- Email: [email protected]
Second-Derivative Spectrophotometric Assay of Mixtures of Dicloxacillin Sodium and Ampicillin Sodium in Pharmaceuticals
Second-Derivative Spectrophotometric Assay of Mixtures of Dicloxacillin Sodium and Ampicillin Sodium in Pharmaceuticals BASILIO MORELLI Received March 25, 1988, from the Universith de li Studi-Bari, Dipartimento di Chimica, Campus Universitario-4, Trav. 200 Re David, Accepted for publication J$ 27, 1988. 70126-8ARI, Italy. Abstract L l ’Zero-crossing’ derivative spectrophotometry has been
used for determining binary mixtures of dicloxacillin Na and ampicillin Na, which are penicillins with closely overlapping absorption spectra. The procedure is rapid, simple, nondestructive, and does not require solutions of equations. Calibration graphs are linear (r = 0.9999),with a zero intercept for up to 60 pg/mL of each antibiotic. Detection limits at the p = 0.05 level of significance were calculated to be 0.29 and 0.31 p,g/mL of dicloxacillin Na and ampicillin Na, respectively. The method was successfully applied to the assay of commercial injections and capsules for these drugs.
Derivative spectrophotometry has only recently become a practical analytical method in the general laboratory because of the rapid progress in microcomputers technology. Fundamental theoretical and practical works of O’Haverl-‘ and Fell and co-worker&1o showed that this technique, if properly understood and applied, is a valuable tool for problem solving in several areas of analytical chemistry. It can lead to quicker and more accurate quantitation of multiple-component mixtures that previously would have required, for example, separation by HPLC.” In recent years, derivative spectrophotometry has received increasing attention with regard to the assay of drugs in their formulations and in systems of clinical and biological interest.12--25 Recent work in our laboratory dealt with the application of derivative spectrophotometry to the determination of binary mixtures of some cephalosporins.*@28In the present paper, we prove that “zero-crossing” second-derivative spectrophotometry is a simple method for determining dicloxacillin Na and ampicillin N a in a mixture. Dicloxacillin Na and ampicillin Na are semisynthetic penicillins with similar absorption spectra. To integrate their respective fields of activity, they are often mixed in some pharmaceuticals which are highly active against both gram-positive and gram-negative bacteria. Hence, these formulations are used for the treatment of several infections such as respiratory, biliary, urinary, odontologic, stomatologic, dermatological, gynecological, and ophthalmic. A quick and reliable procedure was needed to quantitate mixtures of dicloxacillin Na and ampicillin Na, and for quality control of pharmaceutical dosage forms for these drugs. The derivative procedure developed yields accurate and reproducible results.
distilled water. The second-derivative spectra of the mixture against water were then recorded. The absolute value of the derivative at 248.8 and 215.2 nm, for determination of dicloxacillin Na and ampicillin Na, respectively, were measured.
Results and Discussion Spectrophotometric Measurements-The absorption spectra of ampicillin Na (16 pg/mL), dicloxacillin Na (12 pglmL), and a mixture of ampicillin Na and dicloxacillin Na (16 and 12 pg/mL, respectively), in the 192-252-nm wavelength region, are reported in Figure 1. The corresponding first- and second-derivative spectra of ampicillin Na and dicloxacillin Na are represented in Figures 2 and 3, respectively. The zero-order spectra of the two penicillins show extreme similarity, hence the traditional Vierordt’s and modified Vierordt’s methods for assaying binary mixtures gave erroneous results. This confirms Glenn’s limitations whereby unacceptable results are obtained when the absorption spectra of the components are not sufficiently separated.5.23.29.3” The first-derivative spectra of the two antibiotics show significant differences in some areas (Figure 2); nevertheless, some tests showed that the first-derivative spectra of the mixtures are poorly resolved, which prevents, in the present instance, a suitable use of this technique. On the contrary, ‘zero-crossing’1.2second-derivative spectrophotometry offers an extremely valuable means of simultaneously determining both antibiotics in a mixture. In particular, 215.2 nm for determination of ampicillin Na and 248.8 for dicloxacillin Na were selected as optimum working wavelengths in that
1.5
u
1.01 ‘\,
‘\\
\c
Experimental Section Materials-Stock solutions (0.2 mgimL in water) of pure samples of dicloxacillin Na and ampicillin Na (Sigma Chemical) were freshly prepared. Injectable dosage forms and capsules of Ampiplus (A. Menarini, s.a.s.,Italy) were used. Equipment and Instrumental Settings-Equipment and suitable settings were the same as in our earlier paper.ZH Procedure-Suitable aliquots of the penicillin stock solutions were mixed in a 5-mL calibrated flask and diluted to volume with 1042 / Journal of Pharmaceutical Sciences Vol. 77, No. 12, December 1988
I Inm 1 Figure 1-Absorption spectra of (a) ampicillin Na (16 fig/mL), (b) dicloxacillin Na (12 pg/mL), and (c) mixture of ampicillin Na and dicloxacillin Na (16 plus 12 pg/mL, respectively). The reference was water. 0022-3549/88/1200-1042$01.00/0 0 7 988. American Pharmaceutical Association
* 0.25
1I
. ri
-0
4.
