THE EFFECT OF DEVICE CHARACTERISTICS ON NITROGLYCERIN RELEASE FROM A TRANSDERMAL PATCH

List of Tables List of Figures Objectives Introduction Methods Results and Discussion Conclusions Future i,.Jork Bibliography

Instron studies-compression force* as a measure of firmness (*force in kg required to produce 50% deformation).
In-vitro comparison study of trade products.
Release and extraction data from nitroglycerin.  shown that unpredictable release can occur (Hollenberg, 1984;Weber, 1984;Reichek, 1984;Olivari, 1983;Jordan, 1985;Karim, 1983), and evidence suggests that changes in certain physical and chemical characteristics may be responsible. The effects of these factors on controlled drug release have been found to vary between systems.
The purpose of this study was to examine the individual and interactive effects of various device characteristics on the release of nitroglycerin from silicone discs. Data obtained in this study will contribute to the expanding field of drug release from·transdermal delivery systems. The skin is considered the largest organ in the human body and its high accessability make it an excellent target for drug administration.
In order to exert a systemic effect following application of a transdermal device, a drug must pass through one or more cell barriers or membranes. It is therefore necessary to consider the composition of this system and the transport processes involved in the movement of drug molecules across the intact membrane.
The skin includes the tissue that covers the body along with the mucous membranes and modified cutaneous structures. The average surface area of the skin of an adult male is about 20 square feet and may weigh 9 or 10 pounds. Human skin is a stratified tissue composed of three basic layers: the epidermis, the dermis, and the underlying tissue.
The structure of the skin is shown in Figure 1. The functions of the skin include sensation, protection, thermoregulation and secretion. The skin forms an elastic resistant covering which protects man from the environment and prevents the passage of harmful physical' chemical and microbiological agents. (Mullins, 1980).  corneum is the principal barrier to the permeation of drugs through the intact skin (Shaw, 1983). It is composed of 40 percent protein, 40 percent water, and the balance is lipids. This layer is 10-15 um thick, is comprised of dead, dry cells, and it exhibits regional differences in thickness over the body (Michaels, et al, 1975). The cells of the stratum corneum are stacked in overlapping "columns" and are more ordered in the areas of thinner epidermis. Aggregates of keratin inside the cells of the stratum corneum are cross-linked in a characteristic folded configuration (alpha keratin), rendering the fibers elastic.
When stretched, the keratin chain is drawn into a more linear form (beta keratin).
The outermost surface of the stratum corneum is covered in a film made up of emulsified lipids, which act as a seal to bind the cells together. The surface has an acid pH, ranging from 4 to 6.5, depending on the area of the body. The lipid layer between the cells can be considered a bilayer, in that both polar and nonpolar sections are found. The components of this layer have been shown to be directly responsible for retarding penetration of many molecules into the body (Michaels, et al, 1975).
The protein component of the stratum corneum is hydrophilic in nature and can be hydrated to swell the skin using the hydroxyl and amino groups on the molecules. The protein can also be altered by heat or chemicals, such as surfactants, to result in the conversion from alpha to beta configuration. The end result is greatly enhanced permeability of the skin.
The skin as a body organ is constantly undergoing changes. The degree of hydration, thickness, and composition will vary as a function ( 5 of age, race, and environmental conditions. Epidennal cell growth begins in the basal layer of the stratum germinativum. Cells generated in the basal area move toward the skin surface, changing in shape as they move. This process of cell turnover requires up to two months (Cooper, 1985).
In addition to passage through the stratum corneum cells, molecules may also be transported through the follicular orifices and sweat gland ducts which scatter the surface of the skin. These provide alternate pathways for absorption although their significance is questionable.
The contribution of the sweat glands and hair follicles has been investigated by several workers. According to Scheuplein (1972), substances can be absorbed via the appendages by a diffusion process.
He calculated that the route via the glands is important for a brief period immediately after applying the test substance. Since these appendages account for only 0.1 to 1.0 percent of the total skin surface area, the transepidennal route must be the principal route for most materials. (Idson, 1975).

