Physico-chemical characterization of aspirated and simulated human gastric fluids to study their influence on the intrinsic dissolution rate of cinnarizine

To elucidate the critical parameters affecting drug dissolution in the human stomach, the intrinsic dissolution rate (IDR) of cinnarizine was determined in aspirated and simulated human gastric fluids (HGF). Fasted aspirated HGF (aspHGF) was collected from 23 healthy volunteers during a gastroscopic examination. Hydrochloric acid (HCl) pH 1.2, fasted state simulated gastric fluid (FaSSGF), and simulated human gastric fluid (simHGF) developed to have rheological, and physico-chemical properties similar to aspHGF, were used as simulated HGFs. The IDR of cinnarizine was significantly higher in HCl pH 1.2 (952 ± 27 µ g/(cm 2 ⋅ min)) than in FaSSGF pH 1.6 (444 ± 7 µ g/(cm 2 ⋅ min)), and simHGF pH 2.5 (49 ± 5 µ g/(cm 2 ⋅ min)) due to the pH dependent drug solubility and viscosity differences of the three simulated HGFs. The shear thinning behavior of aspHGF had a significant impact on the IDR of cinnarizine, indicating that the use of FaSSGF, with viscosity similar to water, to evaluate gastric drug dissolution, might overestimate the IDR by a factor of 100 – 10.000, compared to the non-Newtonian, more viscous, fluids in the human stomach. The developed simHGF simulated the viscosity of the gastric fluids, as well as the IDR of the model drug, making it a very promising medium to study gastric drug dissolution in vitro.


Introduction
Drug dissolution in the physiological environment of the gastrointestinal tract (GIT) is a prerequisite for drug absorption following oral administration of solid dosage forms.Slow drug dissolution may decrease oral drug absorption (Vardakou et al., 2011), rendering the solubility and dissolution rate of a drug crucial for its in vivo performance (Dressman and Reppas, 2000).Especially in the case of weakly basic drugs, the dissolution rate in the stomach is important, as dissolution may be the rate limiting step for drug absorption (Vertzoni et al., 2005;Vertzoni et al., 2007;Amidon et al., 1995).
Over the last 25 years, increased focus has been directed towards assessing dosage form performance in the GIT by using biorelevant dissolution media to study in vitro dissolution, solubility, and solid phase transition behavior of drug compounds and dosage forms.For highly soluble drugs, i.e. biopharmaceutics classification system (BCS) class I and III drugs, simple pharmacopoeial dissolution media, such as 0.1 M hydrochloric acid (HCl) pH 1.2, are considered suitable for evaluating intra-gastric behavior (Markopoulos et al., 2015).However, to evaluate the dissolution behavior of poorly soluble drugs (BCS class II and IV drugs), biorelevant dissolution media better simulating the actual conditions in the GIT have been recommended to accurately predict in vivo behavior (Klein, 2010;Fagerberg et al., 2010).
It is well known that pH, surfactants, ionic strength, and buffer capacity influence the dissolution rate of a drug in the GIT (Jinno et al., 2000;Dressman et al., 1998;de Smidt et al., 1987;de Smidt et al., 1994;Gibaldi and Feldman, 1970;Gibaldi et al., 1968;Higuchi, 1964).Additionally, hydrodynamics, flow rates and liquid volumes in the GIT have also been shown to influence the drug dissolution rate (Dressman and Reppas, 2000).Dissolution of a solid substance in its own solution may be described by the Noyes Whitney equation (Eq.( 1)) (Noyes and Whitney, 1897).
Where dM/dt is the rate of dissolution, D is the diffusion coefficient, A is the surface area, C S is the saturation solubility, C is the apparent concentration of a drug in the solution at a certain time point, and h is the thickness of the boundary layer around the dissolving particle.
The diffusion coefficient of a drug in solution is described by the Stokes-Einstein equation (Eq.( 2)) (Miller, 1924).
Where D is the diffusion coefficient of the drug in solution, K B is the Boltzman constant, T is absolute temperature, η is the viscosity of the dissolution medium, and r is radius of the diffusion molecule assuming it is a sphere.Together, Eq. ( 1) and Eq. ( 2) predict that an increased viscosity of the dissolution medium will decrease the intrinsic dissolution rate (IDR) of a drug.Additionally, increased media viscosity has been shown to significantly delay tablet disintegration (Anwar et al., 2005;Parojcic et al., 2008;Radwan et al., 2012), which in turn will decrease the drug dissolution rate, as the surface area of a non-disintegrated tablet is much smaller compared to that of a disintegrated tablet.A previous study by Pedersen et al. showed that the rheological profile of the commonly used gastric dissolution media, i.e. 0.1 M HCl pH 1.2 and fasted state simulated gastric fluid (FaSSGF, pH 1.6), differ from that of human gastric fluid (HGF) due to the presences of lipids, mucins and other macromolecules (Pedersen et al., 2013).Mucin molecules are glycoproteins of high molecular weight (2-14⋅10 6 g/mol), and have been shown to be largely responsible for the viscosity and shear thinning properties of mucus and HGF (Pedersen et al., 2013;Allen et al., 1984;Andrews et al., 2009).However, the importance of the content of mucins and HGF viscosity on drug dissolution has not yet been elucidated.Several simulated gastric media, e.g.FaSSGF and FaSSGF-v2, have been developed, but none of these reflect the viscosity of HGF (Erceg et al., 2012;Galia et al., 1998Galia et al., , 1999;;Kostewicz et al., 2002;Nicolaides et al., 1999;Vertzoni et al., 2005).As the viscosity is known to affect the drug dissolution rate, it may be desirable to use a simulated HGF mimicking both the composition and the rheological properties of HGF, to accurately forecast the gastric dissolution behavior of orally administered drugs.With the aim of elucidating the impact of dissolution medium rheology, the IDR of cinnarizine was determined in aspirated and simulated HGF mimicking the composition and the rheological properties of HGF to different degrees.The physico-chemical properties of cinnarizine are shown in Table 1.

