Yu LX, Ellison CD, Conner DP, Lesko LJ and Hussain AS Influence of Drug Release Properties of Conventional Solid Dosage Forms on the Systemic Exposure of Highly Soluble Drugs AAPS PharmSci 2001;
3
(3)
article 24
(https://www.pharmsci.org/scientificjournals/pharmsci/journal/01_24.html).
Influence of Drug Release Properties of Conventional Solid Dosage Forms on the Systemic Exposure of Highly Soluble Drugs
Submitted: July 6, 2001; Accepted: August 21, 2001; Published: September 8, 2001
Lawrence X. Yu1, Christopher D. Ellison1, Dale P. Conner1, Larry J. Lesko1 and Ajaz S. Hussain1
1US Food and Drug Administration, Office of Pharmaceutical Sciences, 5600 Fishers Lane, Rockville, MD 20857
|
Correspondence to:
Lawrence X. Yu
Telephone: 301-827-5246
Facsimile: 301-594-6289
E-mail: yul@cder.fda.gov
|
Keywords:
Small intestinal transit
dissolution
disintegration
absorption modeling
bioequivalence
|
Abstract
This study was designed to theoretically investigate the influence of drug release properties, characterized by the disintegration of a solid dosage form and dissolution of drug particles, on the systemic exposure of highly soluble drugs in immediate release products. An absorption model was developed by considering disintegration of a solid dosage form, dissolution of drug particles, gastrointestinal transit flow, and intestinal absorption processes. The absorption model was linked to a conventional pharmacokinetic model to evaluate the effect of disintegration and dissolution on the peak exposure (Cmax ) and total exposure of area under the curve (AUC). Numerical methods were used to solve the model equations. The simulations show that the effect of disintegration of a dosage form and dissolution of drug particles depend on the permeability of a drug, with a low-permeability drug having a greater effect. To provide similar exposure to an oral solution formulation, a solid dosage form containing a low-permeability drug would need to dissolve more rapidly than a solid dosage form containing a high-permeability drug. It was shown theoretically for poorly permeable drugs that the disintegration rate constant has to be greater than 9 hour -1 (equivalent to approximately 90% in 30 minutes) to make both AUC and Cmax ratios higher than .9, ensuring the confidence interval of .80 to 1.25. The rapid in vitro release requirement of at least 85% dissolved in 30 minutes is sufficient for highly soluble and highly permeable drugs. However, for highly soluble and poorly permeable drugs, the appropriate in vitro release requirement seems to be 90% dissolved in 30 minutes.
Introduction
The US Food and Drug Administration issued a guidance on waiver of in vivo bioavailability and bioequivalence studies for immediate-release, solid, oral dosage forms in August 20001 . The guidance is based on the proposed Biopharmaceutics Classification System (BCS)2 and recommends that sponsors apply for biowaivers for highly soluble and highly permeable drug substances (BCS Class I) in immediate-release, solid, oral dosage forms that exhibit rapid in vitro release . It has been suggested recently that the waiver of in vivo bioequivalence studies should be extended to highly soluble and poorly permeable drugs (BCS Class III)3 . It is unknown, however, what kind of in vitro drug release requirements should be set to ensure that the drug release has no significant effect on in vivo bioavailability.
Despite recent advances in absorption modeling and simulation4-8 , the relationship between in vivo absorption processes and in vitro drug product release has not been well defined; current absorption models account for the dissolution of drug substance, but not for drug product release. Furthermore, drug product release modeling is based mainly on empirical or semi-empirical models9 . Potential drug excipient interactions certainly add to the complexity of drug release modeling10 .
This report aims to establish relationships between absorption processes and drug release properties of a formulation. It investigates how drug substance dissolution and drug product disintegration affect the peak exposure (Cmax ) and total exposure of the area under the curve (AUC) of highly soluble drugs in immediate-release dosage forms using a solution dosage form as a reference. To this end, an in vitro drug release model of a solid dosage form was proposed and an in vivo absorption model was developed. Simulations were conducted to establish appropriate in vitro drug release requirements to ensure in vivo drug release would not cause bioinequivalence for highly soluble and poorly permeable drugs.