Percutaneous Absorption
Due to the nature of the stratum corneum, the process of percutaneous absorption is a passive one rather than an active one.
"The absence of metabolic processes in the 'dead' keratinizing layers precludes any role for active transport" (Malkinson, 1977). This process occurs in two steps. Diffusion through the barrier layer is the first phase, and is affected only by physical factors as detennined by ambient conditions. The second phase is the clearance from the dennis and this depends on blood flow and other factors inherent in the 6 dennal constituents (Idson, 1975). There is a delay period after the drug is placed on the surface of the skin, during which the membrane becomes charged with the penetrant, and then a longer second period during which the chemical penetrates the skin and is removed to the circulation. It is this second period that is governed by Fick's laws of diffusion (Michaels, 1975; It has been shown that removal of the stratum corneum by successive 11 stripping 11 with cellophane tape can significantly increase the skin's permeability. (Grasso, 1972). This process increases transepidennal water loss and can result in a 60-fold increase in penneability (Chi~n, 1984).
Properties inherent in the dosage fonn may also affect the rate of diffusion of a substance through the skin. Both the chemical and physical properties of the test material must be considered. Ideally, the drug should be potent (i.e., be effective in low concentrations) and should have a good hydrophilic/ lipophilic balance (generally 1 or greater) (Chien, 1982). Substances will be more rapidly released from ( 7 vehicles having a low affinity for the penetrant. Pertinent factors here are related to the solubility of the penetrant in the vehicle, the rate of diffusion of the penetrant in the vehicle, and the rate of release of the penetrant from its base (Malkinson, 1977). The pH, melting point, and molecular weight of a compound all contribute to the degree of solubility of a drug in the skin (Shaw, 1983;Kligman 1984;Grasso, 1972). In the formulation of a vehicle for a topical drug application, many factors must be considered. Drug stability, specific product use, site of application and product type must be combined in a dosage form which readily releases the drug when it comes in contact with the skin. For a given concentration of a drug in a certain vehicle, the thermodynamic activity coefficient of the drug in the vehicle may vary from one vehicle to the next (Higuchi, 1960;Cooper, 1984;Horhota, 1979). Penetration is greatest when the drug is present at its maximal solubility concentration.

Factors affecting percutaneous absorption
The permeation of molecules through the skin may be influenced by physiological, physicochemical and thermodynamic factors. Physiologic factors include the age and condition of the skin and the degree of hydration. The condition of the skin, namely intactness, is one of the most important factors affecting penetration. If the stratum corneum is damaged, water loss is increased and permeability is enhanced.
Increased hydration of the skin appears to "open'' the compact substance of the stratum corneum and result in increased permeation of many substances. Hydration of the skin can be enhanced by occlusive dressings, and may result in 4-5-fold increases in permeability (Idson, 8 1975). Physical properties of the vehicle can improve the degree of occlusion produced, leading to hydration of the stratum corneum. The most occlusive vehicles are greases, oils and collodion.
Permeability of the skin appears to decrease with increasing age.
This may be due in part to the greater percentage of water in the body of an infant as compared to the adult (Behl, 1984). In general, the skin of young children is sensitive to chemical irritation and harsh environmental conditions. As they mature, the cells develop resistance to many substances.
The effect of the site of application was noted in a study by Michaels, et al (1975), where it was shown that the permeation of scopolamine was greatest in regions where the stratum corneum is the thinnest. The post-auricular area possesses the thinnest layer, whereas the plantar region has the thickest (Idson, 1975).
Temperature of the skin has also been shown to affect percutaneous absorption. Under normal in-vivo conditions, substances penetrate the skin within a very narrow temperature range. For lipid soluble substances, increasing the skin temperature lowers the viscosity of the tissue lipids, thus reducing the activation energies for diffusion. The dual effects of temperature and humidity have also been shown to increase percutaneous absorption (Idson, 1975;Grasso, 1972).