Volunteers for the study
Human gastric aspirates (aspHGF) were collected from volunteers during gastroscopic examinations at Herlev Hospital, Copenhagen.Volunteers (N = 23) with normal body weight, aged 20 -65 years were included in the study.The volunteers did not eat or drink for six and two hours prior to the study, respectively.Only samples from volunteers that did not have any upper gastrointestinal diseases (known or discovered during the gastroscopic examination) were included in the present study.Smokers, pregnant or breastfeeding women, and volunteers that ingested any medication, food or water on the day of gastric fluid aspiration, were also excluded.The volunteers all gave their written informed consent to the experimental procedure.The study was approved by the Ethical Committee of Denmark and followed the conventions of the Declaration of Helsinki (H-2-2011-073).

Handling of aspiration samples
The aspHGF was collected in volunteers immediately after introduction of a gastroscope.Aspiration was performed using a conventional gastroscope with a build-in suction mechanism.The gastroscopes used during the procedures had a diameter ranging between 9 and 11 mm and allowed visualization of the distribution of fluid in the stomach.The gastroscope was passed through the mouth and the esophagus into the stomach from where fluid was aspirated (aspHGF).No fluid was used to rinse the gastroscope prior the examination and great care was taken not to flush the endoscope with water before the fluid aspiration was performed.Immediately after collection, the aspHGF samples were stored at − 20 ○ C until use (and no longer than 3 months).Prior to characterization and dissolution studies, the pH of all aspHGF samples was measured, and samples with a pH above 5 were excluded from the study.A total of 17 samples were included in the study.These samples were filtrated to remove larger particles and the sampled volumes were pooled.As cinnarizine is a weak base (Table 1), its solubility and IDR in gastric fluid will depend on the pH of the dissolution medium.The pH of the pooled aspHGF was adjusted to 1.6 with HCl, to be able to study the effect of the viscosity difference between aspHGF and FaSSGF, independent of pH (the pH of aspHGF has previously been reported to be 2.5 ± 1.4) (Pedersen et al., 2013).

