- Email: [email protected]
In vitro study of the interaction between quinolones and polyvalent cations
Pharmaceutica Acta Helvetiae 73 Ž1999. 237–245
In vitro study of the interaction between quinolones and polyvalent cations M a Sonia Rodrıguez Cruz, Isabel Gonzalez ´ ´ Alonso ) , Amparo Sanchez–Navarro, ´ a M Luisa Sayalero Marinero Department of Pharmacy and Pharmaceutical Technology, Faculty of Pharmacy, UniÕersity of Salamanca, AÕda. Campo Charro sr n, 37007, Salamanca, Spain Received 3 August 1998; revised 16 September 1998; accepted 30 September 1998
Abstract The aim of the present study was to evaluate the influence of aluminium and iron on the in vitro dissolution kinetics of ciprofloxacin and ofloxacin as well as the usefulness of this type of in vitro data to predict modifications in in vivo absorption processes as a consequence of different factors, such as the widely documented in vivo interaction between quinolones and cations. Fitting of experimental data to different theoretical in vitro dissolution profiles was performed by non-linear regression methods and the statistical moments were calculated from raw experimental data. Analysis of residuals applied to dissolution curves as well as statistical comparison of the estimated parameters were carried out to evaluate the in vitro interaction. The results reveal significative modifications of the dissolution profiles of these quinolones as a consequence of the presence of cations, especially for Fe 2q which decreases 34.7% the maximum amount dissolved for ciprofloxacin and 29.1% for ofloxacin. Al 3q also produces a decrease of the total amount of quinolone dissolved although less relevant than Fe 2q. Analysis of residuals proved to be the best statistical method to evaluate differences between whole dissolution profiles, at least under the experimental conditions used. q 1999 Published by Elsevier Science B.V. All rights reserved. Keywords: Ofloxacin; Ciprofloxacin; Polyvalent cations; Dissolution test; Dissolution kinetics; Interaction
1. Introduction Despite the considerable body of information in the literature about the interaction between quinolones and polyvalent cations ŽKara et al., 1991; Shiba et al., 1992; Wayne Frost et al., 1992; Lomaestro and Bailie, 1993; Maeda et al., 1993; Sanchez Navarro et al., 1994., many ´ aspects related to this process remain to be clarified, including the mechanism or experimental and clinical conditions controlling this interaction. Several authors ŽNix et al., 1989; Polk et al., 1989. have suggested the formation of an insoluble, non-absorbable complex between the quinolone and the corresponding ) Corresponding author. Tel.: q34-23-294-536; fax: q34-23-294-515; e-mail: [email protected]
cation; others ŽRoss and Riley, 1990; Ross and Riley, 1992. have proposed the formation of a ‘micellar phase’ as a consequence of changes in the net charge of the compound and, recently, Macıas et al. Ž1994. have ´ Sanchez ´ found that ofloxacin with metal cations follows the general pattern expected, leading to the formation of salts or coordination compounds, while ciprofloxacin retains the molecular network and no functional group seems to be lost or modified after entering into contact with different polyvalent cations. In contrast, the effect of polyvalent cations on the oral absorption of ciprofloxacin has been reported to be higher than the effect observed with ofloxacin ŽWolfson and Hooper, 1989; Flor et al., 1990.. Additionally, there are some differences among the results found by different authors studying quantitative modifications in the oral absorption of quinolones administered
0031-6865r99r$ - see front matter q 1999 Published by Elsevier Science B.V. All rights reserved. PII: S 0 0 3 1 - 6 8 6 5 Ž 9 8 . 0 0 0 2 9 - 6
238
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
with drugs or food containing cations ŽNix et al., 1990; Martınez-Cabarga et al., 1991; Pertti et al., 1991.. ´ The conclusion to be drawn from the available information is that some type of interaction takes place between quinolones and cations which leads to a decrease in the amount of antibacterial drug absorbed when administered by the oral route, but the interaction process seems to be very sensitive to slight changes in factors related to the experimental andror clinical conditions of the different assays. This would justify the wide dispersion of the results reported in the literature. The present study was therefore carried out to evaluate the influence of the presence of Al 3q and Fe 2q ions on the in vitro dissolution profiles of ciprofloxacin and ofloxacin. Also to analyze the advantages and limitations of different available statistical methods used to compare in vitro dissolution curves. Since drug dissolution is the previous step to drug absorption, the study of the modifications of the dissolution profiles will provide useful information about the possible changes in drug absorption as a consequence of interactions.
2. Materials and methods 2.1. Materials and equipment 2.1.1. Drugs and reagents The quinolone solid dosage formulations assayed were Baycip w 250 mg ŽBayer, Barcelona, Spain. and Tarivid w 200 mg ŽHoescht Farma, Barcelona, Spain.. The cation preparations were: Fero-Gradumet w ŽAbbot Cientifica, Madrid, Spain. and aluminium hydroxide powder ŽMerck, Darmstadt, Germany.. 2.1.2. Equipment Hanson SR 6-Flask Dissolution test station ŽHanson Research, Chatsworth, CA.. U-2000 Spectrophotometer ŽHitachi, Tokyo, Japan.. 5000 Liquid Chromatograph ŽVarian. connected to a Kontron SFM 25 fluorescence detector. 2.2. Methods 2.2.1. In Õitro dissolution assay Dissolution profiles were determined at 37 " 0.58C in 900 ml of phosphate buffer solution, pH s 6. Commercially available dissolution equipment employing the paddle apparatus as described in the USP XXII Ž1989. was used. Rotation speed was maintained at 50 r.p.m. The
paddle was positioned to extend to exactly 2.5 cm above the flask bottom. Samples Ž2 ml. were taken with a graduated syringe at the following times. Ciprofloxacin: 1, 2, 3, 4, 5, 7, 10, 15, 20, 30, 45, 60 and 90 min. Ofloxacin: 5, 10, 15, 20, 30, 45, 60, 90, 120, 180, 240, 300 and 360 min. Each sample was replaced by an equal volume of dissolution media to keep the total volume constant. A correction was made to take into account the cumulative removed volumes when determining the total amount dissolved as a function of time. All samples were suitably filtered, placing a filter at the end of the sample probe, and kept at 58C until immediate analytical determination. The suitability of the paddle apparatus was checked using the USP prednisone and salicylic acid calibrators Žcalibrators for system suitability test of basket and paddle dissolution apparatus. ŽHanson, 1992.. One tablet of the corresponding quinolone was placed in each filled flask Ž6 tablets per run. when establishing the dissolution profiles of the quinolones in the absence of cations. These profiles were considered as reference curves. To evaluate the influence of the different cations on the quinolone dissolution kinetics, the corresponding cation preparation was added to each flask at the same time as the quinolone formulation. For Al 3q, a sufficient amount of aluminium hydroxide powder to obtain 2.5 g of Al 3q was added; when studying the influence of Fe 2q, two 525 mg FeSO4 tablets were included in the dissolution media. The amounts of cations were selected based on the usual clinical doses. Addition of quinolones or cations do not modifies the pH of dissolution media for the sampling period.
