Stability-indicating HPLC-DAD method for the simultaneous determination of fluoroquinolone in combination with a non-steroidal anti-inflammatory drug in pharmaceutical formulation

*Correspondence: N. M. Kassab. Laboratório de Tecnologia Farmacêutica, Faculdade de Ciências Farmacêuticas, Alimentos e Nutrição, Universidade Federal de Mato Grosso do Sul, CEP: 79070-900, Campo Grande, MS, Brazil. E-mail: nmkassab@gmail.com Stability-indicating HPLC-DAD method for the simultaneous determination of fluoroquinolone in combination with a non-steroidal anti-inflammatory drug in pharmaceutical formulation


INTRODUCTION
Fluoroquinolones are a group of synthetic antibacterial agents that are widely used for the treatment of multiple bacterial infections in humans and in veterinary medicine because of their safety, broad antibacterial spectrum, and good tolerance (Chamseddin & Jira, 2011). These agents act by inhibiting DNA gyrase and topoisomerase IV, enzymes involved in bacterial DNA replication, transcription, and recombination (Garmyn et al., 2009).
ENRO has found veterinary applications in the treatment of skin and soft-tissue infections in dogs and cats, uncomplicated and complicated genitourinary tract infections, respiratory (pneumonia) and gastrointestinal diseases, footpad lesions, and mastitis in cattle, sheep, goats, pigs, and poultry. As regular part of animals' diets, the ENRO also reduces mortality and morbidity, improves animal welfare, and promotes growth (Sarmah et al., 2006;Golovnev et al., 2012;Regitano & Leal, 2010). The extensive use of fluoroquinolones as prophylactics or additives in human and veterinary medicine has led to a significant increase in antimicrobial resistance and poses risks to the environment and public health from antimicrobial residues in foods (Sun et al., 2013;Regitano & Leal, 2010).
Piroxicam (PIRO; Figure 1B) is a non-steroidal anti-inflammatory drug (NSAID) used in humans for the treatment of gout, rheumatoid arthritis and osteoarthritis, gynecological diseases, and musculoskeletal disorders. It has anti-inflammatory, analgesic, and antipyretic activity and nonselectively inhibits the activity of the enzyme cyclo-oxygenase, COX-1, and COX-2 (Santos et al., 2011). PIRO is chemically known as 4-hydroxy-2-methyl-N-2-pyridyl-2H-1,2-benzothiazine-3-carboxamide 1,1-dioxide, with the molecular formula C 15 H 13 N 3 O 4 S and a molecular weight of 331.35 g mol −1 (USP, 2008). Clinical studies have demonstrated that the concomitant use of NSAIDs with antimicrobial decreases the severity of clinical symptoms in animals, increases appetite, and hastens the return to productivity, such as in weight gain or milk production, with higher production performance (Bednarek et al., 2003).
However, we are unaware of a method to simultaneously quantify ENRO and PIRO in combination. No analytical method is reported in the official compendiums for the simultaneous determination of ENRO and PIRO in veterinary formulations. In addition, no stability-indicating HPLC method for the simultaneous determination of pharmaceutical combinations is found in the literature.
Therefore, the aim of this study was to develop and validate an isocratic reverse-phase stability-indicating HPLC-DAD method for the simultaneous determination of ENRO and PIRO and their respective degradation products in veterinary formulations.

Chemicals and reagents
The ENRO reference standard (assigned purity of 99.50 %) was kindly donated by Biovet (Vargem Grande Paulista, Brazil). The PIRO reference standard (assigned purity of 100 %) was obtained from a local pharmacy. The injectable solution Zelotril Plus (sample), containing 100 mg mL −1 of ENRO and 12 mg mL −1 of PIRO (declared content), was also purchased locally. All reference substances, as well as the injectable solution, were protected from light throughout the study. HPLC-grade acetonitrile (Vetec, Brazil), analytical-grade phosphoric acid (Synth, Brazil), and fresh ultrapure water from a Milli-Q ® Plus system were used in all analyses.