Preparation of simulated human gastric fluids
Six different simulated HGF were utilized in the present study.HCl (0.1 M) pH 1.2 and FaSSGF were included as these media represent the two most commonly used dissolution media to evaluate the gastric drug dissolution in vitro (Vertzoni et al., 2005).As it has been found that the osmolality of FaSSGF does not mimic that of aspHGF (Erceg et al., 2012;Pedersen et al., 2013;Vertzoni et al., 2007), FaSSGF* (resembling FaSSGF-v2) was prepared with a higher concentration of NaCl as compared to FaSSGF, to ensure the osmolality of the simulated gastric medium was comparable to aspHGF (Table 2).FaSSGF* was included in the study to test if a difference in the dissolution medium osmolality influences the IDR of cinnarizine.To study the impact of the rheological properties of aspHGF, on the IDR of cinnarizine in simulated gastric media, simulated human gastric fluid (simHGF) was utilized.simHGF was developed in a previous study by Pedersen et al., to achieve physicochemical and rheological properties similar to aspHGF (Pedersen et al., 2013).In that study, methylcellulose (MC) was found to be a suitable polymer to obtain a simHGF with similar rheological properties to aspHGF.The addition of MC 20.000 mPa⋅s at a concentration of 0.2% (w/v) to simHGF provided a simulated gastric medium with shear thinning properties similar to aspHGF.To study the effect of the viscosity difference between FaSSGF and simHGF, independent of pH, simHGF* was prepared with a pH of 1.6, similar to FaSSGF.
To study the influence of sheer-thinning and Newtonian fluids on the IDR of cinnarizine, a Newtonian simHGF (N-simHGF) was prepared using MC with a lower viscosity grade (MC 15 mPa⋅s at a concentration of 0.5 %(w/v)) and thereby molecular weight.Table 2 lists the compositions of FaSSGF, FaSSGF*, simHGF, simHGF* and N-simHGF.
FaSSGF and FaSSGF* were prepared by mixing the components listed in Table 2 with purified water.The solution was stirred overnight to ensure complete dissolution of all components, and the pH was adjusted to 1.6 with HCl.
As MC quickly gels at room temperature, simHGF was prepared in three steps; (i) the polymer, MC 20.000 mPa⋅s, was added slowly to ice cold purified water under vigorous stirring to form a 0.25 % (w/v) polymer dispersion, (ii) a two-times concentrated simHGF medium (without MC) was prepared by dissolving double the concentration of the components in Table 2 in purified water, and (iii) the ice cold polymer dispersion was mixed with the concentrated simHGF medium.The final mixture was stirred overnight to achieve homogeneity, and the pH was adjusted to 2.5 (for simHGF) and 1.6 (for simHGF*) with HCl.
N-simHGF was prepared similarly to simHGF, as described above, however, MC 15 mPa⋅s was used as the polymer in a final concentration of 0.5 %(w/v).The pH was adjusted to 2.5 with HCL.
All the prepared simulated gastric fluids were stored at 5 • C for up to 48 h, until used.

