- Email: [email protected]
In situ dissolution testing using potentiometric sensors
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/ejps
In situ dissolution testing using potentiometric sensors Karl Peeters a,∗ , Roy De Maesschalck a , Hugo Bohets b , Koen Vanhoutte a , Luc Nagels c a
Johnson & Johnson Pharmaceutical Research and Development, a Division of Janssen Pharmaceutica NV, Turnhoutseweg 30, B-2340 Beerse, Belgium b Octens, Drie eikenstraat 661, 2650 Edegem, Belgium c University of Antwerp (UIA), Chemistry Department, Groenenborgerlaan 171, 2610 Wilrijk, Belgium
a r t i c l e
i n f o
a b s t r a c t
Article history:
Potentiometric sensors can be used to determine the amount of API dissolved in the dis-
Received 25 January 2008
solution medium in function of time by measuring directly in the dissolution vessel of a
Received in revised form
Paddle (USP type 2) and Basket (USP type 1) apparatus. The prototype potentiometric sen-
21 April 2008
sor instrumentation showed very promising results for a selection of APIs with different
Accepted 22 April 2008
physico-chemical properties. The applicability, benefits and limitations of the prototype
Published on line 3 May 2008
were explored. The applicability of the measurement technique strongly depends on the log(P) of the API. Here, it is shown that measurements can easily be performed for APIs with
Keywords:
a log(P) > 4. Electrode performance however decreases with decreasing log P of the APIs due to
In situ dissolution
decreased drug selectivity in comparison to the excipients and ionic strength of the applied
Potentiometric sensor
dissolution medium. The potentiometric sensors are shown to be insensitive towards undis-
Fiber optics
solved particles and air bubbles as opposed to UV spectrometric measurement where these can lead to severe light scattering. For the tested APIs, the obtained dissolution profiles are very reproducible and show a low variation compared to the measurements using manual sampling and UV or HPLC analysis. The measurements demonstrate that potentiometric sensors are a very promising technology that can become a standard for in situ dissolution measurements. © 2008 Elsevier B.V. All rights reserved.
1.
Introduction
Dissolution testing is one of the most important analytical techniques in the pharmaceutical industry for testing the quality of mainly solid oral drug delivery systems such as tablets and capsules. Over the years, several dissolution apparatus have been standardised as described in the different pharmacopeia. The most applied are the USP type 1 (basket) and 2 (paddle) apparatus. For these apparatus several levels of automated systems are currently available on the market. Automation becomes more important since the number of samples to be analysed increased dramatically in the last years
∗
Corresponding author. Tel.: +32 14603953; fax: +32 14605838. E-mail address: [email protected] (K. Peeters). 0928-0987/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ejps.2008.04.009
and will still increase in the years to come. This is mainly caused by an increased number of stability testing conditions needed for a global registration application and the full characterisation of the dosage form design space as determined in the quality by design (QbD) principles. Samples can be collected by connecting sampling needles to a pump and sample collector. The different types of pumps such as peristaltic, syringe or piston pumps collect the samples at the given times in sample tubes or HPLC vials. An in-line filter can be applied to remove undissolved active and excipients from the sample. The concentration determination is generally performed off-line by transferring the sample tubes or vials manually to
244
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
the UV spectrophotometer or HPLC system. The transfer of the samples to the HPLC system can also be automated (at-line). A next step in automation is performing the measurements on-line. This can be achieved by using a pump in combination with a flow-trough UV cell measurement system. In this way more measurements can be taken, e.g. typically each 2 min. Since the measurement is non-destructive, the medium can also be recirculated to the vessel. On-line measurement using HPLC are also possible when combining, e.g. 6 HPLC systems in parallel. This technology is however not common at the moment. The fastest results can be obtained when no pump or other sampling systems are required and the measurements are performed in-line (in situ). This could be achieved by the introduction of UV fiber optic technology (Cho et al., 1995; Aldridge and Kostek, 1995; Bynum et al., 1999; Johansson et al., 2002; Gray, 2003). Several commercial instruments were successfully introduced on the market such as, e.g. the Rainbow Dynamic Dissolution Monitor by Delphian Technology (pION Inc.) (Schatz et al., 2001), the IO Fiber Optic Dissolution System (C-Technologies), the Opt-diss UV system (Leap Technologies Inc.) or the Erweka-Zeiss ZODIAC (Carl Zeiss) (Lu et al., 2003). The possibility to perform much more measurements, e.g. each 20 s, renders dissolution profiles much more informative. It is possible to observe effects such as the lag period of tablets, i.e. the time needed for wetting of the tablet prior disintegration and dissolution. The fast measurements also allow to measure the dissolution profile of fast dissolving dosage forms such as, e.g. “quicksolve” tablets. This is very valuable information for the formulators, which enables them to further optimise the dosage form (Graffner, 2006). The measurements also provide immediate results, increasing the number of analysis, which can be performed in the same timeframe compared to the use of off-line HPLC. The fiber optic systems are however not applicable for every product since disintegrated dosage forms resulting in a large amount of undissolved particles lead to important scattering of the UV light. Filtration, which is used in off and on-line measurements methods, is not possible. Also air bubbles lead to light scattering and must be avoided by degassing the dissolution media. The light scattering effects can partially be corrected using algorithms such as the baseline correction or second derivative (Schatz et al., 2000). These algorithms are not able to correct when the fiber optic transflectance probe becomes coated with undissolved material. As is the case for traditional UV spectroscopic assay, the presence of UV active excipients, the relatively poor sensitivity for highly potent drugs and the narrow linearity range also can be an issue. To avoid these problems and to maximise the number of in-line dissolution tests, it is important that new technology platforms are introduced for dissolution testing. Potentiometric sensor technology is such a new platform. In this article, dissolution profiles of basic drugs with a log(P) > 4 determined with potentiometric sensors are compared with dissolution profiles using manual sampling and HPLC analysis, and with profiles using UV fiber optics technology. Furthermore, it is demonstrated that the response time of the sensors is fast enough to measure fast releasing products. Also the limitations of the potentiometric sensor are discussed and the possibility to work in very turbid media is demonstrated.
2.