2.2.2. Analytical techniques Samples from all assays, except those on Fe 2q, were analyzed by an UV-spectrophotometric technique at wavelengths of 287 nm and 271 nm for ofloxacin and ciprofloxacin, respectively. Standard curves were prepared in phosphate buffer solution ŽpH s 6. at a concentration range of 0.5–20.0 mgrl for ofloxacin and 0.5–10.0 mgrl for ciprofloxacin. Samples were diluted when necessary to achieve concentrations included within the standard curve range. The UV technique showed adequate linearity, precision and accuracy with coefficients of variation of less than 5% and 2.5% for ofloxacin and ciprofloxacin, respectively. Samples from the Fe 2q assays were analyzed by an HPLC technique. A Varian 5.000 Chromatograph connected to a variable wavelength fluorescence detector was
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
239
Fig. 1. Dissolution profiles of ciprofloxacin and ofloxacin corresponding to in vitro tests carried out under the following experimental conditions: phosphate buffer solution, pH s6, T a s 378C. Each curve is the mean of six experiments.
used. The 15 cm-long column Ž4 mm internal diameter. was packed with Nucleosil w 120 R-18 5m of particle size. The mobile phase consisted of acetonitrile and 0.025 M phosphate buffer solution, pH 3, Ž9:91 vrv. and the flow rate was 2 mlrmin. The fluorescence excitation and emission wavelengths were set at 277 nm and 445 nm for ciprofloxacin and 330 nm and 450 nm for ofloxacin, respectively. 50 ml of the internal standard Žeach quinolone was used as the internal standard of the other one. was added to 200 ml of the sample. The mixture was vigorously shaken for 30 s and 100 ml was injected into the chromotograph. Standard curves were prepared as indicated for UV detection and samples were also diluted if necessary. This HPLC technique also showed adequate linearity, precision and accuracy with coefficients of variation of less than 10% and 8% for ofloxacin and ciprofloxacin, respectively.
Both, UV and HPLC, techniques quantifies free drug in the dissolution media.
2.2.3. Data analysis The dissolution pattern, described as the cumulative amount of drug dissolved, was fitted to different mathematical expressions, commonly related to in vitro dissolution profiles, for immediate release dosage formulations ŽAbdou, 1989.. These expressions are: Ø Zero-order kinetics: dQ dt
sK0
Ž 1.
where: dQrdt s dissolution rate; K o s constant amount of drug dissolved per unit time
Table 1 Mean dissolution parameters of ofloxacin under standard conditions and in presence of two different cations
Ofloxacin 200 mg Oflo 200qAl 3q Oflo 200qFe 2q
MDT Žmin.
CV
Qma x Žmg.
K d Žminy1 .
45.20"13.84 98.41"37.67 36.17"11.41
1.20"2.80ey1 1.25"9.42ey2 1.09"1.05ey1
173.60"11.67 152.11"16.89 123.16"3.58
3.04ey2"6.64ey3 1.77ey2"3.33ey3 4.33ey2"7.20ey3
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
240
Table 2 Mean dissolution parameters of ciprofloxacin under standard conditions and in presence of two different cations
Ciprofloxacin 250 mg Cipro 250 mgqAl 3q Cipro 250 mgqFe 2q
MDT Žmin.
CV
Qma x Žmg.
K d Žminy1 .
8.87"1.69 11.32"4.79 17.32"11.66
1.82"3.74ey1 1.85"3.99ey1 1.04"1.06ey1
219.41"7.74 180.16"46.47 143.24"8.39
2.43ey1"6.69ey2 2.32ey1"9.70ey2 4.76ey1"1.12ey1
Ø First-order kinetics: Q t s Q max Ž 1 y eyK d t .
Ž 2.
where: Q t s amount of drug dissolved at time t; Qmax s maximum amount of drug dissolved; K d s first-order dissolution rate constant Ø Weibull distribution function: Q t s Q max 1 y ey Ž trt d .
b
Ž 3.
where: td s time necessary to achieve 63.2% of total drug dissolved; b s an adimensional parameter defined by the shape of the dissolution curve plot Žrelated to the order of the process.. The fitting was carried out using the PCNONLIN 4.0 program ŽDunne and Stucker, 1992.. The statistical criterion for the selection of the optimum fitting was the minimum AIC value ŽTest MAICE. ŽYamaoka et al., 1977..
In addition to dissolution curve fitting the mean in vitro dissolution time ŽMDT. and the coefficient of variation ŽCV. were calculated using trapezoidal integration ŽPurves, 1992.. The MDT is defined as follows: `
MDT s
H0 t dQrQ
max
Ž 4.
where: Q s amount of drug dissolved at time t; Qmax s maximum amount of drug dissolved. The CV values, which inform about the dispersion of the dissolution process ŽWeiss, 1992., were calculated according to the following expression: CV s
'VDT MDT
Ž 5.
VDT is the variance of the dissolution times, that is the moment of second degree, as the dissolution profile may be considered a distribution function of the residence times
Fig. 2. Dissolution profiles of ofloxacin corresponding to in vitro tests carried out in absence of cations and in presence of Al 3q or Fe 2q. Each curve is the mean of six experiments.
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
241
Fig. 3. Dissolution profiles of ciprofloxacin corresponding to in vitro tests carried out in absence of cations and in presence of Al 3q or Fe 2q. Each curve is the mean of six experiments.
of each drug molecule in the pharmaceutical formulation. VDT was calculated by trapezoidal integration as follows: `
VDT s
H0
2 Ž t y MDT. dQrQmax
Ž 6.
2.2.4.2. Estimation of a ‘ fit factor’ F In order to measure the dissimilarity between the reference curves and those obtained in presence of cations a ‘fit factor’ F was calculated for each added cation. This factor approximates the percent error between two curves and
2.2.4. Statistical analysis 2.2.4.1. ANOVA Statistical moments ŽMDT, CV. as well as other dissolution parameters Ž Qmax , K d . were calculated from each individual dissolution profile and the statistical comparison of each parameter in presence and absence of cations, for both quinolones, was carried out using the test ANOVA at the standard significance level Ž p - 0.05..
Table 3 Modifications of dissolution parameters of ciprofloxacin and ofloxacin calculated as percentage of change of mean parameter values OFLOXACINqAl3q OFLOXACINqFe2q CIPROFLOXACINqAl3q CIPROFLOXACINqFe2q
MDT
Qma x
Kd
≠ 117.72 x 19.98 ≠ 27.62 ≠ 95.26
x 12.38 x 29.06 x 17.89 x 34.72
x 41.78 ≠ 42.43 x 4.53 ≠ 95.88
Fig. 4. Differences between dissolution profiles in absence and presence of each studied cation, quantified by a fit factor value Ž F . calculated from in vitro individual dissolution curves of ciprofloxacin and ofloxacin.
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
242
2.2.4.3. Analysis of the residuals Considering the dissolution profile of the quinolone in absence of cations as the reference curve, the weighted residual was calculated for each sample time and every dissolution curve, according to the well known expression: Qr y Qc Rs Ž 8. Qr Analysis of residuals was carried out for each cation by plotting residuals versus time and examining the scatterplots which inform about the trend in deviations, if they exist.