'0
N
0
4
'0
I
\
U U
N I
-0.25L 200
220
240 a(nml
248.8 nm
260
Flgure 2-First-derivative spectra of (a) ampicillin Na (16 pg/mL) and
- 0.015
I
I
I
(b) dicloxacillin Na (12 pg/mL). The reference was water. Flgure 4-Second-derivative spectra of mixtures of dicloxacillin Na and ampicillin Na (ampicillin Na: 14 pg/mL; dictoxacillin Na: 2, 10, 18, 26, and 34 pg/mL, as shown in curves 1 to 5, respectively). The reference was water.
(zero-crossing
-0.010
200
l----I
220
wavelengths)
A-
240
260
Ilnm)
Flgure 3-Second-derivative spectra of (a) ampicillin Na (16 pg/mL) and (b) dicioxacillin Na (12 pg/ml). The reference was water.
measurements of the absolute value of the total-derivative spectrum taken a t these wavelengths gave the best linear response to analyte concentration. In Figure 4 is reported a typical set of second-derivative spectra of laboratory mixtures of 14 pg/mL of ampicillin Na and increasing concentrations of dicloxacillin Na ranging from 2 to 34 pg/mL. Figure 5 shows the second-derivative spectra of mixtures of 20 pg/mL of dicloxacillin Na plus increasing amounts of ampicillin Na from 6 to 50 pg/mL. The height at 248.8 nm, the zero-crossing wavelength of ampicillin Na, was denoted hl (see Figure 41, and the height at 215.2 nm, the zero-crossing wavelength of dicloxacillin Na, was denoted h2 (see Figure 5). These heights, hl and h2, were proportional to dicloxacillin Na and ampicillin Na concentrations, respectively. Moreover, the values of hl and h2 were not affected by the presence of the other penicillin. An interaction studyg was performed, and the results are displayed by the graphs in Figure 6. The ordinate values HI and H2were obtained from the hl and h2 (mm) measurements taken on the chart recordings of the spectra and normalized as follows:28 H = recorder divisions (h, mm) x scale expansionU00 mm full scale. Figure 6 (A and B) shows that when one component is kept constant (1.2 pg/mL) and the concentration of the other is varied, the H I(at 248.8 nm)
-0.030i
I
220
I
2 40
I
260
hfnm) Flgure 5-Second-derivative spectra of mixtures of dicloxacillin Na and ampicillin Na (dicloxacillin Na: 2 0 pg/mL; ampicillin Na: 6, 14, 22, 30, and 50 pg/mL, as shown in cuwes 1 to 5, respectively). The reference was water.
and H2(at 215.2 nm) values are unaltered up to 60 pg/mL of the second component. Hence, accurate quantitation of the two penicillins was achieved even when the ratio of concentration was 60:1.2. It is also interesting to note distinct isosbestic points in Figure 4 a t 215.2 and 221.6 nm (zero-crossingwavelengths of dicloxacillin Na) and in Figure 5 a t 248.8 nm (zero-crossing wavelength of ampicillin Na) irrespective of the concentration of dicloxacillin Na and ampicillin Na, respectively. Calibration Graphs and Statistical Analysis of ResultsThe linear regression equations calculated for mixtures of dicloxacillin Na and ampicillin Na are assembled in Table I, Journal of Pharmaceutical Sciences / 1043 Vol. 77, No. 12, December 1988
2.3E-04
HI
P
2.1E-04
1.9E-04
e
I
I
I
I
20
40
68
80
-
[AMPICILLIN SODIUMI,pg/m1
5-6E-04
H2
5.4E-04
0
ct
-2.0
.2.0
0
‘4.
x10-s
INTERCEPT
20
40
60
Flgure 7-Joint confidence regions at p = 0.05 level of significance for slope and intercept of regression equations of dicloxacillin Na and ampicillin Na.