Materials and Methods
THEORETICAL
In Vitro Drug Release Model
As shown in Figure 1 , the drug release property of a solid dosage form may be characterized by 2 subprocesses: the liberation of drug particles from a dosage form and the dissolution of drug from the liberated drug particles. It is assumed that the dissolution of drug from the surface of the intact dosage form is negligible. Mathematically, these 2 subprocesses may be expressed as
 ....................(1)
 ....................(2)
 ....................(3)
where Mf , Md , and Ml are the amount of drug remaining in the formulation, in the particles, and in the liquid; Kf and Kd are the drug product disintegration and drug substance dissolution rate constants; V is the volume of dissolution medium; M0 is the dose; and Cs and Cl are the solubility and drug concentration in liquid (Ml = Cl x V).
Equation 1 assumes that the drug particle liberation from the dosage form is a first-order process, as suggested in the literature11 . This is a simplification of a complex process. The factors influencing the liberation of drug particles from dosage forms include formulation and processing factors, such as diluent, disintegrant, binder and granulating agent, lubricant, method of granulation, and compression force. For convenience, we use the term "disintegration" to describe the whole process and we assume that the formulation and process variables are reflected in the disintegration rate constant, Kf . Clearly, the disintegration described here is not equivalent to the disintegration discussed in the in vitro drug release testing.
Equations 2 and 3 are the same as the model proposed in the literature5 . The dissolution rate constant Kd in equations 2 and 3 can be calculated by
 ....................(4)
where D is the diffusion coefficient, M0 is the dose, ρ is the density of the drug, h is the aqueous diffusion layer thickness, and r0 is the initial radius of particles. The model takes no account of particle size distribution, the change of aqueous boundary thickness, and the changes of the surface area per unit weight. However, the dissolution model was able to predict the dissolution of digoxin, griseofulvin, and panadiplon reasonably well5 .
In Vivo Absorption Model
The in vivo absorption model used in this study was developed based on our previously published model5, which not only considers intestinal absorption and dissolution from drug particles, but also drug liberation from dosage forms (ie, disintegration). It is assumed that the model equations of drug release in vivo are similar to those in vitro except that the drug release volume may be different. The assumptions with the model include the following:
1. Absorption from the stomach and colon is insignificant compared with that from the small intestine. The transport across the small intestinal membrane is passive, and the amount of drug transported is equal to its uptake.
2. Liquid and solid drug moving through the small intestine can be viewed as a moving process flowing through a series of segments, each described by a single compartment with linear transfer kinetics from one compartment to the next, and all compartments having different volumes and flow rates but the same residence times8 .
3. Drug product or formulation excipients have no significant impact on gastrointestinal motility and intestinal permeability; disintegration, dissolution, and absorption rate constants are site-independent.
Therefore, incorporating the disintegration process into our previous in vivo model equations4,5 results in the following model equations:
 ....................(5)
 ....................(6)
 ....................(7)
where n = 1, 2, 3, ..., 7, Kt is the transit rate constant12 , and Ka is the absorption rate constant. The overall rate of drug absorption can be calculated by
 ....................(8)
where Ma is the amount of drug absorbed at time t.
In Vivo Pharmacokinetic Model
The rate of drug absorption (equation 8) can be linked to any conventional pharmacokinetic model. Assuming a 1-compartment model with first-order elimination, we have
 ....................(9)
where C is the plasma concentration, V1 is the volume of distribution, and ke is the first-order elimination constant. From equation 9, the AUC and Cmax can then be estimated.
METHODS
Solution of Model Equations
Model equations 1-9 are a typical initial value problem of an ordinary differential equation system. This system was numerically solved by the ADAPT II pharmacokinetic and pharmacodynamic modeling package to estimate the peak exposure and total exposure13 . A subroutine was written to accommodate the model equations.