Analytical procedure
The HPLC analyses were performed in a system consisting of a Dionex ® Ultimate 3000 (Thermo Fisher Scientific, USA), equipped with an Ultimate 3000 RS Variable Wavelength photodiode array detector and Ultimate 3000 pump. The system was connected to a microcomputer running the Chromeleon ® 7.1 Chromatography Data System software.
The experimental conditions were optimized at room temperature (24 + 2 °C) on a Vertisep C18 column (150×4.6 mm, particle size 5 µm) manufactured by Vertical, Inc., USA. All separations were obtained in an isocratic mode using a mobile phase consisting of acetonitrile and water (pH 3.0) in the ratio of 48:52 v/v, respectively. The pH of the water was adjusted to 3.0 using phosphoric acid. The flow rate of the mobile phase was 1.0 mL min -1 and the sample injection volume was 20 μL. The photodiode array detector was set at 278 nm (ENRO) and 360 nm (PIRO).

Preparation of standard stock solutions
The stock solutions of reference standards were prepared in order to optimize the experimental procedure and decrease the chance of errors. Ten mg of each reference standard (ENRO and PIRO) was transferred to 100-mL volumetric flasks. The volumes were completed with acetonitrile and water in the ratio of 50:50 v/v to reach 100 µg mL -1 of each reference standard. These solutions were diluted and used to determine the linearity and accuracy of the method.

Preparation of sample stock solutions
The sample solutions were prepared in the same ratio as the labeled amounts in the veterinary formulation. An aliquot of 5 mL of sample (containing 100 mg mL -1 of ENRO and 12 mg mL -1 of PIRO) was transferred to a 500-mL volumetric flask. The content of the sample was extracted with 500 mL of acetonitrile and water in the ratio of 50:50 v/v under sonication for 10 min, obtaining a stock sample solution containing 1 mg mL -1 of ENRO and 0.120 mg mL -1 of PIRO. This solution was diluted and used to determine the precision and accuracy of the method. For the forced degradation study, 4 mL of the sample stock solution was transferred to a 100-mL volumetric flask and the volume was completed with acetonitrile and water in the ratio of 50:50 v/v (sample stock solution A).

Validation of the HPLC-DAD method
The method was validated based on the ICH and AOAC guidelines, following the validation parameters: specificity, linearity, precision, accuracy, limit of detection (LOD), limit of quantification (LOQ), and robustness (AOAC, 2005;ICH, 2005).

Specificity
The specificity was evaluated through analysis of a placebo solution, as well as by analysis of the drug solutions after forced degradation studies.
Forced degradation studies were performed according to ICH guidelines (ICH, 2000). Aliquots of the sample stock solution were submitted to different stress conditions, and the resulting solutions were analyzed based on the mean of peak areas (n = 3) for the stressed sample solution. The residual amounts of active pharmaceutical ingredients (APIs) and the peak purity data were evaluated.
Acidic and basic degradations were performed by transferring 15 mL of sample stock solution A to 100-mL Erlenmeyer flasks and adding 15 mL of either 0.5 N hydrochloric acid or 0.5 N sodium hydroxide. The solutions were kept under vigorous mechanical stirring for 8 h at room temperature (24 ± 2 °C). After the stress period, the solutions were neutralized.
To promote oxidation, a mixture of 15 mL of sample stock solution A and 15 mL of 3% hydrogen peroxide solution was placed in a 100-mL Erlenmeyer flask and kept at room temperature (24 ± 2 °C) for 8 h.
To evaluate the thermal degradation, a mixture of 15 mL of sample stock solution A and 15 mL of the mobile phase was transferred to an Erlenmeyer flask for 8 h at 60 °C in the dark.
To study the radiation effect, a mixture of 15 mL of sample stock solution A and 15 mL of the mobile phase was transferred to 100-mL Erlenmeyer flasks. The flasks were exposed to two different stress conditions: i) direct sunlight and ii) cool-white fluorescent light (ISO 18909-2006) for up to 62 h and 92 h, respectively. The photodegradation steps were evaluated at room temperature (24 ± 2 °C).
After the stress conditions, aliquots of 5 mL were removed and transferred to 10-mL volumetric flasks, the volumes were completed with the mobile phase, and the solutions were filtered before injection in a Millex 0.45 µm filter (Millipore, Bradford, USA).

Linearity
The linearity was determined via calibration curves (peak area versus concentration). For the curve of ENRO, ten concentration levels were used, and 11 for the PIRO curve, to establish the linear-regression lines by the leastsquares method. All chromatographic determinations were performed in triplicate and at room temperature (24 ± 2 °C). The statistical evaluation was done by analysis of variance (ANOVA).