80
[DICLOXACILLIN SODIUMl.pg/ml Figure 6-lnteraction graphs for ( A ) dicloxacillin Na (1.2 pg/mL) in mixture with ampicillin Na (248.8nm), and (B)ampicillin Na (1.2pg/mL) in mixture with dicloxacillin Na (275.2 nm).
together with the correlation coefficients, the variances, and the detection limits at a level of significance of p = 0.05 for 11 standard samples. Beer’s law is followed for concentrations up to 60 pgimL of each antibiotic. As in a previous paper,*H two different tests of significance of the intercept of lines of regression (H = a + bc) were performed to establish whether the experimental intercept, a, differed significantly from the theoretical value, zero. These tests are very useful in the case of mixtures to verify if the analytical method is free from procedural errors depending on the concentration of one of the two components. The first procedure to estimate the differences a - 0 follows from the determination of the quantities t = a l ~ ~ ~ 8 (s, . 3 1is an estimate of the accuracy of the determination of a ) and their comparison with the tabular data for t-distribution. The values calculated for t are 0.32 for dicloxacillin Na and 1.5 for ampicillin Na (i.e., they do not exceed the 95% criterion of tp = 2.26 for n = 11 samples); this indicates that the intercepts of lines of regression are not significantly different from zero. However, the hypothesis that a = 0 was also tested with a more rigorous procedure requiring the construction of a joint confidence region for the possible value of slope and intercept. The 95% confidence ellipse^^^^^^^^^^^ are displayed in Figure 7. It is clear that points for which the intercept is zero fall well inside the ellipses, thus confirming the hypothesis that a is significantly equal to zero. Figure 8 shows the absolute error, s,,283 in the determination of a given concentration, calculated by means of statistical analysis of the regression equations reported in Table I. The error is mini-
~ 1 6 5 ~ D l C L O X A C l L L l NNa
nm)
(248.8
f
/
:qw ,
.140 0
1
20
60
40
AMPICILLIN N a (215.2 nrni
.175I.169
sc
.
0
1
4
20
40
5
60
Y
pgml Figure 8-Absolute error, s,, in the determination of mixtures of dicloxacillin Na and ampicillin Na, by the second-derivative method (h, and h2 measurements).
Table I-Statistical Analysis of the Determination of Dicloxacillln Na and Ampicillin Na In Mixtures by Derivative Spectrophotometry ~
Penicillin
Wavelength, nm
Dicloxacillin Na Ampicillin Na
248.8 215.2
Regression Equationa
ti,
= 3.71E-06
H2 =
+
1.63E-04c
-5.68E-05+ 4.78E-04C
Correlation Coefficient 0.9999 0.9999
Variance 4.82E-10 4.72E-09
Where c is the concentration of the drug (pg/mL); number of standard specimens, n = 11; level of significance, p = 0.05. 1044 i Journal of Pharmaceutical Sciences Vol. 77, No. 12, December 1988
~
~~
Detection Limit, @/mL 0.29 0.31
mal, in correspondence with -24 pgImL of both penicillins. Useful information about the level of precision that may be obtained with the application of the analytical procedure proposed is drawn from the curves in Figure 9. These represent the confidence limits3I at the p = 0.05 level of significance in the determination of mixtures of dicloxacillin Na and ampicillin Na considered above, which are plotted from calibration data as uncertainty (%) versus concentration (relative error) t,s& (%).2&28 Accuracy and Precision--To test accuracy and precision, five successive determinations on synthetic mixtures of penicillins were carried out. The results reported in Table I1 show that accuracy and precision are very satisfactory. Assay of Dicloxacillin Na and Ampicillin Na in Pharmaceuticals-The method was applied to the recovery of diclox-
16 -*
1284-
c
20
0
\
i
__ A M P I C I L L I N Na (215.2
Pharmaceutical'
Amount of Label Claim, mg
Mean recovery t SD, OIob
Ampiplus injections 241.2 Dicloxacillin Na
101.80 5 0.042 102.92 I+_ 0.045
Ampiplus injections 241.2 Dicloxacillin Na
102.00 ? 0.063 102.78 5 0.055
(Batch No. 1)
(Batch No. 2)
531.5 Ampicillin Na
531.5 Ampicillin Na
Ampiplus capsules 241.2 Dicloxacillin Na
265.7 Ampicillin Na
98.04 f 0.070 103.55 2 0.061
'From A. Menarini, s.a.s., Italy. 'Mean of five determinations, assay as percentage of label claim.
acillin Na and ampicillin Na in injections and capsules containing mixtures of these antibiotics. Although derivative spectrophotometry is potentially useful for the assay of analytes in suspension and in turbid solutions,5 an aqueous extraction of the antibiotics from capsules was preferred because the excipients caused a little opalescence in the samples and less reproducible results. To minimize the variation in the composition of the capsules, the content of five capsules was finely ground and a portion of the powder, equivalent to the stated content of one capsule, was weighed. The extraction was performed with 50 mL of water after stirring for -15 min, followed by filtration of the insoluble residue through Whatman No. 41 filter paper. Injectable dosage forms were in a pure form without excipients, hence prior extraction was unnecessary. The assays were carried out as described in the Experimental Section. Table 111 shows the results of five replicate determinations on injections and capsules of Ampiplus. The results are in good agreement with the labeled content of each penicillin. In conclusion, the method described is simple and confirms that the technique of derivative spectrophotometry, if properly used, is a useful one for the precise, sensitive, and rapid analysis of mixtures of drugs.