Materials and methods
2.1. Experimental set-up electrochemical dissolution testing The dissolution experiments are performed using a Distek 2100C dissolution system (USP apparatus 2/paddle) with modified lids to position the potentiometric sensors in the sampling zone specified in the pharmacopoeia, see Fig. 1. All measurements are performed on a modified eight-channel data acquisition station (D130) from Consort (Belgium) controlled by homemade LabVIEW 7-based software. Each vessel contained a Hamilton Polyplast Pro RX electrode (Switzerland), which was used as reference and a homemade membranebased indicator sensor (Bohets et al., 2007). The indicator sensors are preconditioned before use in the dissolution medium containing 110% of the drug substance. The sensors are currently used as consumables, i.e. new sensors are preconditioned for each API. The sensors can then be used for numerous dissolution runs for the same API (6 month expected live span). The preconditioning time is dissolution medium, temperature and log P dependent. For domperidone (log P = 4.01), the sensors were conditioned overnight in 0.01 M HCl at 37 ◦ C. When the sensors are sufficiently conditioned a
Fig. 1 – Detail of a dissolution vessel with an opened lid containing a potentiometric sensor and a reference electrode.
245
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
Table 1 – Preconditioning conditions for the model drugs Drug
Medium
Temperature (◦ C)
rpm
Loperamide Cinnirazine Domperidone Ketoconazole
0.01 M HCl 0.1 M HCl 0.01 M HCl 0.01 M HCl
37 37 37 37
100 75 75 75
stable signal is obtained. A detailed study of the preconditioning time of these sensors is beyond the scope of this work. An overview of the preconditioning conditions is shown in Table 1. Before the dissolution run, the conditioned sensors are calibrated as explained in more detail in the next sections.
2.2.
Non-electrochemical dissolution parameters
The manual sampling was performed using syringes attached to Distek sampling needles. The filtered samples were manually transferred to sample tubes or HPLC vials. Off-line HPLC analysis was performed using a Waters Alliance 2695 System with a Waters 2996 PDA detector and Empower software. The UV fiber optic measurements were performed on a Sotax AT7 Smart dissolution bath with the UV fiber optic probes of a Rainbow Dynamic Dissolution Monitor (Delphian/pION Inc.) positioned into the hollow shafts. HPLC electron spray mass spectrometry is performed with a Waters Alliance 2695 HPLC coupled with a Waters ZQ2000 mass spectrometer and Empower software.
2.3.
Preconditioning time (h) 36 36 12 –
The formulation and some physicochemical parameters of the analysed APIs are shown in Table 2. Loperamide capsules were tested at 100 rpm in 0.01 M HCl using a four spiral closed sinker. For the manual sampling, a Millipore 25 mm diameter disk filter with a 0.45 m PTFE membrane was used. The samples were analysed with HPLC using a YMC-pack Pro C18 column, 3 m, 100 mm × 4.6 mm i.d. at 35 ◦ C, a flow rate of 1.5 ml/min, injection volume of 100 l, mobile phase 35% 0.5% ammonium acetate in water, 33% acetonitrile and 32% methanol with UV detection at 220 nm. Cinnarizine tablets are tested at 75 rpm in 0.1 M HCl. Dissolution profiles were recorded in situ using UV fiber optic probes with a 5 mm path length at 254 nm. Ketoconazole tablets were tested at 75 rpm in 0.01 M HCl with the fiber optics with a path length of 5 mm at 253 nm. Domperidone melt tablets were measured at 75 rpm in 0.01 M HCl using UV fiber optic probes with a path length of 20 mm at 284 nm with and without Japanese Baskets.
3.
Results and discussion
3.1.
Loperamide
Chemicals and model drugs
Hydrochloric acid is purchased from Acros. Four basic drugs, loperamide, cinnarizine, ketoconazole and domperidone, are selected for this study. They are selected because of their positive charge in the detergent free dissolution medium and their log P > 4. In previous studies (Bohets et al., 2007) it is shown that potentiometric dissolution profiling of basic drugs with log(P) values as low as 1.79 are possible. The performance of the potentiometric sensors, however are very much dependent from the ionic background for drugs with log P values lower than 4 resulting in difficult quantification if possible. The respective reference standards and formulations are produced internally, and therefore were freely available as well as their dissolution method descriptions. Loperamide capsules 2 mg (Imodium® ), cinnarizine tablets 25 mg (Stugeron® ), ketoconazole tablets 200 mg (Nizoral® ), and domperidone instant melt tablets 10 mg (Motilium Instant® ) are analysed for dissolution. All dissolution profiles are obtained in an USP type 2 apparatus at 37 ◦ C.
Fig. 2a shows an example of the in-line response curve of one potentiometric sensor for the dissolution of a 2 mg loperamide capsule in 0.01 M HCl in an USP type 2 apparatus at 100 rpm. When bringing the capsules in the vessels, due to the sinker, they immediately sink to the bottom. Visually it can be seen that the hard gelatine capsules open after about 2 min in the dissolution medium. This lag time can be seen on the profile (Fig. 2b). After the opening of the capsules, the API and excipients come out of the dissolving capsule. The loperamide dissolves very fast, leading to a plateau value after about 15–20 min. In the intercept of Fig. 2a, a potential versus concentration plot of the standard solution is shown, indicating a logarithmic relationship for the concentration range of 0.24–2.1 mg loperamide per 900 ml. The data for this plot are obtained under the same experimental conditions as the dissolution curves. The only difference is that instead of the loperamide capsules, seven subsequent loperamide stan-
Table 2 – Overview of some important physicochemical parameters of the analysed APIs Drug
log P
log D (at pH)
pKa1
Loperamide Cinnirazine Domperidone Ketoconazole
5.13 5.60 4.01 4.40
0.96 (2) – 0.73 (2) −0.96 (2)
8.66 7 7.8 6.51
pKa2 – <3 – 2.94
Label claim (mg) 2 25 10 200
246
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
Fig. 2 – Example of an in-line potentiometric response curve of a 2 mg loperamide formulation of one potentiometric sensor and a potential vs. concentration plot of a loperamide standard (a) and the corresponding dissolution profile with the corresponding potential vs. log c plot (b).
