VOL. 9, PP. 355-366 (1975)
J. BIOMED. MATER. RES.
Biocompatible Implants for the Sustained Zero-Order Release of Narcotic Antagonists ROBERT A. ABRAHAMS and SAMUEL H. RONEL, Hydro Med Sciences, New Brunswick, New Jersey 08909
Summary Implantable, sustained release drug delivery devices offer benefits not obtained through oral ingestion or injection. These include delivery a t a constant therapeutic rate, thus avoiding adverse intermittent and massive dose effects, as well as reliance upon patients taking their prescribed dosages. The drawbacks to their widespread acceptance have been their inability to maintain a zero-order release rate over an extended period of time and poor .biocompatibility. Devices capable of satisfying these requirements have been developed and tested extensively for in vitro release of the narcotic antagonist cyclazocine. By using implant models prepared from Hydron, a hydrophilic polymer known to exhibit excellent tissue compatibility, we have found that the release rate could be precisely regulated by proper geometry, copolymer composition, concentration of ionogenic groups and cross-link density. Devices in such varied forms as capusles,, barrier-film coated tablets and bulk polymerized rods have been tested in vitro for periods approaching 1 year.
INTRODUCTION lcinding a solution t o the narcotic drug abuse problem in this country has attained the status of a national priority,' due to a n estimated 300,000 to a half-million hard drug users.2 The publicity being afforded the methadone maintenance program has been growing steadily more ~ r i t i c a l . ~ . ~ What is required t o alleviate the drug problem is a multifaceted program. It will have t o involve stringent enforcement of restrictions on the trafficking of drugs, and educational programs in the schools t o deter future generations of drug users from developing. Current users will require detoxification, psychological support and a vigorous rehabilitation program. The concept of the drug-free atmosphere of a detoxification center is a good one, but once ambulatory a reforming addict requires some extra support t o help him 355 @ 1975 by John Wiley & Sons, Inc.
ABRAHAMS AND RONEL
356
over emotionally trying episodes, during which time he is most liable to fall back upon drug dependence. The opiate antagonists that have been receiving increasing publicity lately appear to offer an addict desiring to be cured that extra margin of protection. Figure 1 shows some antagonist structures, as compared to morphine. As drugs, these compounds have the effect of blocking the action of the narcotic on the central nervous system by crossing the blood-brain barrier and occupying opiate receptor sites on nerve tissue that narcotics normally 0ccupy.~+6
Sustained Drug Release A major shortcoming to the widespread acceptance of antagonists is that their potency is sh~rt-lived.'-~ An ambulatory patient would
i"'
CYCUZOCINE
Ho
Fig. 1. Structural configuration of the morphine molecule and three of the most widely accepted opiate antagonists.
NARCOTIC ANTAGONIST IMPLANTATION
357
be required to return for an oral dose of the antagonist on a daily basis. Reliance upon this factor would certainly curtail the number of successes. A recent studylo showed a 50% drop-out rate of patients taking cyclazocine orally, as compared with a 20y0 drop-out rate among those on methadone maintenance. However, if a longacting, depot form of antagonist could be developed, for example a device that could be injected or implanted such that it would continuously release a therapeutic quantity of antagonist, the prospects would improve significantly for carrying the addict through emotional crises and give him greater opportunity t o develop a normal life style. The trend toward sustained drug release as a n alternate t o both direct oral ingestion and injection has gathered momentum over the last couple of years as the shortcomings of conventional drug administration became more recognized. These include a constantly varying drug concentration in the system that accompanies intermittent administration. Furthermore, massive dosages are frequently required to get a suitable quantity of drug t o a specific site, with the bulk of t,he drug being distributed throughout the rest of the body where i t may produce undesirable side effects. Implantable sustained release devices, among other benefits, offer the prospect of maintaining a constant blood-drug level which will be within the safe-to-effective range, as well as the capability of direct application a t the site a t which the drug is needed. Finally, such implants remove the necessity of having t o rely upon the patient's cooperation with regard t o administered medication. There are two major difficulties t o be considered when discussing long-term sustained release. The first is the problem of maintaining the release rate within a n effective yet safe range, that is by achieving kinetics that will approach zero-order. The second is t o have a device which will be sufficiently tissue compatible t o ensure that fibrous encapsulation that can gradually inhibit irrigation of the implant surface would be minimal.
