10 to 600 times. Surfactant related increase in solubility(%) resuhs in Table IV showed that the highest percent increase was observed with Caprol 2GO (662%), which has the lowest nifedipine solubility and the lowest percent increase was observed with Capmul Pg-8 (12%), which provided the highest solubility for nifedipine. These results indicated that the effect of surfactant on the solubility ofnifedipine is more dominant, if the solubility of the drug in a lipid is low. The increase in solubility is proportional to the amount surfactant added, Figure 2 and the surfactant concentration and solubility obeyed: All the statistical parameters, the slope (a), intercept SL (solubility of nifedipine in lipid) and correlation coefficient (Rad/) values obtained using Equation2 are tabulated in Table V. Goodness of fit test showed that normal distribution is followed for all ( lipids (p >0.05) and the model is adequate. Data in Table V also  Dissolution profiles of the formulations that contained Cremophor EL showed that addition of 10% (w/w) surfactant in the formulation increased the rate of dissolution, Figure 3. An additional increase of surfactant in the formulation did not show further difference, regardless of the type of lipid. Figure 3 is an example to the dissolution performance of 10 -40% Cremophor EL containing nifedipine lipid formulation.
Overall findings indicated that the presence of 10% Cremophor EL has hidden the variations involved in the dissolution caused by the lipids because micelle-facilitated dissolution takes over when there is sufficient surfactant presents in the formulation.

Effect of Lipids on Nifedipine Dissolution
The dissolution profiles of the formulations containing 10 mg nifedipine, 4.5 g of lipid and no surfactant are given in Figure 4. The highest and complete dissolution of ( nifedipine was obtained with Capmul MCM. Although the highest solubility of nifedipine was obtained in Capmul PG-8, it showed the lowest dissolution rate and extent for nifedipine.
When the dissolution performance of the drug in a lipid series containing increasing fatty acid chain length compared, it was found that increasing chain length increased the dissolution rate and extent, Figure 5. The dissolution rate and the extent of nifedipine given in Density and viscosity were the other important properties of lipids affected dissolution rate and the extent of nifedipine, Tables II and VI. Relatively low density of the lipid resulted in floating of the formulation on surface of the dissolution fluid and slowed the release of the drug. Also as the viscosity of the lipid increased, the mixing time of the formulation was extended hence the dissolution rate and the extent was lowered.
The dissolution performance of nifedipine in the lipids was further evaluated using two parameters; dissolntion extent and rate of nifedipine from each lipid given in Table VI. These values used to investigate the contribution of each factor to dissolution parameters. The statistical evaluations were carried out with the linear regression analyses. Among individual tests, only meaningful relationship was found between dissolution extent and partition coefficient of the drug (Radj2 = 0.991).
Partition coefficient of nifedipine from formulations to the dissolution medium has been given in Figure 6. The drug partitioned in propylene glycol (PG) esters in the highest order. The increasing fatty acid chain from C8 to C18 decreased the partition coefficient from 9.6 to 7.8. Partitioning of drug from lipid to aqueous phase decreased as the solubility of drug increased in the lipid. Dissolution extent was significantly affected by partitioning, Figure 7. As the partition coefficient of drug from increased, ( ( ( dissolution rate decreased producing a correlation coefficient of Rad/= 0.991 between dissolution extent and partition coefficient of nifedipine.
Particle size of the lipid-formulation in the GI fluid can also be an important factor that affects the dissolution of nifedipine. The smallest particle size was obtained from Capmul MCM which provided as small particle as the Miglyol810/CremophorEL mixture did, Table VII. During one hour mixing, the particle size of formulation did not change by time. The linear regression analyses showed that the particle size of each lipid at 60 minutes was correlated well with the dissolution extent at 60 minutes (R 2 adj =0.9758 and p=0.1733 for goodness of fit) and partition coefficient of nifedipine (R 2 adj =0.9835 and p= 0.3146 for goodness of fit). As the particle size increased, dissolution extent decreased and partition coefficient (Po/w) increased, Figure 8. These results also confirm the findings by . As the particle size decreases, the surface area increases so that dissolution enhances.