dard additions are performed yielding a 2.2 mg/900 ml end concentration (110%) in the dissolution medium. After each addition, the potential is evaluated over a period of 10 min. Plotting the potential against the logarithm of the concentration in the same concentration range results in a linear relationship (intercept of Fig. 2b). This is the calibration curve which is used for translating the measured potential into the API concentration and % dissolved. The calculated R2 values are 0.9999 or better for the tested sensors. Because of the good linear fit of the calibration curve, the number of calibration points can be drastically decreased to speed up the calibration procedure. After the first preconditioning, the seven data points were taken to check the electrode quality. As soon as the calibration curve of an electrode shows an R2 value of 0.9995 or better for the linear fit in the discussed concentration range, the electrode quality is considered to be good. The conditioned electrodes with the required quality can now be calibrated using only two data points (e.g. at 30% and 110% of the loperamide label claim). The calibration curve of the six sensors shows an average slope of 60.09 ± 1.91 mV which approaches the theoretical maximum of 61.5 mV for a single charged component at 37 ◦ C according to the Nicholskii–Eisenman equation (Bohets et al., 2007). Additional to the calibration curves, an endpoint calibration is used to correct for the potential drift of the sensors during the dissolution experiment. This endpoint calibration is performed
by determining the sensor potential immediately after the dissolution in the loperamide standard solution with the highest concentration used during calibration curve determination under the same experimental conditions as the dissolution (mostly 110% of the label claim). This leads to the final dissolution profile shown in Fig. 2b. An average of six potentiometric dissolution profiles are compared with the dissolution profiles obtained by manual sampling with HPLC analysis (n = 6) (Fig. 3). Note that UV fiber optic measurements were not possible for
Fig. 3 – Comparison of the dissolution profiles of a 2 mg loperamide formulation obtained by manual sampling with HPLC analysis (n = 6) () and by potentiometric analysis (black full line). The data points after 5, 10, 15, 20, 30, 45 and 60 min are shown together with the standard deviations.
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
3.3.
Fig. 4 – Comparison of the dissolution profiles of a 25 mg cinnarizine formulation determined by potentiometry (n = 6) (), and by fiber optics (n = 6) with (䊉) and without () flow disturbance in the vessel and the corresponding standard deviations.
these capsules due to the low UV absorbance of loperamide at wavelengths above 240 nm. Mathematical comparison of the shown dissolution profiles using the f1 and f2 tests, indicate that they can be considered as similar (f1 = 9.21 and f2 = 61.88 for the data points after 5, 10 and 15 min) (Moore and Flanner, 1996). Furthermore, it can be seen that the standard deviations on the dissolution profiles for the two techniques are similar.
3.2.
Cinnarizine
Similar as for loperamide, dissolution profiles are determined with the potentiometric sensors for a 25 mg cinnarizine tablet formulation. In Fig. 4, the dissolution profiles measured with the potentiometric sensors are compared with the dissolution profiles obtained by the in situ UV measurements using fiber optic probes. The immediate release tablets show a relatively fast dissolution reaching a plateau after about 15–20 min. The average profile of the UV fiber optic measurements is somewhat slower than the average potentiometric dissolution profile. This can be caused by the absence of the reference electrode (present during the potentiometric measurements) in the dissolution vessel. By visual observation, one can see that most powder of the disintegrated tablet remains located at the bottom of the vessel before dissolving. The presence of the relatively thick reference electrode (diameter = 1.2 cm) of the prototype submerged in the dissolution medium creates a more turbulent flow at the bottom of the vessel. This causes the powder of the disintegrated tablet to be more dispersed in the medium and as such a faster dissolution is obtained. When positioning the same reference electrode in the vessels during a dissolution experiment with the UV fiber optics, this causes an equivalent increase in dissolution rate as was observed for the potentiometric analysis (Fig. 4). Despite the disturbance of the hydrodynamics, the f1 –f2 comparison using the first three data points (4, 8 and 12 min) shows that these profiles can be considered as similar because f1 = 5.51 which is lower than 15 and f2 = 70.77 which is higher than 50 (Moore and Flanner, 1996). To circumvent possible issue of disturbing the hydrodynamics in the vessels, a new prototype is under development with miniaturised reference electrodes with a much smaller diameter of 3 mm.
247
Ketoconazole
A third API tested with the potentiometric sensors is ketoconazole. After a first overnight preconditioning phase, the ketoconazole sensors perform well although they do not reach the high quality standard discussed earlier in this work indicating that the preconditioning is not finished. After an extra overnight preconditioning phase in the same ketoconazole solution, the electrode quality decreased dramatically and it decreased further during sensor storage in the dissolution medium. Conditioning failed indicative of contamination of the sensor with a higher log P product. Using HPLC electron spray mass spectrometry, it was shown that ketoconazole slowly deacetylates in the dissolution medium (0.01 M HCl), yielding a higher log P product. This higher log P product interferes with the equilibration of the potentiometric sensor during the preconditioning and storage of the sensor. The result is a sensor conditioned for a higher log P compound. This causes a serious decrease of the sensitivity towards the API (Bohets et al., 2007) and accordingly strongly deviating dissolution results. To circumvent this problem during long-term storage of the sensor for unstable products, the sensors are stored in water-saturated air (of 100% humidity) instead of in the dissolution medium containing the API.
3.4.
Domperidone
A fourth product, domperidone instant melt tablets, is very useful for investigating the response time of the sensors. The instant melt tablets were designed to be released instantaneously in contact with aqueous media. When the instant melt tablets make contact with the 0.01 M HCl dissolution medium, they disintegrate immediately while producing a white foam on top of the medium. The result is an inhomogeneously distributed domperidone in the dissolution medium yielding a peak concentration at the fiber optics probe after 18 s (Fig. 5a). To circumvent this peak concentration in the dissolution profile, Japanese sinker baskets are used to force the instant melt tablets to the bottom of the dissolution vessel. On the bottom of the dissolution vessel it is seen that the instant melt tablets disintegrate within a few seconds and the domperidone dissolves immediately. Since the domperidone dissolves in the medium near the paddle, homogenisation of the solution is much faster circumventing a peak concentration at the fiber optics probe (Fig. 5a). Comparison of this fiber optics dissolution profile with the dissolution profile obtained with the potentiometric sensor shows that at the beginning of the dissolution profile a deviation of about 6 s is noticed (Fig. 5b). This deviation is due to the nature of the data acquisition system giving a time offset (delay 6 s). Internal triggering will be used in the near future to eliminate this offset. At the moment however, the potentiometric dissolution profile is shifted 6 s to the right (Fig. 5) because of this time offset. After correction for the shift, mathematical comparison of the profiles for the first 10 data points obtained by the fiber optics (43 s, 3 s between the first two data points, 5 s for each next data point) show that these profiles are similar giving f1 = 6.60 and f2 = 67.14. This similarity indicates that the lag time of the sensor because of the ion equilibrium is negligible for the domperidone sensors resulting in a fast sensor response. This
248
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
Fig. 5 – Dissolution profiles of a 10 mg domperidone instant tablet with (black) and without (gray) Japanese sinker determined with fiber optics (a) and potentiometric sensors (b) with the corresponding standard deviation (n = 6).
fast sensor response, together with a sampling interval of 4 s makes the potentiometric sensor a good analytical tool for the dissolution profiling of fast releasing products.
3.5.