DlCLOXAClLLlN N a (248.8 nm)
ae
Table Ill-Assay of Dicloxacillin Na and Ampicillin Na in Pharmaceuticals
nm)
12
u
References and Notes - -
I
0
2'0
__7
6'0
4'0
Figure 9-Variation of confidence limits at p = 0.05 level of significance, in the form of uncertainty percent on the concentration (relative error) of dicloxacillin Na and ampicillin Na, by second-derivative methods (h, and h2 measurements).
Table ICReplicate Determinatlons on Synthetic Mixtures of Dlcloxacillln Na and Amplcillln Na
RSD,
N/mL
Mean Value z SD, pglmLa
Dicloxacillin Na and Ampicillin Na
25.0 41.5
25.04 2 0.055 41.35 2 0.074
0.22
Dicloxacillin Na and Ampicillin Na
33.8 7.5
33.74 7.56
?
0.060
2
0.026
0.18 0.34
Mixture
Nominal
Value,
'Mean of five determinations.
%
0.18
O'Haver, T. C.; Green, G. L. Anal. Chem. 1976,48, 312. O'Haver, T. C. Anal. Chem. 1979,51, 91A. O'Haver, T. C. Clin.Chem. 1979,25, 1548. O'Haver, T. C. Anal. Proc. 1982, 19, 22. Fell, A. F. Proc. Analyt. Diu.Chem. SOC.1978, 15, 260. Fell, A. F. J . Pharm. Pharmacol. 1979, 31, 23P. Fell, A. F. UV Spectrom. Group Bull. 1979, 7, 5. Fell, A. F. Anal. Proc. 1980, 17, 512. Fell, A. F.; Smith, G . Anal. Proc. 1982, 19, 28. Fell, A. F. Trends Anal. Chem. 1983,2, 63. Owen, T. Int. Lab. 1987, 17 (8),58. Traveset, J.; Such, V.; Gonzalo, R.; Gelpi, E. J . Pharm. Sci. 1979, 68, 739. 13. Jones, K. G.; Sweeney, G. D. Clin. Chem. 1979,25,71. 14. Balestrieri, C.; Colonna, G.; Giovane, A . ; Irace, G.; Servillo, L. Anal. Btochem. 1980. i06.49. 15. Elsayed, L.; Hassan,.S. M.; Kelany, K. M.; El-Fatatry, H. M. J . Assoc. Off. Anal. Chem. 1980, 63, 992. 16. Jackson, C. M.: Pautler, E.; Lavine. K. Varzan Instr. at Work. NO.UV-14, 1981. 17. Davidson, A. G.; El Sheikh, H. Analyst 1982, 107, 879. 18. El-Din, A. Seif; Korany, M. A.; Abdel-Salam, N. A. Anal. Lett. 1983,16(B12), 891. 19. Tobias, D. Y. J . Assoc. Off. Anal. Chem. 1983, 66, 1450. 20. Korany, M. A.; Wanby, A.M.; Mandour, S. Anal. Lett. 1984, 17(B12), 1373. 21. El-Yazby, F. A.; Barary, M. H. Anal. Lett. 1985, 18fB5/, 629. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
Journal of Pharmaceutical Sciences / 1045 Vol. 77, No. 12,December 1988
22. Hamid, M. E. Abdel; Barary, M. H.; Hassan, E. M.; Elsayed, M. A. Anal. Lett. 1986,19f9-10/,1053. 23. Mohamed, M. E.Anal. Lett. 1986,19(11-12), 1323. 24. Hassan, S. M.; Amer, S. M.; Amer, M. M. Analyst 1987, 112, 1459. 25. Seracu, D. I. Anal. Lett. 1987,20(9/,1417. 26. Morelli, B. Anal. Lett. 1988,21(1), 43. 27. Morelli, B. J . Pharm. Biom.Sci. 1988,6, 199. 28. Morelli, B.J . Pharm. Sci. 1988,77, 615. 29. Glenn, A. L.J . Pharm. Pharmacol. 1960.12.595. 30. Glenn; A. L.J . Pharrn. Pharmacol. 1963;Suppl. 15,123T.
