The Multicomponent
Automated Dissolution System:
An Alternative in the Development and Pharmaceutical Analysis
of Generic Polydrugs
V. Fuerte (1) and M. Maldonado (2)*
(1) Centro de Investigacion y de Estudios, Avanzados del
IPN. Tepepan, Mexico,
(2) Department of Pharmacy, King's College London, United Kingdom
*E-mail address for correspondence: minerva.maldonado@kcl.ac.uk
Abstract
The necessity of assuring the quality of polydrugs, especially
those with low aqueous solubility and in vivo absorption, has
led to the development and evaluation of new techniques that can
reduce the time and cost of analysis. This study examines the
efficiency and accuracy of an automated dissolution system (ADS),
fitted with a simple, integrated, multicomponent detector, for
analysis of generic polydrugs using multiple linear regression
(MLR). Trimethoprim and sulphamethoxazole were chosen as model
drugs for this study and comparison was made with a conventional
analysis based on HPLC. Both analytical systems under study gave
reproducible and accurate results. Analysis of variance showed
that there was no significant statistical difference between the
methods of analysis, nor any statistical difference between the
measured amounts of drug in the three different formulations.
We have demonstrated that low cost instrumentation coupled with
MLR data processing provides satisfactory drug analysis to a standard
at least as good as that achieved using HPLC. Multicomponent ADS
is clearly a convenient and viable alternative to HPLC, and provides
an opportunity to reduce the time and analysis cost of other generic
formulations.
Introduction
n recent decades, the development of new formulations has become
a key function in pharmaceutical companies, principally because
of the need to improve drug efficacy and minimize unwanted side-effects.
An important feature of drug production is the necessity to ensure
appropriate quality control and this requires accurate, often
sophisticated but where possible fast, cost-effective analysis.
The complexities of drug analysis are clearly increased when the
number of drugs or excipients in the formulation increase. Thus
the ease of analysis is especially important for dosage forms
containing more than one drug.
Where drugs have poor absorption characteristics in vivo, or more
especially have poor dissolution characteristics from their chosen
dosage formulation, regulations require drug formulations to be
subject to dissolution analysis. Conventionally, High Pressure
Liquid Chromatography (HPLC) has been chosen for monitoring drug
release from such formulations because of the generally excellent
resolution and sensitivity of the technique. However, analysis
of multiple drug components in a mixture has also been demonstrated
using first and second derivative spectrophotometry [1,2]. More
recently, this type of approach has been incorporated in the development
of UV-Visible Multicomponent Automated Dissolution Systems (ADS).
This instrumentation relies for its analysis on the principles
of absorption additivity. In the study presented here, the analyte
concentration in multicomponent mixtures is determined mathematically
by straightforward multicomponent linear regression. The main
condition to be met is that the number of monitoring wavelengths
must be similar to the number of components present in the solution
[3,4]. An alternative approach is to employ the more complicated
but also more powerful principal component regression (PCR), in
which case spectra from a selected wavelength range are resolved
using an appropriate algorithm [5].
The level of sophistication of these devices is now such that
ADS have been assembled with multiple in-situ fibre optic probes
interfaced to diode array spectrophotometers with PC control and
analysis [6,7]. Given the cost constraints incumbent on analytical
facilities, it is reasonable to question whether such a high level
of sophistication is always necessary. We aim to demonstrate that
excellent results can be achieved with simple, modest and robust
instrumentation. In the system described here, the dissolution
medium is circulated through an in-line filter and delivered to
a tailored spectrophotometer having the capacity to monitor seven
separate flow cells. A multichannel pump circulates the dissolution
medium around seven dissolution flasks and through the seven flow
cells in the sample chamber unit, each having direct UV monitoring
of the two drugs simultaneously. These simple systems have enormous
potential to reduce the cost and time of analysis, and significantly
improve the overall reliability and reproducibility of testing
procedures without reducing the accuracy.
