International Journal of Pharmaceutics 321 (2006) 1–11
`
`Historical Perspectives
`A century of dissolution research: From Noyes and Whitney to the
`Biopharmaceutics Classification System
`Aristides Dokoumetzidis a, Panos Macheras b,∗
`
`a School of Pharmacy and Pharmaceutical Sciences, University of Manchester, Manchester M13 9PL, UK
`b Laboratory of Biopharmaceutics and Pharmacokinetics, School of Pharmacy, University of Athens, Athens 15771, Greece
`Received 2 May 2006; received in revised form 6 July 2006; accepted 7 July 2006
`Available online 15 July 2006
`
`Abstract
`
`Dissolution research started to develop about 100 years ago as a field of physical chemistry and since then important progress has been made.
`However, explicit interest in drug related dissolution has grown only since the realisation that dissolution is an important factor of drug bioavailability
`in the 1950s. This review attempts to account the most important developments in the field, from a historical point of view. It is structured in
`a chronological order, from the theoretical foundations of dissolution, developed in the first half of the 20th century, and the development of a
`relationship between dissolution and bioavailability in the 1950s, going to the more recent developments in the framework of the Biopharmaceutics
`Classification System (BCS). Research on relevant fields of pharmaceutical technology, like sustained release formulations, where drug dissolution
`plays an important role, is reviewed. The review concludes with the modern trends on drug dissolution research and their regulatory implications.
`© 2006 Elsevier B.V. All rights reserved.
`
`Keywords: Drug dissolution; Bioavailability; Drug release
`
`Contents
`
`4.
`
`1.
`2.
`3.
`
`Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`1897–1960: The foundations of dissolution research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`1950–1980: The development of a relationship between dissolution and bioavailability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`3.1.
`1970: Initiation of the official dissolution tests. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`3.2. Research on factors affecting the rate of drug dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`1980s: Dissolution becomes an essential tool for the development and evaluation of sustained release formulations . . . . . . . . . . . . . . . . . .
`4.1. Kinetics of drug release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`4.2.
`In vitro in vivo considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`1980–2000: Emphasis on dissolution as a prognostic tool of oral drug absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`5.
`2000–present: Dissolution in the framework of BCS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`6.
`7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`1
`2
`3
`4
`5
`5
`5
`6
`7
`8
`8
`9
`
`1. Introduction
`
`Oral administration of solid formulations has been the major
`route of drug administration for almost a century. However, it
`
`∗
`
`Corresponding author. Tel.: +30 2107274026; fax: +30 2107274027.
`E-mail address: macheras@pharm.uoa.gr (P. Macheras).
`
`0378-5173/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
`doi:10.1016/j.ijpharm.2006.07.011
`
`was only 50 or so years ago that scientists realised the impor-
`tance of dissolution processes in the physiological availability
`of drugs. In the meanwhile, the study of the dissolution process
`has been developing since the end of the 19th century by phys-
`ical chemists. Therefore, most of the fundamental research in
`the field was not related to drugs at all, and the basic laws for
`the description of the dissolution process were already available
`when interest in drug dissolution started to rise.
`
`LUPIN EX. 1014
`Lupin v. iCeutica
`US Patent No. 8,999,387
`
`Page 1
`
`

`
`2
`
`A. Dokoumetzidis, P. Macheras / International Journal of Pharmaceutics 321 (2006) 1–11
`
`This review attempts to describe the historical evolution of
`drug dissolution. It places particular emphasis on the fundamen-
`tal articles in the field, which shaped the major lines of research
`and regulation policy of the regulatory agencies. Also, paral-
`lel research contributions with significant impact on dissolution
`research are quoted. The present review is structured in chrono-
`logical order, starting from the first dissolution experiment and
`the development of the major models for dissolution of solids,
`moving on to the realization of a relationship between dissolu-
`tion and bioavailability, which initiated the drug related interest
`in dissolution, and progressing to the present applications of dis-
`solution studies, with both their scientific and regulatory aspects.