1046 / Journal of Pharmaceutical Sciences Vol. 77, No. 12, December 1988
31. Nalimov, V. V. The application
of mathematical statistics to chemical analysis; Pergamon: Oxford, 1963;pp 167-189. 32. Mandel, J.; Linning, F. J. Anal. Chem. 1957,29,743.
Acknowledgments The author thanks Drs. Mariateresa Gesmundo and Marina Mariani for their valuable technical assistance.
used for determining binary mixtures of dicloxacillin Na and ampicillin Na, which are penicillins with closely overlapping absorption spectra. The procedure is rapid, simple, nondestructive, and does not require solutions of equations. Calibration graphs are linear (r = 0.9999),with a zero intercept for up to 60 pg/mL of each antibiotic. Detection limits at the p = 0.05 level of significance were calculated to be 0.29 and 0.31 p,g/mL of dicloxacillin Na and ampicillin Na, respectively. The method was successfully applied to the assay of commercial injections and capsules for these drugs.
Derivative spectrophotometry has only recently become a practical analytical method in the general laboratory because of the rapid progress in microcomputers technology. Fundamental theoretical and practical works of O’Haverl-‘ and Fell and co-worker&1o showed that this technique, if properly understood and applied, is a valuable tool for problem solving in several areas of analytical chemistry. It can lead to quicker and more accurate quantitation of multiple-component mixtures that previously would have required, for example, separation by HPLC.” In recent years, derivative spectrophotometry has received increasing attention with regard to the assay of drugs in their formulations and in systems of clinical and biological interest.12--25 Recent work in our laboratory dealt with the application of derivative spectrophotometry to the determination of binary mixtures of some cephalosporins.*@28In the present paper, we prove that “zero-crossing” second-derivative spectrophotometry is a simple method for determining dicloxacillin Na and ampicillin N a in a mixture. Dicloxacillin Na and ampicillin Na are semisynthetic penicillins with similar absorption spectra. To integrate their respective fields of activity, they are often mixed in some pharmaceuticals which are highly active against both gram-positive and gram-negative bacteria. Hence, these formulations are used for the treatment of several infections such as respiratory, biliary, urinary, odontologic, stomatologic, dermatological, gynecological, and ophthalmic. A quick and reliable procedure was needed to quantitate mixtures of dicloxacillin Na and ampicillin Na, and for quality control of pharmaceutical dosage forms for these drugs. The derivative procedure developed yields accurate and reproducible results.
distilled water. The second-derivative spectra of the mixture against water were then recorded. The absolute value of the derivative at 248.8 and 215.2 nm, for determination of dicloxacillin Na and ampicillin Na, respectively, were measured.
Results and Discussion Spectrophotometric Measurements-The absorption spectra of ampicillin Na (16 pg/mL), dicloxacillin Na (12 pglmL), and a mixture of ampicillin Na and dicloxacillin Na (16 and 12 pg/mL, respectively), in the 192-252-nm wavelength region, are reported in Figure 1. The corresponding first- and second-derivative spectra of ampicillin Na and dicloxacillin Na are represented in Figures 2 and 3, respectively. The zero-order spectra of the two penicillins show extreme similarity, hence the traditional Vierordt’s and modified Vierordt’s methods for assaying binary mixtures gave erroneous results. This confirms Glenn’s limitations whereby unacceptable results are obtained when the absorption spectra of the components are not sufficiently separated.5.23.29.3” The first-derivative spectra of the two antibiotics show significant differences in some areas (Figure 2); nevertheless, some tests showed that the first-derivative spectra of the mixtures are poorly resolved, which prevents, in the present instance, a suitable use of this technique. On the contrary, ‘zero-crossing’1.2second-derivative spectrophotometry offers an extremely valuable means of simultaneously determining both antibiotics in a mixture. In particular, 215.2 nm for determination of ampicillin Na and 248.8 for dicloxacillin Na were selected as optimum working wavelengths in that
1.5
u
1.01 ‘\,
‘\\
\c
Experimental Section Materials-Stock solutions (0.2 mgimL in water) of pure samples of dicloxacillin Na and ampicillin Na (Sigma Chemical) were freshly prepared. Injectable dosage forms and capsules of Ampiplus (A. Menarini, s.a.s.,Italy) were used. Equipment and Instrumental Settings-Equipment and suitable settings were the same as in our earlier paper.ZH Procedure-Suitable aliquots of the penicillin stock solutions were mixed in a 5-mL calibrated flask and diluted to volume with 1042 / Journal of Pharmaceutical Sciences Vol. 77, No. 12, December 1988
I Inm 1 Figure 1-Absorption spectra of (a) ampicillin Na (16 fig/mL), (b) dicloxacillin Na (12 pg/mL), and (c) mixture of ampicillin Na and dicloxacillin Na (16 plus 12 pg/mL, respectively). The reference was water. 0022-3549/88/1200-1042$01.00/0 0 7 988. American Pharmaceutical Association
* 0.25
1I
. ri
-0
4.