Penetration Enhancers
A variety of substances are known to enhance the transport of drugs across the skin. Substances included in the formulation of cosmetics, soaps, and detergents may also produce profound local changes which can affect the permeability of the skin to other compounds. In order for 9 these penetration enhancers to be effective, they must modify the stratum corneum in some way. Surfactants increase penetration due to their ability to denature protein (keratin), thus "opening up 11 the skin structure and -increasing its permeability. Other agents improve skin wettability, decrease surface tension or increase solute diffusivity (Keith, 1983;Hwang, 1983;Chien, 1982).
Few agents have been studied as intensively as dimethylsulfoxide (DMSO). In-vitro comparative studies showed DMSO to be superior to other solvents in both enhancing penetration and in favoring dermal retention of a drug (Grasso,1972). In-vivo, DMSO has been found to accelerate the penetration of many substances, such as corticosteroids, salicylic acid, dyes, and antibiotics (Sekura, 1972;Kligman, 1984;Idson,1975). For materials absorbed from the surface, a 50 percent concentration of DMSO can enhance the permeability of dermal connective tissue, apparently through a depolymerizing effect on hyaluronic acid (Malkinson, 1977). Despite the extensive studies, the exact mechanism of action is still not clear. Other possible mechanisms include partial extraction of lipids and protein configurational changes resulting from replacement of water by the vehicle.
Despite its potential in this respect, DMSO can also be quite toxic, causing refractive lens opacities, erythema, edema, and itching, and is noted also for the unpleasant taste and odor associated with its breakdown products (Sekura, 1972). Other, less toxic, agents have also been used as accelerants of drug transport. Surface-active agents (surfactants), especially the anionic surfactants, appear to alter the permeability of the skin to water. Propylene glycol and polyethylene glycol 400 have been known to both increase and decrease permeation of ( 10 substances. Concentrations of 50 percent or greater usually disturb the barrier and increase the loss of water from the skin (Idson, 1985).
Calcium thioglycolate, a commercial depilating agent, was shown to improve the transdennal absorption of theophylline in rats, giving a 40-fold increase in plasma concentration over that of the control (Kushida, 1984).
Recently, another agent, AZONE (Nelson Research and Dev. Co.,Irvine, CA) was shown to enhance the absorption of triamcinolone acetonide, and was effective in a concentration of only 1 percent. This substance has been shown to be more effective than DMSO, and so has been receiving increasing attention in the past year (Chow, 1984;Vaidyanathan, 1985). Chemically, l-dodecylazacycloheptan-2-one, AZONE is a structural analogue of pyrrolidone and methyldecyl sulfoxide. The range of concentrations of AZONE necessary for enhanced penetration is from 0.1 to 5 percent, indicating different partition coefficients for different drugs (Idson, 1985).