`
`2. 1897–1960: The foundations of dissolution research
`
`In 1897, Noyes and Whitney conducted the first dissolu-
`tion experiments and published an article entitled “the rate of
`solution of solid substances in their own solutions” (Noyes and
`Whitney, 1897). Arthur A. Noyes [1866–1936], was a Profes-
`sor of Chemistry at MIT and also served as a president of MIT
`from 1907 to 1909, later moving to Caltech. Together with Willis
`R. Whitney, they studied the dissolution of two sparingly solu-
`ble compounds, benzoic acid and lead chloride. The materials
`were laid around glass cylinders which were submerged into
`vessels containing water. The cylinders were rotated at constant
`speed and under constant temperature. The authors noticed that
`the rate of dissolution is proportional to the difference between
`the instantaneous concentration, C at time t, and the saturation
`solubility, CS, (Fig. 1). This statement can be formulated math-
`ematically as follows:
`= k(CS − C)
`
`(1)
`
`dC
`dt
`
`Fig. 1. Three extracts from the original article of Noyes and Whitney (1897)
`showing the title, the main equation and the concluding statement of the article.
`Reprinted with permission.
`
`Fig. 2. Concentration–time plots of (Noyes and Whitney, 1897) data together
`with plots of Eq. (1) using the original estimates for the values of the constants.
`The data correspond to stick no. 1 for benzoic acid and stick no. 2 for lead
`chloride.
`
`where k is a constant. The experiment configuration ensured that
`the surface of the materials was kept constant during dissolution
`as the materials were in excess of the amount needed to saturate
`the medium. In Fig. 2 plots of these data together with plots of Eq.
`(1) using the original estimates for the values of the constants, are
`shown. The authors attributed the mechanism of dissolution to
`a thin diffusion layer which is formed around the solid surface
`and through which the molecules diffuse to the bulk aqueous
`phase.
`The next development came from Erich Brunner, and Stanis-
`laus von Tolloczko at Gottingen, who published an article in
`1900 based on a series of experiments that extended the condi-
`tions under which Eq. (1) holds and also showed that the rate of
`dissolution depends on the exposed surface, the rate of stirring,
`temperature, structure of the surface and the arrangement of the
`apparatus (Bruner and Tolloczko, 1900). The proposed model
`was derived from Eq. (1) by letting k = k1S:
`= k1S(CS − C)
`
`(2)
`
`dC
`dt
`where S is the surface area. Also, Brunner in 1904 published a
`paper based on the work done in his Ph.D. that studied the prob-
`lem further, trying to find specific relations between the constants
`involved (Brunner, 1904). This work was published together
`with the theoretical work of Walther Nernst [1864–1941], who
`was Professor of Physical Chemistry and the founder and direc-
`tor of the Institute for Physical Chemistry and Electrochem-
`istry at Gottingen where Brunner was working (Nernst, 1904).
`Walther Nernst was one of the major contributors in the field
`of physical chemistry, and received a Nobel Prize in 1920 “in
`recognition of his work in thermochemistry”. The main result of
`this two-part publication of Nernst and Brunner in 1904, which
`was based on the diffusion layer concept and Fick’s second law
`was what is known as the Nernst–Brunner equation, which was
`derived from Eq. (2) by letting k1 = D/(Vh):
`= DS
`(CS − C)
`Vh
`
`dC
`dt
`
`(3)
`
`Page 2
`
`

`
`A. Dokoumetzidis, P. Macheras / International Journal of Pharmaceutics 321 (2006) 1–11
`
`3
`
`where D is the diffusion coefficient, h the thickness of the dif-
`fusion layer and V is the volume of the dissolution medium.
`In 1931 Hixson and Crowell expressed the surface, S of Eq.
`(2) in respect to the weight, w, by letting S to be proportional to
`w2/3, which makes the Eq. (2) applicable to dissolving compact
`objects (Hixson and Crowell, 1931). By this consideration, Eq.
`(2), when integrated yields an equation which relates time to the
`cubic-root of weight and in the special case of sink conditions,
`where small concentrations are considered and the difference
`(Cs − C) can be considered as constant, the cubic-root law takes
`a simple form:
`− w1/3 = k2t
`
`(4)
`
`w1/3
`0
`
`where w0 is the initial weight and k2 a constant. In their paper
`Hixson and Crowell reported that the Noyes–Whitney equation
`in its original form and without any details about the mechanism
`of the process had been sufficiently validated with a wide range
`of experiments, as opposed to the various mechanistic explana-
`tions that had appeared, none of which was entirely satisfactory.
