Validated spectrophotometric methods for simultaneous determination of oxytetracycline associated with diclofenac sodium or with piroxicam in veterinary pharmaceutical dosage form

2.3.1 Selection of wavelengths

2.4.1 Standard stock solution preparation

Standard stock solutions were prepared by analytically weighing 10 mg of each RSS (OTC, DICLO and PIRO) which were separately transferred to 100 mL volumetric flasks, adding approximately 50 mL of acetonitrile. Then, these solutions were taken to the ultrasonic bath during 15 min. Subsequently, the volumes were completed to 100 mL with acetonitrile, obtaining solutions at 100 μg mL⁻¹ of each drug.

2.4.2 Commercial samples stock solutions preparation

Terramicina Mais® (Sample A) and Terracam Plus® (Sample B) were used in the validation experiments of the OTC + DICLO and the OTC + PIRO association, respectively. The specifications declared contents of these commercial products are found in Section 2.1 of this present work.

Using an automatic pipette, 1000 μL of commercial samples A and B were separately transferred into 100 mL volumetric flasks, adding 50 mL of ultrapure water and leaving in ultrasonic bath for 15 min. Subsequently, the volumes were completed with the same solvent, obtaining the commercial sample stock solution A (OTC at 2000 μg mL⁻¹ and DICLO at 100 μg mL⁻¹) and B (OTC at 2200 μg mL⁻¹ and PIRO at 120 μg mL⁻¹).

2.4.3 Analytical parameters

3.1.1 Selection of wavelengths

As previously mentioned, the veterinary pharmaceutical products under study are binaries mixtures. For these reason, it was not possible to quantify directly and simultaneously all drugs by zero order spectrophotometry due to the overlapping of the electronic transition bands of the UV spectra (Fig. 2). However, the direct quantification of OTC by zero order spectrophotometry was possible in the association with DICLO. It was performed at 360 nm, since DICLO does not show significant absorption at this wavelength (Fig. 2a).

Close

Fig. 2 The overlap of absorption's spectra of (a) OTC and DICLO, as well as (b) OTC and PIRO, obtained from the SSR solutions prepared in acetonitrile:water (50:50, v/v, pH 2.5). OTC: oxytetracycline; DICLO: diclofenac sodium and PIRO: piroxicam. * Wavelength used for direct quantification of OTC in association with DICLO.

Export to PPT

The zero order spectra of OTC and DICLO (Fig. 2a), as well as OTC and PIRO (Fig. 2b) were derived for the first order (Fig. 3) in order to eliminate the overlapping issues of spectral bands.

Close

Fig. 3 The overlap of absorption's spectra in the first derivative of (a) OTC and DICLO, as well as (b) OTC and PIRO, obtained from the solutions of SSR prepared in acetonitrile:water (50:50, v/v, pH 2.5). OTC: oxytetracycline; DICLO: diclofenac sodium and PIRO: piroxicam. * Wavelength used for quantification of DICLO and PIRO in their respective associations (OTC zero crossing point); ^** Wavelength used for quantification of OTC in association with PIRO (PIRO zero crossing point).

Export to PPT

The derivative technique used for the quantification of drugs was the zero crossing point, which consists of measuring the absolute value of the absorption amplitude of a component of the mixture at the wavelength in which the absorption derivative of the other component is zero (Donato et al., 2010). It is widely used for the quantification of binary mixtures of drugs in pharmaceutical dosage forms (El-Sayed and El-Salem, 2005). For instance, recently our research group employed this helpful technique for simultaneous determination of gatifloxacin and prednisolone acetate in ophthalmic suspensions (Sversut et al., 2017a, 2017b). Therefore, as observed in Fig. 3, DICLO and PIRO were quantified, in their respective OTC associations, at 298.5 nm (OTC zero crossing point in the both binary mixtures), whereas OTC in association with PIRO was quantified at 339 nm (PIRO zero crossing point).