Hydrophilic Polymers It has been found that a class of hydrophilic polymers, mainly the poly(hydroxyalky1 methacrylates) such as the 2-hydroxyethyl methacrylate shown in Figure 2 and known by the registered trade name of Hydron, can be modified and adapted to satisfy both requirements.
358
ABRAHAMS AND IZONEL 011
Fig. 2. The basic structure of poly(hydroxyethy1 methacrylate), a member of the Hydron family of hydrophilic polymers.
A primary consideration for an implantable device is its tissue compatibilit]y. Many publications have been devoted to this aspect of device design.11-'5 In work published by Levowitz and his colleagues, carried out primarily at Brooklyn Jewish Hospital, it was reportedI6 that Hydron evokes little or no host tissue response. Inflammatory cells and foreign body giant cells are scarce about the implant. The fibrous pseudocapsule is relatively thin and uonadherent t o the polymer. Other ~t,udies'7-'~have similarly demonstrated superior biological tolerance.
Theoretical Considerations The premise behind designing a system which would result in a pseudo zero-order release curve, that is, a controlled quantity of drug release per unit time, starts with Fick's Law of Diffusion:
J(i,x)
=
-D
dc(i) ~
dx
NARCOTIC ANTAGONIST IMPLANTATION
359
where J ( i , x ) is the flux (e.g., g/cmz/sec) of component i in the x-direction; dc(i)/dx is the concentration gradient of component i; D is the diffusion coefficient. From this basic equation one can derive the equation for steady-state material transport through an isotropic membrane.20 dm D*KA _ --_ _ _(Cl - Cd dt
X
is the quantity of material transferred across the membrane per unit time D* is the diffusivity K is the partition coefficient for the transported material between polymer and solution is the internal surface area of the membrane A X is the membrane thickness C1 - Cz is the concentration gradient of the diffusing species across the membrane. dm/dt
If we assume that the drug is being continuously swept from the surface by the flow of interstitial fluids then C2‘v 0 and the release rate will depend directly upon C1. The elution of the drug will remain pseudo zero-order as long as C , is a constant. A delivery device must then be prepared giving consideration to three facts: (1) All the terms on the right side of eq. (2) are constants except for C1. Thus, to provide a constant driving force for the transport across the membrane the concentration of drug within the device should be maintained a t a constant level. A constant concentration can be achieved by maintaining a saturated solution of the drug within the device. This, in turn, can be effected by having a reservoir of solid drug available which will go into solution t o replace the drug eluted through the device wall. ( 2 ) Data gathered in this study indicated that the term A in eq. ( 2 ) is not merely the surface area of the membrane but rather its effective surface area. That is, the quantity of drug released per unit time is directly proportional to the interfacial area between polymer and a saturated solution of the drug. A zero-order device must maintain a constant effective surface area for as long as possible. (3) The structure of the polymer should be formulated to make the diffusion of the drug through its matrix the rate determining step. This can actually be effected by
360
ABRAHAMS AND RONEL
modifying either the polymer or the drug. For example, for a given drug the rate of release can be altered by changing the polarity of the polymers or by modifying the polymer network to either produce steric effects or to control the water content. On the other hand, the drug can be modified to alter its solubility and rate of transport.