Dissolution testing in turbid media
Apart from the response time, also the influence of undissolved dispersed particles in the dissolution medium on the electrode performance is tested. In Fig. 6, a part of the dissolution plateau of a 25 mg cinnarizine tablet in 900 ml 0.1 M HCl measured with the potentiometric sensors is shown. After 6 min, 10 g microcrystalline cellulose is added to the solution. The microcrystalline cellulose added to the vessel does
not dissolve but disperses into the medium causing a white opaque suspension. In this suspension, UV fiber optic measurements are not possible due to the extensive scatter and fouling of the fiber optic windows. However, only a small disturbance of the potential, caused by the manual addition of the microcrystalline cellulose, is noticed when using the potentiometric sensors. After the addition, the potential stabilises to its original value within 10 s indicating the insensitivity towards undissolved particles in the dissolution medium. These measurements demonstrate that the potentiometric sensor technology can probably also be used in complex media such as, e.g. milk, a medium that has been used to simulate the fed stomach and to perform in vivo–vitro correlations in several studies (Nicolaides et al., 1999; Dressman and Reppas, 2000; Jantratid et al., 2008). It is however impossible to measure in milk with fiber optic probes and very complex to measure with HPLC. The technology has also the potential to be used for measuring nanosuspensions for which the particles are too small to or very difficult to filter prior traditional analysis and for which again UV fiber optics are not suitable due to the high degree of light scatter. These promising applications will be further explored.
3.6.
Future plans
Other experiments running with the potentiometric sensors at the moment are the dissolution testing of controlled release formulations and the dissolution testing in media containing certain surfactants. First controlled release tests show good results indicating that depending on the formulation, two to five extra calibration points have to be determined in 24 h. Experiments in media containing surfactants, however showed that the tested sensors were not performing well when conducting dissolution measurements in media containing polysorbate 20 (Tween® 20). Tests in media containing Brij® 35 however showed good results. Also tests in bio-relevant media (e.g. fed state simulated intestine fluid (FeSSIF) and fasted state simulated intestine fluid (FaSSIF)) are planned in the future. Comparable problems as with sur-
Fig. 6 – Plateau of a dissolution profile of a 25 mg cinnarizine tablet in 900 ml 0.1 M HCl before and after (arrow) addition of 10 g microcrystalline cellulose and a picture of the corresponding vessel.
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
factant containing media are expected since in these media, the surfactant sodium taurocholate, is present. Furthermore, dissolution profiling of lower log P drugs are planned and at the moment, a next generation of potentiometric sensors are being developed and have to be explored in further research.
4.
Conclusions
In the present work, it is shown that potentiometric sensors can be used for in situ dissolution testing. The comparison of the dissolution profiles obtained with the potentiometric sensors show indeed good agreement with these of currently available techniques, i.e. manual sampling with HPLC analysis and UV fiber optics. Furthermore, it is shown that the potentiometric sensors have a good response time for fast releasing formulations and no interference with undissolved particles or air bubbles in the dissolution vessel, a clear advantage in comparison with UV fiber optics. Their possible use for, e.g. nanosuspensions or measurements in milk has to be explored. The sensors have the disadvantage that for each API a new set of sensors needs to be preconditioned. Because the potentiometric electrodes are very cheap in comparison to, e.g. fiber optic probes, they may be considered as consumables. The tested prototype needs further development to extend its applicability towards other dissolution media (e.g. media containing frequently used surfactants such as, e.g. polysorbate 20) and the diameter of the reference electrode needs to be decreased to limit as much as possible the disturbance of the flow dynamics in the vessel.
Acknowledgements The authors thank IWT (Institute for the Promotion of Innovation by Science and Technology in Flanders) for financial support (grant 040399), AIC (Antwerps Innovatie Centrum) for financial and organisational support. Furthermore, we thank T. Stockmans, N. Schoonvaere and D. Peeters for their practical help with the dissolution testing and N. Vervoort for his help with the HPLC electron spray mass spectrometry.
249
references
Aldridge, P.K., Kostek, L.J., 1995. In situ fiber optic dissolution analysis. Dissolut. Technol. 2 (4), 10–11. Bohets, H., Vanhoutte, K., De Maesschalck, R., Cockaerts, P., Vissers, B., Nagels, L.J., 2007. Development of in situ selective sensors for dissolution. Anal. Chim. Acta 581, 181–191. Bynum, K.C., Kraft, E., Pocreva, J., Ciurczak, E.W., Palermo, P., 1999. In situ dissolution testing using a UV fiber optic probe dissolution system. Dissolut. Technol. 6 (4), 11. Cho, J.H., Gemperline, P.J., Salt, A., Walker, D.S., 1995. UV/visible spectral dissolution monitoring by in situ fiber-optic probes. Anal. Chem. 67, 2858–2863. Dressman, J.B., Reppas, C., 2000. In vitro–in vivo correlations for lipophilic, poorly water-soluble drugs. Eur. J. Pharm. Sci. 11 (Suppl. 2), S73–S80. Graffner, C., 2006. Regulatory aspects of drug dissolution from a European perspective. Eur. J. Pharm. Sci. 29, 288–293. Gray, V., 2003. Dissolution testing using fiber optics—a regulatory perspective. Am. Pharm. Rev. 6 (2), 26–30. Jantratid, E., Janssen, N., Reppas, C., Dressman, J.B., 2008. Dissolution media simulating conditions in the proximal human gastrointestinal tract: an update. Pharm. Res. 25, 1663–1676. Johansson, J., Cauchi, M., Sundgren, M., 2002. Multiple fiber optic dual-beam UV/vis system with application to dissolution testing. J. Pharm. Bio. Anal. 29, 469–476. Lu, X., Lozano, R., Shah, P., 2003. In situ dissolution testing using different UV fiber optic probes and instruments. Dissolut. Technol. 10 (4), 6–15. Moore, J.W., Flanner, H.H., 1996. Mathematical comparison of dissolution profiles. Pharm. Tech. 20 (6), 64–74. Nicolaides, E., Galia, E., Efthymiopoulos, C., Dressman, J.B., Reppas, C., 1999. Forecasting the in vivo performance of four low solubility drugs from their in vitro dissolution data. Pharm. Res. 16, 1876–1882. Schatz, C., Ulmschneider, M., Altermatt, R., Marrer, S., 2000. Evaluation of the Rainbow Dynamic Dissolution MonitorTM semi-automatic fiber optic dissolution tester. Dissolut. Technol. 7 (4), 8–17. Schatz, C., Ulmschneider, M., Altermatt, R., Marrer, S., Altorfer, H., 2001. Thoughts on fiber optics in dissolution testing. Dissolut. Technol. 8 (2), 1–5.