'0
N
0
4
'0
I
\
U U
N I
-0.25L 200
220
240 a(nml
248.8 nm
260
Flgure 2-First-derivative spectra of (a) ampicillin Na (16 pg/mL) and
- 0.015
I
I
I
(b) dicloxacillin Na (12 pg/mL). The reference was water. Flgure 4-Second-derivative spectra of mixtures of dicloxacillin Na and ampicillin Na (ampicillin Na: 14 pg/mL; dictoxacillin Na: 2, 10, 18, 26, and 34 pg/mL, as shown in curves 1 to 5, respectively). The reference was water.
(zero-crossing
-0.010
200
l----I
220
wavelengths)
A-
240
260
Ilnm)
Flgure 3-Second-derivative spectra of (a) ampicillin Na (16 pg/mL) and (b) dicioxacillin Na (12 pg/ml). The reference was water.
measurements of the absolute value of the total-derivative spectrum taken a t these wavelengths gave the best linear response to analyte concentration. In Figure 4 is reported a typical set of second-derivative spectra of laboratory mixtures of 14 pg/mL of ampicillin Na and increasing concentrations of dicloxacillin Na ranging from 2 to 34 pg/mL. Figure 5 shows the second-derivative spectra of mixtures of 20 pg/mL of dicloxacillin Na plus increasing amounts of ampicillin Na from 6 to 50 pg/mL. The height at 248.8 nm, the zero-crossing wavelength of ampicillin Na, was denoted hl (see Figure 41, and the height at 215.2 nm, the zero-crossing wavelength of dicloxacillin Na, was denoted h2 (see Figure 5). These heights, hl and h2, were proportional to dicloxacillin Na and ampicillin Na concentrations, respectively. Moreover, the values of hl and h2 were not affected by the presence of the other penicillin. An interaction studyg was performed, and the results are displayed by the graphs in Figure 6. The ordinate values HI and H2were obtained from the hl and h2 (mm) measurements taken on the chart recordings of the spectra and normalized as follows:28 H = recorder divisions (h, mm) x scale expansionU00 mm full scale. Figure 6 (A and B) shows that when one component is kept constant (1.2 pg/mL) and the concentration of the other is varied, the H I(at 248.8 nm)
-0.030i
I
220
I
2 40
I
260
hfnm) Flgure 5-Second-derivative spectra of mixtures of dicloxacillin Na and ampicillin Na (dicloxacillin Na: 2 0 pg/mL; ampicillin Na: 6, 14, 22, 30, and 50 pg/mL, as shown in cuwes 1 to 5, respectively). The reference was water.
and H2(at 215.2 nm) values are unaltered up to 60 pg/mL of the second component. Hence, accurate quantitation of the two penicillins was achieved even when the ratio of concentration was 60:1.2. It is also interesting to note distinct isosbestic points in Figure 4 a t 215.2 and 221.6 nm (zero-crossingwavelengths of dicloxacillin Na) and in Figure 5 a t 248.8 nm (zero-crossing wavelength of ampicillin Na) irrespective of the concentration of dicloxacillin Na and ampicillin Na, respectively. Calibration Graphs and Statistical Analysis of ResultsThe linear regression equations calculated for mixtures of dicloxacillin Na and ampicillin Na are assembled in Table I, Journal of Pharmaceutical Sciences / 1043 Vol. 77, No. 12, December 1988
2.3E-04
HI
P
2.1E-04
1.9E-04
e
I
I
I
I
20
40
68
80
-
[AMPICILLIN SODIUMI,pg/m1
5-6E-04
H2
5.4E-04
0
ct
-2.0
.2.0
0
‘4.
x10-s
INTERCEPT
20
40
60
Flgure 7-Joint confidence regions at p = 0.05 level of significance for slope and intercept of regression equations of dicloxacillin Na and ampicillin Na.
