- Email: [email protected]
Nifedipine gastrointestinal therapeutic system
Nifedipine Gastrointestinal Therapeutic System
DAVID R. SWANSON, Ph.D. BRIAN L. BARCLAY, I\?.S.Che. PATRICK S.L. WONG, Ph.D. FELIX THEEUWES, D.Sc. Palo Alto.
California
Convenient once-a-day dosage regimen? are highly desirable in general, and especially for the treatment of asymptomatic diseases such as essential hypertension. Nifedipine is an insoluble, shortacting calcium channel blocker that presents a difficult technical challenge for formulation in a constant 24-hour delivery dosage form. Once-a-day dosage forms have been developed based on the gastrointestinal therapeutic system (GITS) push-pull osmotic pump configuration in three strengths with different drug delivery rates (mg/hour) per dose (mg), as 1.7/30, 3.4/60, and 5.1190. The delivery rates of drug from these systems are controlled by their drug loading, composition of osmotic components, membrane properties, and dimensions. The release rates are independent of pH in the range from gastric pH = 1.2 to intestinal pH = 7.5. The release rates are independent of stirring rate and therefore unlikely to be influenced by motility in the gastrointestinal tract. The drug release rate from the nifedipine GITS dosage form in vivo in the gastrointestinal tract of dogs has been found to be equal to the release rate in vitro, indicating that the in vitro test is predictive of in vivo delivery. Following administration of the nifedipine GITS dosage forms to human subjects, absorption rates, calculated from resulting plasma concentrations, indicate that the cumulative amount of drug absorbed in humans over 24 hours is proportional to the amounts of drug delivered in vitro. Plasma concentrations are therefore predictable and remain relatively constant throughout the 24-hour dosing interval. Nifedipine is a calcium channel blocker used for the treatment of angina pectoris; it has also been used abroad to treat hypertension. Although nifedipine is currently administered three to four times a day orally in capsule form, a convenient once-a-day regimen is highly desirable for treatment of both these conditions. The kinetic elimination half-life of nifedipine is 1.7 hours, and drug plasma concentrations correlate directly with the effect on blood pressure and heart rate [l]. Diastolic blood pressure lowering appears to be a single valued function of plasma concentration with a minimum effective concentration of 15 ngiml. Significant blood pressure lowering-without an associated increase in heart rate-might be achieved by selecting a dosage form that produces a slow increase in plasma concentration or drug absorption. From a kinetic point of view, a zero-order dosage form with slow onset of delivery is ideal, with a duration of delivery of close to 24 hours. From a technical standpoint, the controlled delivery of nifedipine is diffi-
From the ALZA Corporation, Palo Alto, California. Requests for reprints should be addressed to Dr. Felix Theeuwes, ALZA Corporation, 950 Page Mill Road, Palo Alto, California 94303.
December
21, 1987
The
American
Journal
of Medicine
Volume
83
(suppl
6B)
3
SYMPOSIUM
ON CARDIOVASCULAR
Before Operation Semipermeable membrane \
THERAPY-SWANSON
During Operation
Osmotic drug core
1.
Cross section
of the push-pull
DESCRIPTION
Nifedipine GITS was constructed based on the push-pull osmotic pump design reported previously [2-41. As shown in Figure 1, the system is constructed with a bilayer tablet core that contains the drug in the top layer and the osmotic polymeric driving agent in the lower layer. This bilayer tablet is coated with a semipermeable membrane that is drilled on the drug side to allow delivery of the drug formulation through an orifice. The bilayer tablet typically contains 60 to 80 percent of solid drug formulation and 20 to 40 percent of solid osmotic driving formulation. The drug formulation contains the drug combined with osmotic and suspending agents. When the pump is in operation, both the drug and osmotic layers imbibe water across the membrane by the process of osmosis to formulate a suspension form in the drug layer. Simultaneously, by the pulling action of the water in the drug compartment, the drug suspension is pushed out of the drilled orifice by the dispensing action of the expanding osmotic push layer, which results in the total mechanism of drug delivery from the system (Figure 1, right side). This mechanism of operation allows for the delivery of insoluble drugs, such as nifedipine, in suspension in a finely divided form ready for dissolution and absorption. The mass delivery rate from the dosage form, dmidt, can be written by equation I, wherein dV/dt is the total
4
December
21, 1987
The American
Journal
of Medicine
dV dt
C ’
The osmotic volume flow in the osmotic compartment is defined as Q, as indicated in equation II, and the osmotic volume imbibition flow into the drug compartment is defined as F, as indicated by equation III. II.
osmotic pump.
AND THEORY
dm -=-. dt
I.
Expanded push compartment
cult due to the drug’s insolubility (S = 10 pgiml), and because it cannot readily be released by methods based on diffusion. However, drugs can be delivered by an osmotic displacement mechanism independent of their properties, such that the drug is delivered in finely dispersed form ready for dissolution and absorption. The push-pull osmotic pump was developed as a dedicated oral dosage form for the delivery of insoluble drugs in suspension [24]. This article describes the system, and presents the drug delivery characteristics of this dosage form known as the gastrointestinal therapeutic system, or nifedipine GITS. SYSTEM
volume flow from the dosage form times the concentration of drug in suspension in the dispensed formulation (C,).
Delivery orifice
Polymeric push compartment Figure
ET AL
Q=@,
Ill.,
F=c$)D
The concentration in the dispensed drug formulation from the dosage form can also be written as indicated by equation IV, wherein C, is the concentration of solids dispensed from the dosage form, and FD is the fraction of the drug formulated in the drug compartment. IV.
C, = FD . C,
Equation IV implies that the formulated drug fraction in the drug compartment is a constant and is equal to the fraction of drug dispensed in the drug formulation; in other words, it is assumed that the ratio of drug to solids in both the dispensed formulation and the solid formulation are the same. This action is achieved by selecting the proper suspending agents so that there is no shift between the different formulation materials during storage or during operation of the dosage form. By substituting the values for Q, F, and C, from equations II, III, and IV, the total expression for the mass delivery rate follows as indicated in equation V. $
= (Q + F) . FI, . C,
The values of the osmotic flows Q and F can in turn be written explicitly as indicated in equations VI and VII, wherein k is the osmotic membrane permeability coefficient; h is the thickness of the membrane; A,, is the area of the push compartment; rP is the imbibition pressure of the push compartment; A is the total area of the dosage form; and ,‘rD is the imbibition pressure in the drug compartment. VI.
Q =; F =;
A,(H).