Characterization of aspirated and simulated gastric fluids
The pooled aspHGF samples and the various prepared simulated gastric fluids were characterized in terms of pH, osmolality, surface tension, bile salt concentration, protein content, and rheological behavior.The characterization studies were performed to confirm that the simHGF, designed based on a previous batch of aspHGF (Pedersen et al., 2013), did in fact simulate the aspHGF, as well as to help interpret the IDR results.
The pH values were measured by a pH electrode Metrohm (Switzerland) connected to a PHM 220 pH-meter (Radiometer, Brønshøj, Denmark).
Surface tension was determined by the pendant drop method on a KRÜSS DSA100 Drop Shape Analysis System (KRÜSS GmbH, Hamburg, Germany) connected to a Julabo ED-5 Open Bath Circulator (Julabo Labortechnik, Seelbach, Germany).The temperature was kept at 37 • C. A needle with a diameter of 1.825 mm was used and a drop of 10 ± 0.5 μL was analyzed.A measurement was conducted every 10 s for a maximum of 30 min or until the measurements indicated a stabile surface tension.
The total bile salts concentration was quantified using a Total Bile Acids Assay Kit (Diazyme Laboratories Inc., Poway, CA, USA), and following the instructions of the manufacturer.
The total protein content was quantified using a Bicinchoninic Acid protein Kit (Sigma Aldrich, Spruce, St. Louis, USA) with bovine serum albumin as a standard.The assay was carried out following the instructions of the manufacturer.
2.2.4.1.Rheology.Rheological characterization of the pooled aspHGF samples and the different simulated gastric media was conducted on an AR-G2 rheometer, (TA Instruments, Waters Corporation, New Castle, DE, USA) using a cone and plate geometry.All measurements were performed at 37 • C with a 40 mm aluminum steel cone (TA Instruments, Waters Corporation, (New Castle, USA)) at a gap of 33 μm.To limit evaporation, a protective casing, custom fabricated at the Department of Pharmacy, Faculty of Health and Medical Science (Copenhagen, Denmark), was attached onto the fixed plate and the edge of the liquid sample was covered with 5 mL of low-viscous poly(dimethylsilcoxane) oil after lowering of the cone to the measurement gap.
The sample (350 µL) was allowed to equilibrate for 10 min at 37 • C before measurements were conducted.Three consecutive measurements were conducted to determine the rheological properties of the pooled aspHGF and the different simulated gastric media.Inertia dominated measurements were excluded in the data evaluation.
The apparent viscosity of the samples was measured as a function of the shear rate from 0.001 to 1000 s − 1 .The tolerance was set to 5 %.The maximum measuring time for each shear rate was set to 3 min.Measurements not reaching equilibrium within the 3 min were not taken into account.

Dissolution studies
The IDR of cinnarizine in the aspHGF and the different simulated gastric fluids were measured using the μDISS Profiler™ from pION Inc.
(Woburn, MA, USA), containing six dip probes connected to a UV-Vis spectrophotometer.Absorbance of cinnarizine was measured in the wavelength range of 240 -265 nm.The vial holder was preheated to 37 ○ C with a Julabo Open Bath Circulator (Julabo Labortechnik GmbH, Seelbach, Germany).In each experiment (n ˃ 3), IDR was determined in 6 mL of aspHGF or 10 mL of simulated gastric media.The probes were centre-positioned in the vials.A 5 mm path length of the in situ UV probe was used.Each channel was calibrated with its own standard curve prior to the experiment.To calibrate, a stock solution of cinnarizine in ethanol was prepared, and increasing amounts of stock solution were added to the dissolution medium and the absorbance was measured.A Mini-IDR™ compression system (Heath scientific, Buckinghamshire, United Kingdom) was used to prepare miniaturized compacts with a constant surface area.Stainless steel dies containing a cylindrical hole with an area of 0.071 cm 2 were filled with pure drug powder and compressed at a compression pressure of approximately 35 bar for 30 s.Each die was then inserted into a Teflon rotating disk carrier and placed on the bottom of the glass vials.
The dissolution medium was transferred to each vial and the magnetic stirring system was turned on.The stirring rate was 50 rpm or 150 rpm.The absorbance was measured once every 10 s for 20 min.Parafilm was placed over the opening of the vials to avoid evaporation of media.
2.2.5.1.Fitting of curves using μDISS profiler software.To determine the IDR, all dissolution curves were analysed using a second derivative areaunder-curve method.A bi-exponential function (Eq.( 3)) was selected to describe the dissolved concentration as a function of time while accounting for loosely packed powder on the tablet surface.
where C powder and C compact refer to the concentration of drug in the final saturated solution due to the contribution of the powder burst and compact dissolution.The surface areas associated with the loose powder and the compact are A powder and A compact , respectively.P ABL refers to the permeability across the aqueous boundary layer and t 0 is the lag time.
To determine the IDR, the five parameters: C powder , C compact , A powder, A compact and t 0 , was determined.A compact was kept constant at 0.071 cm 2 and should in theory be constant throughout the measurement.The remaining parameters were determined by the software after curve fitting to obtain a low R 2 value, low residuals and a low standard deviation.
where DR max is the maximum dissolution rate and A eff is the effective surface area of drug exposed to the dissolution medium calculated by the μDISS Profiler™ software.