80
[DICLOXACILLIN SODIUMl.pg/ml Figure 6-lnteraction graphs for ( A ) dicloxacillin Na (1.2 pg/mL) in mixture with ampicillin Na (248.8nm), and (B)ampicillin Na (1.2pg/mL) in mixture with dicloxacillin Na (275.2 nm).
together with the correlation coefficients, the variances, and the detection limits at a level of significance of p = 0.05 for 11 standard samples. Beer’s law is followed for concentrations up to 60 pgimL of each antibiotic. As in a previous paper,*H two different tests of significance of the intercept of lines of regression (H = a + bc) were performed to establish whether the experimental intercept, a, differed significantly from the theoretical value, zero. These tests are very useful in the case of mixtures to verify if the analytical method is free from procedural errors depending on the concentration of one of the two components. The first procedure to estimate the differences a - 0 follows from the determination of the quantities t = a l ~ ~ ~ 8 (s, . 3 1is an estimate of the accuracy of the determination of a ) and their comparison with the tabular data for t-distribution. The values calculated for t are 0.32 for dicloxacillin Na and 1.5 for ampicillin Na (i.e., they do not exceed the 95% criterion of tp = 2.26 for n = 11 samples); this indicates that the intercepts of lines of regression are not significantly different from zero. However, the hypothesis that a = 0 was also tested with a more rigorous procedure requiring the construction of a joint confidence region for the possible value of slope and intercept. The 95% confidence ellipse^^^^^^^^^^^ are displayed in Figure 7. It is clear that points for which the intercept is zero fall well inside the ellipses, thus confirming the hypothesis that a is significantly equal to zero. Figure 8 shows the absolute error, s,,283 in the determination of a given concentration, calculated by means of statistical analysis of the regression equations reported in Table I. The error is mini-
~ 1 6 5 ~ D l C L O X A C l L L l NNa
nm)
(248.8
f
/
:qw ,
.140 0
1
20
60
40
AMPICILLIN N a (215.2 nrni
.175I.169
sc
.
0
1
4
20
40
5
60
Y
pgml Figure 8-Absolute error, s,, in the determination of mixtures of dicloxacillin Na and ampicillin Na, by the second-derivative method (h, and h2 measurements).
Table I-Statistical Analysis of the Determination of Dicloxacillln Na and Ampicillin Na In Mixtures by Derivative Spectrophotometry ~
Penicillin
Wavelength, nm
Dicloxacillin Na Ampicillin Na
248.8 215.2
Regression Equationa
ti,
= 3.71E-06
H2 =
+
1.63E-04c
-5.68E-05+ 4.78E-04C
Correlation Coefficient 0.9999 0.9999
Variance 4.82E-10 4.72E-09
Where c is the concentration of the drug (pg/mL); number of standard specimens, n = 11; level of significance, p = 0.05. 1044 i Journal of Pharmaceutical Sciences Vol. 77, No. 12, December 1988
~
~~
Detection Limit, @/mL 0.29 0.31
mal, in correspondence with -24 pgImL of both penicillins. Useful information about the level of precision that may be obtained with the application of the analytical procedure proposed is drawn from the curves in Figure 9. These represent the confidence limits3I at the p = 0.05 level of significance in the determination of mixtures of dicloxacillin Na and ampicillin Na considered above, which are plotted from calibration data as uncertainty (%) versus concentration (relative error) t,s& (%).2&28 Accuracy and Precision--To test accuracy and precision, five successive determinations on synthetic mixtures of penicillins were carried out. The results reported in Table I1 show that accuracy and precision are very satisfactory. Assay of Dicloxacillin Na and Ampicillin Na in Pharmaceuticals-The method was applied to the recovery of diclox-
16 -*
1284-
c
20
0
\
i
__ A M P I C I L L I N Na (215.2
Pharmaceutical'
Amount of Label Claim, mg
Mean recovery t SD, OIob
Ampiplus injections 241.2 Dicloxacillin Na
101.80 5 0.042 102.92 I+_ 0.045
Ampiplus injections 241.2 Dicloxacillin Na
102.00 ? 0.063 102.78 5 0.055
(Batch No. 1)
(Batch No. 2)
531.5 Ampicillin Na
531.5 Ampicillin Na
Ampiplus capsules 241.2 Dicloxacillin Na
265.7 Ampicillin Na
98.04 f 0.070 103.55 2 0.061
'From A. Menarini, s.a.s., Italy. 'Mean of five determinations, assay as percentage of label claim.
acillin Na and ampicillin Na in injections and capsules containing mixtures of these antibiotics. Although derivative spectrophotometry is potentially useful for the assay of analytes in suspension and in turbid solutions,5 an aqueous extraction of the antibiotics from capsules was preferred because the excipients caused a little opalescence in the samples and less reproducible results. To minimize the variation in the composition of the capsules, the content of five capsules was finely ground and a portion of the powder, equivalent to the stated content of one capsule, was weighed. The extraction was performed with 50 mL of water after stirring for -15 min, followed by filtration of the insoluble residue through Whatman No. 41 filter paper. Injectable dosage forms were in a pure form without excipients, hence prior extraction was unnecessary. The assays were carried out as described in the Experimental Section. Table 111 shows the results of five replicate determinations on injections and capsules of Ampiplus. The results are in good agreement with the labeled content of each penicillin. In conclusion, the method described is simple and confirms that the technique of derivative spectrophotometry, if properly used, is a useful one for the precise, sensitive, and rapid analysis of mixtures of drugs.