3. Results Fig. 1 shows the mean in vitro dissolution curves of ciprofloxacin and ofloxacin in phosphate buffer solution
Fig. 5. Scatterplot of residuals for dissolution curves of ofloxacin in presence of cations using the dissolution curve of the quinolone in absence of cations as the reference profile.
quantifies, by a single numerical parameter, the total difference between compared profiles. According to Moore and Flanner Ž1996. this fit factor is defined by the following equation: n
Ý < Qr y Qc < Fs
is1 n
100
Ž 7.
Ý Qr is1
where: n is the total number of samples; Qr and Qc are the amount of drug dissolved under standard conditions Žreference curve. and in presence of the cation, respectively. According to that, the fit factor F is calculated for each cation and each quinolone.
Fig. 6. Scatterplot of residuals for dissolution curves of ciprofloxacin in presence of cations using the dissolution curve of the quinolone in absence of cations as the reference profile.
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
ŽpH s 6.; the corresponding dissolution parameters are included in Tables 1 and 2. The maximum dissolved amount of drug, defined by the asymptote of the fitted curves, are 219.41 " 7.74 mg and 173.60 " 11.67 mg, which represent percentages of 88% and 87% of the theoretical dose for ciprofloxacin and ofloxacin, respectively. The first order rate constant show a value of 2.43e y 1 " 6.69e y 2 miny1 for ciprofloxacin, significantly higher than the 3.04e y 2 " 6.64e y 3 miny1 corresponding to ofloxacin. In agreement with it, the MDT, calculated by trapezoidal integration from raw experimental data, show statistically significant differences between ciprofloxacin Ž8.87 " 1.69 min. and ofloxacin Ž45.20 " 13.84 min.. The dispersion of the dissolution times, measured as the CV, is slightly higher for ciprofloxacin Ž1.82 " 3.74e y 1. than for ofloxacin Ž1.20 " 2.80e y 1.. Figs. 2 and 3 illustrate the effect of the presence of different cations on the in vitro dissolution profiles of ofloxacin and ciprofloxacin, respectively. The cations yield to a decrease of the maximum amount dissolved for ciprofloxacin as well as ofloxacin. Modifications of the first order rate constant are also observed as a consequence of addition of Fe 2q to the dissolution media. Table 3 summarizes the modification of the dissolution parameters caused by presence of each cation; these are expressed as percentage of change calculated from the mean parameter values. Fig. 4 depicts the F values obtained for both quinolones and each cation. This parameter shows the highest value for Fe 2q with ciprofloxacin Ž41.10%., followed by Fe 2q with ofloxacin Ž26.24%.. For Al 3q the values of this factor are similar for both quinolones. Figs. 5 and 6 illustrate the results of the analysis of the residuals by means of the corresponding scatterplot. 4. Discussion Under the same experimental conditions dissolution of ciprofloxacin showed to be much faster than ofloxacin from the commercial dosage forms used. Fig. 1 illustrates these differences which are confirmed by the values of the parameters related to the dissolution rate. MDT for ciprofloxacin is 8.87 " 1.69 min, about one fifth of the corresponding value for ofloxacin Ž45.20 " 13.84 min.. Comparison of the first order dissolution rate constants yields to the same conclusion as this parameter shows a value of 3.04e y 2 " 6.64e y 3 miny1 for ofloxacin and 2.43e y 1 " 6.69e y 2 miny1 for ciprofloxacin. Both, first order kinetics and Weibull distribution function provide an acceptable data fitting with similar AIC values and no statistically significant differences between the estimated dissolution parameters using each function:
243
Q max and K d or td Žeach one calculated as the inverse of the other; td s 1rK d and K d s 1rtd .. Parameters included in Tables 1 and 2 correspond to those values estimated by first order kinetics fitting. Goodness of fitting was additionally tested by comparison of the values of K d obtained by non-linear regression and those calculated from the inverse of MDT, as the relationship K d s 1rMDT comes true for first order kinetic process. Based on it, this relationship can be used as an additional test to verify the goodness of data fitting. According to MDT value, the dissolution rate constants take values of 2.21e y 2 miny1 and 1.13e y 1 miny1 for ofloxacin and ciprofloxacin, respectively, which differ from the estimated values by data fitting, especially for ciprofloxacin Ž1.13e y 1 miny1 versus 2.43e y 1 miny1 .. These results point to pseudo Žor apparent. first order rates as Wagner states for conditions of variable surface area ŽWagner, 1969.. On assessing the influence of polyvalent cations on the dissolution kinetics of these quinolones, Fe 2q proved to be the cation which most significantly decreases Qmax value, followed by Al 3q. Regarding the dissolution rate parameters such as MDT and K d , significative modifications in K d are observed for both quinolones with Fe 2q and also for ofloxacin with Al 3q, nevertheless the changes observed in these parameters do not follow a fixed pattern, but show increases or decreases depending on the drug andror the cation. The interpretation of these results is not easy as the addition of cations to the dissolution media produces the existence of two simultaneous processes: drug dissolution and drug–cation interaction. The characterization of these two kinetic processes by a single first rate constant is a simplistic, non-realistic model, and yields to the estimation of a hybrid constant without a clear physico-chemical meaning. The increase or decrease of this hybrid constant can not be directly associated to changes in the dissolution rate. A similar situation arises when calculating MDT from experimental data. In presence of cations the interaction process takes place and the drug can be in any of the three following states: non-dissolved drug remaining in the dosage form, dissolved drug in the media or drug–cation complex. Under these conditions the parameter calculated using expression number 4 is not the MDT as Q is not the amount of drug dissolved but the amount of free drug in the dissolution media ŽNotice that only free drug is determined.. Because of the limitation showed when using rate parameters to compare dissolution profiles under our experimental conditions, a fit factor F proposed by Moore and Flanner Ž1996. to quantify the total difference between curves, was calculated for each cation, using Eq. Ž7..