Transdennal Drug Delivery
The concept of using topical products to elicit a systemic response is not a new one. As early as 1900, systemic effects were reported following cutaneous application of belladonna, mercury, pilocarpine and cod liver oil (Grasso, 1972). Toxic, and even fatal, effects have also been reported from substances applied topically. Well known compou nds such as salicylic acid, organophosphates, and carbon tetrachloride can all produce toxic results. However, the topical use of therapeutic agents has many advantages over traditional routes of drug administration. For example, the topical administration of ( 11 nitroglycerin in ointment formulations can minimize both the problems of extensive liver metabolism and short duration of activity.
Over the last several years, considerable interest has developed in the area of controlled release of drug substances through the skin.
Controlled or sustained-release systems are being used or tested to deliver an enormous range of drugs, including drugs for the treatment of glaucoma, angina, motion sickness, narcotics addiction, and hypertension. "A well-designed controlled-release delivery system can significantly reduce the frequency of dosing and also maintain a more steady drug concentration in the blood and target tissue cells" (Chien,1982). Transdermal patches allow a drug to enter the body at an approximately uniform rate over an extended period of time. A controlled-release transdermal delivery system can also lower the total daily dosage of a drug, decrease the incidence and severity of side effects, and avoid the variable absorption characteristics of other routes of administration.
After the development of Ciba-Geigy's scopolamine-releasing transdermal delivery system, which releases drug at a continuous rate for 72 hours to control motion sickness, three nitroglycerin transdermal systems were developed independently by three pharmaceutical companies for once-daily medication of angina pectoris.
The first was Transderm-NitroR, developed by Alza and marketed by Ciba-Geigy. It consists of an outer protective layer of aluminized plastic, a drug reservoir containing a nitroglycerin suspension, enclosed in a microporous ethylene-vinyl acetate copolymer diffusion membrane, and a hypoallergenic adhesive layer. The Key Pharmaceuticals product, Nitro-OurR, is a matrix diffusion-controlled delivery system in l l 12 which nitroglycerin/ lactose triturate is homogeneously dispersed in a hydrophilic gel matrix. Finally, Searle's NitrodiscR consists of a solid, silicone-based polj1Tler and is called a "microsealed" delivery system, since nitroglycerin is uniformly dispersed throughout the matrix and diffuses out at a controlled rate (Chien, 1984;Black, 1982).
Before transdermal patches, nitroglycerin was available in the form of sublingual tablets, sustained-release capsules, and ointments. The sublingual tablets allow rapid entry of the drug into the bloodstream but remain effective for less than 30 minutes. The time-released capsules are claimed to work for 8 to 12 hours. The ointments are applied to the skin once or twice a day, and are messy to use. It is difficult to get a uniform application each time a patient uses it. For all three marketed transdermal products, the devices contain sufficient nitroglycerin to maintain delivery of drug for 24 hours. The rate of delivery to the systemic circulation is determined by several factors, including the surface area covered and the characteristics of the skin itself (Fara, 1983). These transdermal systems permit minimal absorption variability while maintaining a constant drug level in the blood (Karim, 1983). Chien, et al (1983) performed skin permeation studies on the abdominal skin of the hairless mouse in order to examine and compare the controlled skin permeation kinetics of nitroglycerin delivered by these three transdermal delivery systems. Results indicated that the release rate profiles adhere fairly well to zero-order kinetics, although different release rates were found between the three systems. The total dose of nitroglycerin delivered at 24 hours by each system was close in magnitude although different technologies were used to manufacture them. 13 Other companies are also working on transdermal nitroglycerin systems (Sanders, 1985). Hereon division of Health-Chem Corp. is awaiting FDA approval for its NitrodermR system. This system consists of polyvinyl chloride copolymers and terpolymers and has a releasecontrol ling membrane. LecTec Corp. of Minnesota is working on a semisolid, hydrophilic gel system containing a polymeric adhesive to improve its strength and tackiness. Nelson Research and Development Co.
of Irvine, California is working on its TranZoneR transdermal patches, which consist of a drug dispersed in various commercially available polymeric matrices along with the company's penetration enhancer, AZONE.
Boehringer Ingelheim recently introduced its Catapres-TTSR (transdermal therapeutic system), which contains the antihypertensive clonidine.
This patch effectively delivers drug for a period of seven days  Nitroglycerin was first synthesized in 1846 by A. Sobrero, by mixing cold concentrated sulfuric and nitric acids with glycerin. The first reported use in medicine was in 1847 by Hering and Davis in Philadelphia (Krantz, 1975).
( 14 Chemically, nitroglycerin is glyceryl trinitrate, or trinitroglycerin, The organic nitrates are dilators of arterial and venous smooth muscle. shown the oral forms to have prolonged pharmacologic activity (Needleman & Johnson, 1980).
Organic nitrates, like other esters, undergo hydrolysis. This reaction occurs more rapidly in alkaline than acid conditions (McNiff,1980). The incubation of nitroglycerin in 4N sodium hydroxfde for fifteen minutes at 37 degrees Centigrade produced almost complete denitration, whereas treatment in 4N hydrochloric acid for six hours produced only a 28 percent decrease (Johnson, 1975).
Pure nitroglycerin is highly explosive and is used in the manufacture of dynamite. Nitroglycerin for use in experimentation can be safely handled when adsorbed onto a suitable diluent such as lactose ( ( 15 (Johnson, 1975 Centigrade for six hours in water produced no detectable denitration (Johnson, 1975).

Identification of nitroglycerin and its breakdown products {Figure
2) has been done by spectrophotometric {McNiff, 1980), colorimetric (Bell, 1964), and chromatographic techniques. Spectrophotometric methods are time-consuming and complex and do not indicate stability, or differentiate nitroglycerin from its degradation products (Baaske, et al, 1979). Gas chromatography (GLC), although it requires a timeconsuming extraction step, has been used for its sensitivity.
Quantitative data are possible even though nitroglycerin is thermally unstable {Wu, 1981).