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/ejps
In situ dissolution testing using potentiometric sensors Karl Peeters a,∗ , Roy De Maesschalck a , Hugo Bohets b , Koen Vanhoutte a , Luc Nagels c a
Johnson & Johnson Pharmaceutical Research and Development, a Division of Janssen Pharmaceutica NV, Turnhoutseweg 30, B-2340 Beerse, Belgium b Octens, Drie eikenstraat 661, 2650 Edegem, Belgium c University of Antwerp (UIA), Chemistry Department, Groenenborgerlaan 171, 2610 Wilrijk, Belgium
a r t i c l e
i n f o
a b s t r a c t
Article history:
Potentiometric sensors can be used to determine the amount of API dissolved in the dis-
Received 25 January 2008
solution medium in function of time by measuring directly in the dissolution vessel of a
Received in revised form
Paddle (USP type 2) and Basket (USP type 1) apparatus. The prototype potentiometric sen-
21 April 2008
sor instrumentation showed very promising results for a selection of APIs with different
Accepted 22 April 2008
physico-chemical properties. The applicability, benefits and limitations of the prototype
Published on line 3 May 2008
were explored. The applicability of the measurement technique strongly depends on the log(P) of the API. Here, it is shown that measurements can easily be performed for APIs with
Keywords:
a log(P) > 4. Electrode performance however decreases with decreasing log P of the APIs due to
In situ dissolution
decreased drug selectivity in comparison to the excipients and ionic strength of the applied
Potentiometric sensor
dissolution medium. The potentiometric sensors are shown to be insensitive towards undis-
Fiber optics
solved particles and air bubbles as opposed to UV spectrometric measurement where these can lead to severe light scattering. For the tested APIs, the obtained dissolution profiles are very reproducible and show a low variation compared to the measurements using manual sampling and UV or HPLC analysis. The measurements demonstrate that potentiometric sensors are a very promising technology that can become a standard for in situ dissolution measurements. © 2008 Elsevier B.V. All rights reserved.
1.
Introduction
Dissolution testing is one of the most important analytical techniques in the pharmaceutical industry for testing the quality of mainly solid oral drug delivery systems such as tablets and capsules. Over the years, several dissolution apparatus have been standardised as described in the different pharmacopeia. The most applied are the USP type 1 (basket) and 2 (paddle) apparatus. For these apparatus several levels of automated systems are currently available on the market. Automation becomes more important since the number of samples to be analysed increased dramatically in the last years
∗
Corresponding author. Tel.: +32 14603953; fax: +32 14605838. E-mail address: [email protected] (K. Peeters). 0928-0987/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ejps.2008.04.009
and will still increase in the years to come. This is mainly caused by an increased number of stability testing conditions needed for a global registration application and the full characterisation of the dosage form design space as determined in the quality by design (QbD) principles. Samples can be collected by connecting sampling needles to a pump and sample collector. The different types of pumps such as peristaltic, syringe or piston pumps collect the samples at the given times in sample tubes or HPLC vials. An in-line filter can be applied to remove undissolved active and excipients from the sample. The concentration determination is generally performed off-line by transferring the sample tubes or vials manually to
244
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
the UV spectrophotometer or HPLC system. The transfer of the samples to the HPLC system can also be automated (at-line). A next step in automation is performing the measurements on-line. This can be achieved by using a pump in combination with a flow-trough UV cell measurement system. In this way more measurements can be taken, e.g. typically each 2 min. Since the measurement is non-destructive, the medium can also be recirculated to the vessel. On-line measurement using HPLC are also possible when combining, e.g. 6 HPLC systems in parallel. This technology is however not common at the moment. The fastest results can be obtained when no pump or other sampling systems are required and the measurements are performed in-line (in situ). This could be achieved by the introduction of UV fiber optic technology (Cho et al., 1995; Aldridge and Kostek, 1995; Bynum et al., 1999; Johansson et al., 2002; Gray, 2003). Several commercial instruments were successfully introduced on the market such as, e.g. the Rainbow Dynamic Dissolution Monitor by Delphian Technology (pION Inc.) (Schatz et al., 2001), the IO Fiber Optic Dissolution System (C-Technologies), the Opt-diss UV system (Leap Technologies Inc.) or the Erweka-Zeiss ZODIAC (Carl Zeiss) (Lu et al., 2003). The possibility to perform much more measurements, e.g. each 20 s, renders dissolution profiles much more informative. It is possible to observe effects such as the lag period of tablets, i.e. the time needed for wetting of the tablet prior disintegration and dissolution. The fast measurements also allow to measure the dissolution profile of fast dissolving dosage forms such as, e.g. “quicksolve” tablets. This is very valuable information for the formulators, which enables them to further optimise the dosage form (Graffner, 2006). The measurements also provide immediate results, increasing the number of analysis, which can be performed in the same timeframe compared to the use of off-line HPLC. The fiber optic systems are however not applicable for every product since disintegrated dosage forms resulting in a large amount of undissolved particles lead to important scattering of the UV light. Filtration, which is used in off and on-line measurements methods, is not possible. Also air bubbles lead to light scattering and must be avoided by degassing the dissolution media. The light scattering effects can partially be corrected using algorithms such as the baseline correction or second derivative (Schatz et al., 2000). These algorithms are not able to correct when the fiber optic transflectance probe becomes coated with undissolved material. As is the case for traditional UV spectroscopic assay, the presence of UV active excipients, the relatively poor sensitivity for highly potent drugs and the narrow linearity range also can be an issue. To avoid these problems and to maximise the number of in-line dissolution tests, it is important that new technology platforms are introduced for dissolution testing. Potentiometric sensor technology is such a new platform. In this article, dissolution profiles of basic drugs with a log(P) > 4 determined with potentiometric sensors are compared with dissolution profiles using manual sampling and HPLC analysis, and with profiles using UV fiber optics technology. Furthermore, it is demonstrated that the response time of the sensors is fast enough to measure fast releasing products. Also the limitations of the potentiometric sensor are discussed and the possibility to work in very turbid media is demonstrated.
2.