244
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
According to the results, Fe 2q modifies the dissolution profile of ciprofloxacin in a 41.1% and only in a 26.24% for ofloxacin; Al 3q affects, from a quantitative point of view, similarly to both quinolones. As a consequence of unweighting the differences between compared profiles, the higher the amount dissolved the stronger its contribution to the total sum of differences and therefore to the F value. In other words the fit factor calculated according to Eq. Ž7. underestimates the differences when those affect to the early phase of the curve which corresponds to the lower amounts of drug dissolved. A ‘weighted’ fit factor will overcome this problem. We propose the analysis of the weighted residuals ŽGabrielsson and Weiner, 1994. as the most suitable statistical method to compare whole dissolution profiles. This type of analysis allows one to ponder and evaluate the modifications from the beginning to the end of the process. The scatterplot of residuals informs about the distribution of differences and bias when they exist. Figs. 5 and 6 show a biased scatterplot of residuals for both quinolones and the two cations showing interesting differences when comparing ofloxacin to ciprofloxacin. The bias is always observed from the early stages for ofloxacin while it does not appear at the beginning for ciprofloxacin but increases with time becoming higher than ofloxacin in the latter stages. These results point out to a quicker but less intense drug–cations interactions for ofloxacin compared to ciprofloxacin. Differences between ofloxacin and ciprofloxacin behaviour in presence of cations, from a physico-chemical point of view, have already been reported in a previous work as referenced in Section 1. Analysis of residuals also permits us to establish the differences between the studied cations. Interaction with Al 3q takes place the fastest while interaction with Fe 2q becomes more evident with time and yields to the highest significative differences in the asymptote of the dissolution curve. The conclusion to be drawn, from the commented results, is that comparison of in vitro dissolution curves provides a useful method to detect factors affecting in vivo absorption processes such as drug–drug andror food–drug interaction. It has been proved for ciprofloxacin and ofloxacin, whose in vitro dissolution profiles in presence of cations change in good agreement with the modifications reported for the oral absorption of these quinolones, administered together with drugs or foods containing Fe 2q or Al 3q. Regarding the statistical method used for comparison, different strategies can be followed, as described and commented in Sections 2.2 and 4, respectively. All of them, in spite of their limitations, bring some information about the changes in dissolution profiles. Nevertheless the
interpretation of changes observed in the fitted parameters and statistical moments must be cautious. Analysis of residuals provides, in any case, the most suitable and recommended method to evaluate the differences between different in vitro dissolution profiles. References Abdou, H.M., 1989. Theory of dissolution. In: Gennaro, A., Migdalof, B., Hasser, G.L., Medwick, T. ŽEds.., Dissolution, Bioavailability and Bioequivalence. MACK Publishing, Pennsylvania, pp. 11–36. Dunne, A.P., Stucker, G.E., 1992. PCNONLIN 4.2 SCI Software-user’s guide. Statistical consultants, Lexington, KY. Flor, S., Guay, D.R.P., Opsahl, T.A., Tack, K., Matzke, G.R., 1990. Effects of magnesium–aluminium hydroxide and calcium carbonate antacids on bioavailability of ofloxacin. Antimicrob. Agents Chemother. 34, 2436–2438. Gabrielsson, J., Weiner, D., 1994. General modelling strategies. In: Gabrielsson, J., Weiner, D. ŽEds.., Pharmacokinetic and Pharmacodynamic Data Analysis. Concepts and Applications. Swedish Pharmaceutical Press, Stockolm, pp. 29–41. Hanson, W.A., 1992. Handbook of Dissolution Testing. Pharmaceutical Technology Publications, Springfield. Kara, M., Hasinoff, B.B., McKay, D.W., Campbell, N.R.C., 1991. Clinical and chemical interactions between iron preparations and ciprofloxacin. Br. J. Clin. Pharmac. 31, 257–261. Lomaestro, B.M., Bailie, G.R., 1993. Effect of multiple staggered doses of calcium on the bioavailability of ciprofloxacin. The Annals of Pharmacother. 27, 1325–1328. Macıas B., Martınez Cabarga, M., Sanchez Navarro, A., ´ Sanchez, ´ ´ ´ Dominguez-Gil Hurle, ´ A., 1994. A physico-chemical study of the interaction of ciprofloxacin and ofloxacin with polyvalent cations. Int. J. Pharm. 106, 229–235. Maeda, Y., Omoda, K., Konishi, T., Takahashi, M., Kihira, K., Hibino, S., Tsukiai, S., 1993. Effects of aluminium-containing antacid on bioavailability of ofloxacin following oral administration of pivaloyloxymethyl ester of ofloxacin as prodrug. Biol. Pharm. Bull. 16, 594–599. Moore, J.W., Flanner, H.H., 1996. Mathematical comparison of dissolution profiles. Pharmaceutical Technology 20 Ž6., 64–74. Martınez-Cabarga, M., Sanchez Navarro, M.A., Colino, C.I., Domınguez´ ´ ´ Gil, A., 1991. Effects of two cations on gastrointestinal absorption of ofloxacin. Antimicrob. Agents Chemother. 35, 2102–2105. Nix, D.E., Watson, W.A., Lener, M.E., Frost, R.W., Krol, G., Goldestein, H., Lettieri, J., Schentag, J.J., 1989. Effect of aluminium and magnesium antacids and ranitidine on the absorption of ciprofloxacin. Clin. Pharmacol. Ther. 46, 700–705. Nix, D., Wilton, J.H., Ronald, B., Distlerath, L., Williams, V.C., Norman, A., 1990. Inhibition of norfloxacin absorption by antacids. Antimicrob. Agents Chemother. 34, 432–435. Pertti, J., Neuvonen, M.D., Kari, T., Kivisto, ¨ M.D., Pasi Lehto, M.B., 1991. Interference of dairy products with the absorption of ciprofloxacin. Clin. Pharmacol. Ther. 50, 498–502. Polk, R.E., Healy, D.P., Sahai, J., Drwal, D., Racht, E., 1989. Effect of ferrous sulfate and multivitamins with zinc on absorption of ciprofloxacin in normal volunteers. Antimicrob. Agents Chemother. 33, 1841–1844. Purves, R.D., 1992. Optimun numerical integration methods for estimation of area-under-the curve ŽAUC. and area uder-the-moment-curve ŽAUMC.. J. Pharmacok. Biopharm. 20 Ž3., 211–226.
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´ Ross, D., Riley, C.M., 1990. Aqueous solubilities of some variously substituted quinolone antimicrobials. Int. J. Pharm. 63, 237–250. Ross, D., Riley, C.M., 1992. Physicochemical properties of the fluoroquinolone antimicrobials: III. Complexation of lomefloxacin with various metal ions and the effect of metal ion complexation on aqueous solubility. Int. J. Pharm. 87, 203–213. Sanchez Navarro, A., Martınez Cabarga, M., Domınguez-Gil Hurle, ´ ´ ´ ´ A., 1994. Oral absorption of ofloxacin administered together with aluminum. Antimicrob. Agents Chemother. 38, 2510–2512. Shiba, K., Sakai, O., Shimada, J., Okazaki, O., Aoki, H., Hadusui, H., 1992. Effects of antacids, ferrous sulfate, and ranitidine on absorption of DR-3355 in humans. Antimicrob. Agents Chemother. 36, 2270– 2274. USP XXII, 1989. The United States Pharmacopeial Convention, 12601 Twinbrook Parkway, Rockville, MD 20852. pp. 1578–1579.
245
Wagner, J.G., 1969. Interpretation of percent dissolved-time plots derived from in vitro testing of conventional tablets and capsules. J. Pharm. Sci. 58 Ž10., 1253–1257. Wayne Frost, R., Lasseter, K.C., Noe, A.J., Shamblen, E.C., Lettieri, J.T., 1992. Effects of aluminum hydroxide and calcium carbonate antacids on the bioavailability of ciprofloxacin. Antimicrob. Agents Chemother. 36, 830–832. Weiss, M., 1992. The relevance of residence time theory to pharmacokinetics. Eur. J. Clin. Pharmacol. 43, 571–579. Wolfson, J.S., Hooper, D.C., 1989. Comparative pharmacokinetics of ofloxacin and ciprofloxacin. Am. J. Med. 87, 31–36. Yamaoka, K., Nakagawa, T., Uno, T., 1977. Application of Akaike’s information criterium ŽAIC. in the evaluation of linear pharmacokinetic equations. J. Pharmacokin. Biopharm. 6, 165–175.