(A - A,(H)).
r&H) TD (H)
Equations I, V, VI, and VII have been presented previously as the equations underlying the operation of the push-pull osmotic pump for the delivery of water-soluble compounds, wherein both the drug and the push compartment contain solutes with constant osmotic pressures. However, for the delivery of an insoluble drug, both the drug and the push compartments have been formulated
Volume
83
(suppl 8B)
SYMPOSIUM
with polymers characterized by imbibition pressures that are not constant, but rather are a function of the degree of hydration H, as defined in equation VIII. Here, WH is the weight of the water imbibed per weight of dry polymer W,). VIII.
H = $:
degree
of hydration
P During operation of the pump, the area of the push compartment expands (Figure l), which therefore indicates that A, is also a function of the degree of hydration. By controlling the various parameters defining the total mass delivery rate in equation V, the push-pull system can be programmed for various delivery rate profiles. For nifedipine, the zero-order delivery profile has been selected. This delivery rate can be programmed for the maximum fraction of drug content delivered at a constant rate by selecting the maximum flow (Q) such that the delivery rate from the system can be written by equation IX, where &jc is the density of the drug compartment. dm ~ = Q . ,%-,c. Fb dt This expression implies that the dosage form dispenses the drug from the drug layer at a rate equal to the speed at which the drug is formulated as a suspension by imbibed water. By comparing equation V with equation IX, the condition for the concentration of solids dispensed from the dosage form can then be derived, as expressed by equation X.
c, = ~Q .
Pdc
F+Q
It is interesting to note from equation IX, and from the expression for Q in equation VI, that the zero-order delivery rate must be programmed based on a constant product of the area times the imbibition pressure (~~?rp)in the push compartment. Based on these design parameters, three different forms of nifedipine GITS have been programmed to deliver 30, 60, and 90 mg of the drug over a period of 24 hours at rates of 1.7, 3.4, and 5.1 mgihour, respectively. The systems are therefore labeled as 1.7130, 3.4160, and 5.1190. Some of the parameters describing the structures of each system are listed in Table I. As indicated in the table, each of the systems contain a 10 percent overage of drug since that amount is consistently retained within the dosage form on expiration of the delivery profile. MATERIALS
AND METHODS
The layered tablet of nifedipine GITS contains nifedipine in the drug layer, along with a high molecular weight polyethylene glycol as the osmotic agent and standard tableting excipients. The osmotic layer contains a similar polymeric osmotic agent and additional standard tableting excipients, including
December
21, 1987
TABLE
I
ON CARDIOVASCULAR
Design
Parameters
Systems
THERAPY-SWANSON
for Nifedipine
1.7/30
Diameter (cm) Drug content (mg) Weight of drug layer (m9) Weight of push layer (ms) Membrane thickness (cm) Membrane permeability (cm”/hour atm) Average zero-order delivery rate (mg/hour)
ET AL
GITS 5.1190
3.4/60
0.79 33
1.03
1.19
66
99
165
330
495
82.5
165
247.5
12 x 10-3
11 x
1o-3
9.9 x 10-3
6.35
x IO-’
6.35
6.35
x 10m7
1.7
3.4
x IO-’
5.1
ferric oxide as a colorant. The tablet is formed by standard granulating and tableting techniques. The rate-controlling semipermeable membrane consists of a combination of a cellulosic material and excipients. This membrane is applied directly to the layered tablet core by standard film-coating techniques. An exit pot-t may be drilled through the membrane on the drug-layer side by either a mechanical drill or a high-speed laser drill; the GITS preparation of nifedipine employs the precision laser method. The minimum membrane orifice diameter is large enough to permit the unrestricted movement of drug and excipients out the orifice (approximately 0.4 mm), whereas the maximum diameter is limited to prevent the uncontrolled leaching of the drug layer out the orifice (1 mm). A final colored overcoat is applied to the system for product identification purposes. This overcoat consists of standard film coat and colorant materials and is applied by aqueous film-coating techniques. The overcoat is strictly cosmetic, and has no effect on the release rate of the dosage form. All materials employed in tableting and coating of the system are listed in the United States Pharmacopeia (USP). Conventional USP dissolution testing procedures are impractical for testing the drug release rate from systems delivering insoluble drugs, such as nifedipine, with a solubility of 10 yglml. A dosage form containing 90 mg of such an insoluble drug would require a g-liter volume test reservoir in order to allow sampling of homogeneous solutions throughout the total delivery period. The methods employed for dissolution testing of the nifedipine GITS system were therefore a modified USP test and a well-published differential method, known as the ALZA method [5]. The ALZA release rate tester (ALZA Corporation, Palo Alto, California) is an automated apparatus designed to measure the drug delivery rate from controlled release systems at selected time intervals. This tester incorporates a series of individual reservoirs in a constant temperature bath with a vertically reciprocating shaking mechanism. The shaking mechanism holds the test system in the fluid reservoir and oscillates the system with a vertical stroke of 2.5 cm at a frequency of 0.5 cycles/second. At the end of a predeter-
The
American
Journal
of Medicine
Volume
83
(suppl
66)
5
SYMPOSIUM
ON CARDIOVASCULAR
THERAPY-SWANSON
ET AL
fm.m-
lime (hour) Figure 2. Comparison of in vitro release rate profiles for three doses of nifedipine GUS tablets. l = 1.7130 (n = 10); A = 3.4160 (n = 10); I = 5.1190 (n = 5); I = range.
90 7-
mg system
. ./ .. I
6.
30 mg system
10
20
30
40
Reciprocal membrane weight (g2 Figure 3. Average zero-order release rate as a function of reciprocal membrane weight for three nifedipine GITS tablets (0 = 1.5-kg batch; A = lo-kg batch; n = 130-kg batch).