`The above approaches can be categorized as various expres-
`sions of the diffusion layer model as a physical explanation for
`dissolution process, where the limiting step has been consid-
`ered to be the diffusion of molecules through a stagnant film of
`liquid around the solid surface. By the 1950s two more alterna-
`tive explanations were available as reviewed by Higuchi (1961).
`The interfacial barrier model, considered that interfacial trans-
`port, rather than diffusion through the film, is the limiting step
`due to a high activation energy level for the former. This model
`was first proposed by Wilderman (1909) and was also consid-
`ered by Zdanovskii (1946), but has not been studied in detail and
`an explicit mathematical description for the dissolution kinetics
`is not available, while variations have also appeared (Miyamoto,
`1933). The third model for dissolution is Danckwerts’ model,
`which appeared in 1951 (Danckwerts, 1951). According to this,
`constantly renewed macroscopic packets of solvent reach the
`solid surface and absorb molecules of solute, delivering them to
`the solution. Combinations of these models were also consid-
`ered. The work of Levich improved the theoretical model of the
`dissolution experiment using rotating disks, taking into account
`the centrifugal force on diffusion (Levich, 1962).
`Despite the advances in in vitro dissolution in chemical engi-
`neering sciences, in the pharmaceutical sciences the concept was
`not used extensively until the early 1950s. Until then the in vivo
`availability of the drug was thought to be determined solely by
`the disintegration of the tablet, ignoring the dissolution process.
`Many in vitro procedures to determine the disintegration time
`of tablets were suggested, at the time, and some of them were
`reviewed by Morrison and Campbell (1965). The first official
`disintegration test for tablets was published in the Pharmacopeia
`◦
`Helvetica in 1934, which used water at 37
`C as the medium and
`periodical shaking, while in the United States Pharmacopeia the
`disintegration test was introduced in the 14th edition in 1950.
`Other methods, developed later, tried to introduce more realistic
`conditions, using, for example, simulated gastric fluids as media
`for the disintegration experiments. One of the most sophisti-
`cated was Filleborn’s method which was published in 1948 and
`
`introduced an artificial stomach with simulated in vivo condi-
`tions, including pH level, peristalsis and the presence of food
`(Filleborn, 1948). In the early 1950s it became clear that disinte-
`gration alone could not account for the physiological availability
`of drugs and in many cases the dissolution rate was, instead, the
`limiting step.
`
`3. 1950–1980: The development of a relationship
`between dissolution and bioavailability
`
`To the best of authors’ knowledge, Edwards in 1951 was the
`first to appreciate that following the oral administration of solid
`dosage forms, if the absorption process of drug from the gastroin-
`testinal tract is rapid, then the rate of dissolution of that drug can
`be the step which controls its appearance in the body. In fact, he
`postulated that the dissolution of an aspirin tablet in the stomach
`and intestine would be the rate process controlling the absorp-
`tion of aspirin into the blood stream (Edwards, 1951). However,
`Nelson in 1957 was the first to explicitly relate the blood levels
`of orally administered theophylline salts to their in vitro disso-
`lution rates (Nelson, 1957). He used a non-disintegrating drug
`pellet, (mounted on a glass side so that only the upper face was
`exposed), placed at the bottom of a 600 mL beaker in such a
`manner that it could not rotate when the dissolution medium
`was stirred at 500 rpm.
`In mid 1960s to early 1970s a number of studies demonstrat-
`ing the effect of dissolution on the bioavailability of a variety
`of drugs were reported in the literature. Two reports were pub-
`lished in 1963 and 1964 drawing attention to the lack of full
`clinical effect for two brands of tolbutamide marketed in Canada
`(Campagna et al., 1963; Levy et al., 1964). These tablets were
`shown to have long disintegration times as well as slow dis-
`solution characteristics (Levy, 1964). Besides, a slight change
`in formulation of an experimental tolbutamide preparation was
`shown to produce significantly lower blood levels and hypo-
`glycemic effect (Varley, 1968). In 1968, Martin et al. (1968)
`reported significant differences in the bioavailability between
`different brands of sodium diphenylhydantoin, chlorampheni-
`col and sulfisoxazole. MacLeod et al. (1972) reported greater
`than 20% difference in peak concentration and area under the
`serum concentration–time curve for three ampicillin products.