In vitro study of the interaction between quinolones and polyvalent cations M a Sonia Rodrıguez Cruz, Isabel Gonzalez ´ ´ Alonso ) , Amparo Sanchez–Navarro, ´ a M Luisa Sayalero Marinero Department of Pharmacy and Pharmaceutical Technology, Faculty of Pharmacy, UniÕersity of Salamanca, AÕda. Campo Charro sr n, 37007, Salamanca, Spain Received 3 August 1998; revised 16 September 1998; accepted 30 September 1998
Abstract The aim of the present study was to evaluate the influence of aluminium and iron on the in vitro dissolution kinetics of ciprofloxacin and ofloxacin as well as the usefulness of this type of in vitro data to predict modifications in in vivo absorption processes as a consequence of different factors, such as the widely documented in vivo interaction between quinolones and cations. Fitting of experimental data to different theoretical in vitro dissolution profiles was performed by non-linear regression methods and the statistical moments were calculated from raw experimental data. Analysis of residuals applied to dissolution curves as well as statistical comparison of the estimated parameters were carried out to evaluate the in vitro interaction. The results reveal significative modifications of the dissolution profiles of these quinolones as a consequence of the presence of cations, especially for Fe 2q which decreases 34.7% the maximum amount dissolved for ciprofloxacin and 29.1% for ofloxacin. Al 3q also produces a decrease of the total amount of quinolone dissolved although less relevant than Fe 2q. Analysis of residuals proved to be the best statistical method to evaluate differences between whole dissolution profiles, at least under the experimental conditions used. q 1999 Published by Elsevier Science B.V. All rights reserved. Keywords: Ofloxacin; Ciprofloxacin; Polyvalent cations; Dissolution test; Dissolution kinetics; Interaction
1. Introduction Despite the considerable body of information in the literature about the interaction between quinolones and polyvalent cations ŽKara et al., 1991; Shiba et al., 1992; Wayne Frost et al., 1992; Lomaestro and Bailie, 1993; Maeda et al., 1993; Sanchez Navarro et al., 1994., many ´ aspects related to this process remain to be clarified, including the mechanism or experimental and clinical conditions controlling this interaction. Several authors ŽNix et al., 1989; Polk et al., 1989. have suggested the formation of an insoluble, non-absorbable complex between the quinolone and the corresponding ) Corresponding author. Tel.: q34-23-294-536; fax: q34-23-294-515; e-mail: [email protected]
cation; others ŽRoss and Riley, 1990; Ross and Riley, 1992. have proposed the formation of a ‘micellar phase’ as a consequence of changes in the net charge of the compound and, recently, Macıas et al. Ž1994. have ´ Sanchez ´ found that ofloxacin with metal cations follows the general pattern expected, leading to the formation of salts or coordination compounds, while ciprofloxacin retains the molecular network and no functional group seems to be lost or modified after entering into contact with different polyvalent cations. In contrast, the effect of polyvalent cations on the oral absorption of ciprofloxacin has been reported to be higher than the effect observed with ofloxacin ŽWolfson and Hooper, 1989; Flor et al., 1990.. Additionally, there are some differences among the results found by different authors studying quantitative modifications in the oral absorption of quinolones administered
0031-6865r99r$ - see front matter q 1999 Published by Elsevier Science B.V. All rights reserved. PII: S 0 0 3 1 - 6 8 6 5 Ž 9 8 . 0 0 0 2 9 - 6
238
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
with drugs or food containing cations ŽNix et al., 1990; Martınez-Cabarga et al., 1991; Pertti et al., 1991.. ´ The conclusion to be drawn from the available information is that some type of interaction takes place between quinolones and cations which leads to a decrease in the amount of antibacterial drug absorbed when administered by the oral route, but the interaction process seems to be very sensitive to slight changes in factors related to the experimental andror clinical conditions of the different assays. This would justify the wide dispersion of the results reported in the literature. The present study was therefore carried out to evaluate the influence of the presence of Al 3q and Fe 2q ions on the in vitro dissolution profiles of ciprofloxacin and ofloxacin. Also to analyze the advantages and limitations of different available statistical methods used to compare in vitro dissolution curves. Since drug dissolution is the previous step to drug absorption, the study of the modifications of the dissolution profiles will provide useful information about the possible changes in drug absorption as a consequence of interactions.
2. Materials and methods 2.1. Materials and equipment 2.1.1. Drugs and reagents The quinolone solid dosage formulations assayed were Baycip w 250 mg ŽBayer, Barcelona, Spain. and Tarivid w 200 mg ŽHoescht Farma, Barcelona, Spain.. The cation preparations were: Fero-Gradumet w ŽAbbot Cientifica, Madrid, Spain. and aluminium hydroxide powder ŽMerck, Darmstadt, Germany.. 2.1.2. Equipment Hanson SR 6-Flask Dissolution test station ŽHanson Research, Chatsworth, CA.. U-2000 Spectrophotometer ŽHitachi, Tokyo, Japan.. 5000 Liquid Chromatograph ŽVarian. connected to a Kontron SFM 25 fluorescence detector. 2.2. Methods 2.2.1. In Õitro dissolution assay Dissolution profiles were determined at 37 " 0.58C in 900 ml of phosphate buffer solution, pH s 6. Commercially available dissolution equipment employing the paddle apparatus as described in the USP XXII Ž1989. was used. Rotation speed was maintained at 50 r.p.m. The
paddle was positioned to extend to exactly 2.5 cm above the flask bottom. Samples Ž2 ml. were taken with a graduated syringe at the following times. Ciprofloxacin: 1, 2, 3, 4, 5, 7, 10, 15, 20, 30, 45, 60 and 90 min. Ofloxacin: 5, 10, 15, 20, 30, 45, 60, 90, 120, 180, 240, 300 and 360 min. Each sample was replaced by an equal volume of dissolution media to keep the total volume constant. A correction was made to take into account the cumulative removed volumes when determining the total amount dissolved as a function of time. All samples were suitably filtered, placing a filter at the end of the sample probe, and kept at 58C until immediate analytical determination. The suitability of the paddle apparatus was checked using the USP prednisone and salicylic acid calibrators Žcalibrators for system suitability test of basket and paddle dissolution apparatus. ŽHanson, 1992.. One tablet of the corresponding quinolone was placed in each filled flask Ž6 tablets per run. when establishing the dissolution profiles of the quinolones in the absence of cations. These profiles were considered as reference curves. To evaluate the influence of the different cations on the quinolone dissolution kinetics, the corresponding cation preparation was added to each flask at the same time as the quinolone formulation. For Al 3q, a sufficient amount of aluminium hydroxide powder to obtain 2.5 g of Al 3q was added; when studying the influence of Fe 2q, two 525 mg FeSO4 tablets were included in the dissolution media. The amounts of cations were selected based on the usual clinical doses. Addition of quinolones or cations do not modifies the pH of dissolution media for the sampling period.