Materials and methods
2.1. Experimental set-up electrochemical dissolution testing The dissolution experiments are performed using a Distek 2100C dissolution system (USP apparatus 2/paddle) with modified lids to position the potentiometric sensors in the sampling zone specified in the pharmacopoeia, see Fig. 1. All measurements are performed on a modified eight-channel data acquisition station (D130) from Consort (Belgium) controlled by homemade LabVIEW 7-based software. Each vessel contained a Hamilton Polyplast Pro RX electrode (Switzerland), which was used as reference and a homemade membranebased indicator sensor (Bohets et al., 2007). The indicator sensors are preconditioned before use in the dissolution medium containing 110% of the drug substance. The sensors are currently used as consumables, i.e. new sensors are preconditioned for each API. The sensors can then be used for numerous dissolution runs for the same API (6 month expected live span). The preconditioning time is dissolution medium, temperature and log P dependent. For domperidone (log P = 4.01), the sensors were conditioned overnight in 0.01 M HCl at 37 ◦ C. When the sensors are sufficiently conditioned a
Fig. 1 – Detail of a dissolution vessel with an opened lid containing a potentiometric sensor and a reference electrode.
245
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
Table 1 – Preconditioning conditions for the model drugs Drug
Medium
Temperature (◦ C)
rpm
Loperamide Cinnirazine Domperidone Ketoconazole
0.01 M HCl 0.1 M HCl 0.01 M HCl 0.01 M HCl
37 37 37 37
100 75 75 75
stable signal is obtained. A detailed study of the preconditioning time of these sensors is beyond the scope of this work. An overview of the preconditioning conditions is shown in Table 1. Before the dissolution run, the conditioned sensors are calibrated as explained in more detail in the next sections.
2.2.
Non-electrochemical dissolution parameters
The manual sampling was performed using syringes attached to Distek sampling needles. The filtered samples were manually transferred to sample tubes or HPLC vials. Off-line HPLC analysis was performed using a Waters Alliance 2695 System with a Waters 2996 PDA detector and Empower software. The UV fiber optic measurements were performed on a Sotax AT7 Smart dissolution bath with the UV fiber optic probes of a Rainbow Dynamic Dissolution Monitor (Delphian/pION Inc.) positioned into the hollow shafts. HPLC electron spray mass spectrometry is performed with a Waters Alliance 2695 HPLC coupled with a Waters ZQ2000 mass spectrometer and Empower software.
2.3.
Preconditioning time (h) 36 36 12 –
The formulation and some physicochemical parameters of the analysed APIs are shown in Table 2. Loperamide capsules were tested at 100 rpm in 0.01 M HCl using a four spiral closed sinker. For the manual sampling, a Millipore 25 mm diameter disk filter with a 0.45 m PTFE membrane was used. The samples were analysed with HPLC using a YMC-pack Pro C18 column, 3 m, 100 mm × 4.6 mm i.d. at 35 ◦ C, a flow rate of 1.5 ml/min, injection volume of 100 l, mobile phase 35% 0.5% ammonium acetate in water, 33% acetonitrile and 32% methanol with UV detection at 220 nm. Cinnarizine tablets are tested at 75 rpm in 0.1 M HCl. Dissolution profiles were recorded in situ using UV fiber optic probes with a 5 mm path length at 254 nm. Ketoconazole tablets were tested at 75 rpm in 0.01 M HCl with the fiber optics with a path length of 5 mm at 253 nm. Domperidone melt tablets were measured at 75 rpm in 0.01 M HCl using UV fiber optic probes with a path length of 20 mm at 284 nm with and without Japanese Baskets.
3.
Results and discussion
3.1.
Loperamide
Chemicals and model drugs
Hydrochloric acid is purchased from Acros. Four basic drugs, loperamide, cinnarizine, ketoconazole and domperidone, are selected for this study. They are selected because of their positive charge in the detergent free dissolution medium and their log P > 4. In previous studies (Bohets et al., 2007) it is shown that potentiometric dissolution profiling of basic drugs with log(P) values as low as 1.79 are possible. The performance of the potentiometric sensors, however are very much dependent from the ionic background for drugs with log P values lower than 4 resulting in difficult quantification if possible. The respective reference standards and formulations are produced internally, and therefore were freely available as well as their dissolution method descriptions. Loperamide capsules 2 mg (Imodium® ), cinnarizine tablets 25 mg (Stugeron® ), ketoconazole tablets 200 mg (Nizoral® ), and domperidone instant melt tablets 10 mg (Motilium Instant® ) are analysed for dissolution. All dissolution profiles are obtained in an USP type 2 apparatus at 37 ◦ C.
Fig. 2a shows an example of the in-line response curve of one potentiometric sensor for the dissolution of a 2 mg loperamide capsule in 0.01 M HCl in an USP type 2 apparatus at 100 rpm. When bringing the capsules in the vessels, due to the sinker, they immediately sink to the bottom. Visually it can be seen that the hard gelatine capsules open after about 2 min in the dissolution medium. This lag time can be seen on the profile (Fig. 2b). After the opening of the capsules, the API and excipients come out of the dissolving capsule. The loperamide dissolves very fast, leading to a plateau value after about 15–20 min. In the intercept of Fig. 2a, a potential versus concentration plot of the standard solution is shown, indicating a logarithmic relationship for the concentration range of 0.24–2.1 mg loperamide per 900 ml. The data for this plot are obtained under the same experimental conditions as the dissolution curves. The only difference is that instead of the loperamide capsules, seven subsequent loperamide stan-
Table 2 – Overview of some important physicochemical parameters of the analysed APIs Drug
log P
log D (at pH)
pKa1
Loperamide Cinnirazine Domperidone Ketoconazole
5.13 5.60 4.01 4.40
0.96 (2) – 0.73 (2) −0.96 (2)
8.66 7 7.8 6.51
pKa2 – <3 – 2.94
Label claim (mg) 2 25 10 200
246
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
Fig. 2 – Example of an in-line potentiometric response curve of a 2 mg loperamide formulation of one potentiometric sensor and a potential vs. concentration plot of a loperamide standard (a) and the corresponding dissolution profile with the corresponding potential vs. log c plot (b).
