CONTRACEPTION
SPERMICIDE
PERMEATION THROUGH BIOCOMPATIBLE
POLYMERS
W. Mark Saltzman and Leonard B. Tena Department of Chemical Engineering, The Johns Hopkins University, Baltimore, MD 21218 ABSTRACT Although spermicides are safe and effective contraceptive/prophylactic agents, they are inconvenient to use. Formulations that provide a controlled release of spermicide may improve user acceptance, and therefore effectiveness. Using a two-chamber diffusion cell, we measured the rates of permeation of nonoxynol-9 (N9), benzalkonium chloride (BC), and chlorhexidine (CH) through films of ethylene-vinyl acetate copolymer (EVAc) and silicone elastomer (SILASTIC). In addition, we encapsulated N9, BC, and CH into solid polymer matrices and measured the rate of spermicide release following immersion in water. We also developed equations for predicting the release rate of spermicide from a vaginal ring containing encapsulated spermicide, and tested these equations using hollow SIlASTIC rings containing pure N9 or BC. N9 diffuses through a thin film of SILASTIC several orders of magnitude slower than through water. The rates of permeation of N9 through EVAc, BC through SILASTIC, and CH through SILASTIC were too Polymer matrices of EVAc or SILASTIC released low to detect over a one-week experiment. N9 at a controlled rate for several days. Based on these measurements, we predict that a vaginal ring containing an inner core of EVAc/NS surrounded by a thin, permeable layer of SILASTIC will provide a controlled, constant release of N9 for over 30 days. Because of its low permeability in SILASTIC, BC is probably not a good spermicide for a long-acting vaginal ring. Because of its low volubility in water, CH is also not a good candidate for controlled release into the vaginal mucus. INTRODUCTION Barrier methods using spermicides provide safe and effective contraception. In addition, spermicides--whether provided in cremes, jellies, foams, or suppositories--are the only currently available female contraceptives that may provide prophylaxis against sexually transmitted disease. Improved formulations for the delivery of spermicides may improve both user acceptance and contraceptive/prophylactic efficacy. In this study, we measured rates of spermicide permeation through biocompatible polymers. This information was used to demonstrate the feasibility of disposable polymeric devices that provide a continuous, controlled release of spermicide. A vaginal ring releasing spermicides would have several potential advantages over the vaginal sponge. Spermicide release from the sponge is poorly controlled and--since mechanical forces can enhance release--user-dependent; the ring would release spermicides in a controlled, device-dependent manner. Because of this, the ring may provide a more consistent level of contraceptive protection over an extended period (24 hr), may be inserted independent of coitus, and--because of the lower detergent dose--may produce less irritation. Since the ring releases spermicide for several weeks, a single ring could be reused many times, thereby reducing cost. These factors may make the ring more acceptable to users than the sponge.’ Of course, the vaginal ring may share some of the benefits and risks of the sponge, including prophylaxis against sexually transmitted diseases and increased risk of candid&is.2 Submitted for publication January 4, 1991 Accepted for publication March 25, 1991
MAY 1991 VOL. 43 NO. 5
497
CONTRACEPTION MATERIALS AND METHODS M&&J,6 Ethylene-vinyl acetate copolymer, EVAc (ELVAX 4OW), was obtained from E.I. duPont de Nemours & Co, Wilmington, DE. Various samples of silicone elastomer were obtained from Dow Corning Medical, Midland, MI: i) SILASTIC@ medical grade sheeting was vulcanized, non-reinforced, and provided as a 0.013cm thick sheet; ii) SILASTIC tubing (MDF-OI 08) was also vulcanized and non-reinforced and provided as tubing with an outer diameter of 0.64 cm and inner diameter of 0.32 cm; and iii) SILASTIC medical grade elastomer (MDX4-4210) was provided as a e-part room temperature curing elastomer (base elastomer and curing agent) with a platinum catalyst. SILASTIC medical adhesive type A (Dow Corning Medical) was used in the fabrication of vaginal rings. Porous cellulose acetate membranes (Supor-200, 0.2 pm pore size, 0.74 porosity) were obtained from Gelman Sciences. The following spermicides and test solutes were used as received: i) toluidine blue-0 (Aldrich Chemicals); ii) fluorescein sodium salt (Sigma Chemical); iii) nonoxynol-9, N9 (GAF Chemicals Corp, IGEPAL CO-630); iv) benzalkonium chloride, BC (Sigma Chemical); v) chlorhexidine diacetate, CH (Sigma Chemical). Analytical grade acetone (Fisher