Current Work An extensive study has been performed on the release of the narcotic antagonist cyclazocine. Both the very water soluble hydrochloride salt, and the sparingly soluble free base derivatives of this drug have been made to elute a t rates on the order of the daily oral dosage. One direction taken by researchers in the development of depot forms of antagonists has been toward biodegradable polymer compositions, such as composites of the antagonist with poly(1actic acid). Limited success has been achieved with formulations derived from poly(1actic acid). However, data obtained follows a first-order release curve21-23rather than the more desirable zero-order release. Unpredictable bursts of drug are also observed throughout, the residence time of the depot22,24 presumably due to fracture of the degrading material resulting in an increased surface area. Furthermore, retrieved implants of such formulations are found to be badly distorted and fragmented. Although tissue reactions a t the site of the depot are reported to be minimall4 inspection of the site reveals decomposition products of the depot and the fate of these residual particles prior t o metabolism is as yet unknown.25
EXPERIMENTAL Three prototype designs successfully achieved the desired release characteristics. These were : (a) A capsular device containing drug, or a drug polymer blend, within the lumen. Shown in Figure 3 taken at two intervals during its active lifetime, this type of device was fabricated by the spin casting of a. tube of poly(hydroxyethy1 methacrylate). An appropriate length (e.g., 1 in.) was cut from the tube and the lumen filled with drug. The devices, which were generally of 3 mm i.d. and 5 mm o.d., were sealed using an adhesive of compatible composition which was polymerized, with acceleration,
NARCOTIC ANTAGONIST IMPLANTATION
361
(b)
Fig. 3. A capsular drug delivery device, photographed at two intervals during its active lifetime. (a) Day 0, (b) Day 98.
362
ABRAHAMS AND RONEL
onto the ends. (b) A tableted device consisting of drug or a drugpolymer blend which is provided with a barrier membrane of appropriake polymeric formulation by dip coating. Figure 4 demonstrates one such device a t two intervals during its active lifetime. (c) A homogeneous rod cast from a solution of cyclazocine in Z-hydroxyethyl methacrylate upon additian of initiator (e.g., 0.2y0 isopropyl percarbonate), within a Teflon tube a t 65°C for 1 hr. The external, diffusion barrier membranes were applied by dip coating a section of appropriate length cut from the molded product in a n alcoholic solution of suitable polymer, such as a copolymer of hydroxyethyl methacrylate and ethoxyethyl methacrylate. Each device was subjected t o a general in vitro procedure, designed t o measure the daily release of drug. This in vitro procedure consisted of placing a n individual device in a vial containing a measured quantity of phosphate buffer solution of pH 7.4, simulating, t o some degree, a physiological environment.26 The vials were then maintained under mild agitation a t 37°C. The buffer solution was changed daily and the concentration of drug in the retained eluent was determined by ultraviolet spectroscopy.
RESULTS AND DISCUSSION Figure 5 graphically demonstrates the dramatic effect that modification of the polymer formulation is capable of producing and the extent t o which release can be controlled. For the three curves of this figure only minor variation in the ionogenic group (methacrylic acid) content produced the observed results. It should be noted that the daily release is quite constant over the time period studied. Results were very reproducible. A similar effect was observed, though not as marked, by modification of the cross-link density of the polymer. Ethylene glycol dimethacrylate was employed in the range of 0.02y0t o 3.0% in the monomer reaction mixture. Another demonstration of the effectiveness of the diffusion barrier is contained in Figure 6, this time for tabletted devices. Each tablet of the free base form of cyclazocine initially weighed about 300 mg. The uncoated device yielded the first-order curve. The other two curves exhibit the zero-order release attainable with a barrier layer. They also demonstrate how the formulation of the coating can be modified to alter the quantity release per unit time.
NARCOTIC ANTAGONIST IMPLANTATION
363
(b)
Fig. 4. A barrier-coated drug device, photographed at two intervals during its active lifetime. (a) Day 18, (b) Day 116.
ABRAHAMS AND RONEL
364
6
% 5 $ 4 a7
2 3
K
*st111elutlng drug
M Y
Fig. 5 . The effect of ionogenic group (MAA) content on the rate of release of cyclaaocine (free base) from a capsular drug delivery device.
The release of 0.7 mg/day maintained by sample 105B was effected through the use of a coating of pure poly(hydroxyethy1 methacrylate) whereas the more rapid release was achieved by increasing the ionogenic group content (methacrylic acid) to 2.4%. Figure 7 provides an illustration of the effectiveness of a barrier membrane on a homogeneous central core (rod). Note the characteristic first-order release curve obtained from a device which
12 $10
e
r e a
e
2
-
6
0
a
2
* still elutlng drug
4 2
DAY Fig. 6. The effect of diffusion barrier membrane composition on the rate of release of cyclaeocine (free base) from a tebletted drug delivery device.
NARCOTIC ANTAGONIST IMPLANTATION I
6o Q -
.