2.2.2. Analytical techniques Samples from all assays, except those on Fe 2q, were analyzed by an UV-spectrophotometric technique at wavelengths of 287 nm and 271 nm for ofloxacin and ciprofloxacin, respectively. Standard curves were prepared in phosphate buffer solution ŽpH s 6. at a concentration range of 0.5–20.0 mgrl for ofloxacin and 0.5–10.0 mgrl for ciprofloxacin. Samples were diluted when necessary to achieve concentrations included within the standard curve range. The UV technique showed adequate linearity, precision and accuracy with coefficients of variation of less than 5% and 2.5% for ofloxacin and ciprofloxacin, respectively. Samples from the Fe 2q assays were analyzed by an HPLC technique. A Varian 5.000 Chromatograph connected to a variable wavelength fluorescence detector was
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
239
Fig. 1. Dissolution profiles of ciprofloxacin and ofloxacin corresponding to in vitro tests carried out under the following experimental conditions: phosphate buffer solution, pH s6, T a s 378C. Each curve is the mean of six experiments.
used. The 15 cm-long column Ž4 mm internal diameter. was packed with Nucleosil w 120 R-18 5m of particle size. The mobile phase consisted of acetonitrile and 0.025 M phosphate buffer solution, pH 3, Ž9:91 vrv. and the flow rate was 2 mlrmin. The fluorescence excitation and emission wavelengths were set at 277 nm and 445 nm for ciprofloxacin and 330 nm and 450 nm for ofloxacin, respectively. 50 ml of the internal standard Žeach quinolone was used as the internal standard of the other one. was added to 200 ml of the sample. The mixture was vigorously shaken for 30 s and 100 ml was injected into the chromotograph. Standard curves were prepared as indicated for UV detection and samples were also diluted if necessary. This HPLC technique also showed adequate linearity, precision and accuracy with coefficients of variation of less than 10% and 8% for ofloxacin and ciprofloxacin, respectively.
Both, UV and HPLC, techniques quantifies free drug in the dissolution media.
2.2.3. Data analysis The dissolution pattern, described as the cumulative amount of drug dissolved, was fitted to different mathematical expressions, commonly related to in vitro dissolution profiles, for immediate release dosage formulations ŽAbdou, 1989.. These expressions are: Ø Zero-order kinetics: dQ dt
sK0
Ž 1.
where: dQrdt s dissolution rate; K o s constant amount of drug dissolved per unit time
Table 1 Mean dissolution parameters of ofloxacin under standard conditions and in presence of two different cations
Ofloxacin 200 mg Oflo 200qAl 3q Oflo 200qFe 2q
MDT Žmin.
CV
Qma x Žmg.
K d Žminy1 .
45.20"13.84 98.41"37.67 36.17"11.41
1.20"2.80ey1 1.25"9.42ey2 1.09"1.05ey1
173.60"11.67 152.11"16.89 123.16"3.58
3.04ey2"6.64ey3 1.77ey2"3.33ey3 4.33ey2"7.20ey3
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
240
Table 2 Mean dissolution parameters of ciprofloxacin under standard conditions and in presence of two different cations
Ciprofloxacin 250 mg Cipro 250 mgqAl 3q Cipro 250 mgqFe 2q
MDT Žmin.
CV
Qma x Žmg.
K d Žminy1 .
8.87"1.69 11.32"4.79 17.32"11.66
1.82"3.74ey1 1.85"3.99ey1 1.04"1.06ey1
219.41"7.74 180.16"46.47 143.24"8.39
2.43ey1"6.69ey2 2.32ey1"9.70ey2 4.76ey1"1.12ey1
Ø First-order kinetics: Q t s Q max Ž 1 y eyK d t .
Ž 2.
where: Q t s amount of drug dissolved at time t; Qmax s maximum amount of drug dissolved; K d s first-order dissolution rate constant Ø Weibull distribution function: Q t s Q max 1 y ey Ž trt d .
b
Ž 3.
where: td s time necessary to achieve 63.2% of total drug dissolved; b s an adimensional parameter defined by the shape of the dissolution curve plot Žrelated to the order of the process.. The fitting was carried out using the PCNONLIN 4.0 program ŽDunne and Stucker, 1992.. The statistical criterion for the selection of the optimum fitting was the minimum AIC value ŽTest MAICE. ŽYamaoka et al., 1977..
In addition to dissolution curve fitting the mean in vitro dissolution time ŽMDT. and the coefficient of variation ŽCV. were calculated using trapezoidal integration ŽPurves, 1992.. The MDT is defined as follows: `
MDT s
H0 t dQrQ
max
Ž 4.
where: Q s amount of drug dissolved at time t; Qmax s maximum amount of drug dissolved. The CV values, which inform about the dispersion of the dissolution process ŽWeiss, 1992., were calculated according to the following expression: CV s
'VDT MDT
Ž 5.
VDT is the variance of the dissolution times, that is the moment of second degree, as the dissolution profile may be considered a distribution function of the residence times
Fig. 2. Dissolution profiles of ofloxacin corresponding to in vitro tests carried out in absence of cations and in presence of Al 3q or Fe 2q. Each curve is the mean of six experiments.
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
241
Fig. 3. Dissolution profiles of ciprofloxacin corresponding to in vitro tests carried out in absence of cations and in presence of Al 3q or Fe 2q. Each curve is the mean of six experiments.
of each drug molecule in the pharmaceutical formulation. VDT was calculated by trapezoidal integration as follows: `
VDT s
H0
2 Ž t y MDT. dQrQmax
Ž 6.
2.2.4.2. Estimation of a ‘ fit factor’ F In order to measure the dissimilarity between the reference curves and those obtained in presence of cations a ‘fit factor’ F was calculated for each added cation. This factor approximates the percent error between two curves and
2.2.4. Statistical analysis 2.2.4.1. ANOVA Statistical moments ŽMDT, CV. as well as other dissolution parameters Ž Qmax , K d . were calculated from each individual dissolution profile and the statistical comparison of each parameter in presence and absence of cations, for both quinolones, was carried out using the test ANOVA at the standard significance level Ž p - 0.05..
Table 3 Modifications of dissolution parameters of ciprofloxacin and ofloxacin calculated as percentage of change of mean parameter values OFLOXACINqAl3q OFLOXACINqFe2q CIPROFLOXACINqAl3q CIPROFLOXACINqFe2q
MDT
Qma x
Kd
≠ 117.72 x 19.98 ≠ 27.62 ≠ 95.26
x 12.38 x 29.06 x 17.89 x 34.72
x 41.78 ≠ 42.43 x 4.53 ≠ 95.88
Fig. 4. Differences between dissolution profiles in absence and presence of each studied cation, quantified by a fit factor value Ž F . calculated from in vitro individual dissolution curves of ciprofloxacin and ofloxacin.
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
242
2.2.4.3. Analysis of the residuals Considering the dissolution profile of the quinolone in absence of cations as the reference curve, the weighted residual was calculated for each sample time and every dissolution curve, according to the well known expression: Qr y Qc Rs Ž 8. Qr Analysis of residuals was carried out for each cation by plotting residuals versus time and examining the scatterplots which inform about the trend in deviations, if they exist.
