THE DEVELOPMENT, CHARACTERIZATION AND EVALUATION OF A NOVEL MULTI-UNIT ERODING MATRIX SYSTEM FOR POORLY SOLUBLE DRUGS

Mechanisms governing the release of drugs from controlled delivery systems are mainly diffusion, osmosis and erosion. For poorly soluble drugs, the existing mechanisms are limited to osmosis and/or matrix erosion. These mechanisms are commonly employed to control drug release from single unit and multi-unit dosage forms. More recently, multiunit dosage forms have gained considerable popularity for controlled release technology due to their advantages over single unit dosage forms. However, the mechanism of polymer controlled surface erosion from a multi-unit dosage form has never been reported in the literature. This study describes the development, characterization and evaluation of a matrix pellet system which releases an insoluble drug via polymer controlled surface erosion mechanism. Extrusion/Spheronization method was used to formulate matrix pellets. The effect of various formulation and process parameters affecting the drug release were characterized by analytical techniques such as Differential Scanning Calorimetry, X-Ray Diffractometry, and Mercury Intrusion Porosimetry. Different insoluble drugs were used as model drugs to demonstrate universal applicability of this novel system. The effect of drug solubility was also investigated on the mechanism of drug release from this system. Solid dispersions of the model insoluble drug was formulated to increase its solubility. It was observed that when the drug properties were changed towards increasing solubility in water, the release mechanism and rate also changed from pure surface erosion to erosion/diffusion. Drug release of nifedipine pellets in vivo occurred for more than 24 hours following zero order kinetics in fasted dogs. Thus it was proved that the approach of controlling drug release by polymer controlled surface erosion mechanism from a multi-unit pellet system is possible and such a system may be beneficial than the current marketed dosage forms of insoluble drugs such as nifedipine.

Formulations prepared to determine the effect of granulation water levels 84 Effect of drug loading on the size of pore necks and pore bases as characterized from the intrusion-extrusion profiles 85 Effect of water required for granulation on pore necks and pore bases as characterized from intrusion-extrusion curves of mercury 86

MANUSCRIPT IV
Composition of pellets prepared with nifedipine and nifedipine:pluronic® F -68 solid dispersions. Correlation of matrix erosion(% w/w) with drug release(%) from pellets 37 Correlation of matrix erosion(% w/w) with drug release(%) at different stirring speeds 38 Correlation of drug release ( % ) with volume reduction by erosion ( % ) of pellets 39

MANUSCRIPT II
Effect of varying polymer ratios on drug released(%) from pellets 59 Effect of different drug loading(% w/w) on drug released (% ) from pellets 60 Effect of granulation water level(% w/w) on drug released (%)from pellets 61 xi i

INTRODUCTION
Release of poorly soluble drugs in a controlled fashion is a challenging task for the pharmaceutical scientist. The mechanisms that are utilized to control release of drugs are mainly diffusion, osmosis and erosion. Alza Corporation has developed the GITS (Gastro Intestinal Therapeutic Systems) system for the release of nifedipine, a sparingly soluble drug, over a period of 24 hours. This is an "Oros" tablet that delivers drug under osmotic pressure differences between the GI fluids and the drug formulation encapsulated in the semi-permeable membrane surrounding the tablet. The release of the drug occurs as a fine suspension from the laser drilled hole bored in the tablet (1,2].
Other approaches used are matrix tablets which release the drug in a controlled fashion.
Low to moderate viscosity grade hydrophilic polymers such as hydroxy propyl cellulose, hydroxy propyl methyl cellulose hydroxy ethyl cellulose, chitosans, alginates etc, have been used for this purpose. One of the drawback of these matrices is that they are single units and bioavailability from such matrices is dependent on gastric retention [3,4] .
Single unit dosage forms of poorly soluble drugs that release the drug by osmosis or erosion are commercially available. However in vivo drug release from such dosage forms may not be predictable and complete due to physiological variations in the gastric retention time and gastric emptying rates. Additionally, the frequency of bowel movements is also a factor that seriously influences bioavailability of drugs from such systems. 2 During the past 20 years there has been a growing interest in multi-unit solid dosage forms such as pellets for controlled drug delivery. Pellets offer significant therapeutic advantages over the traditional single unit dosage forms. Since pellets disperse freely in the GIT, they invariably maximize drug absorption, reduce peak plasma fluctuations, and minimize potential side effects without appreciably lowering the bioavailability of the drug. Pellets also reduce variations in gastric emptying rates and overall transit times.
Thus, intra and inter subject variations of plasma concentrations of the drug, which are common for the single unit dosage forms, are minimized. Another advantage of pellets over single unit dosage forms is that the high local concentrations of therapeutic agents, which may inherently be irritant to the mucosa! membranes, can be avoided. Pellets, when formulated as modified release dosage forms are less susceptible to dose dumping than the reservoir-type single unit formulations [5] .
During the early developmental phase of nifedipine GITS system, 20% of the population in the clinical trials taking nifedipine GITS tablet expelled the tablet intact through the GIT via fecal matter. The pellets on the other hand, due to their small size and large number are dispersed rapidly in the GIT and thus avoid dose dumping or loss of dosage form. Pellets also offer technological advantages over single units such as better flow properties and ease of further processing during tablet compaction or coating for controlled release. Table I shows a partial list of pellet products marketed in the US.
Traditionally coated pellets have been used for controlled release applications. Most of the marketed controlled release pellets available today are coated. More recently, matrix pellets have gained popularity in controlled release technology. Controlled release via matrix pellets avoids the coating process and thus saves time and money. Pellets, manufactured by the pharmaceutical industry, are sized between 500 and 2000 µm.
These can be produced in different ways such as spraying a solution or a suspension of a binder and a drug onto an inert core, building the pellet layer after layer, spraying a melt of fats and waxes from the top into a cold tower (spray congealing) forming pellets as the result of the hardening of molten droplets and spraying a binder solution into the whirling powder using fluidized bed [5]. The most popular method of producing pellets is the Extrusion-Spheronization technique. This process was first reported by Reynolds (1970) and by Conine and Hadley ( 1970) and involves four steps: preparation of the wet mass (granulation), shaping the wet mass into cylinders (Extrusion), breaking up the extrudate and rounding of the particles into spheres (Spheronization) and finally drying of the pellets.
Traditionally, in the Extrusion-Spheronization technique, microcrystalline cellulose (MCC) has been the excipient of choice to prepare matrix pellets. Due to its excellent plasticity, it is widely used as a carrier or filler in the.Extrusion-Spheronization process.
However, MCC forms a non-disintegrating matrix and thus incorporation of a swelling or disintegrating agent is necessary for drug release to occur from such a system. Drug release from such matrices has been studied extensively by O'Conner et al. [6] and it was concluded that drug release occurred by Higuchi' s square root of time equation and followed first order kinetics. Incorporation of a poorly soluble drug in such a matrix system would minimize drug release since the MCC matrix system is non-disintegrating. 4 Therefore, such a system would be inappropriate to formulate controlled release pellets of a poorly soluble drug. Additionally, since the drug is poorly soluble, diffusional release will be negligible. Thus, the only choice remains is that of an eroding pellet, which is a matrix pellet system that erodes from the surface as a function of time and releases the drug which is homogeneously dispersed in the pellet matrix. There is no such system reported in the literature.
Hellar et al. [7] prepared discs of poly (ortho esters) and studied in vitro and in vivo drug release of the highly water insoluble levonorgestrel. Poly (ortho esters) are polymers that erode due to pendent group hydrolysis of the ester groups, however; it is not generally recognized as safe for pharmaceutical applications. Hellar et al. concluded from his study that levonorgestrel release from surface-eroding polymer discs bas three important consequences which are (1) The rate of drug release is directly proportional to drug loading, (2) The lifetime of the delivery device is directly proportional to device thickness, and (3) The rate of drug release is directly proportknal to the total surface area of the disc.
The controlled release systems developed by Hellar et al. using poly (ortbo esters) showed zero order release for months. Drug released in vitro was analyzed by measuring the drug present in the device after periodic time intervals of dissolution and the polymer erosion was detennined by gravimetry. This study demonstrated that an indirect method such as measuring the drug left in the delivery device after dissolution may be employed to quantify drug released and also the use of gravimetry to determine polymer erosion profiles.
Based on the information given above, the specific objectives of this research were,