INTRODUCTION
The use of lipid formulations generated considerable interest in the last two decades because the modem drugs produced are highly lipophilic as the result of High Throughput Screening (HTS). The lipids may improve the bioavailability of drugs through several mechanisms such as by enhancing their solubility/ dissolution August et al., 1997), by homogenously dispersing the drug into the gastric fluid and providing and uniform absorption rate (Yamahira et al, 1979a(Yamahira et al, , b, 1980 and by increasing lymphatic absorption .
Despite numerous studies with lipid-based system to increase the bioavailability of  .
Only tri-glycerides are digested physiologically. The pharmaceutical grade diglycerides are rare. When a mono-glyceride is used in a formulation, it is assumed that it skips the enzymatic hydrolysis step. Therefore, dissolution studies may be predictive of the bioavailability of such systems.
Micelle formation of lipolysis products with bile salts is the crucial step that enhances the solubility and absorption of poorly soluble drugs. The total solubility of compounds such as steroids, griseofulvin, cyclosporine A, danazol, pentazocaine, triamcinolene and diazap""m was found to be pr0portional to taurocho 1 ate concentration, a major bile salt present in the small intestine. A linear relationship was shown between solubilization and partition coefficient of steroids, cyclosporine A, and griseofulvin (Mithani et al., 1996).
A detailed discussion of the lipid digestion process and its impact on drug absorption from the gastro-intestinal tract can be found in several reviews (Eldem and Speiser, 1989;Dumanli et al, 2002  In order to coITectly predict in vivo performance of lipid formulations, more than one evaluation technique such as dissolution, particle size determination and solubility of the drug in the lipids and GI fluids may be needed.
In this study, the bioavailability of nifedipine from different lipid formulations was determined in the beagle dogs. The types of lipids were determined from our previous study in which in vitro dissolution performance of a nifedipine formulation in ( ( mono-glycerides and therefore are not subjected to the lipolysis process. To seek an in vitro and in vivo correlation, the solubility, partitioning and particle size of nifedipine in the given lipids were also determined.

1. Materials
Nifedipine and butamben (an internal standard for HPLC analyses of nifedipine in Scientific., Springfield, NJ. All the chemicals were used as received.

Preparation of Formulations
Nifedipine (100 mg) was dissolved in each lipid vehicle (4.5 g) at 25°C in a shaker.
All formulations were in clear liquid form and there were no changes observed when they were kept at 4°C for a month. The components of the formulations were given in

Dissolution Test
The USP II apparatus with paddle at 50 rpm and 37°± 0.1°C was used to obtain dissolution profiles for the nifedipine lipid capsules. Simulated gastric fluid (SGF) (pHl.2, 900 rnL) was used as the dissolution medium. Since the solubility of nifedipine is low in SGF, Cremophor EL, a nonionic surfactant, was added to the dissolution medium to increase the solubility of drug. Cremophor EL concentration of 3% (w/v) was high enough to improve and maintain the solubility of nifedipine and to differentiate dissolution of nifedipine formulations used. Four capsules containing total 10-mg of drug were placed in capsule cages then dipped into dissolution vessel.
Samples (3 rnL) were withdrawn using a syringe attached to a cannula that has a 10 µ filter at the tip at 3,5,8,10,15,20,40 and 60 minutes. The sample withdrawn was replaced with 3 rnL fresh dissolution medium. The amount of dissolved nifedipine was determined by the HPLC method as explained in Section 2.2.3.1.
The amount of dissolved nifedipine (%)at 60 minutes was taken as the dissolution extent of the formulation.

Partition Coefficient
The partition coefficient of the drug between the formulation and the Simulated

Dosage forms, frequency and method of dosing
The dogs were fasted overnight prior to experiment. Each dog received four 2.5 mg nifedipine capsules (Details of the protocol were given in Table II known to be present in the feed or water.

Blood sampling
Blood samples (lmL) were collected via jugular vein from each dog at 0, 1, 2, 3, 6, 10 and 24 hours and placed into glass tubes containing EDTA used as an anticoagulant.
The tubes were kept at 4°C to prevent decomposition. The plasma was separated from ( the whole blood with centrifugation at 4°C, transferred to screw-cap vials and was kept frozen in a -70°C freezer until assay time.

