UNGAP best practice for improving solubility data quality of orally administered drugs.

An important goal of the European Cooperation in Science and Technology (COST) Action UNGAP (UNderstanding Gastrointestinal Absorption-related Processes, www.ungap.eu) is to improve standardization of methods relating to the study of oral drug absorption. Solubility is a general term that refers to the maximum achievable concentration of a compound dissolved in a liquid medium. For orally administered drugs, relevant information on drug properties is crucial during drug (product) development and at the regulatory level. Collection of reliable and reproducible solubility data requires careful application and understanding of the limitations of the selected experimental method. In addition, the purity of a compound and its solid state form, as well as experimental parameters such as temperature of experimentation, media related factors, and sample handling procedures can affect data quality. In this paper, an international consensus developed by the COST UNGAP network on recommendations for collecting high quality solubility data for the development of orally administered drugs is proposed.

Biorelevant media for transport experiments in the Caco-2 model to evaluate drug absorption in the fasted and the fed state and their usefulness.

In this work we developed and characterized transport media that simulate the composition of micellar phase of intestinal fluids in the fasted and, especially, in the fed state and are appropriate for evaluating intestinal drug permeability characteristics using the Caco-2 model (FaSSIF-TM(Caco) and FeSSIF-TM(Caco), respectively). Media composition was based on FaSSIF-V2 and FeSSIF-V2 and recently reported data on total lipid concentrations in the micellar phase of contents of the upper small intestine in the fasted and the fed state and was adapted for cell culture compatibility. Permeation data were evaluated by compartmental kinetic modeling. Permeability coefficients, P, of hydrophilic drugs were not affected by media composition. In contrast, P values of a series of lipophilic compounds measured with FaSSIF-TM(Caco) and FeSSIF-TM(Caco), and reflecting transport by diffusion were smaller than those obtained with a purely aqueous reference transport medium, aq-TM(Caco), following the rank order aq-TM(Caco)>FaSSIF-TM(Caco)>FeSSIF-TM(Caco). The decline of permeability values was stronger as lipophilicity of the compounds increased. Compared with values estimated using aq-TM(Caco), permeability was reduced, depending on the compound, by more than 20- to 100-fold when measured with FeSSIF-TM(Caco) whereas compound ranking in regard to the permeability characteristics was also affected. The impact of reduced P value on flux through the mucosa, hence on drug absorption, in combination with the drug amount loaded on colloidal particles needs to be taken into consideration in PBPK modeling especially when the food effect is evaluated.

Oral biopharmaceutics tools - time for a new initiative - an introduction to the IMI project OrBiTo.

OrBiTo is a new European project within the IMI programme in the area of oral biopharmaceutics tools that includes world leading scientists from nine European universities, one regulatory agency, one non-profit research organization, four SMEs together with scientists from twelve pharmaceutical companies. The OrBiTo project will address key gaps in our knowledge of gastrointestinal (GI) drug absorption and deliver a framework for rational application of predictive biopharmaceutics tools for oral drug delivery. This will be achieved through novel prospective investigations to define new methodologies as well as refinement of existing tools. Extensive validation of novel and existing biopharmaceutics tools will be performed using active pharmaceutical ingredient (API), formulations and supporting datasets from industry partners. A combination of high quality in vitro or in silico characterizations of API and formulations will be integrated into physiologically based in silico biopharmaceutics models capturing the full complexity of GI drug absorption. This approach gives an unparalleled opportunity to initiate a transformational change in industrial research and development to achieve model-based pharmaceutical product development in accordance with the Quality by Design concept. Benefits include an accelerated and more efficient drug candidate selection, formulation development process, particularly for challenging projects such as low solubility molecules (BCS II and IV), enhanced and modified-release formulations, as well as allowing optimization of clinical product performance for patient benefit. In addition, the tools emerging from OrBiTo are expected to significantly reduce demand for animal experiments in the future as well as reducing the number of human bioequivalence studies required to bridge formulations after manufacturing or composition changes.

Estimation of intragastric drug solubility in the fed state: comparison of various media with data in aspirates.