Introduction
Release of poorly soluble drugs from controlled delivery systems is a challenging task for the pharmaceutical scientist. Alza Corporation has developed a gastrointestinal therapeutic system (GITS) for the release of nifedipine, a poorly soluble drug, over a period of 24 hours. The system is an "Oros" tablet which releases the drug under osmotic pressure differences between the GI fluids and drug concentration in the semi-permeable membrane surrounding the tablet. The release of drug occurs as a fine suspension from the laser drilled GITS device {l). Other approaches for the release of poorly soluble drugs from controlled release erosion matrix tablets employing hydrophilic cellulosic polymers are reported (2,3). These matrices are generally single units and thus may be associated with drawbacks such as irregular bioavailability due to presence of food and dependence on gastric emptying time. Therefore, existing mechanisms for the release of poorly soluble drugs by controlled release are limited to osmosis and/or erosion. Due to their negligible aqueous solubility, diffusion has practically very little or no contribution in the release of such drugs from the controlled delivery system.
More recently, multi-unit dosage forms have gained considerable popularity over conventional single units for controlled release technology. Due to their rapid dispersion in the gastrointestinal tract, they maximize drug absorption, reduce peak plasma fluctuations, minimize potential side effects without lowering drug bioavailability. They also reduce variations in gastric emptying rates and overall transit times. Thus, intra and inter-subject variability of plasma profiles, which are common with single-unit regimens, 15 are minimized. They are also less susceptible to dose dumping than the reservoir or matrix type, single-unit dosage forms (4).
Controlled release of poorly soluble drugs such as nifedipine, ampicillin and isosorbide dinitrate via pellets have been reported (5-9). All these studies primarily employed microcrystalline cellulose as a pellet forming agent. Due to its excellent pellet forming properties, microcrystalline cellulose offers potential advantage in pellet manufacturing by Extrusion/Spheronization technology. Release from such pellets was extensively studied by O'Connor et al (10). It was concluded that drug release follows first order kinetics as described by Higuchi's square root of time equation from such pellets. Since microcrystalline cellulose forms a non-disintegrating matrix when formulated as pellets, incorporation of a poorly soluble drug in such a matrix would only intensify the problems associated with its release. Such a matrix system would often provide no release of the poorly soluble drug at all. This paper reports the formulation of pellets which release a poor! y soluble drug as a result of surface erosion of the matrix pellet. It was postulated that for drug release to occur in zero order fashion, a matrix pellet must erode slowly as function of time from the pellet surface. This will allow the release of homogeneously dispersed drug in the matrix in constant increments as the erosion progresses in the pellets from the surface thus controlling drug release. A schematic representation of such a delivery system is shown in Figure I. 16 2.0

Materials And Methods
The poorly soluble drug used as a model was a thiazole based leukotriene D4 antagonist with a solubility less than 1.3 µg/mL at pH 6.8 (Hoffmann-La Roche Inc., Nutley, NJ).
Eudragit<!I L 100 55 and Eudragit<!I S LOO (Huls America, Inc., Somerset, NJ) were used as release rate controlling polymers and matrix forming agents . Kollidon<!I 90 F (BASF Inc., Parsipanny, NJ) was used as a binder. Avicel<!I PH 101 (FMC Corporation, Philadelphia, PA) was employed to prevent inter-pellet sticking during the spheronization stage. Triethyl citrate (Morflex, Inc., Greensboro, NC) was used as a plasticizer for the Eudragit® polymers. All other chemicals were used as received.

Formulation of Pellets:
Eudragit<!IL 100 55 and Eudragit<!IS LOO powders were mixed in a turbula mixer (Turbula Mixer, Impandex Inc., Maywood, NJ, USA) for 30 minutes. Trietbyl citrate was added to some formulations (Table-I Spheronization was carried out for 20 minutes at 500-1000 rpm. During this period, 5% w/w of total batch size Avicel® PH 101 was sprinkled over the rotating extrudates to prevent the pellets from sticking. Pellets obtained were dried on trays at 50°C for 12 hours . Dried pellets were later sieved to obtain different particle size fractions (Rotap Sieve Shaker, Model RX-29, W.S. Tyler, Inc., OH, fitted with sieve# 8, 10, 12, 14, 16, 18 and 20). The pellets consisted of drug (10.0% w/w), Eudragit®L 100 55 and Eudragit® S 100 (88.0% w/w) and Kollidon®K90F (2.0% w/w). A flow chart of the manufacturing process is presented in Figure 2. The composition of formulations with different polymer ratios is given in Table I.

Determination of Glass Transition Temperature (Tg)
Polymer blends (Eudragit® L 100-55: Eudragit® S 100 in ratio of 1:3) with or without triethyl citrate as a plasticizer were weighed in a DSC aluminium pan. The DSC (Differential Scanning Calorimeter, Seiko Instruments Inc. , Japan, Model SSC5200) was programmed to perform a heat-cool-heat cycle from 0 -200°C. Heating and cooling rates of 10°C/minute was used.

Determination of Matrix Erosion
To study the erosion process of the pellet matrix, three criteria's were monitored, namely; microscopic evaluation of pellets, matrix erosion after dissolution of pellets and volume reduction by erosion of the pellets at different dissolution time intervals.
Pellets were visually inspected, sized and photographed under an optical microscope (Optical Microscope, Nikon HFX,IIA, Japan) before and after matrix erosion and drug release studies. Ten pellets per time interval were evaluated.
Matrix erosion was evaluated by using standard USP dissolution system (Distek, Dissolution System 2100A, USP Apparatus I ,Baskets). Matrix erosion was determined by removing the baskets with pellets at intervals of 2, 4, 6, 8, 10, 12 hours and drying them for 12 hours at 50°C to a constant weight. The difference between the initial and final weight was calculated as percent matrix erosion.
Volume reduction due to erosion of pellets was calculated by using Equation 1.

Equation l
Where, V, is volume (mm 3 ) of a sphere and D is the diameter (mm) of a sphere.
Cumulative percent erosion volume was calculated by dividing the change in volume at time 't' by original volume at time zero. The result of this was multiplied by 100 to obtain percentages. Rate of erosion volume (%/hr) was calculated by dividing cumulative percent erosion volume with the time interval.