dard additions are performed yielding a 2.2 mg/900 ml end concentration (110%) in the dissolution medium. After each addition, the potential is evaluated over a period of 10 min. Plotting the potential against the logarithm of the concentration in the same concentration range results in a linear relationship (intercept of Fig. 2b). This is the calibration curve which is used for translating the measured potential into the API concentration and % dissolved. The calculated R2 values are 0.9999 or better for the tested sensors. Because of the good linear fit of the calibration curve, the number of calibration points can be drastically decreased to speed up the calibration procedure. After the first preconditioning, the seven data points were taken to check the electrode quality. As soon as the calibration curve of an electrode shows an R2 value of 0.9995 or better for the linear fit in the discussed concentration range, the electrode quality is considered to be good. The conditioned electrodes with the required quality can now be calibrated using only two data points (e.g. at 30% and 110% of the loperamide label claim). The calibration curve of the six sensors shows an average slope of 60.09 ± 1.91 mV which approaches the theoretical maximum of 61.5 mV for a single charged component at 37 ◦ C according to the Nicholskii–Eisenman equation (Bohets et al., 2007). Additional to the calibration curves, an endpoint calibration is used to correct for the potential drift of the sensors during the dissolution experiment. This endpoint calibration is performed
by determining the sensor potential immediately after the dissolution in the loperamide standard solution with the highest concentration used during calibration curve determination under the same experimental conditions as the dissolution (mostly 110% of the label claim). This leads to the final dissolution profile shown in Fig. 2b. An average of six potentiometric dissolution profiles are compared with the dissolution profiles obtained by manual sampling with HPLC analysis (n = 6) (Fig. 3). Note that UV fiber optic measurements were not possible for
Fig. 3 – Comparison of the dissolution profiles of a 2 mg loperamide formulation obtained by manual sampling with HPLC analysis (n = 6) () and by potentiometric analysis (black full line). The data points after 5, 10, 15, 20, 30, 45 and 60 min are shown together with the standard deviations.
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
3.3.
Fig. 4 – Comparison of the dissolution profiles of a 25 mg cinnarizine formulation determined by potentiometry (n = 6) (), and by fiber optics (n = 6) with (䊉) and without () flow disturbance in the vessel and the corresponding standard deviations.
these capsules due to the low UV absorbance of loperamide at wavelengths above 240 nm. Mathematical comparison of the shown dissolution profiles using the f1 and f2 tests, indicate that they can be considered as similar (f1 = 9.21 and f2 = 61.88 for the data points after 5, 10 and 15 min) (Moore and Flanner, 1996). Furthermore, it can be seen that the standard deviations on the dissolution profiles for the two techniques are similar.
3.2.
Cinnarizine
Similar as for loperamide, dissolution profiles are determined with the potentiometric sensors for a 25 mg cinnarizine tablet formulation. In Fig. 4, the dissolution profiles measured with the potentiometric sensors are compared with the dissolution profiles obtained by the in situ UV measurements using fiber optic probes. The immediate release tablets show a relatively fast dissolution reaching a plateau after about 15–20 min. The average profile of the UV fiber optic measurements is somewhat slower than the average potentiometric dissolution profile. This can be caused by the absence of the reference electrode (present during the potentiometric measurements) in the dissolution vessel. By visual observation, one can see that most powder of the disintegrated tablet remains located at the bottom of the vessel before dissolving. The presence of the relatively thick reference electrode (diameter = 1.2 cm) of the prototype submerged in the dissolution medium creates a more turbulent flow at the bottom of the vessel. This causes the powder of the disintegrated tablet to be more dispersed in the medium and as such a faster dissolution is obtained. When positioning the same reference electrode in the vessels during a dissolution experiment with the UV fiber optics, this causes an equivalent increase in dissolution rate as was observed for the potentiometric analysis (Fig. 4). Despite the disturbance of the hydrodynamics, the f1 –f2 comparison using the first three data points (4, 8 and 12 min) shows that these profiles can be considered as similar because f1 = 5.51 which is lower than 15 and f2 = 70.77 which is higher than 50 (Moore and Flanner, 1996). To circumvent possible issue of disturbing the hydrodynamics in the vessels, a new prototype is under development with miniaturised reference electrodes with a much smaller diameter of 3 mm.
247
Ketoconazole
A third API tested with the potentiometric sensors is ketoconazole. After a first overnight preconditioning phase, the ketoconazole sensors perform well although they do not reach the high quality standard discussed earlier in this work indicating that the preconditioning is not finished. After an extra overnight preconditioning phase in the same ketoconazole solution, the electrode quality decreased dramatically and it decreased further during sensor storage in the dissolution medium. Conditioning failed indicative of contamination of the sensor with a higher log P product. Using HPLC electron spray mass spectrometry, it was shown that ketoconazole slowly deacetylates in the dissolution medium (0.01 M HCl), yielding a higher log P product. This higher log P product interferes with the equilibration of the potentiometric sensor during the preconditioning and storage of the sensor. The result is a sensor conditioned for a higher log P compound. This causes a serious decrease of the sensitivity towards the API (Bohets et al., 2007) and accordingly strongly deviating dissolution results. To circumvent this problem during long-term storage of the sensor for unstable products, the sensors are stored in water-saturated air (of 100% humidity) instead of in the dissolution medium containing the API.
3.4.
Domperidone
A fourth product, domperidone instant melt tablets, is very useful for investigating the response time of the sensors. The instant melt tablets were designed to be released instantaneously in contact with aqueous media. When the instant melt tablets make contact with the 0.01 M HCl dissolution medium, they disintegrate immediately while producing a white foam on top of the medium. The result is an inhomogeneously distributed domperidone in the dissolution medium yielding a peak concentration at the fiber optics probe after 18 s (Fig. 5a). To circumvent this peak concentration in the dissolution profile, Japanese sinker baskets are used to force the instant melt tablets to the bottom of the dissolution vessel. On the bottom of the dissolution vessel it is seen that the instant melt tablets disintegrate within a few seconds and the domperidone dissolves immediately. Since the domperidone dissolves in the medium near the paddle, homogenisation of the solution is much faster circumventing a peak concentration at the fiber optics probe (Fig. 5a). Comparison of this fiber optics dissolution profile with the dissolution profile obtained with the potentiometric sensor shows that at the beginning of the dissolution profile a deviation of about 6 s is noticed (Fig. 5b). This deviation is due to the nature of the data acquisition system giving a time offset (delay 6 s). Internal triggering will be used in the near future to eliminate this offset. At the moment however, the potentiometric dissolution profile is shifted 6 s to the right (Fig. 5) because of this time offset. After correction for the shift, mathematical comparison of the profiles for the first 10 data points obtained by the fiber optics (43 s, 3 s between the first two data points, 5 s for each next data point) show that these profiles are similar giving f1 = 6.60 and f2 = 67.14. This similarity indicates that the lag time of the sensor because of the ion equilibrium is negligible for the domperidone sensors resulting in a fast sensor response. This
248
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
Fig. 5 – Dissolution profiles of a 10 mg domperidone instant tablet with (black) and without (gray) Japanese sinker determined with fiber optics (a) and potentiometric sensors (b) with the corresponding standard deviation (n = 6).
fast sensor response, together with a sampling interval of 4 s makes the potentiometric sensor a good analytical tool for the dissolution profiling of fast releasing products.
3.5.