MATERIAL AND METHODS
Materials
Sulphamethoxazole, trimethoprim and azetazolamide were USP standards
and supplied by Rockville, USA. Prednisone tablets were also supplied
by Rockville. The three generic polydrug formulations contained
(nominally) the same concentrations of drug and were provided
by three different companies. HPLC solvents for analysis were
supplied by Milinchroph and 0.22 mm Millipore HA and HF filters
were used for filtration of the sample and mobile phase.
HPLC Analysis
A modular Perkin Elmer Model 200 Chromatograph was comprised of
a pump, autoinjector, 30cm C18 reverse phase analytical column,
variable wavelength absorbance detector and computer-based Turbochrom
analytical software. The components of the mobile phase were filtered,
mixed and deaereated under vacuum prior to use. The column was
eluted isocratically with 84.5:14.5:1 water:acetonitrile:glacial
acetic acid at 2 ml/min with detection at 250 nm. Calibration
curves for HPLC analysis were prepared from standard solutions
of the drugs in the mobile phase covering a range of 0 - 60 mg/ml
for sulphamethoxazole and 0 - 12 mg/ml for trimethoprim by analysis
in triplicate of 100 ml standards containing 15.0 mg/ml of the
internal standard azetazolamide. Routine checks on the response
factors were made by daily analysis of a standard solution containing
40.0 mg/ml sulphametoxazole, 8.0 mg/ml of trimethoprim and 15.0
mg/ml azetazolamide.
Analysis of the generic polydrugs involved dissolving tablets
containing 400 mg sulphametoxazole and 80 mg of trimethoprim in
900 ml of 0.10M HCl using the ADS. The conditions were specified
by the USP XXII Pharmacopoeia for this product (USP apparatus
2, paddle, 75 rpm for 60 minutes, 37°C). After an hour, 1.0
ml was removed into 9.0 ml of a 1.5mg/ml solution of internal
standard made up in the mobile phase described earlier. The diluted
samples were filtered and analysed by HPLC.
Multicomponent Automatic Dissolution System
The instrument was fitted with a model AT7 Sotax Dissolutor, 8
channel peristaltic pump, seven vessels, PC directed control through
the Perkin Elmer software and a Lambda 20 UV/Vis spectrophotometer
fitted with a linear 8 cell transporter. The flow-cell pathlength
was 1.0mm. Dissolution conditions were identical to those described
above. The analytical instrumentation was checked for wavelength
accuracy and repeatability. The dissolution apparatus was set
up, calibrated and operated in compliance with the USP compendia
using the recommended
50 mg prednisone tablets.
Calibration curves for the individual drug standards were obtained
by measuring the absorption at 265 and 271 nm. Standards were
prepared in 0.10M HCl in the concentration range 0 - 100 mg/ml
for trimethoprim and 0 -500 mg/ml for sulphamethoxazole. Extinction
coefficients were calculated for the two drugs at both wavelengths
and employed in the multicomponent analysis software. Subsequently,
the appropriate dissolution conditions for the polydrug samples
were established. The analytical method was then validated for
linearity, accuracy and precision [9].
The linearity of the calibration curves were confirmed over a
concentration range equivalent to 10 to 125% dissolution of the
drug. For accuracy, samples were prepared by spiking with drugs
and excipients in the specified volume of dissolution fluid. Accuracy
was determined by testing six samples of each formulation according
to the dissolution method. Specificity was confirmed by comparing
the results of the HPLC and multicomponent analysis.
Design of the study
After construction of appropriate calibration plots, the analytical
methods were further validated by assaying a homogenate of twenty
tablets from each generic brand. Six tablets from each brand were
then separately analysed on two different days using the ADS as
described earlier. Using the UV multicomponent analysis, the amount
of each drug dissolved was initially measured at 2 minute intervals
for the first ten minutes, after which time the drug concentration
was measured at ten minute intervals. After 60 minutes, the solution
was analyzed by HPLC for comparative purposes and the experiment
halted.
RESULTS AND DISCUSSION
The first phase of the study involved the validation of both techniques.