6
December
21, 1987
The
American
Journal
of Medicine
mined time interval (which is set on an electronic controller), the shaking mechanism lifts the test system out of the reservoir, indexes the set of test systems to the next row of unused reservoirs, lowers them into these reservoirs, and starts the reciprocating stirring again for the preset time interval. This change takes less than five seconds. At the end of the final test interval, the systems are removed from the release media and held suspended over another unused reservoir, available for residual drug content determination. The nifedipine contents of each reservoir are solubilized by the volumetric addition of an appropriate solvent. The resulting solution is then assayed for its nifedipine content. The release rate (mgihour) is determined by dividing the amount of nifedipine in each tube by the length in hours of the test interval; the cumulative amount release is then determined by adding together the amounts from the various intervals. The ALZA method has been compared with the USP method for test systems delivering water-soluble drugs. For this case, it has been shown that test results in the ALZA apparatus are identical to those obtained from the USP apparatus [6]. By modifying the USP techniques, a method has been developed that determines the amount of nifedipine released from GITS by an indirect method. Systems are placed in the USP Apparatus II and stirred for predetermined periods of time. At the end of each test interval, the systems are removed from the flask and assayed for the amount of nifedipine remaining unreleased in each. This value is then subtracted from the average initial content, as determined by a separate assay; this therefore determines the amount of nifedipine released during that test interval. This procedure is used for each test time, and the results are used to construct a composite curve for the entire test. The test is repeated for various stirring speeds to determine their influence on release of nifedipine from GITS. Nifedipine was analyzed by high-pressure liquid chromatographic (HPLC) methods with a 266-nm ultraviolet detector. The size of the column was 15 x 0.46 cm, and it contained a 5-mm Altex Ultrasphere ODS packing (Beckman Instruments). The mobile phase was a 30:20:50 mixture of methanol, acetonitrile, and deionized water, respectively. In some cases, solutions of nifedipine in 50:50 polyethylene glycol 400:deionized water mixture were exposed to intense incandescent light for two to four hours. The resulting degradant product was then analyzed by the ultraviolet detector at 282 nm. The results of this procedure were compared with HPLC results, thereby confirming the validity of the alternate method. To compare in vivo system performance with in vitro results, four medium-sized mongrel dogs (two male and two female), fasted overnight, were used in this study. At the beginning of the study, each dog was fed a 150-g portion of dog food (Kal Kan) one-half hour before the first systems were administered. Each dog was fed one 8-g meatball hourly for the next 12 hours, and all dogs were fasted for the last 18 hours of the study. Individually labeled systems were administered with food to each dog at four, eight, 12, 18,24, and 30 hours prior to the sacrifice of each animal. The dogs were then sacrificed at 30 hours after the start of the study, and the systems were recovered from their gastrointestinal tracts. By
Volume
83 (suppl
66)
SYMPOSIUM
ON CARDIOVASCULAR
THERAPY-SWANSON
ET AL
analyzing the residual drug in these systems, the amount released was calculated as the difference between initial and residual content. At the same times as when the dogs were dosed, equal numbers of systems were placed in the in vitro release apparatus. These systems were removed at the same time that the dogs were sacrificed and assayed for residual drug content using an identical analytic procedure. These results were then compared with those obtained in the in vivo experiment. RESULTS
AND COMMENTS
24.00
During the operation of the nifedipine GITS, small quantities of drug were entrapped along the edges of vertical tablet walls and in the edges of the curved top dome. Drug residuals measured in experimental and clinical systems indicated that about IO percent residual drug remains after in vitro testing (9.8 percent, 30 mg; 9.6 percent, 60 mg; 6.5 percent, 90 mg). These results are in agreement with the results of a study performed on dogs, in which the residual amount of drug remaining in the systems after being dosed in vivo agreed directly with results from systems tested in vitro. Based on these results, an overage equivalent to 10 percent of the labeled dose has been incorporated into the final nifedipine GITS. The release rate of nifedipine GITS 1.7130, 3.4160, and 5.1190 are graphed simultaneously in Figure 2. The top part of the figure depicts the release rate as a function of time wherein the bars indicate the range of experimental data for 10 systems each. The lower part of the figure shows superimposed the cumulative amounts released as a function of time for the three forms with the range of experimental data. It is evident from Figure 2 that the dosage forms are equivalent on a normalized mass basis in vitro; that is, two 30-mg systems can be substituted to give the performance of one 60-mg system, and a 30-mg system plus a 60-mg system is equivalent to a 90-mg system. The osmotic imbibition fluxes F and Q in equations VI and VII are inversely proportional to the membrane thickness. Therefore, the release rates indicated in equations V and IX are also inversely proportional to the membrane thickness. For a given formulation composition, dimensions of the dosage form, and membrane composition, the release rates for each system can thus be programmed by controlling the membrane thickness, which in turn is porportional to the membrane weight for each system. Figure 3 shows the average zero-order release rate for each of the three dosage forms as a function of the reciprocal of the membrane weight. Each of the dosage forms shows a relationship that is linear in accordance with the theoretic prediction. Data are indicated for three different batch sizes as scale-up work progressed. The fact that all the data from the three different batch sizes fit on the same straight line indicates that all systems obey the same theoretic model independent of equipment, and that scale-up from small to larger batch sizes can be carried
December
21, 1987
mrne
-.
(hour)
_ ^
Prgure 4. L’ompanson ot tne cumulabve amount of drug released versus time for nifedipine GITS 3.4160 at different batch sizes (0 = 1.5-kg batch; A = IO-kg batch; n = 130kg batch).
out by controlling the composition of the dosage form and membrane thickness. To amplify the results of the scale-up procedures, the data for three batches of 3.5160 systems are graphed as cumulative amounts in Figure 4. This figure shows the average amount released with the range of experimental data for a total of 10 systems each. It is clear from this figure that on superimposition these three figures are nearly identical, illustrating that not only the average zeroorder release rate is independent of batch size, as already expressed in Figure 3, but also that the total release rate profile is independent of the final batch size. The release rate from osmotic dosage forms can be designed to be independent of the orifice size within a minimum to maximum range of sizes. The minimum orifice must be designed sufficiently large to allow drug delivery from the dosage form at the rate in which the volume of water is imbibed through the system. The maximum size must be designed so that diffusional losses through the orifice are small as compared with the pumping rate. Within this minimum-maximum orifice size range of specified area and diameter, the release rate remains independent of the orifice size. For each of the three dosage forms, the release rate as a function of orifice diameter is depicted in Figure 5. The release rate for the three systems has been found to be independent of orifice diameter over a range ratio of maximum to minimum size of at least a factor of five. This feature allows for a high-speed drilling process with low tolerance requirements. From a biopharmaceutical point of view, perhaps the most important characteristic of an oral dosage form is the release rate profile throughout the gastrointestinal tract. It is this release rate that controls the rate of drug absorption and ultimate therapeutic effect. Drug absorption can thus only be controlled by the dosage form if the release rate in vitro closely correlates with the rate of delivery in vivo. Only under this condition does a valid quality control test
The
American
Journal
of Medicine
Volume
83
(suppl 6B)
7
SYMPOSIUM
ON CARDIOVASCULAR
THERAPY-SWANSON
ET AL
apparatus, the residual content of the dosage form was measured at a particular stirring condition. Each of the data points in Figure 7 indicates the cumulative amount released, which was obtained from the initial system content minus the residual amount within each dosage form after the exposure time indicated. Therefore, cumulative amounts released obtained in this fashion contain a larger experimental error; the error associated with the releasing product plus the error in content uniformity of the dosage form. The cumulative amount released was measured for three different stirring conditions at identical time periods. In order to allow superimposition of the cumulative amounts released, the data points were slightly offset from the experimental time measured in order to be able to show the average standard deviations for each set of systems. The actual time data points are those indicated by the circles for the ALZA method. It is clear from this figure that there is no difference in the amount released for this dosage form between any of the stirring conditions indicated. In order to establish a valid quality control procedure in vitro that would be able to safeguard the release rate in vivo, it is necessary that an in vitro-in vivo correlation be established. The environment in the intestinal tract of the dog is a fair replica of that in humans in both pH and motility such that the release rate in the dog can be equated with the rate in humans. Figure 8 shows the cumulative amount released of nifedipine GITS in the gastrointestinal
I- ____ L-L--- ____ -.-I II
Tlf-f--t----J I----‘-‘--0.2
0.4
0.6
------_ I 0.8
1.0
1.2
1.4
Orifice diameter (mm) igure 5. Average zero-order release rate versus orifice diameter of nifedipine GITS (@ = 1.7130; A = 3.4160; n = 5.1190; I = total standard deviation that is the statistical sum of standard deviations between and within systems).