Precision
The intra-day precision (repeatability) was evaluated by analyzing the sample solution at a single concentration within the linearity range of the method. An aliquot of 1.0 mL of sample stock solution was transferred to a 100-mL volumetric flask. The final volumes were completed with the mobile phase to obtain final concentrations of 10.0 and 1.2 μg mL -1 of ENRO and PIRO, respectively. The analyses were performed in six replicates on the same day. To estimate the inter-day precision, the sample solution was prepared fresh at the same concentration level for each drug, and the response was determined in six replicates. The procedure was repeated on three consecutive days. The intra-and interday precisions are expressed in terms of Relative Standard Deviation (%RSD).

Accuracy
The accuracy was determined by recovery studies, using the standard addition method as recommended by AOAC (AOAC, 2005). Recovery was analyzed by adding known amounts of standard solutions to the sample, followed by analysis using the proposed method. Aliquots of standard and sample stock solutions were transferred to 25-mL volumetric flasks and the final volumes were completed with the mobile phase.
The percentage of recovery (R) was calculated by comparing the theoretical and found concentrations, using the following equation: R = [ ( C F -C U ) / C A ] × 100; where C F represents the concentration of the analyte measured in the fortified test sample; C U , the concentration of the analyte measured in the unfortified test sample; and C A , the concentration of the analyte added to the fortified test sample.
Limit of detection (LOD) and limit of quantification (LOQ) LOD was calculated based on the standard deviation of the response and the slope of the calibration curves. It was mathematically evaluated by the equation 3.3 σ/s, where σ represents the standard deviation of the analytical signal and s is the slope of the corresponding calibration curve. The LOQ was determined experimentally by injecting dilute solutions from the standard stock solution.

Robustness
The robustness was evaluated by slightly changing the chromatographic conditions and observing the effect of these changes on the peak area, retention time (t R ), and tailing factor (T f ) of the drugs analyzed. The factors (chromatographic conditions) selected to examine the robustness were the flow rate, percentage of acetonitrile, and pH of the water in the mobile phase. Each factor was investigated at three levels (-1, 0, and + 1) (Sversut et al., 2014). Level 0 refers to the normal chromatographic conditions, i.e., the conditions employed in the proposed method. From this level, the chromatographic conditions were modified to a higher level (+1) or to a lower level (-1). When one condition was changed, the others remained at level 0. Replicate injections (n = 3) of the sample solution were performed under these small changes in the chromatographic conditions. Student's t-test was used for the statistical evaluation.

Choice of stationary phase
Initially, different stationary phases were tested on chromatographic parameters. The chromatographic parameters obtained with the Kinetex C18 and VertiSep C18 were within acceptable limits according to ICH and AOAC (AOAC, 2005;ICH, 2005). However, the VertiSep C18 was chosen for the development and validation of the analytical method because it provided better values for the chromatographic parameters for both drugs.
Although the Kinetex C18 has lower organic-solvent waste, the Vertisep C18 was chosen. Its performance is better in the chromatographic analysis (resolution > 2.0, plates > 2000 and asymmetry of peaks < 2.0), increasing the reliability of the method.