3. Results Fig. 1 shows the mean in vitro dissolution curves of ciprofloxacin and ofloxacin in phosphate buffer solution
Fig. 5. Scatterplot of residuals for dissolution curves of ofloxacin in presence of cations using the dissolution curve of the quinolone in absence of cations as the reference profile.
quantifies, by a single numerical parameter, the total difference between compared profiles. According to Moore and Flanner Ž1996. this fit factor is defined by the following equation: n
Ý < Qr y Qc < Fs
is1 n
100
Ž 7.
Ý Qr is1
where: n is the total number of samples; Qr and Qc are the amount of drug dissolved under standard conditions Žreference curve. and in presence of the cation, respectively. According to that, the fit factor F is calculated for each cation and each quinolone.
Fig. 6. Scatterplot of residuals for dissolution curves of ciprofloxacin in presence of cations using the dissolution curve of the quinolone in absence of cations as the reference profile.
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
ŽpH s 6.; the corresponding dissolution parameters are included in Tables 1 and 2. The maximum dissolved amount of drug, defined by the asymptote of the fitted curves, are 219.41 " 7.74 mg and 173.60 " 11.67 mg, which represent percentages of 88% and 87% of the theoretical dose for ciprofloxacin and ofloxacin, respectively. The first order rate constant show a value of 2.43e y 1 " 6.69e y 2 miny1 for ciprofloxacin, significantly higher than the 3.04e y 2 " 6.64e y 3 miny1 corresponding to ofloxacin. In agreement with it, the MDT, calculated by trapezoidal integration from raw experimental data, show statistically significant differences between ciprofloxacin Ž8.87 " 1.69 min. and ofloxacin Ž45.20 " 13.84 min.. The dispersion of the dissolution times, measured as the CV, is slightly higher for ciprofloxacin Ž1.82 " 3.74e y 1. than for ofloxacin Ž1.20 " 2.80e y 1.. Figs. 2 and 3 illustrate the effect of the presence of different cations on the in vitro dissolution profiles of ofloxacin and ciprofloxacin, respectively. The cations yield to a decrease of the maximum amount dissolved for ciprofloxacin as well as ofloxacin. Modifications of the first order rate constant are also observed as a consequence of addition of Fe 2q to the dissolution media. Table 3 summarizes the modification of the dissolution parameters caused by presence of each cation; these are expressed as percentage of change calculated from the mean parameter values. Fig. 4 depicts the F values obtained for both quinolones and each cation. This parameter shows the highest value for Fe 2q with ciprofloxacin Ž41.10%., followed by Fe 2q with ofloxacin Ž26.24%.. For Al 3q the values of this factor are similar for both quinolones. Figs. 5 and 6 illustrate the results of the analysis of the residuals by means of the corresponding scatterplot. 4. Discussion Under the same experimental conditions dissolution of ciprofloxacin showed to be much faster than ofloxacin from the commercial dosage forms used. Fig. 1 illustrates these differences which are confirmed by the values of the parameters related to the dissolution rate. MDT for ciprofloxacin is 8.87 " 1.69 min, about one fifth of the corresponding value for ofloxacin Ž45.20 " 13.84 min.. Comparison of the first order dissolution rate constants yields to the same conclusion as this parameter shows a value of 3.04e y 2 " 6.64e y 3 miny1 for ofloxacin and 2.43e y 1 " 6.69e y 2 miny1 for ciprofloxacin. Both, first order kinetics and Weibull distribution function provide an acceptable data fitting with similar AIC values and no statistically significant differences between the estimated dissolution parameters using each function:
243
Q max and K d or td Žeach one calculated as the inverse of the other; td s 1rK d and K d s 1rtd .. Parameters included in Tables 1 and 2 correspond to those values estimated by first order kinetics fitting. Goodness of fitting was additionally tested by comparison of the values of K d obtained by non-linear regression and those calculated from the inverse of MDT, as the relationship K d s 1rMDT comes true for first order kinetic process. Based on it, this relationship can be used as an additional test to verify the goodness of data fitting. According to MDT value, the dissolution rate constants take values of 2.21e y 2 miny1 and 1.13e y 1 miny1 for ofloxacin and ciprofloxacin, respectively, which differ from the estimated values by data fitting, especially for ciprofloxacin Ž1.13e y 1 miny1 versus 2.43e y 1 miny1 .. These results point to pseudo Žor apparent. first order rates as Wagner states for conditions of variable surface area ŽWagner, 1969.. On assessing the influence of polyvalent cations on the dissolution kinetics of these quinolones, Fe 2q proved to be the cation which most significantly decreases Qmax value, followed by Al 3q. Regarding the dissolution rate parameters such as MDT and K d , significative modifications in K d are observed for both quinolones with Fe 2q and also for ofloxacin with Al 3q, nevertheless the changes observed in these parameters do not follow a fixed pattern, but show increases or decreases depending on the drug andror the cation. The interpretation of these results is not easy as the addition of cations to the dissolution media produces the existence of two simultaneous processes: drug dissolution and drug–cation interaction. The characterization of these two kinetic processes by a single first rate constant is a simplistic, non-realistic model, and yields to the estimation of a hybrid constant without a clear physico-chemical meaning. The increase or decrease of this hybrid constant can not be directly associated to changes in the dissolution rate. A similar situation arises when calculating MDT from experimental data. In presence of cations the interaction process takes place and the drug can be in any of the three following states: non-dissolved drug remaining in the dosage form, dissolved drug in the media or drug–cation complex. Under these conditions the parameter calculated using expression number 4 is not the MDT as Q is not the amount of drug dissolved but the amount of free drug in the dissolution media ŽNotice that only free drug is determined.. Because of the limitation showed when using rate parameters to compare dissolution profiles under our experimental conditions, a fit factor F proposed by Moore and Flanner Ž1996. to quantify the total difference between curves, was calculated for each cation, using Eq. Ž7..