`In late sixties it was realized that differences in product
`formulation could lead to large differences in speed of onset,
`intensity and duration of drug response. At that time the term
`“bioavailability” was coined to describe either the extent to
`which a particular drug is utilized pharmacologically or, more
`strictly, the fraction of dose reaching the general circulation. The
`most dramatic bioavailability examples have been with digoxin
`in the U.K. and the USA in 1971 and phenytoin in Australia and
`New Zealand in 1968.
`In the former case, different formulations of digoxin yielded
`up to sevenfold differences in serum digoxin levels (Lindenbaum
`et al., 1971). These observations prompted the FDA in collabora-
`tion with the late John Wagner to carry detailed dissolution stud-
`ies on 44 lots from 32 manufacturers of 0.25 mg digoxin tablets
`available in the 1972 North American market-place (Skelly,
`1988). The studies revealed tremendous differences in the dis-
`
`Page 3
`
`

`
`4
`
`A. Dokoumetzidis, P. Macheras / International Journal of Pharmaceutics 321 (2006) 1–11
`
`Fig. 3. Dissolution profiles of three different formulations of digoxin, exhibiting
`large differences, reprinted from (Fraser et al., 1972) with permission.
`
`solution profiles of the digoxin products and substantiated the
`view that either lot-to-lot or amongst brands bioinequivalence
`originates from differences in dissolution rates. Additional dis-
`solution studies conducted in other laboratories confirmed these
`findings (Fraser et al., 1972). In Fig. 3 dissolution profiles of
`different formulations of digoxin are shown from (Fraser et al.,
`1972) exhibiting large differences.
`Phenytoin toxicity occurred in a large number of patients
`when the manufacturer replaced the excipient calcium sulfate
`with lactose in immediate release phenytoin tablets (Tyrer et al.,
`1970). Initially, the lower extent of absorption of phenytoin in
`the presence of calcium sulfate was ascribed to the formation of
`an insoluble calcium-phenytoin salt, Bochner et al. (1972). How-
`ever, Chapron et al. (1979) found no effect when they studied
`the influence of calcium on bioavailability of phenytoin admin-
`istering calcium gluconate before, with and after a single dose
`of 300 mg of phenytoin. These results indicated that the higher
`hydrophilicity of lactose compared to calcium sulfate, promoted
`the dissolution rate of phenytoin resulting in higher bioavail-
`ability and consequently higher concentrations of phenytoin in
`plasma, exceeding its narrow therapeutic range of 10–20 ␮g/mL.
`The results of this study are shown in Fig. 4. A decade later,
`loss of seizure control occurred in a patient on phenytoin was
`related to altered dissolution characteristics caused by the phys-
`ical changes of phenytoin capsules (Cloyd et al., 1980).
`
`3.1. 1970: Initiation of the official dissolution tests
`
`All of the above bioavailability concerns prompted the intro-
`duction of dissolution requirements in tablet and capsule mono-
`graphs in pharmacopeias. Of equal significance was the recog-
`nition of the immense value of dissolution testing as a tool for
`quality control. Thus, equivalence in dissolution behaviour was
`sought in light of both the bioavailability and quality control
`considerations throughout the last 35 years.
`As mentioned above a number of studies mainly in the USA
`during the 20-year period 1950–1970 shed light on the impor-
`tance of pharmaceutical ingredients and processes in regard to
`the dissolution–bioavailability relationship. As a result of these
`developments, the basket-stirred-flask test (USP apparatus 1)
`
`Fig. 4. Plot of blood phenytoin concentrations, reprinted with permission from
`(Tyrer et al., 1970), including the original legend.
`
`was adopted as an official dissolution test in 6 monographs of
`the United States Pharmacopeia (USP) and National Formulary
`(NF) in 1970. Due to the continuous intense interest in the sub-
`jects of dissolution and gastrointestinal absorption, an explosion
`in the number of monographs of the dissolution requirements
`in subsequent USP/NF editions was noted (Table 1). Remark-
`able events during this evolution are the adoption of the paddle
`method (USP apparatus 2) in 1978, the publication of a gen-
`eral chapter on Drug Release in USP 21 (1985), the presence
`of 23 monographs for modified-release dosage forms in USP
`22-NF 18 (1990), the adoption of the reciprocating cylinder
`(USP apparatus 3) for extended-release products in 1991 and
`the adoption of the flow-through cell in (USP apparatus 4) for
`extended-release products in 1995.