The suitability of various media to forecast the solubility of ketoconazole and dipyridamole in the fed stomach at various periods after meal administration was evaluated. Solubilities were measured with the shake-flask method in gastric fluids aspirated 30, 60 and 120 min after administration of 500 ml Ensure plus to healthy fasted adults, in three sets of simulated gastric fluids based on milk, and in simple aqueous buffered media. Simple aqueous buffered media vastly underestimated the intragastric solubility of model compounds in the fed state. When using undigested milk-based media, the solubilities of model compounds in aspirates were also underestimated by a factor of 2.5-27. Solubility in milk digested with pepsin was useful for estimating the intragastric solubility of ketoconazole (within 20%) but overestimated the intragastric values of dipyridamole by a factor of 2-19. For both drugs, the solubility in milk digested with pepsin and lipase predicted the solubility in aspirates collected 60 min after meal administration, whereas at other times it overestimated the intragastric solubility (by a factor of <5). Both the use of biorelevant media and simulation of intragastric digestion are necessary for the prediction of drug solubility in the fed stomach. Milk digested with pepsin and lipase enabled the estimation of the intragastric solubility of dipyridamole and ketoconazole at 1 h after meal intake. Simulation of vesicle/micellar structures seems to be key for the prediction of intragastric solubility in the fed stomach.

A comparative study of different release apparatus in generating in vitro-in vivo correlations for extended release formulations.

The importance of hydrodynamics in the development of in vitro-in vivo correlations (IVIVCs) for a BCS Class II compound housed in a hydrophilic matrix formulation and for a BCS Class I compound housed in an osmotic pump formulation was assessed. In vitro release data were collected in media simulating the fasted state conditions in the stomach, small intestine and the ascending colon using the USP II, the USP III and the USP IV release apparatuses. Using the data collected with the USP II apparatus, the plasma profiles were simulated and compared with human plasma profiles obtained after administration of the same dosage forms to healthy fasted volunteers. Data obtained with the USP III and USP IV apparatuses were directly correlated with the deconvoluted human plasma profiles. In vitro hydrodynamics affected the release profile from the hydrophilic matrix. For both formulations, based on the values of the difference factor, all three apparatuses were equally useful in predicting the actual in vivo profile on an average basis. Although some hydrodynamic variability is likely with low solubility drugs in hydrophilic matrices, the hydrodynamics of USP II, III and IV may all be adequate as a starting point for generating IVIVCs for monolithic dosage forms in the fasted state.

Simulation of gastric lipolysis and prediction of felodipine release from a matrix tablet in the fed stomach.

The importance of intragastric lipolysis to felodipine release from a hydrophilic, extended release tablet in the fed stomach was assessed in USP II apparatus with the tablet fixed on a steel wire above the paddle. The release medium, homogenized long-life milk, was gradually digested with shots of acidic solutions of pepsin over the course of the experiment in absence and in presence of biorelevant concentrations of a lipase that was similar to human gastric lipase. Percentage tablet erosion at specific times in the same media was measured in separate experiments. The data were compared to published data for intragastric release in fed healthy adults. In all cases, felodipine release occurred under sink conditions. Lipase facilitated felodipine release from the eroded polymer, bringing the release profile closer to the in vivo data. Likewise, the relationship between tablet erosion and amount of released felodipine reflected the in vivo data only when lipase was added to the medium. It was concluded that modelling intragastric lipolysis is necessary in order to simulate felodipine release from the extended release tablets in the fed stomach.

Hydroxypropylmethylcellulose significantly lowers blood cholesterol in mildly hypercholesterolemic human subjects.

OBJECTIVE: To determine the cholesterol-lowering efficacy of hydroxypropylmethylcellulose (HPMC) in mildly hypercholesterolemic humans. SUBJECTS: Trial one: entry mean (range) total serum cholesterol values of eight female and four male subjects were 6.48 (5.57-7.51) mmol l(-1) (250 (215-290) mg dl(-1)) and 6.60 (5.57-7.64) mmol l(-1) (255 (215-295) mg dl(-1)), respectively. Trial two: corresponding values for 20 women and 20 men were 5.96 (5.43-6.48) mmol l(-1) 230 (210-250) mg dl(-1)) and 6.05 (5.46-6.63) mmol l(-1) 233 (211-256) mg dl(-1)), respectively. RESULTS: Trial one: HPMC decreased (P< or =0.05) total and LDL-cholesterol 9.3 and 15.3% (medium), 16.9 and 23.5% (high) and 13.8 and 19.4% (ultra-high), respectively, over placebo. Trial two: total and LDL-cholesterol decreased (P< or =0.05) throughout the 8 weeks, with mean (weeks 4-8) reductions of 7 and 8% at 5 g day(-1), and 12 and 15% at 15 g day(-1), respectively, over placebo. Adverse effects were minimal. Trial one: medium, high and ultra-high viscosity HPMC at 15 g day(-1) for 1 week each;1-week wash-out between treatments. Trial two: ultra-high viscosity HPMC at 5 or 15 g day(-1) for 8 weeks. CONCLUSIONS: HPMC soluble fiber, especially high-viscosity grades, significantly lowers cholesterol at well-tolerated doses, showing promise as a treatment of hypercholesterolemia.