>
J. BIOMED. MATER. RES.
Biocompatible Implants for the Sustained Zero-Order Release of Narcotic Antagonists ROBERT A. ABRAHAMS and SAMUEL H. RONEL, Hydro Med Sciences, New Brunswick, New Jersey 08909
Summary Implantable, sustained release drug delivery devices offer benefits not obtained through oral ingestion or injection. These include delivery a t a constant therapeutic rate, thus avoiding adverse intermittent and massive dose effects, as well as reliance upon patients taking their prescribed dosages. The drawbacks to their widespread acceptance have been their inability to maintain a zero-order release rate over an extended period of time and poor .biocompatibility. Devices capable of satisfying these requirements have been developed and tested extensively for in vitro release of the narcotic antagonist cyclazocine. By using implant models prepared from Hydron, a hydrophilic polymer known to exhibit excellent tissue compatibility, we have found that the release rate could be precisely regulated by proper geometry, copolymer composition, concentration of ionogenic groups and cross-link density. Devices in such varied forms as capusles,, barrier-film coated tablets and bulk polymerized rods have been tested in vitro for periods approaching 1 year.
INTRODUCTION lcinding a solution t o the narcotic drug abuse problem in this country has attained the status of a national priority,' due to a n estimated 300,000 to a half-million hard drug users.2 The publicity being afforded the methadone maintenance program has been growing steadily more ~ r i t i c a l . ~ . ~ What is required t o alleviate the drug problem is a multifaceted program. It will have t o involve stringent enforcement of restrictions on the trafficking of drugs, and educational programs in the schools t o deter future generations of drug users from developing. Current users will require detoxification, psychological support and a vigorous rehabilitation program. The concept of the drug-free atmosphere of a detoxification center is a good one, but once ambulatory a reforming addict requires some extra support t o help him 355 @ 1975 by John Wiley & Sons, Inc.
ABRAHAMS AND RONEL
356
over emotionally trying episodes, during which time he is most liable to fall back upon drug dependence. The opiate antagonists that have been receiving increasing publicity lately appear to offer an addict desiring to be cured that extra margin of protection. Figure 1 shows some antagonist structures, as compared to morphine. As drugs, these compounds have the effect of blocking the action of the narcotic on the central nervous system by crossing the blood-brain barrier and occupying opiate receptor sites on nerve tissue that narcotics normally 0ccupy.~+6
Sustained Drug Release A major shortcoming to the widespread acceptance of antagonists is that their potency is sh~rt-lived.'-~ An ambulatory patient would
i"'
CYCUZOCINE
Ho
Fig. 1. Structural configuration of the morphine molecule and three of the most widely accepted opiate antagonists.
NARCOTIC ANTAGONIST IMPLANTATION
357
be required to return for an oral dose of the antagonist on a daily basis. Reliance upon this factor would certainly curtail the number of successes. A recent studylo showed a 50% drop-out rate of patients taking cyclazocine orally, as compared with a 20y0 drop-out rate among those on methadone maintenance. However, if a longacting, depot form of antagonist could be developed, for example a device that could be injected or implanted such that it would continuously release a therapeutic quantity of antagonist, the prospects would improve significantly for carrying the addict through emotional crises and give him greater opportunity t o develop a normal life style. The trend toward sustained drug release as a n alternate t o both direct oral ingestion and injection has gathered momentum over the last couple of years as the shortcomings of conventional drug administration became more recognized. These include a constantly varying drug concentration in the system that accompanies intermittent administration. Furthermore, massive dosages are frequently required to get a suitable quantity of drug t o a specific site, with the bulk of t,he drug being distributed throughout the rest of the body where i t may produce undesirable side effects. Implantable sustained release devices, among other benefits, offer the prospect of maintaining a constant blood-drug level which will be within the safe-to-effective range, as well as the capability of direct application a t the site a t which the drug is needed. Finally, such implants remove the necessity of having t o rely upon the patient's cooperation with regard t o administered medication. There are two major difficulties t o be considered when discussing long-term sustained release. The first is the problem of maintaining the release rate within a n effective yet safe range, that is by achieving kinetics that will approach zero-order. The second is t o have a device which will be sufficiently tissue compatible t o ensure that fibrous encapsulation that can gradually inhibit irrigation of the implant surface would be minimal.