`It should also be noted that the first guidelines for dissolution
`testing of solid dosage forms were published in 1981 as a joint
`report of the Section for Official Laboratories and Medicines
`
`Table 1
`Number of monographs in the US Pharmacopeia and the National Formulary
`which require dissolution or release tests
`
`Edition/year
`
`Monographs for
`immediate-release
`dosage forms
`
`Monographs for
`modified-release
`dosage forms
`Extended
`
`USP 18-NF 13/1970
`USP 19-NF 14/1975
`USP 20-NF 15/1980
`USP 21-NF 16/1985
`USP 22-NF17/1990
`USP 23-NF18/1995
`USP 24-NF19/2000
`USP 29-NF24/2006
`
`6
`12
`60
`400
`462
`501
`552
`619
`
`–
`–
`–
`1
`18
`6
`26
`38
`
`Delayed
`
`–
`–
`–
`–
`5
`25
`14
`14
`
`Page 4
`
`

`
`A. Dokoumetzidis, P. Macheras / International Journal of Pharmaceutics 321 (2006) 1–11
`
`5
`
`Control Services and the Section of Industrial Pharmacists of
`the FIP (FIP, 1981).
`
`3.2. Research on factors affecting the rate of drug
`dissolution
`
`During the early stages of drug dissolution research
`(1950–1960) and in particular after dissolution was established
`to be an important factor in the bioavailability of certain drugs,
`the detailed study of factors affecting the dissolution rate were
`studied extensively.
`The degree of agitation is one of the important factors deter-
`mining dissolution. Generally, higher stirring rates result in
`higher dissolution rates. This was studied quantitatively as well
`and several publications appeared, that gave experimental evi-
`dence of a power law relationship between dissolution rate and
`stirring rate (Wurster and Taylor, 1965). Under certain condi-
`tions this power-law collapsed to an almost linear relationship.
`Dissolution rate depends also directly on solubility, as the
`Noyes–Whitney equation (Eq. (1)) suggests. This became of
`particular importance as the influence of solubility on bioavail-
`ability was considered to come primarily from its influence on
`dissolution rather than saturation of GI fluids. This is so, because
`sink conditions were considered to prevail inside the intestines,
`at least for highly permeable drugs (Wurster and Polli, 1961;
`Gibaldi and Feldman, 1967). It was also realized that solubil-
`ity can be affected by the presence of solubilizing agents in the
`dissolution medium either by partitioning of the drug into the
`micelles of a surfactant or complexation of the drug with one
`or more substances. The seminal articles of Bates et al. (1966)
`on griseofulvin dissolution and Tao et al. (1974) on cholesterol
`dissolution in bile salt solutions can be considered as the ini-
`tiatory studies on drug dissolution in micellar solutions. Also,
`in 1968 the publication of the book “solubilization by surface-
`active agents and its applications in chemistry and the biological
`sciences” marked the new very rapidly growing field (Elworthy
`et al., 1968). A method called “solid dispersion formulation”
`was also developed in order to enhance the dissolution rate
`of sparingly soluble compounds. The drug is dispersed in an
`inert hydrophilic carrier, which promotes the dissolution of drug
`through its high wettability. Dispersion of chloramphenicol in
`urea is one of the first classic examples (Chiou, 1971).
`Another factor that influences the dissolution rate is the sur-
`face exposed in the solvent. This is primarily affected by the
`particle size, meaning the smaller the particles, and therefore in
`greater number, the higher their total exposed surface compared
`to larger but fewer particles of the same total mass. The effect
`is especially dramatic with poorly soluble compounds as, for
`example, digoxin which showed 100% increase in bioavailabil-
`ity when its particle size was reduced from 100 ␮m to approxi-
`mately 10 ␮m (Jounela et al., 1975). Studies on the effect of par-
`ticle size were reviewed by Levy (1963). However, the relation-
`ship of particle size–surface area–dissolution rate is not always
`straightforward. Finholt (1974) clearly demonstrated that if the
`drug is hydrophobic and the dissolution medium has poor wet-
`ting properties, reduction of particle size may lead to a smaller
`effective surface area and a slower dissolution rate. Finholt
`
`(1974) reported that when granules containing phenacetin in dif-
`ferent particle sizes were prepared using gelatine as a hydrophilic
`diluent their dissolution rate was found to increase as the particle
`size was progressively decreased. On the contrary, when simple
`phenacetin particles were tested for their dissolution in 0.1N
`HCl, the dissolution rate increased as the particle size increased.