Model Parameters
Table 1 shows the values of model parameters used in the simulation. The dose of the model drug is assumed to be 100 mg. Based on the dose, the BCS boundary value of solubility that determines the solubility membership of a compound was calculated to be .4 mg/mL. The fraction of dose absorbed of 90% corresponds to an absorption rate constant of .82 h-1 based on prediction model equation 3. The elimination rate was assumed to be .23 h-1 , which corresponds to a half-life of 3 hours. The volume of distribution was assumed to be 100 L. The volumes of the gastrointestinal tract compartments were taken from the literature14 .
In vivo absorption was simulated by considering high (4.1 h-1 ) and low (0.082 h-1 ) absorption rate constants, which correspond to high and low permeabilities of 1.0 x 10-3 and 2.0 x 10-5 cm/s based on Ka = 2Peff /R, where R is the small intestinal radius 1.75 cm. The permeability of 2.0 x 10-5 cm/s corresponds to the fraction of dose absorbed of .23, implying the maximum bioavailability of 23%.
In Vitro Drug Release and In Vivo Absorption
Simulations of in vitro drug release were conducted for highly soluble drugs. The simulated in vitro drug release was then compared with the in vivo absorption to estimate the in vitro drug release rate necessary to provide similar exposure to a solution dosage form. The in vitro drug release was characterized by the percent dissolved at 30 minutes. In vivo absorption was described by the peak and total exposure ratios of a solid dosage form relative to an oral solution to identify where a solid dosage form would perform like a solution in vivo.
The effect of gastric emptying was not considered, and the drug product was assumed to be administered directly into the duodenum. The stomach acts like a regulator, and gastric emptying of dosage forms varies significantly15 . In general, the longer the gastric emptying time, the less the effect of disintegration and dissolution on absorption, provided that the test and reference formulations are emptied at the same rate. Thus, ignoring the effect of gastric emptying represents the worst-case scenario and provides a conservative estimate of in vitro release rate.
Results
In Vitro Drug Release
Figure 2 (A) shows a drug release profile of a solid dosage form containing a drug with the solubility of .4 mg/mL in the dissolution volume of 900 mL, commonly used in USP dissolution apparatus I and II. Figure 2 (A) also shows the amount of drug particles in the dissolution medium and the amount of solid drug remaining in the tablet. For a particular product, we can vary the disintegration and dissolution constants to achieve the desired dissolution profiles. Therefore, the drug release model covers both drug substance and drug product attributes.
Figure 2 (B) shows the effect of disintegration rate constant on the in vitro drug release profile, where the dissolution rate constant was kept constant (Kd = 1600 mL/h). Because each product uses the same drug substance, we can reasonably assume that each has a similar dissolution rate constant Kd , provided that the manufacturing process will not alter the particle size of drug substance and that there is no drug excipient interaction. This simulation is for products where the particle size specification is the same for all drug substances. Consequently, the difference in dissolution profiles is mainly caused by the differences in dosage forms, reflected in the disintegration rate constant.
Figure 2 (C) shows the percent of dose dissolved at 30 minutes as a function of disintegration and dissolution constants, Kf and Kd , respectively. Figure 2 (C) gives the levels of Kf and Kd needed to ensure rapid drug release in vitro (at least 85% in 30 minutes). It remains to be shown how dosage forms affect in vivo pharmacokinetics. In other words, how do the observed differences of in vitro drug release translate into in vivo?