Data analysis
All data are expressed as mean and standard deviation (SD), except for the rheological characterization data, which is expressed by a single determination.To determine if there was any statistical significant difference between the IDR of cinnarizine in the different simulated gastric media, the results were analyzed by a single sided analysis of variance followed by a Bonferroni posttest.

Results and discussion
The term biorelevant dissolution media has been used to describe the already existing simulated gastric and intestinal media developed based on in vivo chemical characterization results for both gastric and intestinal fluids (e.g.SGF, FaSSGF, and FaSSIF) (Dressman et al., 1998;Galia et al., 1998Galia et al., , 1999;;Vertzoni et al., 2005).However, as these media do not mimic the physical properties, such as the rheological behavior of the human fluids, they can only be regarded as "chemically relevant" dissolution media.Since both the chemical and physical characteristics of the in vivo GIT fluids may affect drug dissolution, the rheological characteristics of the human GIT fluids, and the simulated biorelevant dissolution media should be taken into account when studying drug dissolution.E.g., it is hypothesized that shear thinning behavior, can affect the IDR of a drug, as the viscosity of the dissolution media changes with different motility patterns or stirring rates used during measurement.In this study the IDR of cinnarizine was studied in aspHGF, as well as in various dissolution media simulating HGF to various degrees.Prior to the dissolution studies, the dissolution media were characterized with respect to pH, osmolality, surface tension, bile salt concentration, protein content, and rheological behavior.

Characterization of the aspirated and simulated gastric fluids
Table 3 lists the measured pH, osmolality, surface tension, bile salt concentration, and protein content of the pooled aspHGF, FaSSGF, FaSSGF*, simHGF, simHGF* and N-simHGF.As a reference, the values originally reported by Pedersen et al., and used to design the simHGF (Pedersen et al., 2013), are also reported in Table 3.
When comparing the present aspHGF with the reference aspHGF (Pedersen et al., 2013), no significant differences are observed for the osmolality, surface tension, bile salt concentration and protein content (Table 3).The pH was different, however, this difference was induced in order to be able to compare the IDR in aspHGF to that in FaSSGF at the same pH.
The osmolality, surface tension, and bile salt content of simHFG, N-simHGF, and simHGF* correspond to that of aspHGF (reference), and aspHGF (pooled), respectively (Table 3).The protein content of each of the simulated gastric media was kept constant at 0.1 g/L, which is equivalent to the protein content of FaSSGF and corresponding to the pepsin content (Table 2).The protein content determined by Pedersen et al. (2013) and in the pooled aspHGF in the present study was significantly higher (4.9 ± 1.0 g/L and 3.6 ± 0.7 g/L, respectively) compared to the protein content in the simulated gastric media (Table 3) (Pedersen et al., 2013).However, it was decided not to increase the protein content, as the observed small difference was not expected to influence the viscosity of the media, but could possibly change the surface tension in the media, which was unwanted.
The osmolality of FaSSGF was significantly lower than the osmolality in aspHGF (reference and pooled), simHGF and simHGF* (Table 3).However, FaSSGF* had an osmolality which was identical to those of aspHGF and simHGF.The surface tension of FaSSGF and FaSSGF* was higher than that of aspHGF and simHGF due to the relatively low amount of bile salts added to this medium.