Postprandial evolution in composition and characteristics of human duodenal fluids in different nutritional states.

The purpose of this study was to assess the changes in duodenal composition in three nutritional states: fasted, fed, and fat-enriched fed state. Two isocaloric meals were administered to healthy subjects on nonconsecutive days. Subsequently, duodenal samples were collected every 30 min after which they were characterized with respect to pH, lipolytic products, bile salts, phospholipids, osmolality, and surface tension. The resulting time profiles displayed fluctuating patterns, which reflect high inter- and intrasubject variability. Duodenal composition was not altered by the higher fat percentage of the fat-enriched liquid meal. Monoglycerides, amounting from 5% to 88% of total lipids, were the dominant lipolytic species, followed by free fatty acids. Within 30 min after meal administration, individual intraduodenal concentrations of lipid products were 0.0-5.5, 1.0-14.9, and 3.1-22.4 mg/mL in fasted, fed, and fat-enriched fed state, respectively. The corresponding values for bile salts were 2.0-9.0, 6.9-9.3, and 4.4-30.3 mM and for phospholipids 0.06-2.4, 2.6-5.7, and 1.4-9.3 mM, respectively. Specific trends though, were not detected. This study illustrates the variable intraluminal conditions that can result after food intake. As intraduodenal events (e.g., intraduodenal dissolution) affect absorption of poorly water soluble and/or highly lipophilic drugs, this variability may possibly contribute to the highly variable drug plasma-time profiles often observed.

Estimating drug solubility in the gastrointestinal tract.

Solubilities measured in water are not always indicative of solubilities in the gastrointestinal tract. The use of aqueous solubility to predict oral drug absorption can therefore lead to very pronounced underestimates of the oral bioavailability, particularly for drugs which are poorly soluble and lipophilic. Mechanisms responsible for enhancing the luminal solubility of such drugs are discussed. Various methods for estimating intra-lumenal solubilities are presented, with emphasis on the two most widely implemented methods: determining solubility in fluids aspirated from the human gastrointestinal tract, and determining solubility in so-called biorelevant media, composed to simulate these fluids. The ability of the biorelevant media to predict solubility in human aspirates and to predict plasma profiles is illustrated with case examples.

Optimization and validation of a high-performance liquid chromatographic method with UV detection for the determination of ketoconazole in canine plasma.

An isocratic high-performance liquid chromatographic method with detection at 240 nm was developed, optimized and validated for the determination of ketoconazole in canine plasma. 9-Acetylanthracene was used as internal standard. A Hypersil BDS RP-C18 column (250 mm x 4.6 mm, 5 microm particle size), was equilibrated with a mobile phase composed of methanol, water and diethylamine 74:26:0.1 (v/v/v). Its flow rate was 1 ml/min. The elution time for ketoconazole and 9-acetylanthracene was approximately 9 and 8 min, respectively. Calibration curves of ketoconazole in plasma were linear in the concentration range of 0.015-10 microg/ml. Limits of detection and quantification in plasma were 5 and 15 ng/ml, respectively. Recovery was greater than 95%. Intra- and inter-day relative standard deviation for ketoconazole in plasma was less than 3.1 and 4.7%, respectively. This method was applied to the determination of ketoconazole plasma levels after administration of a commercially available tablet to dogs.

Biowaiver monographs for immediate release solid oral dosage forms: acetaminophen (paracetamol).