In Vivo Absorption
For highly soluble and highly permeable drugs, oral drug absorption is complete before it reaches the end of the small intestine. Thus, the dosage form or disintegration may not reduce the AUC, although it may alter the onset of absorption and affect the Cmax . For the purpose of simulation, Figures 3 (A) and (B) show the effect of Kf and Kd on AUC and Cmax ratios using an oral solution as a reference (that is, AUC [tablet]/AUC[solution] and Cmax [tablet]/Cmax [solution]) for highly soluble and highly permeable compounds. Figure 3 (A) shows that disintegration has limited effect on AUC. An example of such a compound is acetaminophen-only a small portion of small intestine is required to reach complete absorption. Even with the disintegration rate constant of 1 hour-1 , which corresponds to a disintegration time of 1 hour (h), there are still 2.2 hours of transit time left for absorption based on the mean small intestinal transit time of 3.32 hours (4). Consequently, the simulation outcome shown in Figure 3 (A) is as expected. Nevertheless, disintegration does affect the Cmax , as shown in Figure 3 (B). This implies that prolonging disintegration will reduce the Cmax while keeping the AUC similar for highly soluble and highly permeable drugs. Figures 3 (A) and (B) show that the disintegration can be the absorption rate-limiting step; however, the current rapid drug release criterion (>85% at 30 minutes) is more than sufficient to prevent this from occurring. For example, for a drug with solubility .4 mg/mL, Kd was calculated to be 1200 mL/h from equation 4, based on the dose of 100 mg, diffusion coefficient of 5 x 10-6 cm2 /sec, density of 1200 mg/mL, aqueous diffusion layer of 30 mm, and particle size of 25 mm. Thus, from Figures 3 (A) and (B), Kf has to be greater than 2.5 to make both the AUC and Cmax ratios higher than .9, ensuring the confidence interval of .8 to 1.25. Based on the Kd and Kf values, Figure 2 (C) gives the percent dissolved at 30 minutes around 60%, well below the criterion of 85%.
Figures 4 (A) and (B) show the effect of disintegration and dissolution rate constants on the AUC and Cmax ratios for highly soluble and poorly permeable drugs. Example drugs are acyclovir and enalaprilate (in terms of permeability). Figures 4 (A) and (B) show that AUC and Cmax are more sensitive to the disintegration and dissolution rate constants for poorly permeable drugs. In contrast to the highly soluble and highly permeable drugs (Figure 3 ), where the Cmax is more sensitive than AUC for highly soluble and poorly permeable drugs, AUC is more sensitive than Cmax . Using the same example above but reducing the permeability from 1.0 x 10-3 cm/s to 2.0 x 10-5 cm/s, the disintegration rate constant has to be greater than 9 h-1 to make both AUC and Cmax ratios higher than .9, ensuring the confidence interval of .8 to 1.25. From Figure 2 (C), we can see that the percent dissolved at 30 minutes is now around 89%, based on Kd of 1200 mL/h and Kf of 9 h-1 . Therefore, compared with highly permeable drugs, higher in vitro drug release requirements are needed. The current rapid drug release criterion (85% dissolved at 30 minutes) is not likely sufficient to ensure bioequivalence in vivo between solid and solution formulations.
Model Assumptions
This model assumes the same absorption rate constant throughout the small intestine, suggesting that the simulation outcome applies only to passively transported drugs. It is further assumed that in vitro and in vivo dissolution have the same disintegration and dissolution rate constants. Because the disintegration may be improved in vivo with the presence of bile acids, the assumption of the same disintegration rate constant in vivo and in vitro may be a conservative estimate.
The model further assumes that there is no significant absorption from the stomach or colon for passively transported drugs. It should be noted that many passively transported drugs could be absorbed from the colon and that the situation actually can be improved if the compound is absorbed from the colon. Therefore, from a regulatory point of view, the assumption is more conservative.
Equation 4 shows that the aqueous diffusion layer thickness is the most likely factor that can make the Kd difference between in vivo and in vitro because this is the only factor that depends on the external environment (ie, hydrodynamics). Recent experimental evidence suggests that the dissolution results in vitro are much lower than expected in vivo16 ; therefore, the assumption of the same disintegration and dissolution rate constants in vitro and in vivo may also be relatively conservative.