exist to safeguard the performance and therapeutic response of the dosage form. It is well recognized that pH, between 1 in the stomach to 7.5 in the intestine, and motility in the intestinal tract are the most variable parameters encountered by the dosage form. The release rate in vitro must be independent of these conditions in order to be able to produce a predictable in vivo delivery rate profile. Figure 6 shows the release rate of nifedipine GITS as a function of time in both gastric and intestinal fluids. The top part of the figure shows the superimposition of the release rate profiles in gastric and intestinal fluids as a function of time, while the lower portion of the figure shows the cumulative amounts released as a function of time. It is clear in this figure that the release rate from the systems in either medium are indistinguishable and independent of pH. The release rates from GITS are conventionally measured by the ALZA method [5] because it gives the release rate in a differential fashion and with a higher degree of accuracy than the conventional USP method. In order to measure the release rate as a function of stirring conditions, the cumulative amounts of drug released as a function of time were measured by a modified USP method. It is conventionally assumed that motility patterns in the intestinal tract can be equated with stirring conditions in the range of up to 150 revolutions/minute. The release rate from nifedipine GITS was therefore measured in the modified USP apparatus with the paddle stirring at 50, 100, and 150 revolutions/minute. Since the amount, of drug released from the nifedipine GITS system cannot be dissolved in the total volume of fluid contained in the USP
8
December
21, 1987
The
American
Journal
of Medicine
III
I,,,,,,
Time
$
(hour)
0.0 ’
igure 6. Average zero-order release rate and cumulative amount released for nifedipine GITS (A = artificial gastric fluid, (n = 5); l = artificial intestinal fluid, (n = 5); I = range).
Volume
83
(suppl 68)
SYMPOSIUM
ON CARDIOVASCULAR
THERAPY-SWANSON
ET AL
looQO.g
ao-
1 2 E
70-
[
50-
L9 + E 3
4030-
60-
zoloO12 Time (hour) owl,,
Figure 7. The cumulative percent released for nifedipine GlTS 5.1190 as a function of stirring rates using the modified USP paddle method (A = LISP, 50 revolutionsiminute; q = USP, 100 revolutionslminute; V = LISP, 150 revolutions/minute; l = ALZA release method; I = standard deviation). Data are offset from actual time measured indicated for ALZA method.
tract of dogs as a function of time as compared with the amount released in vitro. This figure shows that the environment, with its variability in pH and motility (such as the intestine of the dog), has no influence on the performance of the dosage form since the amounts released in both the in vitro and in vivo media are identical. The environmental conditions in the intestinal tracts of the dog and humans are substantially the same; therefore, based on these data one can expect the dosage form to behave in the intestinal tract of humans as it performs in vitro. CONCLUSIONS Nifedipine GITS was designed based on the push-pull system configuration in order to deliver drug in a zeroorder controlled predictable fashion as a function of time. Cumulative amounts released in vitro in subdivided forms are additive and interchangeable with the full-strength form.
I
2
4
6
8
I
I
I
I
I
I
10 12 14 16 18 20 22 24 Residence time (hour)
I
I
I
26
28
30
Figure 8. Comparison of in vivo and in vitro cumulative amount released for nifedipine GITS 1.7130 (0 = in vivo release, [n = as indicated]; A = in vitro release, (n = 4); I = standard deviation).
The release rates for nifedipine GITS are controlled by membrane composition and weight, and independent of batch size and scale of equipment. They are independent of the orifice diameter within a size range of a factor of five, and are independent of pH within the physiologic range of 1.2 to 7.5. The release rates of nifedipine GITS are also independent of the stirring rate at or below 150 revolutions/minute, in the physiologic analogous range. In vivo, the release rate has been found to be equal to that in vitro, which qualifies the latter test method as a valid quality control procedure to safeguard performance of the dosage form in vivo. A number of bioavailability studies in humans have been carried out comparing nifedipine GITS with the standard capsule form. Based on the plasma concentrations achieved in these studies, the rate of absorption was found to be equal to the rate of release in vitro at any time for the duration of the delivery period. Publications on that subject are presently in progress.
REFERENCES 1.
2.
3.
Kleinbloesem CH, van Brummelen P, van de Linde JA, Voogd PJ, Breimer DD: Nifedipine: kinetics and dynamics in healthy subjects. In: Nifedipine: clinical pharmacokinetics and haemodynamic effects (thesis). University of Leiden, 1985; 4760. Theeuwes F: Novel drug delivery systems. In: Prescott LF, Nimmo WS, eds. Drug absorption: proceedings of the Edinburgh International Conference, September 1979. Balgowlah, Australia: ADIS Press, 1981; 157-176. Cortese R, Theeuwes F: Osmotic device with hydrogel driving
December
21, 1987
4.
5. 6.
The
member. US. patent no. 4,327,725, issued May 4, 1982. Assigned to ALZA Corporation, Palo Alto, California. Wong P, Barclay B, Deters J, Theeuwes F: Osmotic device with dual thermodynamic activity. US. patent no. 4,612,008, issued September 16, 1986. Assigned to ALZA Corporation, Palo Alto, California. Theeuwes F: The elementary osmotic pump. J Pharm Sci 1975; 64: 1987-1991. Theeuwes F, Swanson D, Wong P, et al: Elementary osmotic pump for indomethacin. J Pharm Sci 1983; 72: 253-257.