Literature data are reviewed on the properties of acetaminophen (paracetamol) related to the biopharmaceutics classification system (BCS). According to the current BCS criteria, acetaminophen is BCS Class III compound. Differences in composition seldom, if ever, have an effect on the extent of absorption. However, some studies show differences in rate of absorption between brands and formulations. In particular, sodium bicarbonate, present in some drug products, was reported to give an increase in the rate of absorption, probably caused by an effect on gastric emptying. In view of Marketing Authorizations (MAs) given in a number of countries to acetaminophen drug products with rapid onset of action, it is concluded that differences in rate of absorption were considered therapeutically not relevant by the Health Authorities. Moreover, in view of its therapeutic use, its wide therapeutic index and its uncomplicated pharmacokinetic properties, in vitro dissolution data collected according to the relevant Guidances can be safely used for declaring bioequivalence (BE) of two acetaminophen formulations. Therefore, accepting a biowaiver for immediate release (IR) acetaminophen solid oral drug products is considered scientifically justified, if the test product contains only those excipients reported in this paper in their usual amounts and the test product is rapidly dissolving, as well as the test product fulfils the criterion of similarity of dissolution profiles to the reference product.

Sensitive and simple liquid chromatographic method with ultraviolet detection for the determination of nifedipine in canine plasma.

An isocratic high-performance liquid chromatographic method with detection at 240 nm was developed, optimized and validated for the determination of nifedipine in canine plasma. Liquid-liquid extraction was used as the sample preparation technique. Carbamazepine was used as internal standard. A Hypersil BDS RP-C18 column (250 mm x 4.6 mm, 5 microm) was equilibrated with a mobile phase composed of water and methanol, 45:55 (v/v). Its flow rate was 1 ml min(-1). The elution time for nifedipine and carbamazepine was approximately 12 and 8 min, respectively. Calibration curves of nifedipine in plasma were linear in the concentration range of 1-200 ng ml(-1). Limits of detection and quantification in plasma were 0.5 and 1.5 ng ml(-1), respectively. Recovery was greater than 98%. Intra- and inter-day relative standard deviation for nifedipine in plasma was less than 8.5 and 10%, respectively. This method was applied to the determination of nifedipine plasma levels after administration of commercially available soft gelatine capsules to dogs.

Optimized determination of lycopene in canine plasma using reversed-phase high-performance liquid chromatography.

An isocratic high-performance liquid chromatographic method with detection at 472 nm was developed, optimized and validated for the determination of lycopene in canine plasma. Ethyl-beta-apo-8'-carotenoate was used as internal standard. A Hypersil BDS RP-C18 column (150 mm x 4.6 mm), 5 microm particle size, was equilibrated with a mobile phase composed of acetonitrile and methanol (50:50, v/v). Its flow rate was 1.5 ml/min. The elution time for lycopene and ethyl-beta-apo-8'-carotenoate was approximately 11 and 5 min, respectively. Calibration curves of lycopene were linear in the concentration range of 3-200 ng/ml in plasma. Limits of detection and quantification in plasma were 1 and 4 ng/ml, respectively. Recovery was greater than 97%. Intra- and inter-day relative standard deviation for lycopene in plasma was less than 1.8 and 3.1%, respectively. This method was applied to the determination of lycopene plasma levels after single dose administration to dogs.

Biorelevant dissolution testing to predict the plasma profile of lipophilic drugs after oral administration.

PURPOSE: To quantitatively compare in vitro dissolution data in biorelevant and compendial media, to investigate whether in vitro differences are reflected in the simulated plasma profile and to specify under which circumstances prediction of the plasma profile of orally administered lipophilic drugs can be achieved. METHODS: Previously published dissolution data from seven products of four lipophilic drugs were compared using the first order model, the RRSBW distribution, and a model based on the Noyes-Whitney theory. Simulated plasma profiles were then obtained using a model-dependent approach. Simulated and observed plasma profiles were compared with the difference factor, f1. RESULTS: No model consistently provided the best fit to the in vitro data, which varied significantly with medium composition. Prediction of the plasma profile was possible (9.6 < or = f1 < or = 34.2) in seven out of eleven cases. CONCLUSIONS: Although prediction of the plasma profile of lipophilic drugs solely on the basis of in vitro data remains an ambitious target, this study shows that the plasma profile of a lipophilic drug can be predicted with appropriate in vitro dissolution data, provided that the absolute bioavailability of the drug is known and the drug has dissolution limited absorption.

In vitro-in vivo correlations for lipophilic, poorly water-soluble drugs.