Scientific) and methylene chloride (Aldrich Chemicals) were used as received. of Polymer Films EVAc was purified by washing in distilled water overnight, washing in a Soxhlet extractor with water for one day then acetone for three days. The washed polymer was dissolved in methylene chloride (5 to 20% w/v). Solutions of EVAc in methylene chloride (5%,10%, and 20%) were coated onto a porous membrane support (Gelman Supor 200) with a photoresist spincoater (Headway Research, TX) by applying 1 ml of EVAc/methylene chloride solution and spinning for 30 set at 50 to 8,000 rpm. Medical grade SILASTIC MDX4-4210 was prepared according to manufacturer’s specifications. Both components were degassed under vacuum for 2 to 6 hr before mixing. The components were mixed in a 1O:i ratio (elastomer:curing agent), degassed for 1 hr, pressed between two glass plates that were separated by thin glass shims, and cured for 24 hr at 2O’JC. Films were checked for homogeneity by light microscopy (400x). ent of diffusion coefficients in a two-s The polymer film or membrane for studv was olaced between the chambers of a stirred diffusion cell (Vanauard International, with a-chamber volume of 3.3 ml and chamber orifice of diameter 1 cm). ‘A sample with low concentration of the solute was placed in one chamber (sample side); a sample with high concentration was placed in the other (donor side). At fixed intervals after filling the diffusion cell reservoirs, small samples were taken from the sample side. The concentration of solute in the sample was determined by UWvisible light absorption (N9, BC, CH, toluidine blue) or fluorescence spectrophotometry (fluorescein). The thin films of EVAc and SILASTIC were tested for pinhole defects by mounting them in the diffusion cell with phosphate buffered water in the sample side chamber and toluidine blue in buffered water in the donor side chamber for several hours. If toluidine blue was detected in the sample side during this period, the film was discarded. Diffusion coefficients were calculated using a pseudosteady-state approachs,4? Ill
v Pm= g
a zln
c,-c, [ C-Cl 1
where C is the concentration of solute in the sample chamber, c’ is the concentration in the donor chamber, Co is the initial concentration in the sample chamber, Co’ is the initial concentration in the donor chamber, Pm is the permeability of the membrane to the solute, A is the total area available for diffusion, and V is the volume of the donor and supply chambers (assumed equal).
498
MAY 1991 VOL. 43 NO. 5
CONTRACEPTION The permeability, Pm, is related to the diffusion coefficient membrane, Dim, by:
PI
of the solute i in the polymer
Dim = P,,,L/K
where L is the.membrane thickness and K is the partition coefficient.2 For diffusion through macroporous membranes, solute transport occurs through water-filled pores in the membrane, so K is equal to the membrane porosity E. The diffusion coefficient of the solute i in the membrane, Df:m, is related to the free diffusion coefficient of the solute in water, D,:
[31
Q, = Q:mz
where z is the tortuosity of the water-filled pore space within the membrane. through 3 for solute diffusion through a macroporous membrane gives: [4]
Dfw = 2
Combining eq. 1
i $lnr$)
The diffusion coefficient for each test solute in water was determined by performing experiments with cellulose acetate membranes and calculating DiW from eq. 4. For diffusion through thin, nonporous films of EVAc or SILASTIC, the permeability of the test solute through the film was calculated from diffusion cell data with eq. 1. These experiments were characterized by a modified permeability defined as PmL=Df:mK. . * coemt for N9 m SILASUI; Dry slabs of SILASTIC (thickness -2 mm and diameter -1 cm) were weighed and placed in a large volume of pure N9. Periodically, the surface of each slab was wiped clean of spemticide and the slab was reweighed to determine the amount of N9 in the slab. The samples were incubated for several weeks until the mass of the slab no longer changed with time. Once the slab was completely swollen with the spermicide, it was placed in water, incubated for several days, and the concentration of spermicide in the water was periodically measured. Controlled rely matna EVAc matrices (thickness 1.5-3.7 mm and diameter 5 cm) containing spermicide were fabricated by solvent evaporation. A known mass of EVAc was dissolved in methylene chloride solution to make a 10% (w/v) polymer solution. Spermicide was added to this solution. The polymerlspermicide solution was homogenized, poured into a 5cm glass mold at -8WC, allowed to solidify for 10 min, covered with a wide mouthed glass jar, maintained at -20°C for 2 d and under vacuum for an