American
Journal
of Medicine
Volume
83
(suppl 68)
9
DAVID R. SWANSON, Ph.D. BRIAN L. BARCLAY, I\?.S.Che. PATRICK S.L. WONG, Ph.D. FELIX THEEUWES, D.Sc. Palo Alto.
California
Convenient once-a-day dosage regimen? are highly desirable in general, and especially for the treatment of asymptomatic diseases such as essential hypertension. Nifedipine is an insoluble, shortacting calcium channel blocker that presents a difficult technical challenge for formulation in a constant 24-hour delivery dosage form. Once-a-day dosage forms have been developed based on the gastrointestinal therapeutic system (GITS) push-pull osmotic pump configuration in three strengths with different drug delivery rates (mg/hour) per dose (mg), as 1.7/30, 3.4/60, and 5.1190. The delivery rates of drug from these systems are controlled by their drug loading, composition of osmotic components, membrane properties, and dimensions. The release rates are independent of pH in the range from gastric pH = 1.2 to intestinal pH = 7.5. The release rates are independent of stirring rate and therefore unlikely to be influenced by motility in the gastrointestinal tract. The drug release rate from the nifedipine GITS dosage form in vivo in the gastrointestinal tract of dogs has been found to be equal to the release rate in vitro, indicating that the in vitro test is predictive of in vivo delivery. Following administration of the nifedipine GITS dosage forms to human subjects, absorption rates, calculated from resulting plasma concentrations, indicate that the cumulative amount of drug absorbed in humans over 24 hours is proportional to the amounts of drug delivered in vitro. Plasma concentrations are therefore predictable and remain relatively constant throughout the 24-hour dosing interval. Nifedipine is a calcium channel blocker used for the treatment of angina pectoris; it has also been used abroad to treat hypertension. Although nifedipine is currently administered three to four times a day orally in capsule form, a convenient once-a-day regimen is highly desirable for treatment of both these conditions. The kinetic elimination half-life of nifedipine is 1.7 hours, and drug plasma concentrations correlate directly with the effect on blood pressure and heart rate [l]. Diastolic blood pressure lowering appears to be a single valued function of plasma concentration with a minimum effective concentration of 15 ngiml. Significant blood pressure lowering-without an associated increase in heart rate-might be achieved by selecting a dosage form that produces a slow increase in plasma concentration or drug absorption. From a kinetic point of view, a zero-order dosage form with slow onset of delivery is ideal, with a duration of delivery of close to 24 hours. From a technical standpoint, the controlled delivery of nifedipine is diffi-
From the ALZA Corporation, Palo Alto, California. Requests for reprints should be addressed to Dr. Felix Theeuwes, ALZA Corporation, 950 Page Mill Road, Palo Alto, California 94303.
December
21, 1987
The
American
Journal
of Medicine
Volume
83
(suppl
6B)
3
SYMPOSIUM
ON CARDIOVASCULAR
Before Operation Semipermeable membrane \
THERAPY-SWANSON
During Operation
Osmotic drug core
1.
Cross section
of the push-pull
DESCRIPTION
Nifedipine GITS was constructed based on the push-pull osmotic pump design reported previously [2-41. As shown in Figure 1, the system is constructed with a bilayer tablet core that contains the drug in the top layer and the osmotic polymeric driving agent in the lower layer. This bilayer tablet is coated with a semipermeable membrane that is drilled on the drug side to allow delivery of the drug formulation through an orifice. The bilayer tablet typically contains 60 to 80 percent of solid drug formulation and 20 to 40 percent of solid osmotic driving formulation. The drug formulation contains the drug combined with osmotic and suspending agents. When the pump is in operation, both the drug and osmotic layers imbibe water across the membrane by the process of osmosis to formulate a suspension form in the drug layer. Simultaneously, by the pulling action of the water in the drug compartment, the drug suspension is pushed out of the drilled orifice by the dispensing action of the expanding osmotic push layer, which results in the total mechanism of drug delivery from the system (Figure 1, right side). This mechanism of operation allows for the delivery of insoluble drugs, such as nifedipine, in suspension in a finely divided form ready for dissolution and absorption. The mass delivery rate from the dosage form, dmidt, can be written by equation I, wherein dV/dt is the total
4
December
21, 1987
The American
Journal
of Medicine
dV dt
C ’
The osmotic volume flow in the osmotic compartment is defined as Q, as indicated in equation II, and the osmotic volume imbibition flow into the drug compartment is defined as F, as indicated by equation III. II.
osmotic pump.
AND THEORY
dm -=-. dt
I.
Expanded push compartment
cult due to the drug’s insolubility (S = 10 pgiml), and because it cannot readily be released by methods based on diffusion. However, drugs can be delivered by an osmotic displacement mechanism independent of their properties, such that the drug is delivered in finely dispersed form ready for dissolution and absorption. The push-pull osmotic pump was developed as a dedicated oral dosage form for the delivery of insoluble drugs in suspension [24]. This article describes the system, and presents the drug delivery characteristics of this dosage form known as the gastrointestinal therapeutic system, or nifedipine GITS. SYSTEM
volume flow from the dosage form times the concentration of drug in suspension in the dispensed formulation (C,).
Delivery orifice
Polymeric push compartment Figure
ET AL
Q=@,
Ill.,
F=c$)D
The concentration in the dispensed drug formulation from the dosage form can also be written as indicated by equation IV, wherein C, is the concentration of solids dispensed from the dosage form, and FD is the fraction of the drug formulated in the drug compartment. IV.
C, = FD . C,
Equation IV implies that the formulated drug fraction in the drug compartment is a constant and is equal to the fraction of drug dispensed in the drug formulation; in other words, it is assumed that the ratio of drug to solids in both the dispensed formulation and the solid formulation are the same. This action is achieved by selecting the proper suspending agents so that there is no shift between the different formulation materials during storage or during operation of the dosage form. By substituting the values for Q, F, and C, from equations II, III, and IV, the total expression for the mass delivery rate follows as indicated in equation V. $
= (Q + F) . FI, . C,
The values of the osmotic flows Q and F can in turn be written explicitly as indicated in equations VI and VII, wherein k is the osmotic membrane permeability coefficient; h is the thickness of the membrane; A,, is the area of the push compartment; rP is the imbibition pressure of the push compartment; A is the total area of the dosage form; and ,‘rD is the imbibition pressure in the drug compartment. VI.
Q =; F =;
A,(H).