Assay of nif edipine in plasma
Nifedipine in all samples was assayed using a modified version of the HPLC method described by Mehta et al. (2002). In this method, a vortex mixer (Scientific Industries Inc., Bohemia, NY) was used to equilibrate the frozen samples at room temperature.
Methanol (100 µl) containing 2 µg/mL butamben (as the internal standard) and acetonitrile (2 mL The calibration graph was obtained by adding nifedipine solution in methanol to drugfree plasma. Nifedipine plasma solutions were prepared at concentration of 0, 10,20,40,80,120,160,200 and 240 ng/mL and stored in a -70°C freezer for pending analysis. The samples we 1 ~processed as describ2d above. The ratios c ~the peak area of nifedipine to that of butamben were used to construct a calibration graph, Appendix , Figure16. CV for precision of the experiment was within 2%.

Pharmacokinetic analysis
Pharmacokinetic parameters were calculated from nifedipine plasma concentration.
Maximum plasma concentration (Cmax), time of occurrence of Cmax (Tmax) and area under the curve (AUC 0 _ 24 h) was calculated using WinNonlin software by Pharsight Corporation (Mountain View, CA).

Statistical Analyses
Results for plasma concentrations were presented as mean ± SEM (standard error mean). Statistical comparisons of pharmacokinetic parameters were performed using t-test for each pair. JMP Statistical Software version 4.0.4 (SAS Institute Inc, Cary, NC) has been used for analysis.

RESULTS AND DISCUSSION
The assay demonstrated that nifedipine lipid formulations contained 99-102 % of the original nifedipine loading. The plasma concentrations obtained after dosing each animal with respective formulations are presented in Table III. One of the female dogs (F4) vomited the Formulation A 10 minutes after the administration. Therefore, the results obtained from this dog excluded from further calculations. Table IV shows the mean pharmacokinetic parameters (Cm:-.x, T max, AUCo-24 h) of the respective formulations and Figure 1, the plasma concentration profiles obtained during 10 hours following administration. As seen in Table IV  The bioavailability of nifedipine from formulation Dis followed by C>B>A. The reason for higher bioavailability with Miglyol810 (glyceryl tri-caprylate) compared r with CapmulMCM (glyceryl mono-caprylate/caprate) can be explained by Iipolysis. ' Miglyol810 is a tri-glyceride and broken down to mono-glyce1ide and fatty acids during lipolysis whereas CapmulMCM is a mono-glyceride and stays as it is during the digestion process. Although both form micelles with bile salts, digestion of Miglyol810 will generate higher amount of fatty acids. Therefore, the amount of micelles produced in the presence of bile salts with Miglyol810 will be higher than the   Table V. This may be one for the reasons of the differences observed in dissolution, Figure 2. Similarly, the formulation that was dispersed to larger particle size emulsion performed poorer bioavailability, Figure 1. The particle size effect may be true for the formulations that contain only mono-glycerides. However, it is not relevant for comparison of Formulations C and B. If the particle size is a factor, the lower bioavailability should be obtained for Formulation C containing Miglyol 810.

(
The effect of solubility of nifedipine in lipids on bioavailability also showed an inverse effect for the formulations without surfactant. The highest nifedipine solubility was obtained in CapmulPG-8 and this formulation provided the lowest bioavailability.
Miglyol810 provided the lowest solubility for nifedipine, but the highest bioavailability among the lipids excluding the formulation with surfactant, Table V.
This property affects the partitioning of nifedipine from the formulations to the dissolution medium. Partitioning values were also included in the same table. It provided the similar effect on bioavailability of nifedipine as dissolution does.

Bioavailability of nifedipine obtained with different lipid formulations in beagle dogs
showed that the type lipid and surfactant used in the formulation plays an important role. Dissolution is a good predictor for the in vivo performance of nifedipine lipid formulation when it was formulated with non-digestible lipids (mostly monoglycerides).
Although, the solubility of nifedipine in lipids, particle size of the formulation in dissolution medium and partitioning of the drug from the formulation to dissolution medium seem to affect the in vivo performance of the drug, dissolution rate and extent of the formulation and digestibility of the lipid used in the formulations played the major roles in bioavailability of nifedipine from lipid-based formulations.