Choice of mobile phase
Some mobile phases were tested in a Vertisep C18 column to provide the best chromatographic separation of drugs. The effects of the different compositions of the mobile phases on the chromatographic parameters are summarized in Table I.
The reported values of pK a for ENRO were pK a1 = 6.0, corresponding to the carboxyl group, and pK a02 = 8.8, corresponding to the basic piperazinyl group and the isoelectric pH = 7.32 (Chakravarthy et al., 2015). The mobile phase pH affects the acid-base behavior of this drug and can reduce the retention time in the stationary phase. In acidic pH, the N-ethyl group (tertiary alkylamine) accepts a proton, and in basic pH the carboxyl group (carboxylic acid) donates a proton, showing the amphoteric character of the ENRO (Beale, Block, 2011). At pH 3.0 the tertiary alkylamine group remained ionized, leading to lower retention time in the reverse stationary phase. The use of a buffered mobile phase is dispensable because the proposed pH value is at least 2 units below the pK a , providing a completely ionized tertiary alkylamine group (Beale, Block, 2011).
PIRO has reported pK a values (USP, 2008) of 3.79 and 4.76 and is poorly soluble in polar solvents (Batlouni, 2010). However, the amine and hydroxyl groups in its structure allow the use of a mobile phase with a lower proportion of organic solvent, which optimizes the method.
PIRO was more retained in the stationary phase when mobile phases composed of a larger amount of water (mobile phase F; Table I) were used, and when acetonitrile was substituted for methanol (mobile phase H; Table I). The pH adjustment of the mobile phases secured the ionization of the proton acceptor groups. However, pH adjustment alone was not sufficient; it was necessary to optimize the proportion of the organic modifier in the mobile phase.
Comparing mobile phases B and F (Table I), both are composed of acetonitrile and water at pH 3.0. It was clear that reducing the proportion of the organic modifier (acetonitrile) in mobile phase F significantly affected the retention time of PIRO, and therefore the chromatographic parameters could not be determined. When comparing the mobile phase B with D and E, in mobile phases B and E all the chromatographic parameters evaluated in Table I are acceptable, which is not confirmed with the value found in theoretical plates in mobile phase D. However, mobile phase B contains a lower proportion of organic solvent, which makes the method advantageous. In mobile phase C, with a higher pH than in mobile phases B and D, and only acetonitrile and water were used at the same proportions as in mobile phase B, there was a reversal of the elution, in which PIRO had a lower retention time than ENRO. This was due to a lower degree of ionization of the tertiary alkylamine group of ENRO in the less-acidic pH.
In these studies, suitable separation with high resolutions, satisfactory theoretical plates, and optimum peak symmetry were achieved with mobile phase B, acetonitrile:water (48:52 v/v, pH 3.0). Under the optimized chromatographic conditions, it was possible to obtain efficient separation of APIs in the veterinary formulations. Peak identity was confirmed by the retention time and by the reference-spectra match factor. Sharp, symmetrical peaks of ENRO and PIRO were obtained at retention times of 1.33 and 4.89 min, respectively ( Figure 2). All values for the reference-spectra match factor were higher than 950, indicating the high similarity between the analytes and library reference spectra (Agilent Technologies, 2008).