Methods for in-vitro release studies
Many studies have been published, which carry out permeability measurements on cadaver skin and animal skin (Keith, 1983). The most widely reported in-vitro method for evaluating percutaneous absorption of drugs uses a membrane mounted between two fluid-filled chambers  . The device, a diffusion cell, allows drug permeation to be monitored from the side exposed to the stratum corneum (donor) through to the dermal side (receptor) (Bronaugh, 1985). The Franz diffusion cell (figure 3) simulates clinical conditions by maintaining the receptor compartment at 37 degrees Centigrade while allowing the donor compartment to be exposed to ambient temperatures and humidities. The receptor compartment typically contains a propylene glycol or buffer solution to solubilize the drug and to simulate sink conditions. Hairless mouse skin or human cadaver epidermis may be used as the test membrane (Chien, 1983;Swarbrick, 1984;Behl, 1984;Cavey, 1985). Silicone elastomers such as polydimethylsiloxane have also been used successfully as transport membranes (Hsieh, 1985;Flynn & Roseman, 1971  ( ( 19 complexities, dissolution testing has been used only as a guide to the formulator in the early stages of drug product design, and 11 • • • as a quality control device to ensure process and batch-to-batch consistency for a particular formulation of a particular manufacturer" (Cooper, 1984). At the present time, there are no official criteria for the regulation of dissolution testing of transdermal products, because of the wide range of products available. This method is used by a number of companies who manufacture transdermals for assessing the release characteristics of their devices. A distinct advantage of this method is that it eliminates the step of the transport of drug across a membrane (a difficult factor to control) and instead allows evaluation of the release characteristics of the device. Four solutions (100, 135, 150, and 200 mcg/ml) were prepared using 10% methanol in water to solubilize the nitroglycerin/lactose triturate.
The solutions were vortexed for two minutes, left to stand for ten minutes, and then vortexed for one minute longer. HPLC was used to quantitate the samples.

Dissolution
Drug release studies were performed using conventional dissolution equipment. To assess in-vitro drug release from the vehicle, the dissolution apparatus ( A hole in the tape slightly larger than the disc but smaller than the

Extractions
In order to evaluate the total drug content per disc, an extraction procedure was done using two discs from each of the twenty-seven batches of silicone discs containing nitroglycerin. Using a metal spatula, the discs were cut into 5x5, 6x6, or 7x7 pieces, respectively for the 1. Six flasks at a time were placed on a reciprocal shaker and shaken for ( 28 two hours. A single sample was removed from each flask and filtered before assaying by HPLC.

Penetration Enhancer Study
A formulation was chosen from the twenty-seven batches of silicone discs containing nitroglycerin which showed relatively slow release characteristics. This formulation was composed of the high concentration of nitroglycerin (30 mg/ml), the 1.25 inch diameter, and the medium firmness (silicone to catalyst ratio of 1 to 2). Two batches of this formulation were prepared, one containing 10 percent and the other 20 percent dimethylsulfoxide (DMSO). The pol}11ler components were mixed and preheated, as before. The nitroglycerin/lactose triturate was stirred in, and finally the DMSO was added. Due to the small batch size, these formulations were injected into 1.25 inch molds rather than poured into sheets and cut. The volume injected was 2.5 ml, in order to maintain an average thickness of 0.3 cm. These were covered with aluminum foil and allowed to cure for 24 hours.

Comparative studies
The results of the nitroglycerin study were expected to indicate which parameters have the greatest impact on drug release from this system. The degree to which these effects can be applied to transdermal systems in general was tested by evaluating their contribution to the release of scopolamine from the same system. Scopolamine was chosen for this portion of the study since it is reported to possess percutaneous absorption properties (Chandrasekaran, et al, 1984), and since it is ( 29 already marketed as a transdennal drug delivery system (Transdenn-Scop, Ciba-Geigy).

Preparation of Scopolamine Base
Scopolamine free base was used rather than the hydrobromide salt since its physical characteristics (better lipid solubility) make it more suitable for transdennal delivery. The free base was prepared by dissolving 1.00 gram of scopolamine hydrobromide in 20 ml of methanol and 40 ml of chlorofonn. This was reacted with 20 ml of a 1 percent sodium bicarbonate solution (pH= 8.0) in a separatory funnel. This mixture was shaken and vented several times to complete the reaction.
The methanol phase was separated off, leaving the scopolamine base dissolved in chlorofonn. The chlorofonn was allowed to evaporate by air, leaving the clear, viscous liquid scopolamine base. Several "batches" of the base were prepared and combined. The purity of the free base was verified by HPLC prior to use, using the hydrobromide salt as a reference.