244
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´
According to the results, Fe 2q modifies the dissolution profile of ciprofloxacin in a 41.1% and only in a 26.24% for ofloxacin; Al 3q affects, from a quantitative point of view, similarly to both quinolones. As a consequence of unweighting the differences between compared profiles, the higher the amount dissolved the stronger its contribution to the total sum of differences and therefore to the F value. In other words the fit factor calculated according to Eq. Ž7. underestimates the differences when those affect to the early phase of the curve which corresponds to the lower amounts of drug dissolved. A ‘weighted’ fit factor will overcome this problem. We propose the analysis of the weighted residuals ŽGabrielsson and Weiner, 1994. as the most suitable statistical method to compare whole dissolution profiles. This type of analysis allows one to ponder and evaluate the modifications from the beginning to the end of the process. The scatterplot of residuals informs about the distribution of differences and bias when they exist. Figs. 5 and 6 show a biased scatterplot of residuals for both quinolones and the two cations showing interesting differences when comparing ofloxacin to ciprofloxacin. The bias is always observed from the early stages for ofloxacin while it does not appear at the beginning for ciprofloxacin but increases with time becoming higher than ofloxacin in the latter stages. These results point out to a quicker but less intense drug–cations interactions for ofloxacin compared to ciprofloxacin. Differences between ofloxacin and ciprofloxacin behaviour in presence of cations, from a physico-chemical point of view, have already been reported in a previous work as referenced in Section 1. Analysis of residuals also permits us to establish the differences between the studied cations. Interaction with Al 3q takes place the fastest while interaction with Fe 2q becomes more evident with time and yields to the highest significative differences in the asymptote of the dissolution curve. The conclusion to be drawn, from the commented results, is that comparison of in vitro dissolution curves provides a useful method to detect factors affecting in vivo absorption processes such as drug–drug andror food–drug interaction. It has been proved for ciprofloxacin and ofloxacin, whose in vitro dissolution profiles in presence of cations change in good agreement with the modifications reported for the oral absorption of these quinolones, administered together with drugs or foods containing Fe 2q or Al 3q. Regarding the statistical method used for comparison, different strategies can be followed, as described and commented in Sections 2.2 and 4, respectively. All of them, in spite of their limitations, bring some information about the changes in dissolution profiles. Nevertheless the
interpretation of changes observed in the fitted parameters and statistical moments must be cautious. Analysis of residuals provides, in any case, the most suitable and recommended method to evaluate the differences between different in vitro dissolution profiles. References Abdou, H.M., 1989. Theory of dissolution. In: Gennaro, A., Migdalof, B., Hasser, G.L., Medwick, T. ŽEds.., Dissolution, Bioavailability and Bioequivalence. MACK Publishing, Pennsylvania, pp. 11–36. Dunne, A.P., Stucker, G.E., 1992. PCNONLIN 4.2 SCI Software-user’s guide. Statistical consultants, Lexington, KY. Flor, S., Guay, D.R.P., Opsahl, T.A., Tack, K., Matzke, G.R., 1990. Effects of magnesium–aluminium hydroxide and calcium carbonate antacids on bioavailability of ofloxacin. Antimicrob. Agents Chemother. 34, 2436–2438. Gabrielsson, J., Weiner, D., 1994. General modelling strategies. In: Gabrielsson, J., Weiner, D. ŽEds.., Pharmacokinetic and Pharmacodynamic Data Analysis. Concepts and Applications. Swedish Pharmaceutical Press, Stockolm, pp. 29–41. Hanson, W.A., 1992. Handbook of Dissolution Testing. Pharmaceutical Technology Publications, Springfield. Kara, M., Hasinoff, B.B., McKay, D.W., Campbell, N.R.C., 1991. Clinical and chemical interactions between iron preparations and ciprofloxacin. Br. J. Clin. Pharmac. 31, 257–261. Lomaestro, B.M., Bailie, G.R., 1993. Effect of multiple staggered doses of calcium on the bioavailability of ciprofloxacin. The Annals of Pharmacother. 27, 1325–1328. Macıas B., Martınez Cabarga, M., Sanchez Navarro, A., ´ Sanchez, ´ ´ ´ Dominguez-Gil Hurle, ´ A., 1994. A physico-chemical study of the interaction of ciprofloxacin and ofloxacin with polyvalent cations. Int. J. Pharm. 106, 229–235. Maeda, Y., Omoda, K., Konishi, T., Takahashi, M., Kihira, K., Hibino, S., Tsukiai, S., 1993. Effects of aluminium-containing antacid on bioavailability of ofloxacin following oral administration of pivaloyloxymethyl ester of ofloxacin as prodrug. Biol. Pharm. Bull. 16, 594–599. Moore, J.W., Flanner, H.H., 1996. Mathematical comparison of dissolution profiles. Pharmaceutical Technology 20 Ž6., 64–74. Martınez-Cabarga, M., Sanchez Navarro, M.A., Colino, C.I., Domınguez´ ´ ´ Gil, A., 1991. Effects of two cations on gastrointestinal absorption of ofloxacin. Antimicrob. Agents Chemother. 35, 2102–2105. Nix, D.E., Watson, W.A., Lener, M.E., Frost, R.W., Krol, G., Goldestein, H., Lettieri, J., Schentag, J.J., 1989. Effect of aluminium and magnesium antacids and ranitidine on the absorption of ciprofloxacin. Clin. Pharmacol. Ther. 46, 700–705. Nix, D., Wilton, J.H., Ronald, B., Distlerath, L., Williams, V.C., Norman, A., 1990. Inhibition of norfloxacin absorption by antacids. Antimicrob. Agents Chemother. 34, 432–435. Pertti, J., Neuvonen, M.D., Kari, T., Kivisto, ¨ M.D., Pasi Lehto, M.B., 1991. Interference of dairy products with the absorption of ciprofloxacin. Clin. Pharmacol. Ther. 50, 498–502. Polk, R.E., Healy, D.P., Sahai, J., Drwal, D., Racht, E., 1989. Effect of ferrous sulfate and multivitamins with zinc on absorption of ciprofloxacin in normal volunteers. Antimicrob. Agents Chemother. 33, 1841–1844. Purves, R.D., 1992. Optimun numerical integration methods for estimation of area-under-the curve ŽAUC. and area uder-the-moment-curve ŽAUMC.. J. Pharmacok. Biopharm. 20 Ž3., 211–226.
M a S. Rodrıguez Cruz et al.r Pharmaceutica Acta HelÕetiae 73 (1999) 237–245 ´ Ross, D., Riley, C.M., 1990. Aqueous solubilities of some variously substituted quinolone antimicrobials. Int. J. Pharm. 63, 237–250. Ross, D., Riley, C.M., 1992. Physicochemical properties of the fluoroquinolone antimicrobials: III. Complexation of lomefloxacin with various metal ions and the effect of metal ion complexation on aqueous solubility. Int. J. Pharm. 87, 203–213. Sanchez Navarro, A., Martınez Cabarga, M., Domınguez-Gil Hurle, ´ ´ ´ ´ A., 1994. Oral absorption of ofloxacin administered together with aluminum. Antimicrob. Agents Chemother. 38, 2510–2512. Shiba, K., Sakai, O., Shimada, J., Okazaki, O., Aoki, H., Hadusui, H., 1992. Effects of antacids, ferrous sulfate, and ranitidine on absorption of DR-3355 in humans. Antimicrob. Agents Chemother. 36, 2270– 2274. USP XXII, 1989. The United States Pharmacopeial Convention, 12601 Twinbrook Parkway, Rockville, MD 20852. pp. 1578–1579.
245
Wagner, J.G., 1969. Interpretation of percent dissolved-time plots derived from in vitro testing of conventional tablets and capsules. J. Pharm. Sci. 58 Ž10., 1253–1257. Wayne Frost, R., Lasseter, K.C., Noe, A.J., Shamblen, E.C., Lettieri, J.T., 1992. Effects of aluminum hydroxide and calcium carbonate antacids on the bioavailability of ciprofloxacin. Antimicrob. Agents Chemother. 36, 830–832. Weiss, M., 1992. The relevance of residence time theory to pharmacokinetics. Eur. J. Clin. Pharmacol. 43, 571–579. Wolfson, J.S., Hooper, D.C., 1989. Comparative pharmacokinetics of ofloxacin and ciprofloxacin. Am. J. Med. 87, 31–36. Yamaoka, K., Nakagawa, T., Uno, T., 1977. Application of Akaike’s information criterium ŽAIC. in the evaluation of linear pharmacokinetic equations. J. Pharmacokin. Biopharm. 6, 165–175.