Specificity
The excipients present in the pharmaceutical dosage forms did not affect the analysis because no peak   was observed at drug retention times, which proves the specificity of the method.
The specificity of the method was also evaluated by checking the peak purity of all analytes after the forced degradation studies (Figure 3). Chromatographic peak purity data were evaluated from the spectral analysis report supplied by the photodiode array detection. Peak purity match factor (PPM) values higher than 950 indicate a homogeneous peak (Maio et al., 2006). Almost all PPM values for API peaks in chromatograms of stressed standard solutions were higher than 950. The peak purity value was lower than 950 in basic and sunlight degradation for ENRO after 8 and 62 hours, respectively. The peak purity value for PIRO was lower than 950 only in sunlight degradation after 8 hours.
Almost all stress conditions were sufficient to degrade the drugs. In studies of acidic, basic and oxidative degradation, PIRO was more stable than ENRO. However, in studies of degradation by sunlight and fluorescent light, ENRO was more stable than PIRO. In the study of thermal degradation, both drugs were equally stable.
Under acidic hydrolysis ( Figure 3A), ENRO and PIRO were degraded up to 23.63% and 2.5%, respectively. Under basic hydrolysis ( Figure 3B), PIRO suffered no degradation, while ENRO was totally degraded (PPM = 694), precluding determination of the content. Starek et al. (2009) observed two degradation products of PIRO under acidic conditions (HCl 1 N; 60 and 120 °C for 1 h) and only one product under basic conditions (NaOH 1 N; 60 and 120 °C for 1 h). To identify these degradation products, 1 H NMR and LC-MS-MS were carried out, and the spectrum data of the first product showed the presence of signals corresponding to pyridine-2-amine. The second product was identified as 2-methyl-2H-benzo[e][1,2]thiazin-4(3H)-one 1,1-dioxide in acidic hydrolysis. In basic hydrolysis, only the product pyridine-2-amine was identified.
A similar result was seen in the study of Modhave et al. (2011), where in acidic and basic media (HCl 0.1 N and NaOH 0.5 N; 80 °C for 24 h), hydrolysis occurred in the amide group of PIRO, resulting in the formation of the same degradation products reported by Starek et al. (2009).
In a study of fluoroquinolones in which ENRO was subjected to acid medium, it lost the carboxylic acid functional group, generating a degradation product which was subsequently identified by 1 H-NMR (Allah, 2004).
Our results showed that ENRO is more reactive under acid hydrolysis conditions because decarboxylation in ENRO is faster than the hydrolysis of the amide group in PIRO (competitive reactions). In basic conditions, this is more evident, and ENRO undergoes complete degradation.
Under oxidative conditions ( Figure 3C), ENRO was degraded up to 29.3% and PIRO was degraded up to 9.5%. The HP1 and HP2 peaks are not degradation products, because injection of hydrogen peroxide alone resulted in the same retention time and a similar profile.
A study with ENRO and its degradation products found that ENRO was stable in all the degradation conditions except in oxidation degradation, where slight degradation was observed (0.86%) when performed by adding 30% peroxide solution (H 2 O 2 ) at 70 °C for 1 hour (Chakravarthy, Sailaja, Kumar, 2015).
Another stability-indicating study with ENRO in combination with amoxicillin found no degradation for ENRO. There was no evidence of degradation of the drug when exposed to acidic (HCl 0.1 N for 4 h; room temperature), alkaline (NaOH 0.1 N; 65 °C for 7 days), oxidative (H 2 O 2 0.3% for 7 days; room temperature), thermal (70 °C for 3 weeks), and photo (UV light for 3 days) stress conditions (Batrawi, Wahdan, Al-Rimawi, 2017).
The oxidation reactions of drugs using H 2 O 2 involves electron-transfer mechanisms. Groups such as amines, sulfides, phenols, double bonds, benzylic carbon, allylic carbon, and α-carbon attached to a heteroatom can undergo oxidation to N-oxides, hydroxylamine, sulfoxides, sulfones, quinones, hydroperoxides, hydroxides, and ketones ENRO and PIRO can undergo oxidation with H 2 O 2 , the former because of the amine group (piperazine), and the latter because of the enol group (Charde et al., 2013).
Under thermal degradation ( Figure 3D), ENRO and PIRO were not degraded; they retained the content of 100% and PPM of 1000. Fluoroquinolones are chemically stable and resistant to increased temperatures because of the stability of the quinolone ring (Babić, Perisa, Skorić, 2013). A study with PIRO showed that 2 hours of heating at 100 °C produced a gradual loss of absorbance, indicating degradation of the molecule, possibly due to hydrolysis of PIRO by cleavage of the amide bond attached to the 2-pyridine ring (Ahmad, Aminuddim, Nazin, 2011).
Under photodegradation, when sunlight was the stressor agent ( Figure 3E), ENRO was degraded up to 1.35% and PIRO was totally degraded (PPM = 700), precluding determination of the content after 8 hours of exposure. However, after 62 hours of exposure, ENRO was also totally degraded (PPM = 584).
When white fluorescent light was the stressor agent ( Figure 3F), ENRO was degraded less than 1% and PIRO 63.58% after 92 hours of exposure. Light-stress conditions can induce photo-oxidation by free radical mechanisms (Charde et al., 2013).
Several types of stress-produced oxidative processes via free radicals have been reported. The hydroxyl radical, for example, typically reacts through two competitive mechanisms, hydrogen removal or hydroxylation. According to Santoke et al. (2009), the addition of the hydroxyl radical to the aromatic ring in fluoroquinolones occurs because the aromatic ring of quinolones is activated, due to the presence of donor groups such as amines bonded to the aromatic ring (e.g., Piperazine) leading to the formation of phenols. The cyclopropane functional group present in ENRO has been shown to have low reactivity with the hydroxyl radical. Defluorination can occur at the carbon-fluorine position, which undergoes rapid HF elimination to form phenoxyl radicals. The replacement of hydrogen and piperazine groups also occurs in ENRO, which can also be replaced by the hydroxyl radical.
A n o t h e r s t u d y o f p h o t o d e g r a d a t i o n w i t h fluoroquinolones, using an irradiation spectrum similar to sunlight, observed changes in the piperazine ring and the carboxylic acid ester group, and also that the ethyl group of ENRO was first removed (Burhenne et al., 1997). Another study found that photolysis of ENRO involved the fragmentation of C-F and C-COOH bonds, suggesting that photolysis of ENRO was accompanied by decarboxylation, defluorination and also N-dealkylation processes (Li, Niu, Wang, 2011). Lin et al. (2010) observed that exposure of ENRO to fluorescent light caused the formation of ciprofloxacin as an intermediate, in pond water and sediment slurry; the study also showed that natural-light irradiation had a greater effect on the degradation of fluoroquinolones than fluorescent and ultraviolet light. Lin et al. (2010) also observed that the half-life of ENRO was 14.1 days (pond water) and 71 days (sediment slurry) upon exposure to fluorescent light.
The difference in stability between fluoroquinolone ENRO and oxicam PIRO during the photodegradation process can be explained by the presence of an enol group in PIRO. The enol group is very susceptible to reaction with singlet oxygen, forming dioxetane intermediates (Modhave et al., 2011). The subsequent ring cleavage and the transacylation process could explain the formation of N-methylsaccharine and N-(2-pyridyl) oxamic acid, the main products detected by HPLC-MS analysis in the studies of Lemp, Zanocco and Gunther (2001) and Miranda, Vargas and Serrano (1991). Dragomiroiu et al. (2015) observed that exposure of PIRO under different temperature and luminosity conditions showed that direct light exposure leads to significant degradation of the drug (about 24.76%) when the stock solution was placed in direct sunlight for 4 hours.
The largest amounts of degradation products (5 degradation-product peaks) were generated under sunlight conditions after 62 hours of exposure for the ENRO and PIRO combination. Smaller amounts of degradation products were obtained under thermal and fluorescent light conditions, where no impurity peak was observed.