Dissolution Studies:
Since the drug is poorly soluble, drug release from the pellets was determined by an indirect procedure which involved determination of drug left in the pellets after dissolution by UV analysis. The difference between initial and final amount of drug present in the pellets after dissolution was calculated as percent drug release.

Pellet Processing by Extrusion/Spheronization:
Extrusion with Eudragit®L 100 55 and Eudragit®S 100 as pellet forming agents was satisfactory and pellets of uniform shape and size were obtained ( Figure 3).
Spheronization occurs by rotation of the extrudates at high speeds on a friction plate within a vertical cylinder. During this stage each individual pellet rotates on its own axis due to centrifugal force. This action results in liquid migration from the interstices between particles to the surface of the sphere which may be accompanied by migration of ingredients in the formulation. lf the drug is soluble in the granulating liquid, then on drying may lead to non homogeneous distribution of ingredients in the pellets (11).
The drug and the polymers used in this study were insoluble which prevented them from solubilizing or retaining moisture within the pellet matrix, resulting in the migration of moisture alone towards the pellet surface. This action created inter-pellet adherence during the spheronization process. Inter-pellet adherence was eliminated by sprinkling 5% w/w of Avicel®PH I 0 I on the extrudates during the spheronization step.

Characterization of Pellets:
Release profiles of the pellets ( 1.2 mm) prepared with and without Uiethyl citrate as plasticizer is shown in Figure 4. It was observed that 70 to 100 % drug release was obtained within six hours from these pellets. Pellets with 1: 1 and 1 :3 ratios of Eudragit® L 100 55: Eudragit® S 100 were formulated. Pellets within each of the two formulation ratios containing plasticizer showed enhanced drug release rates when compared to pellets without plasticizer. This effect was consistent when the polymer ratio of the pellets were increased. The increased drug release from the pellets containing plasticizer may be the result of increased dissolution rate of the polymers after plasticization.  Table   2. Polymer blends with plasticizer showed a significant reduction in glass transition temperature and enthalpy. Glass transition temperature of both the polymers were reduced by about 60% indicating that the polymer blend became more amorphous after plasticization, therefore its solubility was increased.

Characterization of Matrix Erosion and Mechanism_ of Drug Release:
Microscopic studies showed that the pellets during drug release were reduced in size as a function of time while maintaining a constant surface geometry ( Figure 6A thru F). To extend the release period to more than six hours, 2.0 mm pellets were formulated. Figure   21 7 shows the extent of matrix erosion and drug release from the pellets. Matrix erosion and drug release occurred simultaneously (Figure 7). This correlation of matrix erosion with drug release holds true at stirring rates of 25, 50 and 100 rpm as demonstrated by Figure 8. These findings prove that drug release was a direct consequence of matrix erosion and was stirring rate independent. Figure 9, shows the correlation of drug released with percent volume reduction by erosion. It indicates a direct relationship between drug release and volume reduction by erosion. Volume reduction depends on the diameter of the pellets. As the pellet erodes with time the pellet diameter reduces due to which erosion volume increases to maintain a constant rate of drug release (Table 3). Table 3 shows the changes in pellet volume, cumulative % erosion volume and rate of erosion volume as a function of dissolution time. The rate of erosion volume from Table 3 was observed to be constant up to 10 hours. This indicated that pellets eroded from the surface with consequent size reduction without affecting the erosion volume. Thus drug release following zero order kinetics was obtained.
These discussions explain the zero order release and matrix erosion profiles achieved from pellets and provide strong evidence for a surface erosion mechanism and for negligible diffusional release of the drug.

22
Unifonn matrix pellets were obtained by using Eudragit®L 100 55 and Eudragit® S 100 as pellet forming agents. Pellets of satisfactory quality without microcrystalline cellulose in the matrix can be fonnulated.
As hypothesized, multi-unit pellet system fonnulated for controlled release of a poorly soluble drug by polymer controlled surface erosion mechanism were developed and characterized. These pellets reduced in size as a result of polymer controlled surface erosion of the drug and provided zero order controlled release up to 12 hours. \ \,       Microscopical ~valuation of matrix erosion and size reduction of pellets (magnification: 5X).

Introduction
The design and evaluation of a novel multi-unit erosion matrix that releases a poorly soluble drug by matrix erosion for 12 hours was reported earlier [ l]. Several authors have reported factors such as polymer type, drug concentration, drug solubility, pelletization technique used, influencing drug release rate [2][3][4][5][6][7][8][9]. All these factors were evaluated for osmotically or diffusion controlled pellets employing microcrystalline cellµlose as the principal pellet forming agent and release rate governing polymer in the pellet.
The pellets used in this study were manufactured by Extrusion/Spheronization technique, therefore any change in the formulation or process parameters may influence matrix erosion and drug release from the pellets [10]. The aim of this study was to investigate the influence of the most critical formulation variables (ratio of polymers used and drug loading) and process variables (water required for granulation, pellet size and spheronization time) on matrix erosion and drug release from the pellets. Previously, the linear relationship between matrix erosion and drug release at various dissolution stirring rates was described [I]. It was concluded that in such systems, matrix erosion and drug 43 release occurred simultaneously, thus matrix erosion can be monitored to predict drug release from the pellets.

Materials and methods
The poorly soluble drug used as a model was a thiazole based leukotriene 0 4 antagonist with aqueous solubility < (polyvinylpyrrolidone) as a binder (2% w/w), Triethyl citrate as plasticizer for Eudragits ( 15% w/w of total Eudragit content), deionized water for granulation (70% w/w).