Although in the first OTC derivative spectra there are other zero crossing points (about 210, 220, 230, 270 and 360 nm), the 298.5 nm was chosen for DICLO and PIRO analysis, since in the others wavelengths were observed deviations in the linear correlation of the Lambert-Beer laẃs. Besides that, below 260 nm interference of adjuvants occurs (see the following selectivity result section). For the same reason, zero crossing points in the PIRO spectrum, such as at 275 nm, were not chosen for OTC analysis.

Furthermore, as mentioned in Section 2.3, the first derivative spectra were pre-processed using the Stavizky and Golay (1964) smoothing filter with 15 to 20 point window and second order polynomial. This tool reduces high frequency spectral noise by subdividing the sample space into odd subintervals values. Then, applying the least squares method, it is obtained a polynomial function, where the center point of the interval is replaced by the value calculated through the function (Press et al., 2007).

3.2.1 Linearity, LOQ and LOD

Table 4 shows the parameters obtained from the analytical curves of each drug under study. All analytical curves were linear (r > 0.9996) in the employed concentration range. Moreover, the values of LOQ and LOD indicated that the proposed spectrophotometric methods have presented adequate sensitivity for the simultaneous or isolated determination of OTC, DICLO and PIRO in pharmaceutical products.

The proposed spectrophotometric methods for determination of OTC (zero and first derivative orders) were more sensitive than those developed by Jelikić-Stankov et al. (1989), as well as by Gallego and Arroyo (2002), since these authors found LOQ values of 200 and 185 μg mL⁻¹, respectively. Besides that, our proposed methods have a broader linear concentration range than others works previously mentioned, which employed concentration ranges of 2.5 to 34.78 μg mL⁻¹ (Jelikić-Stankov et al., 1989) and of 2 to 36 μg mL⁻¹ (Gallego and Arroyo, 2002). Furthermore, higher linear correlation coefficients were obtained for DICLO and PIRO compared to those found in others spectrophotometric methods described in the literature for the quantification of these drugs in pharmaceutical products (Basan et al., 2001; Sznitowska and Stokrocka, 2007).

Table 5 depicts the linearity statistical analysis results. The results obtained for the Cochrańs test allowed the application of the Ordinary Least Squares method and indicate that y-axis values (dependent variable) are homoscedastic, since the values of C_calculated < C_critical. The p-value of the linear regression, obtained by the ANOVA test of the analytical curves, was highly significant (p < 0.0001), which confirms the hypothesis of proportionality between the variables x (concentration) and y (signal).

The residues of each observation are given by the difference between the observed value and the fitted value of the regression line. Statistical analysis of the residues gives us information about why the model did not fit well with the data, i.e., nonlinearity. This occurs due to the presence of outliers and nonconstant variance of the residues (heteroscedasticity) (Chan, 2004). As shown in Table 5, all residues have had a normal distribution according to the results obtained in the Shapiro-Wilk test (p > 0.05).

The graphs of the residues obtained from the linear regression of each drug for assessment of the linearity of the method are shown in Fig. 4, in which we could observe values close to zero, indicative of the absence of outliers. Besides that, the scatter of the points shows no clear pattern. Then, we could conclude that the residues variance is constant.

Close

Fig. 4 Residues obtained from the curves' linear regression of (a) DICLO-UV first derivative order at 298.5 nm, (b) OTC-UV first derivative order at 339 nm, (c) PIRO-UV first derivative order at 198.5 nm and (d) OTC-UV zero order at 360 nm.

Export to PPT

3.2.2 Precision

Table 6 shows the results obtained from the precision assay of the analyzed drugs through the proposed methods. The values obtained for repeatability (intraday precision) showed RSD below the recommended maximum limit of 5.0% (ICH, 2005). The difference between variances obtained from five determinations carried out on two consecutive days for each concentration level was shown to be non-significant (p > 0.05) by the Student́s t-test with α = 5%. Therefore, as averages of the concentrations obtained on days 1 and 2 did not differ significantly, it was confirmed the intermediate accuracy of the proposed method.

Contrary to what was done in this work, where we verified 3 levels of concentrations of each drug in the evaluation of precision, Gallego and Arroyo (2002) checked only one concentration level (16 mg mL⁻¹) for repeatability assay of OTC, and these authors found a RSD value of 0.61%. Basan et al. (2001) employed two levels of concentration (8 and 15 µg mL⁻¹) to analyze the repeatability and intermediate precision of PIRO, reaching a RSD value lower than 1.50%.