additional 2 d. SILASTIC matrices (thickness 1.5-2.0 mm and diameter 5 cm) containing spermicide were fabricated by mixing the spermicide with the elastomer and degassing for 72 hr. The curing agent was added to the degassed mixture and the resulting viscous solution was poured into a leveled glass mold. The slab was allowed to cure for 3 d at room temperature. The slabs were then cut into discs (radius 5.1 mm). Each disc was placed in 20-ml of phosphate buffered water and incubated on a shaking table at 37°C. Periodically, the buffered water was completely replaced and the concentration of spermicide in the water was determined by spectrophotometry. The diffusion coefficient for solute transport in a porous polymer was calculated from these desorption measurements. For desorption from a slab of thickness L, the total mass of solute released from the matrix is given by: Mt = 4CoVm~~ 19 where DfIeff is the effective diffusion coefficient for solute diffusion in the polymer slab, Co is the total concentration of solute in the polymer matrix, Vm is the volume of the matrix, and t is time following initiation of desorpti0n.s
MAY 1991 VOL. 43 NO. 5
499
CONTRACEPTION . . . B Silicone tubes, 0.64 cm x 0.32 cm x 17 cm (O.D. x I.D. x length), were filled with spermicide by injection of 1.4 ml pure N9 or BC . The ends of the tubes were sealed to form a 5.4 cm ring by cleaning the surfaces of residual spermicide, connecting the two ends with a thin glass plug (diameter slightly greater than 0.32 cm made by stretching, scoring, breaking, and smoothing a glass stir rod), and sealing the connection with silicone adhesive. In preliminary experiments, rings were filled with an aqueous solution of fluorescein. These rings were immersed in water; they had no leaks. Each ring was placed in 25 ml of buffered water and incubated on a shaking table at 37%. Periodically, the water was replaced and the concentration of solute in the water determined. The rate of spermicide release was predicted from solution to the diffusion equation in cylindrical coordinates (see below). Assuming steady-state for a hollow cylindrical tube with inner radius a, outer radius b, and length L=2nR, where the wall of the cylinder (acr
SPERMICIDE
PERMEATION THROUGH BIOCOMPATIBLE
POLYMERS
W. Mark Saltzman and Leonard B. Tena Department of Chemical Engineering, The Johns Hopkins University, Baltimore, MD 21218 ABSTRACT Although spermicides are safe and effective contraceptive/prophylactic agents, they are inconvenient to use. Formulations that provide a controlled release of spermicide may improve user acceptance, and therefore effectiveness. Using a two-chamber diffusion cell, we measured the rates of permeation of nonoxynol-9 (N9), benzalkonium chloride (BC), and chlorhexidine (CH) through films of ethylene-vinyl acetate copolymer (EVAc) and silicone elastomer (SILASTIC). In addition, we encapsulated N9, BC, and CH into solid polymer matrices and measured the rate of spermicide release following immersion in water. We also developed equations for predicting the release rate of spermicide from a vaginal ring containing encapsulated spermicide, and tested these equations using hollow SIlASTIC rings containing pure N9 or BC. N9 diffuses through a thin film of SILASTIC several orders of magnitude slower than through water. The rates of permeation of N9 through EVAc, BC through SILASTIC, and CH through SILASTIC were too Polymer matrices of EVAc or SILASTIC released low to detect over a one-week experiment. N9 at a controlled rate for several days. Based on these measurements, we predict that a vaginal ring containing an inner core of EVAc/NS surrounded by a thin, permeable layer of SILASTIC will provide a controlled, constant release of N9 for over 30 days. Because of its low permeability in SILASTIC, BC is probably not a good spermicide for a long-acting vaginal ring. Because of its low volubility in water, CH is also not a good candidate for controlled release into the vaginal mucus. INTRODUCTION Barrier methods using spermicides provide safe and effective contraception. In addition, spermicides--whether provided in cremes, jellies, foams, or suppositories--are the only currently available female contraceptives that may provide prophylaxis against sexually transmitted disease. Improved formulations for the delivery of spermicides may improve both user acceptance and contraceptive/prophylactic efficacy. In this study, we measured rates of spermicide permeation through biocompatible polymers. This information was used to demonstrate the feasibility of disposable polymeric devices that provide a continuous, controlled release of spermicide. A vaginal ring releasing spermicides would have several potential advantages over the vaginal sponge. Spermicide release from the sponge is poorly controlled and--since mechanical forces can enhance release--user-dependent; the ring would release spermicides in a controlled, device-dependent manner. Because of this, the ring may provide a more consistent level of contraceptive protection over an extended period (24 hr), may be inserted independent of coitus, and--because of the lower detergent dose--may produce less irritation. Since the ring releases spermicide for several weeks, a single ring could be reused many times, thereby reducing cost. These factors may make the ring more acceptable to users than the sponge.’ Of course, the vaginal ring may share some of the benefits and risks of the sponge, including prophylaxis against sexually transmitted diseases and increased risk of candid&is.2 Submitted for publication January 4, 1991 Accepted for publication March 25, 1991
MAY 1991 VOL. 43 NO. 5
497
CONTRACEPTION MATERIALS AND METHODS M&&J,6 Ethylene-vinyl acetate copolymer, EVAc (ELVAX 4OW), was obtained from E.I. duPont de Nemours & Co, Wilmington, DE. Various samples of silicone elastomer were obtained from Dow Corning Medical, Midland, MI: i) SILASTIC@ medical grade sheeting was vulcanized, non-reinforced, and provided as a 0.013cm thick sheet; ii) SILASTIC tubing (MDF-OI 08) was also vulcanized and non-reinforced and provided as tubing with an outer diameter of 0.64 cm and inner diameter of 0.32 cm; and iii) SILASTIC medical grade elastomer (MDX4-4210) was provided as a e-part room temperature curing elastomer (base elastomer and curing agent) with a platinum catalyst. SILASTIC medical adhesive type A (Dow Corning Medical) was used in the fabrication of vaginal rings. Porous cellulose acetate membranes (Supor-200, 0.2 pm pore size, 0.74 porosity) were obtained from Gelman Sciences. The following spermicides and test solutes were used as received: i) toluidine blue-0 (Aldrich Chemicals); ii) fluorescein sodium salt (Sigma Chemical); iii) nonoxynol-9, N9 (GAF Chemicals Corp, IGEPAL CO-630); iv) benzalkonium chloride, BC (Sigma Chemical); v) chlorhexidine diacetate, CH (Sigma Chemical). Analytical grade acetone (Fisher Scientific) and methylene chloride (Aldrich Chemicals) were used as received. of Polymer Films EVAc was purified by washing in distilled water overnight, washing in a Soxhlet extractor with water for one day then acetone for three days. The washed polymer was dissolved in methylene chloride (5 to 20% w/v). Solutions of EVAc in methylene chloride (5%,10%, and 20%) were coated onto a porous membrane support (Gelman Supor 200) with a photoresist spincoater (Headway Research, TX) by applying 1 ml of EVAc/methylene chloride solution and spinning for 30 set at 50 to 8,000 rpm. Medical grade SILASTIC MDX4-4210 was prepared according to manufacturer’s specifications. Both components were degassed under vacuum for 2 to 6 hr before mixing. The components were mixed in a 1O:i ratio (elastomer:curing agent), degassed for 1 hr, pressed between two glass plates that were separated by thin glass shims, and cured for 24 hr at 2O’JC. Films were checked for homogeneity by light microscopy (400x). ent of diffusion coefficients in a two-s The polymer film or membrane for studv was olaced between the chambers of a stirred diffusion cell (Vanauard International, with a-chamber volume of 3.3 ml and chamber orifice of diameter 1 cm). ‘A sample with low concentration of the solute was placed in one chamber (sample side); a sample with high concentration was placed in the other (donor side). At fixed intervals after filling the diffusion cell reservoirs, small samples were taken from the sample side. The concentration of solute in the sample was determined by UWvisible light absorption (N9, BC, CH, toluidine blue) or fluorescence spectrophotometry (fluorescein). The thin films of EVAc and SILASTIC were tested for pinhole defects by mounting them in the diffusion cell with phosphate buffered water in the sample side chamber and toluidine blue in buffered water in the donor side chamber for several hours. If toluidine blue was detected in the sample side during this period, the film was discarded. Diffusion coefficients were calculated using a pseudosteady-state approachs,4? Ill
v Pm= g
a zln
c,-c, [ C-Cl 1
where C is the concentration of solute in the sample chamber, c’ is the concentration in the donor chamber, Co is the initial concentration in the sample chamber, Co’ is the initial concentration in the donor chamber, Pm is the permeability of the membrane to the solute, A is the total area available for diffusion, and V is the volume of the donor and supply chambers (assumed equal).