2.4
In vitro release studies:

Results and discussions
Several studies report the influence of formu lation and process variables on drug release from pellets formulated by Extrusion/Spheronization process [2][3][4][5][6][7][8][9]. However, the results of these studies are specific to the formulation and utilize either microcrystalline cellulose (MCC) or MCC with various hydrophilic or hydrophobic in combination. Drug release from such matrices is predominantly characterized by first order kinetics due to the presence of microcrystalline cellulose used as the matrix [11]. Tapia et. al. [2] studied the effect of chitosan on drug release from matrix pellets manufactured by Extrusion/Spheronization and concluded that drug delivery occurred by gel formation of chitosan through diffusion process. Gel formation was found to be a direct function of polymer ratio.
The rate controlling polymers used in this study were Eudragit® L 100 55 and Eudragit® S 100. These polymers dissolve above pH 5.5 and 7.0 respectively. Some of their popular 47 commercial uses include tablet and pellet coatings to achieve controlled or sustained release.
The effect of increased Eudragit® S 100 content on drug release from 2.0 mm pellets is shown in Figure 1. It was observed that rate of drug relea3e decreased as the ratio of Eudragit® S 100 increased in the formulation without any significant change in the release kinetics. Figure 2 shows the effect of drug loading on drug release. Matrix erosion data was used to compare the effects of drug loading with that of placebo pellets. The same figure demonstrates that drug release from pellets with 5, 10 and 20% w/w drug loading was similar to that of placebo pellets which strongly indicated that the drug release mechanism was matrix erosion controlled up to 20% w/w drug loading. However, above 20% w/w drug loading, the release rates were found to decrease as the drug load increased up to 40% w/w. The reason for this finding may be hydrophobicity of the drug incorporated into the matrix.
The influence of the amount of granulation liquid on the drug release rate from pellets made by Extrusion/Spheronization has been the topic of many publications (Baert et al. 48 [4], Jerwanska et al. [5]). Baert et al and his co workers demonstrated that slower release rate was the result of increasing amounts of granulating liquid. They correlated the effects of granulation liquid with the differences in hardness, density and structure of the pellets, whereas Jerwanska et.al and his co-workers, through their study concluded that rate of drug release increased with increasing granulation liquid level due to an increase in porosity obtained after drying. They also correlated these results with differences in hardness of the pellets.
The effect of the granulation water level on the matrix pellets prepared by employing Eudragit® L 100 55 and Eudragit® S 100 as the rate controlling and pellet fonning agents is shown in Figure-3. Increased granulation water levels had a direct effect on the drug release rates. These findings are similar to the findings of Jerwanska et al [5]. However, there seemed to be no significant difference in the release rates above 65% w/w granulation water level. This can be explained by the effect of moisture content on the degree of liquid saturation of the extrudates. Jerwanska et al [5], proposed that for a continuous extrusion process, adequate water is required to bridge the particles together until liquid saturation in the granulation is achieved. This is necessary to deform the granulation to form extrudates and consequently shape them in to spheres by spheronization. If the granulation water level is below the liquid saturation point the 49 spheres obtained will be hard and less porous leading to decreased drug release rates.
Above the liquid saturation point the hardness and porosity of the pellets are not significantly affected.
In order to investigate the most critical spheronization times which would have an effect on drug release, pellets were spheronized for 2,5,8,10,20 Table 3. From Table 3, the pellet hardness changes with spheronization time up to about IO minutes with maximum hardness recorded for pellets spheronized at 8 minutes, where after the hardness decreases up to 20 minutes. No significant difference in the pellet hardness from 20 to 40 minutes was observed. This may be explained by the densification process occurring during the spheronization step. As spheronization time progresses from zero to time 't', the extrudates are cut into uniform particles and shaped into spheres due to the centrifugal and frictional forces present in the spheronizer during operation. These forces act on each and every particle making them more dense and more spherical with time.
However, after a critical period no further densification occurs with increase in spheronization time. Data from Table 3 indicates that the pellet densification process 50 talces about 10 minutes above which very minor changes in densification occur. Thus a spheroinzation time of 2, I 0 and 20 minutes was selected to study the effects of time on drug release.

Conclusions
This study shows the effects of various formulation (ratio of polymers used and drug loading) and process (granulation water level, pellet size and spheronization time) parameters on drug release by surface erosion from multi-unit matrix pellets. Each parameter evaluated, demonstrated a change in drug release from the pellets. Increased 5 I amounts of Eudragit® S 100 retarded the rate of matrix erosion and drug release from the pellets. The drug loading had no influence on drug release mechanism up to the 20% w/w level above which increasing levels of drug up to 40% w/w retarded matrix erosion.
Granulation water level at 65 % w/w had a significant effect on the rate of matrix erosion and drug release as compared to the formulation with 60% w/w granulation water level.
Above 65% w/w, there was no significant effect on the rate of matrix erosion and drug release.
Matrix erosion and drug release rates can be optimized by processing the pellets at different spheronization times. Thus, by optimizing the formulation and process variables pellets that can release a poorly soluble drug by polymer controlled, surface erosion mechanism for 12 hours following zero order kinetics. 52

ABSTRACT
Controlled release erosion matrix pellets were prepared by a Extrusion/Spheronization technique. The effect of drug loading, water required for granulation and spheronization time-on porosity parameters (intrusion-extrusion isotherms, pore size distribution, total pore surface area, mean pore diameter, shape and morphology of pores) and drug release rates were investigated. Porosity parameters were detennined by using mercury intrusion porosimetry. In vitro release was performed in phosphate buffer pH 6.8 using USP XXIl Apparatus I (baskets, at 50 rpm) by UV spectrophotometery. The drug loading was found to have a profound effect on the porosity parameters. Pellets with low drug loading showed increased pore surface area, with small mean pore diameters and an increased number of total pores. Whereas pellets with high drug loading had decreased pore surface area with bigger mean pore diameters and a decrease in the total number of pores.
With high drug loading, drug release rate was found to be decreased. Water required for granulation had a direct effect on the total porosity of the pellets. Dissolution studies showed that release rates were directly related to the water required for granulation.
Spheronization time from 2 to 10 minutes had a pronounced effect on porosity parameters and release rates. No changes in porosity parameters and release rates were observed from lO to 20 minutes of spheronization time. It was shown that each porosity parameter investigated was well correlated with drug release rates and thus it is important to study the effect of porosity parameters in evaluating the In vitro performance of multi-unit erosion matrix for controlled release of a poorly soluble drug.

INTRODUCTION
Porosity is a measure of void spaces in a material and can be generally calculated by using a number of techniques such as density, gas adsorption, water displacement and porosimetry (I). Determination of pore structures of solids can provide important information on disintegration, dissolution, adsorption and diffusion of drugs (2). Pore size measurements provide information on the actual pore structures, including pore diameter and volume, and can be determined by gas adsorption and mercury porosimetry.
The gas adsorption method is limited to pore diameters smaller than 2000 Angstroms, whereas mercury porosimetry is capable of measuring larger pores and inter-particle spaces (3). Thus mercury porosimetry is a suitable technique to determine a broad range of pores of a sample.
The method is based on intrusion of mercury into the pores of a solid sample and is quantified by the Washburn Equation (4).
Pr= -2 yCos 0 where P =pressure (psi), r =pore radius (µm), y =surface tension of mercury (dynes/cm) and 0 = the contact angle of mercury. This equation holds true only when the surface tension and contact angle of mercury are kept constant and shape of the pores is assumed to be circular. 66 By mercury penetration under pressure, one can determine the size and quantity of void spaces and pores in porous materials. In addition, mercury expelled from pores as a function of decreasing pressure provides information about the shape and structure of the pores (5). In porosimetry, voids are defined as spaces between particles or the several pieces constituting the specimen, whereas cracks, crevices, holes and fissures within the specimen, whether a single piece or a powder, are termed as pores (6).
Mercury porosimetry has been extensively used in porosity determination of granules (7-II) , tablets ( 12-17) and pharmaceutical powders ( 18,19). The development, characterization and evaluation of a novel multi-unit erosion matrix for a poorly soluble drug was reported in our previous study (20). In which, matrix pellets of a model poorly soluble drug (thiazole based leukotriene antagonist, aqueous solubility < 1.23 µg/mL) was pelletized with Eudragit® L 100 55 and S 100 used as release rate controlling polymers.
The pellets were prepared by Extrusion/Spheronization technique and the effect of formulation (drug load, water required for granulation) and process (spheronization time) variables on drug release were studied (21). In this paper we have used mercury intrusion porosimetry to understand the effect of formulation and process variables on drug release behavior relative to the changes in porosity parameters.

MATERIALS AND METHODS
A thiazole based leukotriene D4 antagonist ( Composition of pellets formulated to determine the effects of drug loading are given in Table I.

Water required/or granulation:
Composition of pellets formulated to study the effects of granulation water level are given in Table 2.