(A - A,(H)).
r&H) TD (H)
Equations I, V, VI, and VII have been presented previously as the equations underlying the operation of the push-pull osmotic pump for the delivery of water-soluble compounds, wherein both the drug and the push compartment contain solutes with constant osmotic pressures. However, for the delivery of an insoluble drug, both the drug and the push compartments have been formulated
Volume
83
(suppl 8B)
SYMPOSIUM
with polymers characterized by imbibition pressures that are not constant, but rather are a function of the degree of hydration H, as defined in equation VIII. Here, WH is the weight of the water imbibed per weight of dry polymer W,). VIII.
H = $:
degree
of hydration
P During operation of the pump, the area of the push compartment expands (Figure l), which therefore indicates that A, is also a function of the degree of hydration. By controlling the various parameters defining the total mass delivery rate in equation V, the push-pull system can be programmed for various delivery rate profiles. For nifedipine, the zero-order delivery profile has been selected. This delivery rate can be programmed for the maximum fraction of drug content delivered at a constant rate by selecting the maximum flow (Q) such that the delivery rate from the system can be written by equation IX, where &jc is the density of the drug compartment. dm ~ = Q . ,%-,c. Fb dt This expression implies that the dosage form dispenses the drug from the drug layer at a rate equal to the speed at which the drug is formulated as a suspension by imbibed water. By comparing equation V with equation IX, the condition for the concentration of solids dispensed from the dosage form can then be derived, as expressed by equation X.
c, = ~Q .
Pdc
F+Q
It is interesting to note from equation IX, and from the expression for Q in equation VI, that the zero-order delivery rate must be programmed based on a constant product of the area times the imbibition pressure (~~?rp)in the push compartment. Based on these design parameters, three different forms of nifedipine GITS have been programmed to deliver 30, 60, and 90 mg of the drug over a period of 24 hours at rates of 1.7, 3.4, and 5.1 mgihour, respectively. The systems are therefore labeled as 1.7130, 3.4160, and 5.1190. Some of the parameters describing the structures of each system are listed in Table I. As indicated in the table, each of the systems contain a 10 percent overage of drug since that amount is consistently retained within the dosage form on expiration of the delivery profile. MATERIALS
AND METHODS
The layered tablet of nifedipine GITS contains nifedipine in the drug layer, along with a high molecular weight polyethylene glycol as the osmotic agent and standard tableting excipients. The osmotic layer contains a similar polymeric osmotic agent and additional standard tableting excipients, including
December
21, 1987
TABLE
I
ON CARDIOVASCULAR
Design
Parameters
Systems
THERAPY-SWANSON
for Nifedipine
1.7/30
Diameter (cm) Drug content (mg) Weight of drug layer (m9) Weight of push layer (ms) Membrane thickness (cm) Membrane permeability (cm”/hour atm) Average zero-order delivery rate (mg/hour)
ET AL
GITS 5.1190
3.4/60
0.79 33
1.03
1.19
66
99
165
330
495
82.5
165
247.5
12 x 10-3
11 x
1o-3
9.9 x 10-3
6.35
x IO-’
6.35
6.35
x 10m7
1.7
3.4
x IO-’
5.1
ferric oxide as a colorant. The tablet is formed by standard granulating and tableting techniques. The rate-controlling semipermeable membrane consists of a combination of a cellulosic material and excipients. This membrane is applied directly to the layered tablet core by standard film-coating techniques. An exit pot-t may be drilled through the membrane on the drug-layer side by either a mechanical drill or a high-speed laser drill; the GITS preparation of nifedipine employs the precision laser method. The minimum membrane orifice diameter is large enough to permit the unrestricted movement of drug and excipients out the orifice (approximately 0.4 mm), whereas the maximum diameter is limited to prevent the uncontrolled leaching of the drug layer out the orifice (1 mm). A final colored overcoat is applied to the system for product identification purposes. This overcoat consists of standard film coat and colorant materials and is applied by aqueous film-coating techniques. The overcoat is strictly cosmetic, and has no effect on the release rate of the dosage form. All materials employed in tableting and coating of the system are listed in the United States Pharmacopeia (USP). Conventional USP dissolution testing procedures are impractical for testing the drug release rate from systems delivering insoluble drugs, such as nifedipine, with a solubility of 10 yglml. A dosage form containing 90 mg of such an insoluble drug would require a g-liter volume test reservoir in order to allow sampling of homogeneous solutions throughout the total delivery period. The methods employed for dissolution testing of the nifedipine GITS system were therefore a modified USP test and a well-published differential method, known as the ALZA method [5]. The ALZA release rate tester (ALZA Corporation, Palo Alto, California) is an automated apparatus designed to measure the drug delivery rate from controlled release systems at selected time intervals. This tester incorporates a series of individual reservoirs in a constant temperature bath with a vertically reciprocating shaking mechanism. The shaking mechanism holds the test system in the fluid reservoir and oscillates the system with a vertical stroke of 2.5 cm at a frequency of 0.5 cycles/second. At the end of a predeter-
The
American
Journal
of Medicine
Volume
83
(suppl
66)
5
SYMPOSIUM
ON CARDIOVASCULAR
THERAPY-SWANSON
ET AL
fm.m-
lime (hour) Figure 2. Comparison of in vitro release rate profiles for three doses of nifedipine GUS tablets. l = 1.7130 (n = 10); A = 3.4160 (n = 10); I = 5.1190 (n = 5); I = range.
90 7-
mg system
. ./ .. I
6.
30 mg system
10
20
30
40
Reciprocal membrane weight (g2 Figure 3. Average zero-order release rate as a function of reciprocal membrane weight for three nifedipine GITS tablets (0 = 1.5-kg batch; A = lo-kg batch; n = 130-kg batch).