Rheological characterization of the aspirated and simulated gastric fluids
Fig. 1A shows that the shear viscosity profile of the pooled aspHGF and the simHGF was similar and both within the shear viscosity range previously measured in aspHGF by Pedersen et al. (Pedersen et al., 2013).Pooled aspHGF and simHGF showed shear thinning behavior from 0.01 s − 1 to 177 s − 1 and 0.003 s − 1 to 100 s − 1 , respectively, followed by a plateau at higher shear rates.At shear rates above 177 s − 1 the shear viscosity of aspHGF was lower than that of simHGF (Fig. 1B).In a study by Bennett et al. the shear rates corresponding to the mixing pattern of the antrum were determined using magnetic resonance imaging and locust bean gum.It was found that shear rates in the range of 30-60 s − 1 corresponded the forces applied to drug delivery systems (locust bean gum) in the antrum of the fasted stomach (indicated in Fig. 1B) (Bennett et al., 2009).The shear viscosity of simHGF was measured to be 2.7 ± 0.2 mPa⋅s at a shear rate of 56.2 s − 1 .At this shear rate, the shear viscosity of pooled aspHGF was lower than the simHGF, i.e. 2.0 ± 0.1 mPa⋅s.However, both simHGF and pooled aspHGF had a higher shear viscosity than FaSSGF (1.1 ± 0.0 mPa⋅s) and 0.1 M HCl (1.1 ± 0.3 mPa⋅s).The rheological profile of simHGF* was identical to that of simHGF (and was therefore not shown in Fig. 1).N-simHGF containing 0.5 %(w/v) MC 15 mPa⋅s showed Newtonian behavior at shear rates above 5.6 s − 1 , where its viscosity was independent of the shear rate.The viscosity of N-simHGF at a shear rate of 56.2 s − 1 was 2.0 ± 0.0 mPa⋅s.
The osmolality of dissolution media has previously been shown to impact on drug dissolution (Jantratid et al., 2008;Streng et al., 1984).Ionic strength can have an impact on the interaction of the protonated compounds and taurocholate by facilitating the formation of insoluble salts and this interaction might have a significant impact on the solubility and thereby the dissolution profile (Erceg et al., 2012;Reppas and Vertzoni, 2012;Streng et al., 1984;Vertzoni et al., 2005).FaSSGF* was developed with an osmolality identical to aspHGF, in order to evaluate whether an increased osmolality would affect the IDR of cinnarizine.The IDR of cinnarizine measured in FaSSGF* (426 ± 7 µg/(cm 2 ⋅min)) and FaSSGF (444 ± 7 µg/(cm 2 ⋅min)) were similar, and thus no effect of osmolality was observed (Fig. 2).
No significant differences were observed between the IDR of cinnarizine in simHGF* and aspHGF due to similar contents, characteristics and rheological behavior of these two media (Figs. 1, 2, and Table 3).However, the IDR of cinnarizine in aspHGF and simHGF* was significantly lower (approximately 50 %) than in FaSSGF and FaSSGF*, presumably due to the differences in the viscosity (Figs. 1 and 2).This is in accordance with the Noyes Whitney and Stoke Einstein equation (Eq.( 1) and ( 2)) as the viscosity of simHGF* and pooled aspHGF was measured to be approximately twice the viscosity of FaSSGF and FaSSGF*.