`The situation was altered returning to normality, when a surface
`active agent Tween 80 was added to the dissolution medium.
`The anomalous behaviour was attributed to the better wetting
`of larger particles in comparison to the smaller particles, which
`floating on the medium exposed a smaller surface area to the
`medium. The addition of surface active agent restored the normal
`situation by improving the wetting of particles. Similar results
`were obtained with phenobarbital and aspirin (Finholt, 1974).
`During this period an important contribution to the math-
`ematical modelling of dissolution curves was published by
`Langenbucher (1972). He observed that if one plots the quantity
`−ln(1− m) versus time on a log–log plot, where m is the accu-
`mulated fraction of dissolved material, the curve looks linear,
`and one can then perform linear regression. This is equivalent
`(cid:1)
`(cid:2)
`to fitting a Weibull equation to the dissolution data:
`−(t − T )b
`m = 1 − exp
`
`(5)
`
`a
`
`where t is time, T a lag time, a a scale constant and b is a shape
`constant.
`
`4. 1980s: Dissolution becomes an essential tool for the
`development and evaluation of sustained release
`formulations
`
`The first mention of a constant release oral medication is
`quoted in a British patent almost 70 years ago (Lipowski, 1934).
`In 1952, Smith Kline and French introduced the first time-
`released medicine, Dexedrine (dextroamphetamine sulfate). It
`was marketed and used in a Spansule—a novel form of drug
`delivery (Blythe et al., 1959). Since then the term sustained
`release is in common usage to describe orally administered
`products that modulate the time course of drug concentration
`in the body by releasing the drug over extended time periods.
`The selection of a drug candidate for the design of a sustained
`release system depends on various criteria such as short bio-
`logical half-life (t1/2), narrow therapeutic index, efficient GI
`absorption, small daily dose and marketing benefits. Theeuwes
`and Bayne were the first to derive in 1977 a relationship between
`t1/2, the optimum therapeutic range blood level, Cmax − Cmin,
`and the dosing interval, T, assuming a one-compartment model
`with repetitive intravenous injections at pseudo-steady state
`(Theeuwes and Bayne, 1977):
`T ≤ 1.44 · t1/2 ln
`Cmax
`Cmin
`
`(6)
`
`4.1. Kinetics of drug release
`
`Since late 1970s the development of sustained release deliv-
`ery systems evolved rapidly. The basic performance requirement
`
`Page 5
`
`

`
`6
`
`A. Dokoumetzidis, P. Macheras / International Journal of Pharmaceutics 321 (2006) 1–11
`
`of these systems is that they release drug in vivo according to a
`predictable rate. The kinetics of drug release follows the opera-
`tive release mechanism of the system, e.g., diffusion through
`inert matrix, diffusion across membrane or hydrophilic gel,
`osmosis, ion-exchange, etc. By far, diffusion is the principal
`release mechanism, since apart from the diffusion-controlled
`systems, diffusion takes place at varying degrees in both chem-
`ically and swelling-controlled systems.
`Solute release models preceded the development of drug
`delivery systems by many years. In fact, the mathematical mod-
`elling of drug release from diffusion-controlled systems relies
`on the Higuchi model published in 1961 (Higuchi, 1961). He
`analyzed the kinetics of release from an ointment assuming that
`the drug is homogeneously dispersed in the planar matrix and
`the medium into which it is released acts as a perfect sink under
`pseudo steady-state conditions. Higuchi derived Eq. (7) for the
`cumulative amount q(t) of drug released at time t:
`√
`= K
`t
`
`(7)
`
`q(t)
`q∞
`where q∞ is the cumulative amount of drug released at infinite
`−1/2
`time and K is a composite constant with dimension time
`related to drug diffusional matrix as well as the design charac-
`teristics of the system. Due to the approximate nature of Eq. (7),
`its use for the analysis of release data is recommended only for
`the first 60% of the release curve (q(t)/q∞)≤ 0.60).