6
December
21, 1987
The
American
Journal
of Medicine
mined time interval (which is set on an electronic controller), the shaking mechanism lifts the test system out of the reservoir, indexes the set of test systems to the next row of unused reservoirs, lowers them into these reservoirs, and starts the reciprocating stirring again for the preset time interval. This change takes less than five seconds. At the end of the final test interval, the systems are removed from the release media and held suspended over another unused reservoir, available for residual drug content determination. The nifedipine contents of each reservoir are solubilized by the volumetric addition of an appropriate solvent. The resulting solution is then assayed for its nifedipine content. The release rate (mgihour) is determined by dividing the amount of nifedipine in each tube by the length in hours of the test interval; the cumulative amount release is then determined by adding together the amounts from the various intervals. The ALZA method has been compared with the USP method for test systems delivering water-soluble drugs. For this case, it has been shown that test results in the ALZA apparatus are identical to those obtained from the USP apparatus [6]. By modifying the USP techniques, a method has been developed that determines the amount of nifedipine released from GITS by an indirect method. Systems are placed in the USP Apparatus II and stirred for predetermined periods of time. At the end of each test interval, the systems are removed from the flask and assayed for the amount of nifedipine remaining unreleased in each. This value is then subtracted from the average initial content, as determined by a separate assay; this therefore determines the amount of nifedipine released during that test interval. This procedure is used for each test time, and the results are used to construct a composite curve for the entire test. The test is repeated for various stirring speeds to determine their influence on release of nifedipine from GITS. Nifedipine was analyzed by high-pressure liquid chromatographic (HPLC) methods with a 266-nm ultraviolet detector. The size of the column was 15 x 0.46 cm, and it contained a 5-mm Altex Ultrasphere ODS packing (Beckman Instruments). The mobile phase was a 30:20:50 mixture of methanol, acetonitrile, and deionized water, respectively. In some cases, solutions of nifedipine in 50:50 polyethylene glycol 400:deionized water mixture were exposed to intense incandescent light for two to four hours. The resulting degradant product was then analyzed by the ultraviolet detector at 282 nm. The results of this procedure were compared with HPLC results, thereby confirming the validity of the alternate method. To compare in vivo system performance with in vitro results, four medium-sized mongrel dogs (two male and two female), fasted overnight, were used in this study. At the beginning of the study, each dog was fed a 150-g portion of dog food (Kal Kan) one-half hour before the first systems were administered. Each dog was fed one 8-g meatball hourly for the next 12 hours, and all dogs were fasted for the last 18 hours of the study. Individually labeled systems were administered with food to each dog at four, eight, 12, 18,24, and 30 hours prior to the sacrifice of each animal. The dogs were then sacrificed at 30 hours after the start of the study, and the systems were recovered from their gastrointestinal tracts. By
Volume
83 (suppl
66)
SYMPOSIUM
ON CARDIOVASCULAR
THERAPY-SWANSON
ET AL
analyzing the residual drug in these systems, the amount released was calculated as the difference between initial and residual content. At the same times as when the dogs were dosed, equal numbers of systems were placed in the in vitro release apparatus. These systems were removed at the same time that the dogs were sacrificed and assayed for residual drug content using an identical analytic procedure. These results were then compared with those obtained in the in vivo experiment. RESULTS
AND COMMENTS
24.00
During the operation of the nifedipine GITS, small quantities of drug were entrapped along the edges of vertical tablet walls and in the edges of the curved top dome. Drug residuals measured in experimental and clinical systems indicated that about IO percent residual drug remains after in vitro testing (9.8 percent, 30 mg; 9.6 percent, 60 mg; 6.5 percent, 90 mg). These results are in agreement with the results of a study performed on dogs, in which the residual amount of drug remaining in the systems after being dosed in vivo agreed directly with results from systems tested in vitro. Based on these results, an overage equivalent to 10 percent of the labeled dose has been incorporated into the final nifedipine GITS. The release rate of nifedipine GITS 1.7130, 3.4160, and 5.1190 are graphed simultaneously in Figure 2. The top part of the figure depicts the release rate as a function of time wherein the bars indicate the range of experimental data for 10 systems each. The lower part of the figure shows superimposed the cumulative amounts released as a function of time for the three forms with the range of experimental data. It is evident from Figure 2 that the dosage forms are equivalent on a normalized mass basis in vitro; that is, two 30-mg systems can be substituted to give the performance of one 60-mg system, and a 30-mg system plus a 60-mg system is equivalent to a 90-mg system. The osmotic imbibition fluxes F and Q in equations VI and VII are inversely proportional to the membrane thickness. Therefore, the release rates indicated in equations V and IX are also inversely proportional to the membrane thickness. For a given formulation composition, dimensions of the dosage form, and membrane composition, the release rates for each system can thus be programmed by controlling the membrane thickness, which in turn is porportional to the membrane weight for each system. Figure 3 shows the average zero-order release rate for each of the three dosage forms as a function of the reciprocal of the membrane weight. Each of the dosage forms shows a relationship that is linear in accordance with the theoretic prediction. Data are indicated for three different batch sizes as scale-up work progressed. The fact that all the data from the three different batch sizes fit on the same straight line indicates that all systems obey the same theoretic model independent of equipment, and that scale-up from small to larger batch sizes can be carried
December
21, 1987
mrne
-.
(hour)
_ ^
Prgure 4. L’ompanson ot tne cumulabve amount of drug released versus time for nifedipine GITS 3.4160 at different batch sizes (0 = 1.5-kg batch; A = IO-kg batch; n = 130kg batch).
out by controlling the composition of the dosage form and membrane thickness. To amplify the results of the scale-up procedures, the data for three batches of 3.5160 systems are graphed as cumulative amounts in Figure 4. This figure shows the average amount released with the range of experimental data for a total of 10 systems each. It is clear from this figure that on superimposition these three figures are nearly identical, illustrating that not only the average zeroorder release rate is independent of batch size, as already expressed in Figure 3, but also that the total release rate profile is independent of the final batch size. The release rate from osmotic dosage forms can be designed to be independent of the orifice size within a minimum to maximum range of sizes. The minimum orifice must be designed sufficiently large to allow drug delivery from the dosage form at the rate in which the volume of water is imbibed through the system. The maximum size must be designed so that diffusional losses through the orifice are small as compared with the pumping rate. Within this minimum-maximum orifice size range of specified area and diameter, the release rate remains independent of the orifice size. For each of the three dosage forms, the release rate as a function of orifice diameter is depicted in Figure 5. The release rate for the three systems has been found to be independent of orifice diameter over a range ratio of maximum to minimum size of at least a factor of five. This feature allows for a high-speed drilling process with low tolerance requirements. From a biopharmaceutical point of view, perhaps the most important characteristic of an oral dosage form is the release rate profile throughout the gastrointestinal tract. It is this release rate that controls the rate of drug absorption and ultimate therapeutic effect. Drug absorption can thus only be controlled by the dosage form if the release rate in vitro closely correlates with the rate of delivery in vivo. Only under this condition does a valid quality control test
The
American
Journal
of Medicine
Volume
83
(suppl 6B)
7
SYMPOSIUM
ON CARDIOVASCULAR
THERAPY-SWANSON
ET AL
apparatus, the residual content of the dosage form was measured at a particular stirring condition. Each of the data points in Figure 7 indicates the cumulative amount released, which was obtained from the initial system content minus the residual amount within each dosage form after the exposure time indicated. Therefore, cumulative amounts released obtained in this fashion contain a larger experimental error; the error associated with the releasing product plus the error in content uniformity of the dosage form. The cumulative amount released was measured for three different stirring conditions at identical time periods. In order to allow superimposition of the cumulative amounts released, the data points were slightly offset from the experimental time measured in order to be able to show the average standard deviations for each set of systems. The actual time data points are those indicated by the circles for the ALZA method. It is clear from this figure that there is no difference in the amount released for this dosage form between any of the stirring conditions indicated. In order to establish a valid quality control procedure in vitro that would be able to safeguard the release rate in vivo, it is necessary that an in vitro-in vivo correlation be established. The environment in the intestinal tract of the dog is a fair replica of that in humans in both pH and motility such that the release rate in the dog can be equated with the rate in humans. Figure 8 shows the cumulative amount released of nifedipine GITS in the gastrointestinal
I- ____ L-L--- ____ -.-I II
Tlf-f--t----J I----‘-‘--0.2
0.4
0.6
------_ I 0.8
1.0
1.2
1.4
Orifice diameter (mm) igure 5. Average zero-order release rate versus orifice diameter of nifedipine GITS (@ = 1.7130; A = 3.4160; n = 5.1190; I = total standard deviation that is the statistical sum of standard deviations between and within systems).