Influence of Shear-thinning and Newtonian fluids on the IDR of cinnarizine
In order to evaluate the effect of the shear thinning properties of aspHGF and simHGF on the IDR of cinnarizine, the IDR was determined in the shear thinning media simHGF, as well as in the Newtonian fluids N-simHGF and FaSSGF.To evaluate the shear thinning behavior, the IDR experiments were performed at two different shear stresses, using stirring rates of 50 rpm and 150 rpm.Fig. 3 shows the measured IDRs of cinnarizine in simHGF, N-simHGF and FaSSGF, with stirring rates of 50 and 150 rpm.
As shown in Fig. 3, the IDRs of cinnarizine measured in simHGF at shear rates of 50 rpm (49 ± 5 µg/(cm2⋅min)) and 150 rpm (67 ± 2 µg/ (cm2⋅min)) were significantly different (p < 0.001), due to the shear thinning properties of simHGF (Fig. 1).The corresponding viscosity differences in the simHGF at the different shear rates were measured to  indicates the shear rate interval, which is comparable to the forces applied to drug delivery systems in the antrum of the fasted stomach (Bennett et al., 2009).be approximately 2.7 mPa⋅s and 2.2 mPa⋅s at 50 rpm and 150 rpm, respectively (Fig. 1B).There was no significant difference in the IDR of cinnarizine measured in N-simHGF at shear rates of 50 rpm (64 ± 9 µg/ (cm2⋅min)) and 150 rpm (69 ± 1 µg/(cm2⋅min)).This is in agreement with theory, as the viscosity is independent on the shear rate in a Newtonian medium and thereby the diffusion coefficient is constant, resulting in an unchanged IDR.A similar trend was observed when measuring the IDR of cinnarizine in the Newtonian FaSSGF where the IDR was determined to be 444 ± 7 µg/(cm2⋅min) and 429 ± 6 µg/ (cm2⋅min) at shear rates of 50 rpm and 150 rpm, respectively.The motility pattern varies with location in the human stomach even in the fasted state (Bennett et al., 2009;Hasler, 2009).In the proximal part of the stomach (fundus) slow and sustained contractions are present (0.5 cm/s proximal body), while in the antrum distally of the stomach contractions occur more often (4 cm/s) (Hasler and Yamada, 2009).It has been found that the fundus motility correspond to non or very slow shear rates of approximately 0.1 s − 1 whereas the antrum motility corresponds to shear rates of 30-60 s − 1 (Bennett et al., 2009).The motility differences in the stomach have been shown to affect the viscosity of aspHGF due to the shear thinning properties (Allen et al., 1984;Pedersen et al., 2013).Thus, different location of a dosage form in the stomach after administration might have a significant impact on the dissolution rate of the drug and thereby on the absorption especially for poorly soluble drugs.Large differences of the dissolution behavior are expected to occur depending on whether the drug dissolution will be in the fundus or antrum of the stomach due to differences in motility pattern and thus, large viscosity differences.Using FaSSGF to evaluate fundic drug release might overestimate the IDR by a factor of 100-10.000,due to viscosity differences between aspHGF and FaSSGF in the range 100-10.000mPa⋅s under conditions as in the fundus.

Conclusions
When evaluating the dissolution rate of weakly basic drugs in the fasted stomach, parameters such as pH, viscosity and motility pattern should be considered.Even small pH differences in simulated human gastric fluids lead to significant differences in the IDR of the weakly basic model drug, cinnarizine.The different motility patterns present in the human stomach might induce significant differences in dissolution of weakly basic drugs due to the shear thinning properties of aspHGF.Thus, the use of Newtonian FaSSGF, with viscosity similar to water, to evaluate drug dissolution might overestimate the fundic drug IDR by a factor of 100-10.000,compared to the non-Newtonian, more viscous, fluid in the human stomach.Therefore, the application of media with viscosity as gastric fluids are recommended to evaluate gastric drug dissolution in vitro.

Fig. 3 .
Fig. 3. Intrinsic dissolution rates obtained for cinnarizine in simHGF, N-simHGF and FaSSGF, with stirring rates of 50 and 150 rpm.Please note that simHGF 2.5 and the FaSSGF 1.6 at 50 rpm are also included in Fig. 2, and are added to Fig. 3 to facilitate comparison.

Table 1
Physico-chemical characteristics of the model compound cinnarizine.

Table 2
Composition of the utilized simulated human gastric media.
ǂ Due to batch variations, the concentration of the MC 20.000 mPa⋅s was changed from 0.2 %(w/v) to 0.125 %(w/v) to obtain a rheological profile of simHGF comparable to aspHGF in the present study, see Fig.1.P.B.Pedersen et al.

Table 3
Composition and characteristics of the aspirated and the simulated gastric fluids.
a Set values.