Although several routes of administration can be considered for new drug entities, the most popular remains the oral route. To predict the in vivo performance of a drug after oral administration from in vivo data, it is essential that the limiting factor to absorption can be modelled in vitro. In the case of BCS class II drugs dissolution is rate-limiting to absorption, so the use of biorelevant dissolution tests can be used to predict differences in bioavailability among different formulations and dosing conditions. To achieve an a priori correlation, the composition, volume and hydrodynamics of the contents in the gastrointestinal lumen following administration of the dosage form must be accurately simulated. Four media have been chosen/developed to model composition of the gastric and intestinal contents before and after meal intake. These are SGF, milk, FASSIF and FeSSIF, which model fasted and fed state conditions in the stomach and small intestine respectively. Using these media, excellent correlations have been obtained with the following poorly soluble drugs: danazol, ketoconazole, atovaquone and troglitazone. In all cases, fed vs. fasted state effects can be predicted from dissolution data and, where several formulations were available for testing, dissolution tests could also be used to determine which would have the best in vivo performance.

Longitudinal versus radial effects of hydroxypropylmethylcellulose on gastrointestinal glucose absorption in dogs.

Many water soluble fibers have been shown to favorably affect the postprandial glucose profile in humans. Hydroxypropylmethylcellulose (HPMC), a fiber which has been shown to increase glucose tolerance in dogs and noninsulin dependent diabetics, was chosen to study the luminal interactions which mediate this effect. The ability of HPMC to influence upper gastrointestinal (GI) viscosity, transit, and water flux of 5% and 20% glucose solutions was studied in five female dogs fistulated at the proximal duodenum and/or midjejunum. HPMC elevated intraluminal viscosity, with a linear relationship existing between input and luminal viscosity. The ability to modify intraluminal viscosity was greater for isoosmotic (5%) glucose solutions than for hyperosmotic (20%) glucose solutions. HPMC also modified the transit profile of isoosmotic (5%) glucose solutions at midgut by both increasing the lag times before the onset of chyme recovery from 5.5+/-3.1 min to 9-55 min (depending on the viscosity of the administered solution) and decreasing the first-order transit rate constants from 0.115+/-0.07 min-1 to 0.014-0.035 min-1. By contrast, the transit profile of hyperosmotic (20%) glucose solutions was not significantly affected. Net cumulative water flux across the gut wall was not significantly affected in either case by the presence of HPMC. These results, in combination with the amount of glucose recovered from midgut fistula, suggest that following the administration of glucose solutions, HPMC effects on blood glucose levels are mediated by mechanisms which relate to the increased intraluminal viscosity but vary according to the input glucose load. For isoosmotic glucose loads, both the decreased upper GI transit rate and hindered radial movement play a role. Although HPMC modifies glucose absorption from hyperosmotic solutions, this study shows that luminal effects occurring before midgut are modest.

Forecasting the in vivo performance of four low solubility drugs from their in vitro dissolution data.

PURPOSE: To assess the usefulness of biorelevant dissolution tests in predicting food and formulation effects on the absorption of four poorly soluble, lipophilic drugs. METHODS. Dissolution was studied with USP Apparatus II in water, milk, SIFsp, FaSSIF, and FeSSIF. The in vitro dissolution data were compared on a rank order basis with existing in vivo data for the tested products under fasted and fed state conditions. RESULTS: All drugs/formulations showed more complete dissolution in bile salt/lecithin containing media and in milk than in water and SIFsp (USP 23). Comparisons of the in vitro dissolution data in biorelevant media with in vivo data showed that in all cases it was possible to forecast food effects and differences in absorption between products of the same drug with the physiologically relevant media (FaSSIF, FeSSIF and milk). Differences between products (both in vitro or in vivo) were less pronounced than differences due to media composition (in vitro) or dosing conditions (in vivo). CONCLUSIONS: Although biorelevant dissolution tests still have issues which will require further refinement, they offer a promising in vitro tool for forecasting the in vivo performance of poorly soluble drugs.

The effect of HPMC--a cholesterol-lowering agent--on oral drug absorption in dogs.