Conclusion
A novel absorption model was developed by incorporating gastrointestinal transit flow, intestinal absorption, and drug release properties of solid dosage forms. The results of this simulation suggest that the current rapid drug release criterion (>85% dissolved in less than 30 minutes in .1 HCl, pH 4.5, and pH 6.8 buffers) may ensure the bioequivalence of solid dosage forms containing highly soluble and highly permeable drugs, but not highly soluble and poorly permeable drugs. The appropriate in vitro release requirements seem to be 90% dissolved in 30 minutes for highly soluble and poorly permeable drugs.
Acknowledgements
We would like to thank Drs. Mei-Ling Chen, Donna Volpe, and Christine Maupin for their constructive comments and suggestions.
References
1.
Guidance for industry, Waiver of In Vivo Bioavailability and Bioequivalence Studies for Immediate Release Solid Oral Dosage Forms Based on a Biopharmaceutics Classification System. August 2000, CDER/FDA. ww.fda.gov/cder/guidances/index.htm
2.
Amidon GL, Lennernas H, Shah VP, Crison JR. A theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. Pharm Res. 1995;12:413-420.
[PUBMED]
3.
Blume HH, Schug BS. The biopharmaceutics classification system (BCS): Class III drugs¾better candidates for BA/BE waiver? Eur J Pharm Sci. 1999;9:117-121.
[PUBMED]
4.
Yu LX, Gatlin LA, Amidon GL. Predicting gastrointestinal drug absorption in humans. In: Amidon GL, Lee PI, Topp EM, eds. Transport Processes in Pharmaceutical Systems. New York, New York, Marcel Dekker, Inc; 1999:377-409.
5.
Kaus LC, Gillespie WR, Hussain AS, Amidon GL. The effect of in vivo dissolution, gastric emptying rate, and intestinal transit time on the peak exposure and total exposure on drugs with different gastrointestinal permeabilities. Pharm Res. 1999;16:272-280.
[PUBMED]
6.
Yu LX. An integrated absorption model for determining dissolution, permeability, and solubility limited absorption. Pharm Res. 1999;16:1884-1888.
[PUBMED]
7.
Norris DA, Leesman GD, Sinko PJ, Grass GM. Development of predictive pharmacokinetic simulation models for drug discovery. J Control Rel. 2000;65:55-62.
[PUBMED]
8.
Kalampokis A, Argyrakis P, Macheras P. A heterogeneous tube model of intestinal drug absorption based on probabilities concepts. Pharm Res. 1999;16:1764-1769.
[PUBMED]
9.
Polli JE. Dependence of in vitro-in vivo correlation analysis acceptability on model selections. Pharm Dev Technol. 1999;4:89-96.
[PUBMED]
10.
Rohrs BR, Thamann TJ, Gao P, Stelzer DJ, Bergren MS, Chao RS. Tablet dissolution affected by a moisture mediated No solid-state interaction between drug and disintegrant. Pharm Res. 1999;16:1850-1856.
[PUBMED]
11.
Abdou HM. Dissolution, Bioavailability & Bioequivalence. Easton, Pennsylvania, Mack Publishing Co; 1989:39.
12.
Yu LX, Amidon GL. A compartmental absorption and transit model for estimating oral drug absorption. Int J Pharm. 1999;186:119-125.
[PUBMED]
13.
D'Argenio DZ, Schumitzky A. Adapt II: Mathematical Software for Pharmacokinetic/Pharmacodynamic Systems Analysis. Los Angeles, CA: University of Southern California, Los Angeles; 1992.
14.
Yu LX, Amidon GL. Saturable small intestinal drug absorption in humans: modeling and explanation of the cefatrizine data. Eur J Pharm Biopharm. 1998;45:199-203.
[PUBMED]
15.
Davis SS, Hardy JG, Fara JW. Transit of pharmaceutical dosage forms through the small intestine. Gut. 1986;27:886-892.
[PUBMED]
16.
Bonlokke L, Hovgaad L, Kristensen HG, Knutson L, Lindall A, Lennernas H. A comparison between direct determination of in vivo dissolution and the deconvolution technique in humans. Eur J Pharm Sci. 1999;8:19-27.
[PUBMED]
| 