Spheronization time:
Pellets were processed at 2, IO and 20 minutes spheronization times . Formulation composition maintained constant for this study were the drug load ( IO % w/w), polymer ratio (I : 3) same as in Table 2, Kollidon® 90F as binder (2 % w/w), triethyl citrate as plasticizer (15 % w/w of total Eudragit®L 100 55 and Eudragit®S 100) and water for granulation (70 % w/w of the total batch size).

Drug release studies:
69 It was shown in our previous study that pellets prepared with the model poorly soluble drug, released the drug as a direct function of matrix erosion (20). In vitro drug release was determined by using USP XXII Apparatus I with baskets at 50 rpm (Distek Inc., NJ, USA) in 500 mL of pH 6.8 phosphate buffer at 37 .0 ± 0.5° C.

Mercury intrusion porosimetry:
Porosity parameters such as intrusion-extrusion isotherms, pore size distribution, total pore surface area, mean pore diameters, shape and morphology of the pores were The mean pore diameter (D'mean) was calculated by Eq 4.

Vtot D'mean=4s
3 4 Pore morphology was characterized from the intrusion-extrusion profiles of mercury in the pellets as described by Orr et. al. (6).

RESULTS AND DISCUSSION
Effect of Drug Loading: 7 1 The intrusion volume of mercury is a function of •total porosity. In Figure 1 the cumulative intrusion volume was plotted agai nst pore diameters showing the intrusionextrusion profile of pellets with different drug loading. The intrusion and extrusion curves form a hysteresis indicating that majority of the pores present in the pellets were ink-well type pores that had smal l openings with broad bases. Although no particular trend was observed in the intrusion profiles with respect to drug loading, the intrusion volume of mercury was significantly lower for 30 and 40% w/w than the 5, 10 and 20% w/w drug loading ( Figure !). Figure 2 shows the incremental intrusion volume as a function of the pore diameter of the pellets with increasing drug loading. From Figures I and 2, the number of pores and mean pore diameters of the pellets can be characterized. The data indicates that as the drug loading increased from 0-10% w/w, the mean pore diameter increased with the total number of pores essentially remaining constant whereas, with 30 and 40% w/w drug loads the mean pore diameters increased and the total number of pores decreased. Figure 3 shows the effect of drug loading on the total pore surface area and mean pore diameter of pellets; they seem to have an inverse relationship as expected. Table 3 lists the calculated ranges of pore necks and pore bases as a function of increasing drug loading as characterized from Figures 1 and 2. The data from Table 3 indicates that pore bases were nearly twice the size of pore necks at all levels of drug loading; indicating that all pores have large bases with relatively small necks. This difference 72 becomes more apparent as drug loading increases above 30% w/w. This interpretation is supported by the relation of drug loading, total pore surface area and the mean pore diameters of the pellets as shown in Figure 3. The results indicate that with increasing drug concentration the pores became wider with larger necks and thus reduced in number.
These changes are illustrated schematically in Figure 4. Figure 5 shows the dissolution profiles of the pellets with different drug loading. Drug release from these pellets occurred via surface erosion. Therefore theoretically, the nature of pores present at the surface of the pellet must influence the erosion rate rather than the total porosity of the pellet matrix during the dissolution process. In pellets with high drug loads, the total polymer content is relatively low. Since the weight fraction of drug per unit weight of the drug-polymer mixture is high, the drug particles associate to form drug agglomerates (22) and this agglomeration tendency of the drug at high drug loads will reduce the number of pores and thus total pore surface area is reduced. Such a system during dissolution will have a low contact surface area with the dissolution media.
However, in pellets with low drug loads, the weight fraction of polymer per unit weight of the drug-polymer mixture is high, therefore chances of drug agglomeration are less resulting in more pores with smaller mean pore diameters and increased total pore surface area. Thus, the increase in mean pore diameter and decrease in total pore surface area of pellets with high drug loading were primarily due to agglomeration of the drug particles.
As it is discussed above, because of the existence of larger pores, the surface area of contact between the dissolution medium and pellets with high drug load is reduced, which 73 reduces pellet hydration and consequently the erosion rates. This was confirmed by the dissolution profiles . given in Figure 5.

Effect of Water Required for Granulation:
The intrusion-extrusion profiles of mercury for the percent water added to the granulation are shown by plotting cumulative intrusion volume against pore diameter in Figure 6.
The total intrusion volume was found to be a direct function of granulation water level.
This indicated that total porosity of the pellets increased with the addition of water for granulation from 60-70% w/w. These findings are similar to the results obtained by other researchers (23-26). Figure 7 is a plot of incremental intrusion volume against pore diameter which shows the pore size distribution of pellets with different granulation water levels. All pores present are between 0.01-0.l µmin size. Table 4 summarizes the results of granulation water level on the range of pore necks and pore bases. The pore base being the average width of the ink-well type pores inside the pellet matrix. From Figure 7 and Table 4 it is evident that increasing the granulation water level from 60 to 65% w/w increased the total number of pores, but the pore necks and bases were not affected indicating that the water levels used in the study increases the porosity without affecting the morphology of pores.
When the granulation water level was increased from 65 to 70% w/w, the pore neck and pore base ranges remain narrow but the number of pores increase, resulting in overall increase in the porosity of the pellets. 74 Figure 8 shows the effect of total pore surface area and mean pore diameters against granulation water levels. The data indicate that the total pore surface area increases without any significant change in the mean pore diameter as a function of increased granulation water levels. This finding also strongly supports the fact that with the addition of more granulation water, the number of pores increased without any change in the mean pore diameters. These changes are illustrated schematically in Figure 9. The dissolution profile of pellets formulated at different granulation water levels are given in Figure 10. The dissolution rates increase with the increase in porosity and total pore surface area of the pellets with 60, 65 and 70% w/w water for granulation. This increase in the porosity and total pore surface area of the pellets increased the dissolution contact area of the medium with the pellet surface resulting in faster hydration and consequently caused higher erosion rates.

Spheronization Time:
The sphericity of a pellet is a function of spheronization time. The longer they are spheronized more spherical pellets are produced. The circular motion of the friction plate in the spheronizer, shape the sphagetti like extrudates into smal ler and uniform granu les.

75
Eventually, the collision of these granules with the friction plate and the walls of the spheronizer change their shape into small spheres or pellets as a function of time. This transformation may be analogous to tablet compaction. "The term compactability is the abi lity of the bed of particles to cohere into or form a compact of a defined mechanical strength"(26). In compacting a tablet, the force applied by the upper punch has a direct relation with the compactability of the tablet. It is also generally observed that after a critical force no further increase can change the degree of compaction. Similarly, during spheronization, the pellet is compacted up to a critical strength above which no more compaction is observed. The change in porosity parameters of tablets as a function of compaction force are reported (12-17). However, for pellets no information showing the changes in porosity parameters as a function of spheronization time is reported. Therefore, it was important to elucidate this process with respect to the change in porosity parameters, particularly because the dissolution rates of the pellet were a function of spheronization time.
To understand the changes occurring in porosity with spheronization, the pellets were processed at three different spheronization times, 2, 10 and 20 minutes. Figure 11 shows the total intrusion volume against pore diameters as a function of spheronization time.
The data indicate that porosity was not significantly affected by spheronization at 2, l O and 20 minutes. Figure 12 shows the plot of incremental intrusion volume against pore diameters which demonstrates that the pores increased with 2 to 10 minute spheronization time. However, 76 after 10 minutes, no change in the pore size distribution was observed upto 20 minutes. Figure 13 confirms these findings by demonstrating no change in the total pore surface area and mean pore diameter from 10 to 20 minutes.
In summary, following the argument given earlier, processing period from 2 to 10 minutes increased the pores, total pore surface area and decreased pore diameters, beyond this time up to 20 minutes none of the porosity parameters changed. Figure 14 shows the effect of spheronization time on dissolution profiles of pellets which were processed for 2, 10 and 20 minutes. The dissolution rates of pellets processed at 10 and 20 minutes were same. However, pellets processed at 2 minutes spheronization time showed faster dissolution rates. Figure 15 shows a schematic representation of the effect of spheronization time on the porosity of the pellets.