The objective of this study was to evaluate the effects which hydroxypropylmethylcellulose (HPMC) may exert on oral drug absorption, in cases where this soluble fiber is administered to regulate blood lipid levels. Studies were conducted in vitro and in healthy female mongrel dogs using two different grades of HPMC, i.e. K8515 HPMC and ultra high molecular weight (UHMW) HPMC. The maximum plasma concentration, Cmax, of paracetamol and both the Cmax and the area under the concentration-time curve, AUC, of cimetidine were significantly decreased by the coadministration of 10 g of K8515 HPMC or 7.5 g of UHMW HPMC dissolved in 500 mL normal saline under fasting conditions. No statistically significant effects were observed on hydrochlorothiazide or mefenamic acid absorption. Based on in vitro data and previous studies it appears that reductions in gastric emptying and dissolution rate of paracetamol account for the effect observed in vivo. For cimetidine, a drug which can be absorbed from both the small and the large intestine, the indigestibility of HPMC in the colon in addition to the great reduction of dissolution rate led to reductions of both the Cmax and AUC values. The long Tmax values, even in the absence of HPMCs and the more modest reduction of the dissolution rate of hydrochlorothiazide by the HPMCs are thought to have precluded the observation of any significant alterations in the in vivo absorption profile. Owing to its erratic absorption, no statistically based conclusion could be drawn about the effects of coadministered HPMC on the oral absorption of the poorly soluble mefenamic acid. It is concluded that the effects of HPMCs on drug absorption in dogs are most pronounced for compounds with absorption profiles that are dependent on gastric emptying, i.e. compounds that are highly water soluble and that exhibit short Tmax values. Compounds with long absorption profiles appear to be less susceptible to changes in absorption behavior due to coadministration of HPMCs.

Effect of elevated viscosity in the upper gastrointestinal tract on drug absorption in dogs.

The objectives of these studies were, first, to determine the effect of elevated luminal viscosity on the gastrointestinal absorption of four model drugs and, second, to identify the key processes influencing drug absorption under elevated viscosity conditions. Studies were conducted in vitro and in healthy female mongrel dogs under fasting conditions. In the canine model, both the rate and extent of paracetamol and hydrochlorothiazide absorption were significantly decreased by the coadministration of 15 g guar gum dissolved in 500 ml normal saline. In the case of cimetidine, the rate but not extent of absorption was decreased. Owing to the high variability in the data, no statistically based conclusion could be drawn about the effects of coadministered guar gum on the oral absorption of the poorly soluble mefenamic acid. Based on the in vitro data, it appears that substantial reductions in the dissolution rate of paracetamol, hydrochlorothiazide and cimetidine account for the effects observed in vivo. It is concluded that the effect of an elevation in the intraluminal viscosity on drug absorption is greatest for highly soluble drugs, and results from a combination of a decrease in dissolution rate and gastric emptying rate.

Evaluation of various dissolution media for predicting in vivo performance of class I and II drugs.

PURPOSE: In this paper we seek to verify the differences in dissolution behavior between class I and class II drugs and to evaluate the suitability of two new physiologically based media, of Simulated Gastric Fluid (SGF) and of milk for their ability to forecast trends in the in vivo performance of class II compounds and their formulations. METHODS: Dissolution behavior of two class I drugs, i.e. acetaminophen and metoprolol, and of three class II drugs, i.e. danazol, mefenamic acid and ketoconazole, was studied with USP Apparatus 2 in water, SGF, milk, Simulated Intestinal Fluid without pancreatin (SIFsp) and in two media simulating the small intestinal contents in the fed (FeSSIF) and fasted (FaSSIF) states, respectively. RESULTS: Class I powders dissolved rapidly in all media tested. Acetaminophen dissolution in milk was slow from one tablet formulation, in all other cases dissolution was more than 85% complete in 15 minutes. The dissolution rate of metoprolol was shown to be dependent on formulation and manufacturing method, and one of the three tablet formulations did not meet compendial specifications (80%/30 minutes). Dissolution behavior of class II drugs was greatly affected by choice of medium. Dissolution from a capsule formulation of danazol proved to be dependent on the concentration of solubilizing agents, with a the 30-fold increase in percentage dissolved within 90 minutes upon changing from aqueous media without surfactants to FaSSIF. Use of FeSSIF or milk as the dissolution medium resulted in an even greater increase in percentage dissolved, 100 and 180-fold respectively. Dissolution of the weak acid mefenamic acid from a capsule formulation is dependent on both pH and bile salt concentration, which leads to an offset between increased bile salt concentration and lower pH in the fed state compared to the fasted state medium. The weak base ketoconazole showed complete dissolution from a tablet formulation in Simulated Gastric Fluid without pepsin (SGFsp) within 30 minutes, 70% dissolution in 2 hours under fed state simulated upper jejunal conditions but only 6% dissolution in 2 hours under fasted state conditions. CONCLUSIONS: As predicted, dissolution of class II drugs proved to be in general much more dependent on the medium than class I drugs. With the array of compendial and physiological media available, it should be possible to design a suitable set of tests to predict the in vivo dissolution of both class I and II drugs from immediate release formulations.