Hydrophilic Polymers It has been found that a class of hydrophilic polymers, mainly the poly(hydroxyalky1 methacrylates) such as the 2-hydroxyethyl methacrylate shown in Figure 2 and known by the registered trade name of Hydron, can be modified and adapted to satisfy both requirements.
358
ABRAHAMS AND IZONEL 011
Fig. 2. The basic structure of poly(hydroxyethy1 methacrylate), a member of the Hydron family of hydrophilic polymers.
A primary consideration for an implantable device is its tissue compatibilit]y. Many publications have been devoted to this aspect of device design.11-'5 In work published by Levowitz and his colleagues, carried out primarily at Brooklyn Jewish Hospital, it was reportedI6 that Hydron evokes little or no host tissue response. Inflammatory cells and foreign body giant cells are scarce about the implant. The fibrous pseudocapsule is relatively thin and uonadherent t o the polymer. Other ~t,udies'7-'~have similarly demonstrated superior biological tolerance.
Theoretical Considerations The premise behind designing a system which would result in a pseudo zero-order release curve, that is, a controlled quantity of drug release per unit time, starts with Fick's Law of Diffusion:
J(i,x)
=
-D
dc(i) ~
dx
NARCOTIC ANTAGONIST IMPLANTATION
359
where J ( i , x ) is the flux (e.g., g/cmz/sec) of component i in the x-direction; dc(i)/dx is the concentration gradient of component i; D is the diffusion coefficient. From this basic equation one can derive the equation for steady-state material transport through an isotropic membrane.20 dm D*KA _ --_ _ _(Cl - Cd dt
X
is the quantity of material transferred across the membrane per unit time D* is the diffusivity K is the partition coefficient for the transported material between polymer and solution is the internal surface area of the membrane A X is the membrane thickness C1 - Cz is the concentration gradient of the diffusing species across the membrane. dm/dt
If we assume that the drug is being continuously swept from the surface by the flow of interstitial fluids then C2‘v 0 and the release rate will depend directly upon C1. The elution of the drug will remain pseudo zero-order as long as C , is a constant. A delivery device must then be prepared giving consideration to three facts: (1) All the terms on the right side of eq. (2) are constants except for C1. Thus, to provide a constant driving force for the transport across the membrane the concentration of drug within the device should be maintained a t a constant level. A constant concentration can be achieved by maintaining a saturated solution of the drug within the device. This, in turn, can be effected by having a reservoir of solid drug available which will go into solution t o replace the drug eluted through the device wall. ( 2 ) Data gathered in this study indicated that the term A in eq. ( 2 ) is not merely the surface area of the membrane but rather its effective surface area. That is, the quantity of drug released per unit time is directly proportional to the interfacial area between polymer and a saturated solution of the drug. A zero-order device must maintain a constant effective surface area for as long as possible. (3) The structure of the polymer should be formulated to make the diffusion of the drug through its matrix the rate determining step. This can actually be effected by
360
ABRAHAMS AND RONEL
modifying either the polymer or the drug. For example, for a given drug the rate of release can be altered by changing the polarity of the polymers or by modifying the polymer network to either produce steric effects or to control the water content. On the other hand, the drug can be modified to alter its solubility and rate of transport.