CONCLUSIONS
This study demonstrated that the changes in porosity parameters (intrusion-extrusion isotherms, pore size distribution, total pore surface area, mean pore diameter, pore shape and morphology) of pellets made with insoluble drug substance is affecting drug release rates with erosion controlled mechanism when the drug loading, granulation water level and spheronization time are modified. By increasing the granulation water level, the number of pores are increased without affecting the mean pore diameter. The total porosity of the pellets was increased with 77 higher granulation water level. This increases the erosion rate of pellets leading to faster dissolution of the drug.
With spheronization time, the porosity parameters are affected depending on the time.
Up to I 0 minutes of spheronization time, the number of pores increased with total increase in surface area and decrease in pore diameter. No significant increase in porosity parameters was observed when the spheronization time was further increased from I 0 to 20 minutes. This difference is reflected by erosion rate and dissolution profiles.
Thus, the study of porosity parameters is important in characterizing and predicting the !!! vitro performance of multi-unit matrix pellets.

ACKNOWLEDGMENTS
I would like to express my gratitude to late Mr. Jaques Tussounion from Hoffmann La-Roche Inc., Nutley, NJ 07110 for reviewing and giving valuable suggestions while I was writing this manuscript. Financial support from Hoffmann La-Roche Inc., Nutley, NJ 07110 is deeply appreciated.

Drug Load Kollidon®90F
Eudragit®L 100 Eudragit®S 100 *Plasticizer ...      Nifedipine (N) and nifedipine:Pluronic® F-68 solid dispersion (SD) pellets were characterized for drug release mechanisms from a multi-unit erosion matrix system for controlled release. N was micronized using a jet mill. SD with Pluronic® F-68 was prepared by the fusion method. N and SD were characterized by particle size analysis, solubility, DSC and XRD studies. Samples were subsequently processed into matrix pellets by Extrusion/Spheronization using Eudragit® L 100 SS and Eudragit® S 100 as release rate controlling polymers. Drug release mechanisms from pellets were characterized by microscopy and mercury intrusion porosimetry. DSC and XRD studies indicated no polymorphic changes in N after micronization and also confirmed the formation of SD of N with Pluronic® F-68. Pellets of N showed a 24 hour drug release profile following zero order kinetics. Pellets of SD showed a 12 hour release profile following first order kinetics. Aqueous solubility of N after SD formation was found to be increased by IO folds. Due to increased solubility of N in SD, the drug release mechanism was found to be changed from pure surface erosion to erosion/diffusion mechanism thereby altering the release rate/kinetics.

Introduction
Nifedipine is a poorly water-soluble drug and when administered orally in the crystalline form has poor bioavailability. For poorly soluble drugs, dissolution is the rate-limiting step for gastrointestinal absorption of the drug from solid dosage forms. Since dissolution rate is directly proportional to surface area, decreased particle size may increase the dissolution rate. Numerous attempts have been made to modify the dissolution characteristics of drugs to attain more rapid and complete absoqition (1)(2)(3)(4)(5).
Controlled release Oros® tablets of nifedipine are commercially available. The drug releases in the form of a microfine suspension through a laser drilled hole in the tablet via osmosis following zero order kinetics for 24 hours. Osmotic controlled release multi-unit pellets and granules of nifedipine have also been reported (6).
The mechanism of polymer controlled surface erosion that provides a constant delivery of a poorly soluble drug via multi-unit erosion matrix was reported in our previous study (7). In such a system the drug release was found to be proportional to matrix erosion. Hence, matrix erosion could be used to predict drug release. This system consisted of Eudragit® Release mechanisms of a drug from solid dosage forms may be related to the porosity.
Porosity is a result of the presence of voids and pores in a sample where voids are the inter particulate spaces and pores are typically the crevices, cracks and fissures located in the particle (11). The porosity can be characterized by mercury porosimetry. The pore structure of a solid can provide valuable information regarding its dissolution and diffusion properties (12). Therefore, porosity and pore size distribution measurements have been extensively used to study tablets ( 13-18), granules (I 9-23) and pharmaceutical powders (24,25). Void porosity can be characterized by low pressure mercury 106 porosimetry (upto 30 psi) and is detemtined by calculating the pore volume diameter. In contrast, pores are analyzed by high pressure mercury porosimetry (upto 30,000 psi).
According to this method, the cumulative volume of mercury intruded is a function of porosity, increased volumes indicate an increased porosity.
The present study was undertaken to develop, characterize and evaluate the multi-unit erosion matrix as described previously (7) with nifedipine and nifedipine:Pluronic® F-68 solid dispersion. A physical characterization of nifedipine solid dispersion by particle size analysis, aqueous solubility, DSC and XRD studies were conducted before they were pelletized. Later, pellets containing nifedipine or nifedipine:Pluronic® F-68 solid dispersions were prepared by a Extrusion/Spheronization technique. The effect of porosity parameters (cumulative intrusion volume, pore size distribution, pore volume diameter, total intrusion volume and total pore surface area) on dissolution time of the pelletized nifedipine and nifedipine:Pluronic® F-68 solid dispersion were detemtined to better explain the mechanism of drug release from controlled release matrix pellets and to detemtine the differences that were introduced by the nifedipine:Pluronic® F-68 solid dispersions.

Materials and methods
Nifedipine (

Particle size detennination
Particle size determination was carried out with Master Sizer X, Malvern Instruments Inc., Southborough, MA, USA. An excess amount of drug was suspended in Tween 80 in 100 mL of distilled water and was sonicated for 30 seconds for a thorough dispersion. This suspension was circulated at medium speed for particle size distribution studies.

Preparation of nifedipine: Pluronic® F-68 solid dispersions
Solid dispersion with different drug:pluronic ratios were prepared by the fusion method (26). The required amount of Pluronic® F-68 was weighed accurately and heated to 100° C until it formed a transparent melt. Nifedipine (mean particle size: 2.31 µm) was added to this melt in small portions with a constant stirring rate of 750 rpm . The temperature of the mixture was kept constant at 100° C. This mixture was stirred for 45 minutes until a clear transparent melt was formed. The melt was then poured on to a glass plate and was 108 allowed to solidify at room temperature. The solid mass was powdered and uniformly mixed in a mortar and 80/100 mesh (150-180 µm) particles were used for pelletization.