Linearity
Linear calibration curves were evaluated in the concentration ranges from 0.25 to 16.0 μg mL -1 for ENRO and 0.125 to 8.0 μg mL -1 for PIRO. The correlation coefficients were 0.9997 and 0.9999 for ENRO and PIRO, respectively. The statistical ANOVA evaluation indicated a significant linear regression. These results indicated a linear correlation between the peak areas and drug concentrations.

Precision
For ENRO and PIRO, the values of %RSD were lower than 0.91% (repeatability) and 0.82% (intermediate precision). Therefore, the proposed method has good precision for the simultaneous determination of these drugs.

Accuracy
The mean recovery values were 99.79 ± 1.11% and 99.40 ± 1.24% for ENRO and PIRO, respectively. These recovery values are within acceptable limits [50] (100 ± 2%) and suggest good accuracy of the proposed method.

Limit of detection (LOD) and limit of quantification (LOQ)
The LOD values were 0.096 and 0.017 μg mL -1 for ENRO and PIRO, respectively (signal to noise ratio of 3:1). The LOQ were 0.25 and 0.125 μg mL -1 for ENRO and PIRO, respectively. These results indicated the sensitivity of the method.

Robustness
The results for robustness indicated that small changes in the chromatographic conditions did not significantly modify the peak areas, retention time, and tailing factors of the drugs.
As seen in Table II, a 1% change in acetonitrile, 0.3% in the pH and 3% in the flow rate of the mobile phase only slightly affected these chromatographic parameters and did not compromise the analyses of the drugs. ± 0.14 ± 1.50 ± 1.38 ± 1.25 ± 0.74 ± 1.62 a Three factors (A, B and C) were slightly changed at three levels (+1, 0, −1); each time a factor was changed from level 0, with the other factors remaining at level 0. b Retention time. c Tailing factor.

Application of the method
The proposed HPLC-DAD method is suitable for simultaneous determination of ENRO and PIRO in veterinary dosage formulations in routine analysis. For the commercial formulation Zelotril Plus ® (100 mg mL -1 ENRO and 12 mg mL -1 PIRO), the contents found were 99.91 ± 0.81% of ENRO and 96.62 ± 0.08% of piroxicam.

CONCLUSIONS
A stability-indicating HPLC-DAD method was developed and validated for simultaneous determination of enrofloxacin and piroxicam in veterinary formulations. This method showed specificity, precision, accuracy, sensitivity, and robustness. In addition, the method was successfully applied to separate the active pharmaceutical ingredients from their forced degradation products. The results suggest that the proposed method can be used in the routine quality-control analysis.