Preparation of Scopolamine discs and release studies
Silicone discs containing scopolamine were prepared using two different concentrations of drug, 1% and 2%. The ratio of silicone to catalyst was 1 to 2, and the diameters chosen were 1.00 and 1.25 inches.
The polymer/ catalyst mixture was preheated at 60 degrees Centigrade for 12 to 15 minutes and the drug was carefully stirred in. The polymer/ drug mixture was poured into pans and allowed to cure for 24 hours ( ( 30 before cutting the discs. The finished discs were weighed and the thickness was calculated.
Release studies were perfonned using the same dissolution procedure as the nitroglycerin studies. Three discs each from the total of four batches were run. Samples were taken at 0. 5, 1, 2, 4, 6, 8, 12, and 24 hours and refrigerated until analysis.

Scopolamine Assay
The release of scopolamine was analyzed using the  in thickness of gel between the two ends of a pan appeared to be most responsible for the variation in individual weights of silicone discs.
75 percent of t he discs that were used in the release studies varied in weight less than 5 percent. The remaining discs were within 10 percent of mean weights.
The actual thickness of the finished discs was not physically measured, but was calculated for each disc based on the weight and known surface area of the disc. It was found that 10.0 ml of uncured silicone weighed 10.0 grams, therefore the conversion factor of 1.0 gram/cm 3 was used in the volume and thickness calculations. Table 4   The relative firmness of the discs was quantified using the Instron hardness tester. This instrument produced reliable values for comparison, using a standard procedure. The compression force in kilograms required to produce a 50 percent deformation of the discs was measured at several points on two discs from each medium size (1.25 inch diameter) batch (Table 5). Compression force as a function of firmness ( Figure 5) and of concentration ( Figure 6) were plotted. In general, as the total amount of catalyst was increased, the compression force also increased. Likewise, increasing the drug concentration produced an increase in the compression force, indicating a firmer disc (30 mg/cm 3 > 20 mg/cm 3 > 10 mg/cm 3 ). In Figure 5 the discs having the highest concentrations of drug showed the highest compression rates and increased in a linear fashion as the catalyst ratio increased (r= 0.997). The 10 and 20 mg/cm 3 discs produced increased compressions relative to concentration and catalyst levels, however there appeared to be a leveling effect from the catalyst (r= 0.983 and 0.939 for the 20 and 10 mg/cm 3 concentrations).
When compression force was plotted as a function of concentration, the results were similar in that higher ratios of catalyst and higher concentrations produced higher compressional forces. The slopes increased as a function of catalyst level and the linearity was better than generally seen in Figure 5 (R= 0.998, 0.978, 0.962).
In any case, we can conclude that higher catalyst levels and concentrations of suspended material produce firmer discs. The significance of this finding was tested by evaluating the effect of firmness of release of drug from the discs.  In order to initially assess the in-vitro release method, three commercial transdermal products were tested. The products were Key

~
Pharmaceutical 's NitroDurR 10 cm 2 , Searle's NitrodiscR and Ciba's Transderm-NitroR-5. Each of these systems is designed to release 5 mg of nitroglycerin through the skin over 24 hours. Fluxes and rates of release for the first two hours and the final 12 hours (Table 6) were ---~ calculated. Overall, it was found that the rate of nitroglycerin release from the commercial products was higher during the first two hours than in the last 12 hours, with the exception of the Transderm-NitroR rate which changed dramatically during the last 12 hours (Figure 9). The Searle and Key products showed essentially no change in rate during the last 12 hours. Initial flux rates for the Key, Searle, and Ciba products were 1.07, 0.28, and 0.11 mg/cm 2 /hr, respectively, and 0.014, 0.003, and 0.065 mg/cm 2 /hr in the final 12 hours.
Total milligrams and percent of nitroglycerin released over 24 hours varied according to the formulation. NitroDurR and NitrodiscR demonstrated decreasing rates with time due to their relatively firmer systems which retarded the migration of nitroglycerin inside the disc to the surface.