498
MAY 1991 VOL. 43 NO. 5
CONTRACEPTION The permeability, Pm, is related to the diffusion coefficient membrane, Dim, by:
PI
of the solute i in the polymer
Dim = P,,,L/K
where L is the.membrane thickness and K is the partition coefficient.2 For diffusion through macroporous membranes, solute transport occurs through water-filled pores in the membrane, so K is equal to the membrane porosity E. The diffusion coefficient of the solute i in the membrane, Df:m, is related to the free diffusion coefficient of the solute in water, D,:
[31
Q, = Q:mz
where z is the tortuosity of the water-filled pore space within the membrane. through 3 for solute diffusion through a macroporous membrane gives: [4]
Dfw = 2
Combining eq. 1
i $lnr$)
The diffusion coefficient for each test solute in water was determined by performing experiments with cellulose acetate membranes and calculating DiW from eq. 4. For diffusion through thin, nonporous films of EVAc or SILASTIC, the permeability of the test solute through the film was calculated from diffusion cell data with eq. 1. These experiments were characterized by a modified permeability defined as PmL=Df:mK. . * coemt for N9 m SILASUI; Dry slabs of SILASTIC (thickness -2 mm and diameter -1 cm) were weighed and placed in a large volume of pure N9. Periodically, the surface of each slab was wiped clean of spemticide and the slab was reweighed to determine the amount of N9 in the slab. The samples were incubated for several weeks until the mass of the slab no longer changed with time. Once the slab was completely swollen with the spermicide, it was placed in water, incubated for several days, and the concentration of spermicide in the water was periodically measured. Controlled rely matna EVAc matrices (thickness 1.5-3.7 mm and diameter 5 cm) containing spermicide were fabricated by solvent evaporation. A known mass of EVAc was dissolved in methylene chloride solution to make a 10% (w/v) polymer solution. Spermicide was added to this solution. The polymerlspermicide solution was homogenized, poured into a 5cm glass mold at -8WC, allowed to solidify for 10 min, covered with a wide mouthed glass jar, maintained at -20°C for 2 d and under vacuum for an additional 2 d. SILASTIC matrices (thickness 1.5-2.0 mm and diameter 5 cm) containing spermicide were fabricated by mixing the spermicide with the elastomer and degassing for 72 hr. The curing agent was added to the degassed mixture and the resulting viscous solution was poured into a leveled glass mold. The slab was allowed to cure for 3 d at room temperature. The slabs were then cut into discs (radius 5.1 mm). Each disc was placed in 20-ml of phosphate buffered water and incubated on a shaking table at 37°C. Periodically, the buffered water was completely replaced and the concentration of spermicide in the water was determined by spectrophotometry. The diffusion coefficient for solute transport in a porous polymer was calculated from these desorption measurements. For desorption from a slab of thickness L, the total mass of solute released from the matrix is given by: Mt = 4CoVm~~ 19 where DfIeff is the effective diffusion coefficient for solute diffusion in the polymer slab, Co is the total concentration of solute in the polymer matrix, Vm is the volume of the matrix, and t is time following initiation of desorpti0n.s
MAY 1991 VOL. 43 NO. 5
499
CONTRACEPTION . . . B Silicone tubes, 0.64 cm x 0.32 cm x 17 cm (O.D. x I.D. x length), were filled with spermicide by injection of 1.4 ml pure N9 or BC . The ends of the tubes were sealed to form a 5.4 cm ring by cleaning the surfaces of residual spermicide, connecting the two ends with a thin glass plug (diameter slightly greater than 0.32 cm made by stretching, scoring, breaking, and smoothing a glass stir rod), and sealing the connection with silicone adhesive. In preliminary experiments, rings were filled with an aqueous solution of fluorescein. These rings were immersed in water; they had no leaks. Each ring was placed in 25 ml of buffered water and incubated on a shaking table at 37%. Periodically, the water was replaced and the concentration of solute in the water determined. The rate of spermicide release was predicted from solution to the diffusion equation in cylindrical coordinates (see below). Assuming steady-state for a hollow cylindrical tube with inner radius a, outer radius b, and length L=2nR, where the wall of the cylinder (acr