Dissolution testing in turbid media
Apart from the response time, also the influence of undissolved dispersed particles in the dissolution medium on the electrode performance is tested. In Fig. 6, a part of the dissolution plateau of a 25 mg cinnarizine tablet in 900 ml 0.1 M HCl measured with the potentiometric sensors is shown. After 6 min, 10 g microcrystalline cellulose is added to the solution. The microcrystalline cellulose added to the vessel does
not dissolve but disperses into the medium causing a white opaque suspension. In this suspension, UV fiber optic measurements are not possible due to the extensive scatter and fouling of the fiber optic windows. However, only a small disturbance of the potential, caused by the manual addition of the microcrystalline cellulose, is noticed when using the potentiometric sensors. After the addition, the potential stabilises to its original value within 10 s indicating the insensitivity towards undissolved particles in the dissolution medium. These measurements demonstrate that the potentiometric sensor technology can probably also be used in complex media such as, e.g. milk, a medium that has been used to simulate the fed stomach and to perform in vivo–vitro correlations in several studies (Nicolaides et al., 1999; Dressman and Reppas, 2000; Jantratid et al., 2008). It is however impossible to measure in milk with fiber optic probes and very complex to measure with HPLC. The technology has also the potential to be used for measuring nanosuspensions for which the particles are too small to or very difficult to filter prior traditional analysis and for which again UV fiber optics are not suitable due to the high degree of light scatter. These promising applications will be further explored.
3.6.
Future plans
Other experiments running with the potentiometric sensors at the moment are the dissolution testing of controlled release formulations and the dissolution testing in media containing certain surfactants. First controlled release tests show good results indicating that depending on the formulation, two to five extra calibration points have to be determined in 24 h. Experiments in media containing surfactants, however showed that the tested sensors were not performing well when conducting dissolution measurements in media containing polysorbate 20 (Tween® 20). Tests in media containing Brij® 35 however showed good results. Also tests in bio-relevant media (e.g. fed state simulated intestine fluid (FeSSIF) and fasted state simulated intestine fluid (FaSSIF)) are planned in the future. Comparable problems as with sur-
Fig. 6 – Plateau of a dissolution profile of a 25 mg cinnarizine tablet in 900 ml 0.1 M HCl before and after (arrow) addition of 10 g microcrystalline cellulose and a picture of the corresponding vessel.
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 4 ( 2 0 0 8 ) 243–249
factant containing media are expected since in these media, the surfactant sodium taurocholate, is present. Furthermore, dissolution profiling of lower log P drugs are planned and at the moment, a next generation of potentiometric sensors are being developed and have to be explored in further research.
4.
Conclusions
In the present work, it is shown that potentiometric sensors can be used for in situ dissolution testing. The comparison of the dissolution profiles obtained with the potentiometric sensors show indeed good agreement with these of currently available techniques, i.e. manual sampling with HPLC analysis and UV fiber optics. Furthermore, it is shown that the potentiometric sensors have a good response time for fast releasing formulations and no interference with undissolved particles or air bubbles in the dissolution vessel, a clear advantage in comparison with UV fiber optics. Their possible use for, e.g. nanosuspensions or measurements in milk has to be explored. The sensors have the disadvantage that for each API a new set of sensors needs to be preconditioned. Because the potentiometric electrodes are very cheap in comparison to, e.g. fiber optic probes, they may be considered as consumables. The tested prototype needs further development to extend its applicability towards other dissolution media (e.g. media containing frequently used surfactants such as, e.g. polysorbate 20) and the diameter of the reference electrode needs to be decreased to limit as much as possible the disturbance of the flow dynamics in the vessel.
Acknowledgements The authors thank IWT (Institute for the Promotion of Innovation by Science and Technology in Flanders) for financial support (grant 040399), AIC (Antwerps Innovatie Centrum) for financial and organisational support. Furthermore, we thank T. Stockmans, N. Schoonvaere and D. Peeters for their practical help with the dissolution testing and N. Vervoort for his help with the HPLC electron spray mass spectrometry.
249
references
Aldridge, P.K., Kostek, L.J., 1995. In situ fiber optic dissolution analysis. Dissolut. Technol. 2 (4), 10–11. Bohets, H., Vanhoutte, K., De Maesschalck, R., Cockaerts, P., Vissers, B., Nagels, L.J., 2007. Development of in situ selective sensors for dissolution. Anal. Chim. Acta 581, 181–191. Bynum, K.C., Kraft, E., Pocreva, J., Ciurczak, E.W., Palermo, P., 1999. In situ dissolution testing using a UV fiber optic probe dissolution system. Dissolut. Technol. 6 (4), 11. Cho, J.H., Gemperline, P.J., Salt, A., Walker, D.S., 1995. UV/visible spectral dissolution monitoring by in situ fiber-optic probes. Anal. Chem. 67, 2858–2863. Dressman, J.B., Reppas, C., 2000. In vitro–in vivo correlations for lipophilic, poorly water-soluble drugs. Eur. J. Pharm. Sci. 11 (Suppl. 2), S73–S80. Graffner, C., 2006. Regulatory aspects of drug dissolution from a European perspective. Eur. J. Pharm. Sci. 29, 288–293. Gray, V., 2003. Dissolution testing using fiber optics—a regulatory perspective. Am. Pharm. Rev. 6 (2), 26–30. Jantratid, E., Janssen, N., Reppas, C., Dressman, J.B., 2008. Dissolution media simulating conditions in the proximal human gastrointestinal tract: an update. Pharm. Res. 25, 1663–1676. Johansson, J., Cauchi, M., Sundgren, M., 2002. Multiple fiber optic dual-beam UV/vis system with application to dissolution testing. J. Pharm. Bio. Anal. 29, 469–476. Lu, X., Lozano, R., Shah, P., 2003. In situ dissolution testing using different UV fiber optic probes and instruments. Dissolut. Technol. 10 (4), 6–15. Moore, J.W., Flanner, H.H., 1996. Mathematical comparison of dissolution profiles. Pharm. Tech. 20 (6), 64–74. Nicolaides, E., Galia, E., Efthymiopoulos, C., Dressman, J.B., Reppas, C., 1999. Forecasting the in vivo performance of four low solubility drugs from their in vitro dissolution data. Pharm. Res. 16, 1876–1882. Schatz, C., Ulmschneider, M., Altermatt, R., Marrer, S., 2000. Evaluation of the Rainbow Dynamic Dissolution MonitorTM semi-automatic fiber optic dissolution tester. Dissolut. Technol. 7 (4), 8–17. Schatz, C., Ulmschneider, M., Altermatt, R., Marrer, S., Altorfer, H., 2001. Thoughts on fiber optics in dissolution testing. Dissolut. Technol. 8 (2), 1–5.