exist to safeguard the performance and therapeutic response of the dosage form. It is well recognized that pH, between 1 in the stomach to 7.5 in the intestine, and motility in the intestinal tract are the most variable parameters encountered by the dosage form. The release rate in vitro must be independent of these conditions in order to be able to produce a predictable in vivo delivery rate profile. Figure 6 shows the release rate of nifedipine GITS as a function of time in both gastric and intestinal fluids. The top part of the figure shows the superimposition of the release rate profiles in gastric and intestinal fluids as a function of time, while the lower portion of the figure shows the cumulative amounts released as a function of time. It is clear in this figure that the release rate from the systems in either medium are indistinguishable and independent of pH. The release rates from GITS are conventionally measured by the ALZA method [5] because it gives the release rate in a differential fashion and with a higher degree of accuracy than the conventional USP method. In order to measure the release rate as a function of stirring conditions, the cumulative amounts of drug released as a function of time were measured by a modified USP method. It is conventionally assumed that motility patterns in the intestinal tract can be equated with stirring conditions in the range of up to 150 revolutions/minute. The release rate from nifedipine GITS was therefore measured in the modified USP apparatus with the paddle stirring at 50, 100, and 150 revolutions/minute. Since the amount, of drug released from the nifedipine GITS system cannot be dissolved in the total volume of fluid contained in the USP
8
December
21, 1987
The
American
Journal
of Medicine
III
I,,,,,,
Time
$
(hour)
0.0 ’
igure 6. Average zero-order release rate and cumulative amount released for nifedipine GITS (A = artificial gastric fluid, (n = 5); l = artificial intestinal fluid, (n = 5); I = range).
Volume
83
(suppl 68)
SYMPOSIUM
ON CARDIOVASCULAR
THERAPY-SWANSON
ET AL
looQO.g
ao-
1 2 E
70-
[
50-
L9 + E 3
4030-
60-
zoloO12 Time (hour) owl,,
Figure 7. The cumulative percent released for nifedipine GlTS 5.1190 as a function of stirring rates using the modified USP paddle method (A = LISP, 50 revolutionsiminute; q = USP, 100 revolutionslminute; V = LISP, 150 revolutions/minute; l = ALZA release method; I = standard deviation). Data are offset from actual time measured indicated for ALZA method.
tract of dogs as a function of time as compared with the amount released in vitro. This figure shows that the environment, with its variability in pH and motility (such as the intestine of the dog), has no influence on the performance of the dosage form since the amounts released in both the in vitro and in vivo media are identical. The environmental conditions in the intestinal tracts of the dog and humans are substantially the same; therefore, based on these data one can expect the dosage form to behave in the intestinal tract of humans as it performs in vitro. CONCLUSIONS Nifedipine GITS was designed based on the push-pull system configuration in order to deliver drug in a zeroorder controlled predictable fashion as a function of time. Cumulative amounts released in vitro in subdivided forms are additive and interchangeable with the full-strength form.
I
2
4
6
8
I
I
I
I
I
I
10 12 14 16 18 20 22 24 Residence time (hour)
I
I
I
26
28
30
Figure 8. Comparison of in vivo and in vitro cumulative amount released for nifedipine GITS 1.7130 (0 = in vivo release, [n = as indicated]; A = in vitro release, (n = 4); I = standard deviation).
The release rates for nifedipine GITS are controlled by membrane composition and weight, and independent of batch size and scale of equipment. They are independent of the orifice diameter within a size range of a factor of five, and are independent of pH within the physiologic range of 1.2 to 7.5. The release rates of nifedipine GITS are also independent of the stirring rate at or below 150 revolutions/minute, in the physiologic analogous range. In vivo, the release rate has been found to be equal to that in vitro, which qualifies the latter test method as a valid quality control procedure to safeguard performance of the dosage form in vivo. A number of bioavailability studies in humans have been carried out comparing nifedipine GITS with the standard capsule form. Based on the plasma concentrations achieved in these studies, the rate of absorption was found to be equal to the rate of release in vitro at any time for the duration of the delivery period. Publications on that subject are presently in progress.
REFERENCES 1.
2.
3.
Kleinbloesem CH, van Brummelen P, van de Linde JA, Voogd PJ, Breimer DD: Nifedipine: kinetics and dynamics in healthy subjects. In: Nifedipine: clinical pharmacokinetics and haemodynamic effects (thesis). University of Leiden, 1985; 4760. Theeuwes F: Novel drug delivery systems. In: Prescott LF, Nimmo WS, eds. Drug absorption: proceedings of the Edinburgh International Conference, September 1979. Balgowlah, Australia: ADIS Press, 1981; 157-176. Cortese R, Theeuwes F: Osmotic device with hydrogel driving
December
21, 1987
4.
5. 6.
The
member. US. patent no. 4,327,725, issued May 4, 1982. Assigned to ALZA Corporation, Palo Alto, California. Wong P, Barclay B, Deters J, Theeuwes F: Osmotic device with dual thermodynamic activity. US. patent no. 4,612,008, issued September 16, 1986. Assigned to ALZA Corporation, Palo Alto, California. Theeuwes F: The elementary osmotic pump. J Pharm Sci 1975; 64: 1987-1991. Theeuwes F, Swanson D, Wong P, et al: Elementary osmotic pump for indomethacin. J Pharm Sci 1983; 72: 253-257.
American
Journal
of Medicine
Volume
83
(suppl 68)
9