`In late 1960s, Wang et al. published an article which can be
`considered as the initiator of the realization that two apparently
`independent mechanisms of transport, a Fickian diffusion and
`a case II transport, contribute in most cases to the overall drug
`release (Wang et al., 1969). The former mechanism is governed
`by Fick’s law, while the latter reflects the influence of polymer
`relaxation on the molecules’ movement in the matrix (Enscore et
`al., 1977). Some years later, Fu et al. (1976) used a mechanistic
`model to study the release of a drug homogeneously distributed
`in a cylinder. In reality, Fu et al. solved Fick’s second law equa-
`tion assuming constant cylindrical geometry and no interaction
`between drug molecules.
`In 1985, a date which marks the initial rapid phase of growth
`of delivery systems, Peppas (1985) introduced a semi-empirical
`equation (the so-called power law) to describe drug release from
`polymeric devices in a generalized way:
`= K1tn
`
`(8)
`
`q(t)
`q∞
`where K1 is a constant reflecting the structural and geometric
`−n units
`characteristics of the delivery system expressed in time
`and n is a release exponent the value of which is related to the
`underlying mechanism(s) of drug release (Ritger and Peppas,
`1987). Again, valid estimates for K1 and n can be derived
`from the fitting of Eq. (8) to the first 60% of the experimen-
`tal release data. Detailed discussions of the assumptions of the
`derivations of Eqs. (7) and (8) in relation to their valid appli-
`cations to real data can be found in literature (Siepmann and
`Peppas, 2001; Macheras and Iliadis, 2006). Since Eqs. (7) and
`(8) enjoy a wide applicability in the analysis of drug release
`studies, caution should be exercised for their proper use in rela-
`
`tion to the elucidation of the release mechanisms (Rinaki et al.,
`2003b).
`Through the years a plethora of mechanistic release models
`have been published in literature (Siepmann and Peppas, 2001;
`Macheras and Iliadis, 2006). Although the mechanistic models
`are more physically realistic, their mathematical complexity is
`their main disadvantage for wide use. In recent years, Monte
`Carlo simulations following the pioneering work of Bunde et
`al. (1985) were used to study drug release from Euclidean
`(Siepmann et al., 2002, 2004; Kosmidis et al., 2003b) or frac-
`tal spaces (Kosmidis et al., 2003a). The work of (Kosmidis et
`al., 2003a,b) demonstrated that the Weibull function (Eq. (5)),
`is the most powerful tool for the description of release kinetics
`in either Euclidean or fractal spaces. Based on these findings, a
`methodology was developed (Papadopoulou et al., 2006) for the
`elucidation of release mechanisms using the entire set of data
`and the estimate for the exponent b of time.
`
`4.2. In vitro in vivo considerations
`
`The major objective in the design of an oral controlled release
`formulation is to achieve little or no effect of the GI environment
`upon the rate of drug release. This is a rather difficult goal since
`the formulation traverses a varying milieu: from a pH close to
`1 in the fasted stomach through the duodenum (pHs 4–5) and a
`gradually increasing intestinal pH reaching the alkaline region in
`the distal section of the intestinal tract. In parallel, these formula-
`tions can be dosed either in presence or absence of food and the
`dramatic physiological changes, e.g., pH, bile and pancreatic
`secretions can influence the rate of drug release. Overall, this
`complex-heterogeneous GI environment has a greater impact
`on drug dissolution for controlled release formulations than that
`observed with conventional preparations. Based on this realiza-
`tion a separate general chapter, Drug Release (cid:6)724(cid:7) was adopted
`in the USP 21-NF 16 as early as 1985 providing methodology
`and acceptance criteria for extended-release and delayed-release
`products (see Table 1).
`Dilantin®, an extended-release product of Parke Davis was
`the first to have an approved dissolution specification attached
`to it as a condition of lot-to-lot approval by the FDA. Shah et
`al. (1983) proposed a dissolution window over time to distin-
`guish the two types of Dlantin® formulations (100 and 300 mg)
`and ensure lot-to-lot bioequivalence. During the same time, two
`quinidine gluconate formulations, Quinaglute Duratabs® (Inno-
`vator brand, Berlex) and an unapproved and marketed prod-
`uct were found to have quite similar dissolution characteristics
`despite of the fact that they were bio-inequivalent (Prasad et al.,
`1982). The similarity of dissolution profiles was justified in 0.1N
`HCl a