Solubility of nifedipine and nifedipine in Pluronic 119 F-68 solid dispersion
Solubility of nifedipine alone and nifedipine in the Pluronic® F-68 solid dispersion ( 1: 1) was determined by placing an excess amount of sample in amber glass vials with 10 mL deionized water. The samples were then subsequently allowed to equilibrate at 25° C in an incubator shaker for 24 hours. Samples were filtered and the filtrate was analyzed for nifedipine by an HPLC method. A Waters 600E multi solvent delivery system (Waters

Differential scanning calorimetry (DSC) and X-ray diffraction (XRD) studies
DSC was carried out with a Seiko Instruments Inc., Japan, Model SSC5200 system.
Approximately 10 mg of sample was placed in a hermetically sealed aluminium pan and was scanned at the rate of 10° C/min from 0 to 200° C. Qualitative powder X-ray diffraction was performed by a Scintag X-Ray Diffractometer System, CA, USA by using nickel filtered copper potassium alpha radiation.

Detemzination of In Vitro drug release
In vitro dissolution was performed using USP XXII Apparatus I in 500 mL of pH 6.8 phosphate buffer with ionic strength of 0.05 M, at 50 rpm and 37.0 ± 0.5° C (Distek Inc., NJ, USA) . Pellets obtained after dissolution were characterized for their shape and structure by an optical microscope by Nikon HFX, IA, Japan. Transverse sections of pellets obtained after 2 and 4 hour dissolution times were analyzed for the distribution of drug in the matrix.

7 Determination of porosity parameters
Pellet dissolution time as a function of cumulative intrusion volume of mercury, pore size distribution, pore volume diameter, total intrusion volume and total pore surface area were determined by mercury intrusion porosimetry. A Micromeritics PoreSizer Model 9320, Micromeritics Inc., Norcross, GA, USA was used 1·or the determinations. Each sample was measured in triplicate.

Results and discussion
Results of particle size determination are tabulated in Table II A linear relationship of drug release via matrix erosion of a poorly soluble drug, similar to nifedipine, was described in our earlier study (7). The validity of this matrix erosion hypothesis was tested with nifedipine and nifedipine:Pluronic® F-68 solid dispersion pellets. The in vitro release profiles of nifedipine pellets before and after micronization and nifedipine:Pluronic® F-68 solid dispersion pellets are shown in Figure 4. Pellets prepared with nifedipine of three different particle sizes provided a zero order 24 hour drug release profile. On the other hand, drug release from the pellets prepared with nifedipine:Pluronic® F-68 solid dispersions was changed from zero to first order and the 112 release rates had significantly increased compared to the pellets prepared with nifedipine alone. Drug release rates from the solid dispersion pellets was increased as Pluronic® F-68 increased from 0.5 to 1.0 part in the solid dispersions. Dissolution from these pellets followed first order kinetics for about 12 hours for both the strengths. From Figure 4 it can also be concluded that particle size differences of nifedipine did not significantly influence the release pattern and rates from nifedipine pellets.
In order to understand the underlying release mechanism, the pellets collected at different time intervals during dissolution testing were analyzed under the microscope. Figure 5 shows pellets prepared with nifedipine:Pluronic® F-68 (I: 1) solid dispersion after 12 hours of dissolution. The size of the pellets was decreased due to surface erosion.
Nifedipine pellets also eroded in a similar fashion over a period of 24 hours. Both these pellets maintained their geometrical shape but were reduced in size. Furthermore, pellets of nifedipine and nifedipine:Pluronic® F-68 (1:1) solid dispersion that were removed from the dissolution medium on the 2 and 4 hours of dissolution were dried at 50° C for 12 hours and transverse sections of these pellets were investigated. After 4 hours the pellets became very soft which made it impossible to obtain the transverse. Transverse sections of nifedipine pellets (Figures 6a and 6b) showed that the drug remained unifomtly distributed in the matrix at 2 and 4 hours, whereas nifedipine:Pluronic® F-68 ( 1: 1) solid dispersion pellets showed release of the drug from the core by diffusion. The increased aqueous solubility of drug in the solid dispersion explains the enhanced erosion and release rates from nifedipine:Pluronic® F-68 solid dispersion pellets as compared to nifedipine pellets. Increased aqueous solubility had also increased the release of drug 113 from the pellets of solid dispersion which occurred by erosion and simultaneous diffusion from the matrix. Whereas release of drug from nifedipine pellets was purely by erosion mechanism.
To further confirm the release mechanisms of both the pellets, their porosity parameters were measured and determined by mercury intrusion porosimetry. The porosities were determined after the pellets were exposed to 2. 4, 6 and 8 hours of dissolution media.  Figure 11 shows the total pore surface area against dissolution time. The total pore surface area of nifedipine:Pluronic® F-68 solid dispersion pellets increased linearly from 2 to 8 hours of dissolution time.
This maybe due to the formation of voids and pores as nifedipine and pluronic was diffusing out of the matrix. However, it is postulated that the total pore surface area is being reduced during dissolution because the size of the pellets becomes smaller. Such a phenomenon can only occur if surface erosion is the only mechanism of release which in fact was observed with nifedipine pellets. Their total surface area decreased linearly with dissolution time (Figure 11). This confirms that surface erosion is the release mechanism of nifedipine pellets. In addition, the results demonstrated in Figure 11 strongly indicate that upon incorporation of a poorly soluble drug like nifedipine in erosion matrix pellet systems, a zero order release for 12-24 hours as described previously (7) is obtained.
However, a change in the physical properties and solubility of the drug as it occurs with nifedipine: Pluronic® F-68 solid dispersions alters the release profile and kinetics.

Conclusions
In conclusion, controlled release of nifedipine (poorly soluble drug) following zero order kinetics for 24 hours from a multi-unit erosion matrix was achieved. It was proved that multi-unit erosion matrix systems as described earlier (7)       Nifedipine is commercially available as soft gelatin capsules and tablets for short term and extended treatments. Controlled release nifedipine is available as an extended release film coated tablet and also as a GITS system. The extended release film coated tablet contains a tablet core coated by a slow releasing layer comprising of the drug and the hydrophilic polymers such as hydroxypropylcellulose and hydroxypropylmethylcellulose.
The outer slow releasing layer provides the initial drug release followed by rapid drug release from the tablet core. Drug release from such a tablet typically follows first order kinetics. One of the most desirable outcome in controlled drug delivery is to achieve zero order kinetics in vivo so as to obtain a constant therapeutic effect of the drug for a maximum duration. This is achieved by the nifedipine GITS system for controlled delivery.
14 1 The GITS system releases finely powdered nifedipine in a suspension form into the gastrointestinal lumen at a controlled rate over a 24 hour period. The release mechanism involves a "push-pull" process. As water is absorbed across the semi-permeable membrane surrounding the bilayer tablet, nifedipine particles become suspended in solution and are then "pushed" into the intestinal tract as the osmotically active polymers expand. Hydration of the GITS tablet occurs for approximately 2 hours before substantial amounts of nifedipine is detected in plasma. Dose dumping of nifedipine does not occur from the GITS system however approximately 10% of the total GITS tablet content remains unabsorbed after the tablet is emptied [2]. The dosage forms described above are examples to current nifedipine formulations that are available commercially for controlled delivery.
The development, characterization and evaluation of a novel multi-unit erosion matrix pellet system of nifedipine was described elsewhere [3]. It was designed to release a poorly soluble drug by surface erosion as a consequence of the polymer erosion from the matrix pellets. The drug release mechanism from this system is illustrated schematically in Figure !. In vitro evaluation of this system in pH 6.8 phosphate buffer demonstrated zero order drug release in 24 hours [ 4].
The purpose of this study was to determine the bioavailability and pharmacokinetic parameters such as Cmax. T """'' AUC 0-24 h· and MRT 0-24 h of nifedipine from this novel erosion for separation of plasma proteins after drug extraction from the blood samples was used.
Turbo Yap® LV Evaporator with nitrogen gas pressure of 1.0 bar (Zymark Corporation, Hopkinton, MA) was used as a sample concentrator for the assay.