The chromatogram in Figure
1 clearly shows the peaks relating to the two drugs of interest
and the internal standard are well separated and essentially symmetrical.
Figures 2a -2d
show the calibration curves obtained for trimethoprim and sulfamethoxazole
by HPLC and multicomponent analysis. Although both systems give
essentially linear correlations, the multicomponent analysis seems
to be more reproducible.
Results of the drug assays for twenty tablet homogenates of the different brands are summarized in Table 1.
|
|||
|
HPLC |
Analysis |
Multicomponent |
Analysis |
|
Drug |
Assay ± VC |
Drug |
Assay ± VC |
|
SMX A |
100.6 ± 0.3 |
SMX A |
100.3 ± 0.4 |
|
SMX B |
98.9 ± 1.4 |
SMX B |
98.7 ± 0.2 |
|
SMX C |
101.4 ± 2.1 |
SMX C |
101.5 ± 0.6 |
|
TMP A |
103.9 ± 0.1 |
TMP A |
98.9 ± 1.0 |
|
TMP B |
103.0 ± 1.7 |
TMP B |
97.6 ± 1.7 |
|
TMP C |
103.8 ± 0.2 |
TMP C |
97.5 ± 0.9 |
It is clear that all the brands were within the assay limits
established for this product by the USP XXII (97-103%) when using
the multicomponent analysis. The HPLC analyses gave a broadly
similar result, however the trimethoprim assays were slightly
higher than expected in all three brands. This may reflect the
fact that the HPLC method could be further optimized. The variation
coefficient was less than 3% for the HPLC system and less than
2% for the multicomponent analysis indicating that both methods
are highly reproducible.
Table 2 shows the results of the second phase of the study which involved assaying individually six tablets on two different days for each brand.
| Table 2. Analysis of whole tablets from the three generic brands. Each value represents the mean average of six determinations, with analyses performed on two different days. VC = variation coefficient. | |||
| HPLC | Analysis | Multicomponent | Analysis |
| Drug | Assay ± VC | Drug | Assay ± VC |
| SMX A | 100.2 ± 0.1 | SMX A | 103.7 ± 0.2 |
| 102.5 ± 0.1 | 104.3 ± 0.1 | ||
| SMX B | 99.6 ± 1.1 | SMX B | 98.9 ± 1.4 |
| 95.8 ± 0.3 | 95.1 ± 2.3 | ||
| SMX C | 104.1 ± 3.6 | SMX C | 105.2 ± 1.0 |
| 101.2 ± 1.3 | 103.0 ± 0.2 | ||
| TMP A | 104.1 ± 3.7 | TMP A | 104.1 ± 3.73 |
| 103.5 ± 0.4 | 104.1 ± 1.26 | ||
| TMP B | 104.1 ± 1.8 | TMP B | 104.7 ± 4.61 |
| 103.8 ± 2.5 | 99.2 ± 1.3 | ||
| TMP C | 102.7 ± 0.3 | TMP C | 104.3 ± 1.4 |
| 103.7 ± 0.8 | 95.8 ± 2.0 | ||
Although the drug assays gave values that were more variable than those presented in Table 1, this was expected given that the analyses were conducted on single tablets rather than on the homogenates. The assay results provided using the two analytical techniques were in close agreement in most cases.
Table 3 summarizes the analysis of variance (by balanced design) and shows that there are no significant statistical differences among drugs, brands and systems.
| Table 3. Analysis of variance (by balanced design) for the data shown in Table 2. | |||||
|
Source |
Degrees of freedom |
Sum of Square |
Media Square |
Fisher (F) |
Probability (P) |
|
Drugs |
1 |
13.1 |
13.1 |
1.74 |
0.202 |
|
Brands |
2 |
31.9 |
15.9 |
2.13 |
0.146 |
|
Methods |
1 |
2.1 |
2.1 |
0.28 |
0.603 |
|
Error |
19 |
142.3 |
7.5 |
||
|
Total |
23 |
189.3 |
|||
The variation between brands appears to be the most significant
while the least significant was the variation between systems.