Disc Extractions
The results of the disc extraction study are shown in Table 7. An overall mean of 60 percent extraction was found. The effects of size and concentration on the total percent of drug extracted from the discs appeared to be insignificant.
The fact that only 60 percent of the theoretical drug load was extracted leads to several possible conclusions which include incomplete extractions, drug decomposition, standard decomposition, and physical/ chemical interaction of the drug within the silicone.
The first three possibilities were examined during the method and assay validation processes. In developing the extraction process, a second extraction of each disc using fresh fluid was attempted, but no   Samples of silicone discs containing nitroglycerin were tested at random, but were selected according to their weight. It was important to reduce the interfering effect of disc thickness for these studies, therefore discs were chosen which were within 5 percent of variation in thickness. However, due to the small batch size, it was occasionally necessary to select a disc which varied up to 10 percent from the mean.   in concentration. This appears to be due to a function of the driving force of diffusion and the total drug load. As the total drug load was increased, a smaller percentage of drug was actually released since some drug must remain inside the matrix as the driving force.  In comparing batches having the same value of f 1 , we see a common factor of surface area. For example, batches 9, 12 and 27 have a flux rate of 0.172 mg/cm 2 /hr, and have the same surface area of 11.94 cm 2 .
Similarly, batches 7, 10 and 25 have f 1 1 s of 0.242 mg/cm 2 /hr and have a common surface area of 5.07 cm 2 . Since no other common formulation factors were noted between these batches, it appears that the initial flux of nitroglycerin from these systems is most influenced by the surface area of the discs.
On the other hand, f 2 appeared to be influenced the most by the concentration of drug in the disc. For example, if we examine batches 3, 10, 19 and 20, which had f 2 1 s of 0.063, 0.062, 0.059 and 0.061 mg/cm 2 /hr, respectively, we find a common factor of low drug concentration (10mg/cm 3 ).
The ratios of f 1 to f 2 are also listed in Table 9. As can be seen from the data, no generalizations can be made with regards to increases in any formulation parameters, suggesting a possible interactive effect of these parameters.
Analysis of variance (Figure 20) of the release data showed surface area (P< 0.01) and concentration (P< 0.01) to be significant factors · in the total release of nitroglycerin from the silicone discs. The firmness of the discs had no significant effect on the release of drug from these systems. Table 10   where a subtle decrease in total milligrams released was noted as the firmness increased. These two groups were both of the small surface area (5.07 cm 2 ) and the low and medium (10 and 20 mg/cm 3 ) concentrations.

Comparative
Scopolamine base could not be uniformly dispersed into this silicone system. The base appeared as droplets of various sizes which did not dissolve when broken with a glass stirring rod. For this reason, reproducible release characteristics could not be expected. This is evident in the results of the release studies which show virtually no release of scopolamine from any sample during the 24 hour time period.
The scopolamine base is a sticky, viscous liquid and is difficult to handle. Accurate weighing and mixing are essential in an evaluation of this type, but can not be guaranteed.
Vigorous mixing with a glass stirring rod reduced the droplet size, however this method of mixing introduced air bubbles, which are undesirable and could not be removed. The scopolamine base soon settled back to the bottom of the beaker.

Penetration enhancer study
The addition of dimethylsulfoxide (DMSO) to silicone/ nitroglycerin formulations was expected to improve the release of nitroglycerin from systems which showed relatively slow release characteristics. However, at the levels tested, this did not occur. In fact, the addition of DMSO to these formulations prevented the cross-linking of the pol}111er, and Due to the lack of resources in the area of penetration enhancers, this portion of the study was not pursued. Other researchers in the field of transdermal drug release have indicated that the application of a penetration enhancer directly to the surface of the device (attempting to directly affect the stratum corneum) is a more successful approach than the encorporation of the penetration enhancer into the matrix (Sanders, 1985). (

FUTURE WORK
This study provided insight into how the release of nitroglycerin is affected by the design of the disc. Several possibilities exist for further investigation in the following areas: 1. In retrospect, a wider range of firmness levels could probably more clearly define the significance of this factor on drug release.
2. To evaluate a system using a more potent drug, i.e., one which would require a smaller drug load in the device, such as timolol or clonidine.
3. To investigate the possibility of encorporating a more compatible penetration enhancer into these systems. Compatible penetration enhancers are currently being studied by other investigators and will find an increased use in commercial transdennal products.
4. To compare this in-vitro method of drug release study to one utilizing diffusion cells with either excised skin or silastic membranes.