Current Work An extensive study has been performed on the release of the narcotic antagonist cyclazocine. Both the very water soluble hydrochloride salt, and the sparingly soluble free base derivatives of this drug have been made to elute a t rates on the order of the daily oral dosage. One direction taken by researchers in the development of depot forms of antagonists has been toward biodegradable polymer compositions, such as composites of the antagonist with poly(1actic acid). Limited success has been achieved with formulations derived from poly(1actic acid). However, data obtained follows a first-order release curve21-23rather than the more desirable zero-order release. Unpredictable bursts of drug are also observed throughout, the residence time of the depot22,24 presumably due to fracture of the degrading material resulting in an increased surface area. Furthermore, retrieved implants of such formulations are found to be badly distorted and fragmented. Although tissue reactions a t the site of the depot are reported to be minimall4 inspection of the site reveals decomposition products of the depot and the fate of these residual particles prior t o metabolism is as yet unknown.25
EXPERIMENTAL Three prototype designs successfully achieved the desired release characteristics. These were : (a) A capsular device containing drug, or a drug polymer blend, within the lumen. Shown in Figure 3 taken at two intervals during its active lifetime, this type of device was fabricated by the spin casting of a. tube of poly(hydroxyethy1 methacrylate). An appropriate length (e.g., 1 in.) was cut from the tube and the lumen filled with drug. The devices, which were generally of 3 mm i.d. and 5 mm o.d., were sealed using an adhesive of compatible composition which was polymerized, with acceleration,
NARCOTIC ANTAGONIST IMPLANTATION
361
(b)
Fig. 3. A capsular drug delivery device, photographed at two intervals during its active lifetime. (a) Day 0, (b) Day 98.
362
ABRAHAMS AND RONEL
onto the ends. (b) A tableted device consisting of drug or a drugpolymer blend which is provided with a barrier membrane of appropriake polymeric formulation by dip coating. Figure 4 demonstrates one such device a t two intervals during its active lifetime. (c) A homogeneous rod cast from a solution of cyclazocine in Z-hydroxyethyl methacrylate upon additian of initiator (e.g., 0.2y0 isopropyl percarbonate), within a Teflon tube a t 65°C for 1 hr. The external, diffusion barrier membranes were applied by dip coating a section of appropriate length cut from the molded product in a n alcoholic solution of suitable polymer, such as a copolymer of hydroxyethyl methacrylate and ethoxyethyl methacrylate. Each device was subjected t o a general in vitro procedure, designed t o measure the daily release of drug. This in vitro procedure consisted of placing a n individual device in a vial containing a measured quantity of phosphate buffer solution of pH 7.4, simulating, t o some degree, a physiological environment.26 The vials were then maintained under mild agitation a t 37°C. The buffer solution was changed daily and the concentration of drug in the retained eluent was determined by ultraviolet spectroscopy.
RESULTS AND DISCUSSION Figure 5 graphically demonstrates the dramatic effect that modification of the polymer formulation is capable of producing and the extent t o which release can be controlled. For the three curves of this figure only minor variation in the ionogenic group (methacrylic acid) content produced the observed results. It should be noted that the daily release is quite constant over the time period studied. Results were very reproducible. A similar effect was observed, though not as marked, by modification of the cross-link density of the polymer. Ethylene glycol dimethacrylate was employed in the range of 0.02y0t o 3.0% in the monomer reaction mixture. Another demonstration of the effectiveness of the diffusion barrier is contained in Figure 6, this time for tabletted devices. Each tablet of the free base form of cyclazocine initially weighed about 300 mg. The uncoated device yielded the first-order curve. The other two curves exhibit the zero-order release attainable with a barrier layer. They also demonstrate how the formulation of the coating can be modified to alter the quantity release per unit time.
NARCOTIC ANTAGONIST IMPLANTATION
363
(b)
Fig. 4. A barrier-coated drug device, photographed at two intervals during its active lifetime. (a) Day 18, (b) Day 116.
ABRAHAMS AND RONEL
364
6
% 5 $ 4 a7
2 3
K
*st111elutlng drug
M Y
Fig. 5 . The effect of ionogenic group (MAA) content on the rate of release of cyclaaocine (free base) from a capsular drug delivery device.
The release of 0.7 mg/day maintained by sample 105B was effected through the use of a coating of pure poly(hydroxyethy1 methacrylate) whereas the more rapid release was achieved by increasing the ionogenic group content (methacrylic acid) to 2.4%. Figure 7 provides an illustration of the effectiveness of a barrier membrane on a homogeneous central core (rod). Note the characteristic first-order release curve obtained from a device which
12 $10
e
r e a
e
2
-
6
0
a
2
* still elutlng drug
4 2
DAY Fig. 6. The effect of diffusion barrier membrane composition on the rate of release of cyclaeocine (free base) from a tebletted drug delivery device.
NARCOTIC ANTAGONIST IMPLANTATION I
6o Q -
.
>