The most notable contrasting feature which distinguishes the multicomponent
analysis from the HPLC method is that of analysis time. HPLC is
discontinuous and invasive whereas multicomponent analysis used
in combination with the ADS gives direct in-situ monitoring of
the dissolution process. In the present study a low cost, relatively
unsophisticated multicomponent ADS is able to deliver the dissolution
profile of both drugs simultaneously in a short space of time.
This is illustrated by Figures
3a and 3b where, for reasons of clarity, only a selection
of the available data points have been plotted. The same analysis
of samples by HPLC would take up to 6 hours.
CONCLUSIONS
We have demonstrated the utility of the UV/Visible multicomponent
analysis for routine analysis of tabletted pharmaceuticals, especially
polydrugs. This system when used to its full potential is capable
of giving a complete profile of the drugs' release, and may prove
to be at its most versatile when dealing with formulations containing
more than the two drugs used in the present study.
To ensure good results with the multicomponent ADS, it is essential
to calibrate the system, determine its suitability and validate
the protocol. Particular attention should be focused on the filtration
process, particularly for those brands where the excipients interfere
with the measured absorbances. This can occur when the particle
size is small enough to pass through the pores of the filter causing
light scattering effects. Matters can be further complicated if
the insoluble excipients contain chromophores, but these two problems
can be largely eliminated through careful selection of the filter
porosity. In conclusion, we believe that low cost Multicomponent
ADS is a highly effective analytical tool offering considerable
advantages over competing chromatographic methods. We expect the
technique to become increasingly important in the next few years,
particularly in the analysis of generic polydrug formulations
where there are likely to be considerable time and cost benefits.
Acknowledgements
The authors wish to thank Dr. Angelina Peña (Procter &
Gamble Co), Dr. G. Castanèda and Dr. E. Hong for constructive
review and suggestions. We also thank To Ing. Octavio Morales
(Perkin-Elmer de Mexico) for kindly providing instrumentation
and technical support.
References
[1] J.M. Hoover, R.A. Soltero, P.C. Bansal, J. Pharm. Sci. 76
(1987) 242-244.
[2] J.L. Murtha, T.N. Julian, G.W. Radebaugh, J. Pharm. Sci. 77
(1988) 715-718.
[3] W.C. Brown, J.R. Obremski, Anal. Chem. 54 (1982) 1472-1475.
[4] S.D. Funs, J.M. Harris, Anal. Chem. 57 (1985) 2680-2684.
[5] S-C, Lo, S.M. Donahue, C.W. Brown, J. Pharm. Sci. 82 (1993)
350-354.
[6] C-S, Chen, C.W. Brown, Pharm. Res. 11 (1994) 979-983.
[7] J.H. Cho, P.J. Gemperline, A. Salt, D.S. Walker, Anal. Chem.
67 (1995) 2858-2863.
[8] P.J. Gemperline, J-H. Cho, B. Baker, B. Batchelor, D.S. Walker,
Analytica
Chimica Acta 345 (1997) 155-159.
[9] C. Riley, T. Rosanske, In: Development and Validation of Analytical
Methods
(1996) p 169-208. Pub. Elsevier.
Notes
1. Fibre optic probe deals with turbid system. Our deals with
filtered samples.
2. One third of runs were lost due to bubble formation on the
probe surface - drawback.
3. N.H. Anderson, D. Johnston, P.R. Vojvodic, J. Pharm. Biomed.
Anal. 1990, 8, 987-989. Used multicomponent UV analysis to analyse
one unspecified drug in the presence of an interfering excipient
(titanium dioxide) which gives light scattering but still satisfies
the requirements for multicomponent analysis.
4. C.W. Brown, J. Lin, Appl. Spectrosc. 1993, 47, 615-618.
5. C.-S. Chen, C.W. Brown, Pharm. Res. 1994, 11, 979-983.
6. M. Josefson, E. Johansson, A. Tortensson, Anal. Chem. 1988,
60, 2666-2671.