Formulation of pellets
Eudragit®L 100 55 and Eudragit®S 100 powders were mixed in a turbula mixer for 30 minutes. Triethyl citrate was added as a plasticizer and the resultant mixture was triturated in a mortar for 5 minutes. Drug and polyvinyl pyrrolidone (Kollidon®K90F) used as a binder, were added and mixed for 30 minutes in a turbula mixer. This mixture was then granulated in a mortar with deionized water and later extruded at 40 rpm screw speed. The extrudates were immediately transferred into a rotating plate in the spheronizer. Spheronization was carried out for 10 minutes at 800-1000 rpm. During this period, 5% w/w of total batch size Avicel® PH 101 was sprinkled over the rotating 144 extrudates to prevent pellets from sticking. Pellets obtained were dried on trays at 50°C for 12 hours. The pellets consisted of nifedipine (20.0% w/w), Eudragit®L 100 55 and Eudragit® S 100 (78.0% w/w total in ratio of 1:3 respectively) and Kollidon®K90F (2.0% w/w). Granulation water level used was 58% w/w of the total batch size. Pellets (150 mg) were filled in a size 2 blue colored capsule before they were administered to the animals.

Assay of nifedipine in pellets
Nifedipine content of the pellets was determined by UV spectrophotometry. Environmental Monitoring and Support Laboratory. No contaminants expected to interfere with the study were known to be present in the feed or water.
Each dog received one 30 mg nifedipine erosion matrix pellets capsule or IO plus 20 mg Adalat® soft gelatin capsules in fasted state. Following a one week washout period, each dog received a different formulation in phase two. The experimental protocol details are given in Table I.

Blood sampling
Blood samples (6 mL) were taken from each dog at 0, 1, 2, 4, 6, 8, 12, 16, 20 and 24 hours after dosing for the nifedipine erosion matrix pellets. Blood samples from dogs who received Adalat® soft gelatin capsules were collected at 0, 0.5, l, 2, 4, 6, 8 and 12 hours after dosing. The samples collected were transferred into test tubes containing lithium heparin, used as an anticoagulant, and to prevent decomposition they were placed in an ice bucket prior to centrifugation. Plasma was separated after cold centrifugation and was frozen in amber glass vials at -20° C under yellow light before analysis.

Assay of Nifedipine in Plasma
Nifedipine in all samples was assayed using a modified version of the HPLC method described by Miyazaki et al [5].

Processing Blood Samples for HPLC
Methanol (I 00 µI) containing 2 µg/mL butamben, used as an internal standard and acetonitrile (2 mL) were added to 0.5 mL of plasma in a test tube and were agitated in a vortex mixer for 30 minutes. After centrifugation at 4000 rpm for 20 minutes, 2 mL of the supernatant was transferred into a test tube containing l mL of distilled water, to this solution 4.5 mL of acetone-chloroform mixture ( l: l v/v) was added. This mixture was agitated for l hour on a vortex mixture to ensure complete extraction of nifedipine into the organic phase and was then centrifuged at 4000 rpm for 20 minutes to separate the organic and aqueous phases. The aqueous phase was discarded and 5 mL of the organic phase was transferred to a fresh test tube, and was reduced to dryness in a sample concentrator under nitrogen at 45° C for 30 minutes. The residue was dissolved in 100 µI of the mobile phase and 20 µI of the solution was injected into the HPLC system.

Chromatographic Conditions
HPLC pump used was a Waters multi- Before mixing, the buffer was brought to pH 6.1 with 50% phosphoric acid. Run time used was 30 minutes and the flow rate was 0.8 mL/min at column pressure of approximately 1200 psi. The wavelength of detection was 237 nm. 14 8

Calibration Graph
Standard solutions containing 0.05, 0.1, 0.2, 0.4, 0.6, 0 .8, 1.0 and 10.0 µg/rnL nifedipine in methanol that contained 2 µg/rnL butamben (internal standard) were prepared under yellow light. The standard solution (JOO µI) was added to 0.5 rnL of drug free plasma and the samples were processed as described above. The ratios of the peak height of nifedipine to that of butamben were used to construct a calibration graph. Stock solutions of both nifedipine and the internal standard ( 1 mg/rnL in methanol) were stored in complete darkness; these solutions were freshly prepared every 2 weeks. Precision obtained using the described technique was ±5%.

Pharmacokinetic Analysis
The most suitable model to describe the pharmacokinetics of nifedipine was determined by fitting the data to a hierarchy of models using WinNonlin software.  Figure 2 shows the nifedipine plasma concentration profile for 24 hours following administration of the pellets and the immediate release capsules. The mean T max for nifedipine erosion matrix pellets from Adalat® capsules contain nifedipine in the solubilized form in a polyethylene glycol based co-solvent system. The bioavailability from Adalat® 20 mg soft gelatin capsules was reported earlier by Sallam et.al. [6]. Accordingly, the lower AUC obtained with Adalat® soft gelatin capsules might be due the precipitation of the poorly soluble nifedipine in the gastric fluid. As a result the particle size of nifedipine may also have increased, which can be the cause of reduced nifedipine absorption. 150 Nifedipine release from the matrix pellets is governed by the polymer controlled surface erosion process. In this mechanism, drug release occurs in a constant fashion in the form of a microfine suspension in the gastrointestinal tract and thus is readily available for a prolonged period. lt is also interesting to observe that the nifedipine plasma concentrations were obtained one hour after administration without any significant lag time, Figure 2. The pellet matrix contains Eudragit® L 100 55 and Eudragit® S LOO polymers which dissolve at pH 5.5 and pH 7.0 respectively. Considering that the pellets were very small multi-unit systems (particle size: 2.00 mm), they are expected to have a small gastric residence time after which exposure to pH 5.5 and higher pH's may have caused the pellets to release the drug. The most significant effect that is shown in Figure   2 is that nifedipine release from the multi-unit pellets continued for over 24 hours. Thus, the elimination rate constants could not be calculated for this period

Conclusions
Controlled delivery of nifedipine via polymer controlled surface erosion of nifedipine provided zero-order drug release both in vitro and in vivo for 24 hours. Bioavailability from the controlled release pellet system was four times more than the conventional immediate release Adalat® soft gelation capsules of nifedipine.
Thus it was demonstrated that the surface erosion mechanism may be used in pellets to obtain a controlled release system that delivers a poorly soluble drug like nifedipine effectively and in a constant fashion .

LINEARITY:
The linearity of nifedipine in the mobile phase was determined by simple linear regression. Figure 6 depicts the standard curve and linear regression of nifedipine in mobile phase.