DlCLOXAClLLlN N a (248.8 nm)
ae
Table Ill-Assay of Dicloxacillin Na and Ampicillin Na in Pharmaceuticals
nm)
12
u
References and Notes - -
I
0
2'0
__7
6'0
4'0
Figure 9-Variation of confidence limits at p = 0.05 level of significance, in the form of uncertainty percent on the concentration (relative error) of dicloxacillin Na and ampicillin Na, by second-derivative methods (h, and h2 measurements).
Table ICReplicate Determinatlons on Synthetic Mixtures of Dlcloxacillln Na and Amplcillln Na
RSD,
N/mL
Mean Value z SD, pglmLa
Dicloxacillin Na and Ampicillin Na
25.0 41.5
25.04 2 0.055 41.35 2 0.074
0.22
Dicloxacillin Na and Ampicillin Na
33.8 7.5
33.74 7.56
?
0.060
2
0.026
0.18 0.34
Mixture
Nominal
Value,
'Mean of five determinations.
%
0.18
O'Haver, T. C.; Green, G. L. Anal. Chem. 1976,48, 312. O'Haver, T. C. Anal. Chem. 1979,51, 91A. O'Haver, T. C. Clin.Chem. 1979,25, 1548. O'Haver, T. C. Anal. Proc. 1982, 19, 22. Fell, A. F. Proc. Analyt. Diu.Chem. SOC.1978, 15, 260. Fell, A. F. J . Pharm. Pharmacol. 1979, 31, 23P. Fell, A. F. UV Spectrom. Group Bull. 1979, 7, 5. Fell, A. F. Anal. Proc. 1980, 17, 512. Fell, A. F.; Smith, G . Anal. Proc. 1982, 19, 28. Fell, A. F. Trends Anal. Chem. 1983,2, 63. Owen, T. Int. Lab. 1987, 17 (8),58. Traveset, J.; Such, V.; Gonzalo, R.; Gelpi, E. J . Pharm. Sci. 1979, 68, 739. 13. Jones, K. G.; Sweeney, G. D. Clin. Chem. 1979,25,71. 14. Balestrieri, C.; Colonna, G.; Giovane, A . ; Irace, G.; Servillo, L. Anal. Btochem. 1980. i06.49. 15. Elsayed, L.; Hassan,.S. M.; Kelany, K. M.; El-Fatatry, H. M. J . Assoc. Off. Anal. Chem. 1980, 63, 992. 16. Jackson, C. M.: Pautler, E.; Lavine. K. Varzan Instr. at Work. NO.UV-14, 1981. 17. Davidson, A. G.; El Sheikh, H. Analyst 1982, 107, 879. 18. El-Din, A. Seif; Korany, M. A.; Abdel-Salam, N. A. Anal. Lett. 1983,16(B12), 891. 19. Tobias, D. Y. J . Assoc. Off. Anal. Chem. 1983, 66, 1450. 20. Korany, M. A.; Wanby, A.M.; Mandour, S. Anal. Lett. 1984, 17(B12), 1373. 21. El-Yazby, F. A.; Barary, M. H. Anal. Lett. 1985, 18fB5/, 629. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
Journal of Pharmaceutical Sciences / 1045 Vol. 77, No. 12,December 1988
22. Hamid, M. E. Abdel; Barary, M. H.; Hassan, E. M.; Elsayed, M. A. Anal. Lett. 1986,19f9-10/,1053. 23. Mohamed, M. E.Anal. Lett. 1986,19(11-12), 1323. 24. Hassan, S. M.; Amer, S. M.; Amer, M. M. Analyst 1987, 112, 1459. 25. Seracu, D. I. Anal. Lett. 1987,20(9/,1417. 26. Morelli, B. Anal. Lett. 1988,21(1), 43. 27. Morelli, B. J . Pharm. Biom.Sci. 1988,6, 199. 28. Morelli, B.J . Pharm. Sci. 1988,77, 615. 29. Glenn, A. L.J . Pharm. Pharmacol. 1960.12.595. 30. Glenn; A. L.J . Pharrn. Pharmacol. 1963;Suppl. 15,123T.
1046 / Journal of Pharmaceutical Sciences Vol. 77, No. 12, December 1988
31. Nalimov, V. V. The application
of mathematical statistics to chemical analysis; Pergamon: Oxford, 1963;pp 167-189. 32. Mandel, J.; Linning, F. J. Anal. Chem. 1957,29,743.
Acknowledgments The author thanks Drs. Mariateresa Gesmundo and Marina Mariani for their valuable technical assistance.