3.2.3 Accuracy

The values obtained in the recovery test are presented in Table 7. The amounts of recovered OTC, DICLO and PIRO were within the acceptable limits of 98–102% for all three concentration levels analyzed (AOAC, 2005). Therefore, the proposed methods have adequate accuracy for the simultaneous or isolated determination of OTC and DICLO, as well as OTC and PIRO.

Similar to what we found, Prasad and Rao (2010) achieved the recovery limit of 98–102% for OTC, using a visible spectrophotometric method for simultaneous determination of OTC in pharmaceutical samples. Didamony and Amin (2004), who proposed an adaptation of a color reaction for spectrophotometric determination of DICLO and PIRO in pure form and in pharmaceutical formulations, also reached a mean percentage recoveries comparable those obtained in our study. The method proposed by El-Didamony and Amin (2004) gave a mean percentage of 99.8 ± 1.2% for DICLO and 100.3 ± 0.8% for PIRO.

3.2.4 Selectivity

Fig. 5A and B, that showing the UV-zero and first derivative orders spectra of the sample solution A (Terramicina Mais®), respectively, revealed that the adjuvants commonly present in veterinary injectable solutions have absorption below 260 nm. However, they do not show analytical signal at the wavelengths at which the drugs are analyzed, that is, in order zero at 360 nm (OTC) and at the first derivative and 298.5 nm (DICLO). Therefore, they do not interfere in the spectrophotometric analysis of OTC and DICLO in the pharmaceutical products.

Close

Fig. 5 The (a) UV zero and (b) UV first derivative orders spectra of commercial sample solution A overlapping the placebo solution. Commercial sample A: Terramicina Mais®, containing 40 μg mL⁻¹ OTC and 14 μg mL⁻¹ of DICLO. 298.5 and 360 nm were the wavelengths used to determination of DICLO and OTC in the first derivative and zero orders, respectively.

Export to PPT

Fig. 6 shows the first derivative order spectrum of commercial sample solution B overlapping the placebo solution spectrum. It is possible to verify that the main adjuvants present in the analyzed commercial samples do not present analytical signal in the first derivative order at the wavelengths where OTC and PIRO are analyzed, 339 and 298.5 nm, respectively. These results confirm the selectivity of the proposed methods.

Close

Fig. 6 The first derivative spectrum of commercial sample solution B overlapping the placebo solution spectrum. Commercial sample B: Terracam Plus®, containing 44 μg mL⁻¹ of OTC and 14.4 μg mL⁻¹ of PIRO. 298.5 and 339 nm were the wavelengths used to determination of PIRO and OTC, respectively.

Export to PPT

3.2.5 Robustness

Fig. 7 shows the absolute values of the Effects observed to the robustness test calculated by Eq. (4). The critical values $(\sqrt{2S)}$ $(\sqrt{2S})$ for OTC and DICLO were, respectively, 1.17 and 0.10 (Fig. 7a). For the association of OTC and PIRO, the critical values were, respectively, 2.59 and 0.12 (Fig. 7b).

Close

Fig. 7 Effects observed to the robustness test related to combination of (a) OTC + DICLO and (b) OTC + PIRO. A/a to G/g letters represent the selected factors and its variation (A: proportion of acetonitrile, B: acetonitrile supplier, C: pH of the solvent mixture, D: wavelength, E: sonication time, F: nylon syringe filter, G: Room temperature).

Export to PPT

No change in the parameters selected for the proposed methods resulted in absolute values of Effect above the critical values for each drug, indicating the method remained robust in face of the evaluated variations. However, the effect that most approached the critical value was due to the variation in sonication time (E) to PIRO. This can be explained by the fact that the increase in sonication time has influenced the solubility of the drug and, consequently, in the analytical signal observed in the first derivative order.

Table 8 shows the obtained relative amount values for the drugs' analyzed during the robustness test. The relative contents' values at the altered conditions were close to those obtained under the normal conditions, remaining in an acceptable range of 90 to 105% relative to the theoretical concentration.