Yehia et al.: Journal of aoaC international Vol. 99, no. 1, 2016 73

DRUG FORMULATIONS AND CLINICAL METHODS

Stability Study and Kinetic Monitoring of Cefquinome Sulfate Using Cyclodextrin-Based Ion-Selective Electrode: Application to Biological SamplesAli M. YehiA, RehAM M. ARAfA, SAMAh S. AbbAS, and SAwSAn M. AMeRCairo University, Faculty of Pharmacy, Analytical Chemistry Department, El-Kasr El-Aini St, Cairo, Egypt, 11562

Received August 6, 2015. Accepted by SW September 22, 2015.Corresponding author’s e-mail: aliyehia00@yahoo.com or

ali.yehia@pharma.cu.edu.egDOI: 10.5740/jaoacint.15-0185

Two novel cefquinome sulfate (CFQ)-selective electrodes were performed with dibutyl sebacate as a plasticizer using a polymeric matrix of polyvinyl chloride. Sensor 1 was prepared using sodium tetraphenylborate as a cation exchanger without incorporation of ionophore, whereas 2-hydroxy propyl β-cyclodextrin was used as ionophore in sensor 2. A stable, reliable, and linear response was obtained in concentration ranges 3.2 × 10−5 to 1 × 10−2 mol/L and 1 × 10−5 to 1 × 10−2 mol/L for sensors 1 and 2, respectively. Both sensors could be sufficiently applied for quantitative determination of CFQ in the presence of degradation products either in bulk powder or in pharmaceutical formulations. Sensor 2 provided better selectivity and sensitivity, wider linearity range, and higher performance. Therefore it was used successfully for accurate determination of CFQ in biological fluids such as spiked plasma and milk samples. Furthermore, an online kinetic study was applied to the CFQ alkaline degradation process to estimate the reaction rate and half-life with feasible real-time monitoring. The developed sensors were found to be fast, accurate, sensitive, and precise compared with the manufacturer’s reversed-phase chromatographic method.

Cephalosporins are widely used antibiotics in veterinary medicine for the prevention and treatment of bacterial infections (1). Cefquinome sulfate (CFQ) is an

aminothiazolyl fourth generation cephalosporin, {6R-[6α, 7β(Z)]}-1-[(7-{[(2-amino-4-thiazolyl)-(methoxyimino)acetyl]amino}-2-carboxy-8-oxo-5-thia-1-azabicyclo[4.2.0]oct-2-en-3-yl)methyl]-5, 6, 7, 8-tetrahydroquinolinium sulfate (Figure 1). Introducing aminothiazolyl methoxyimino moiety into the acyl side chain extends the activity against both Gram-negative and -positive bacteria, making CFQ more resistant to β-lactamases inhibition. Moreover, this group enhances cefquinome bioavailability compared with older generations of cephalosporins (2). CFQ is mainly used to treat clinical mastitis, which is a disease caused by Staphylococcus aureus (Gram-positive bacteria; 3). Parenteral and intramammary combination therapy for Escherichia coli mastitis has been suggested (4–6). Therefore, determination of CFQ in biological tissues and

fluids is an application of utmost importance. Several methods for this application have been suggested in the literature and most of them were chromatographic methods with the aid of solid-phase extraction as a pretreatment and purification step for biological samples (7–10). The cited methods succeeded in determination of CFQ. However, biological sample treatment prior to its analysis is always a laborious process. Another dilemma with biological application is the low concentrations of drugs present in the samples. For the studied drug, the initial plasma drug concentration following intravenous administration is about 8.51 µg/mL (1.35 × 10−5 mol/L; 2). On the other hand, the maximum residue limit (MRL) for CFQ in milk is 20 µg/kg according to Annex I Regulation No. 2377/90/EEC [European Economic Community] (11).

Generally, cephalosporin has been susceptible to degradation even in solution (12, 13) or in solid states (14). CFQ, without exception, shows low stability toward acidic and alkaline hydrolysis, along with its susceptibility to photodegradation and thermal decomposition (15). In this work, stability study of CFQ was performed using two membrane-selective electrodes, one without ionophore (sensor 1) and the other with ionophore (sensor 2). The ability of the proposed sensors to selectively determine CFQ in the presence of its degradation products as well as the sensors’ application to directly determine CFQ in the form of dosages and different biological samples was evaluated. Moreover, sensor 2 was used to monitor the online kinetic study of the degradation process.

Experimental

Materials and Chemicals

Pure standard.—CFQ was provided by Intervet, Schering-Plough Animal Health (Munich, Germany). Its purity was 100.32% according to the manufacturer’s reversed-phase chromatographic method (RP-HPLC). A C18 stainless steel column (250 × 4.6 mm, 5 μm) was used for the stationary phase. The mobile phase consisted of a mixture of 90 mL acetonitrile, 12 mL phosphoric acid, and 3.45 g sodium perchlorate monohydrate in 1 L water. The pH adjusted to 3.6 using diethylamide, a flow rate of 1.0 mL/min, and UV detection at 270 nm.

Pharmaceutical formulations.—Two market samples were used in this work. CobactanTM 2.5% suspension batch No. A445A02 (Intervet Schering-Plough Animal Health) was labeled to contain 25 mg CFQ per milliliter. The other market sample was Cobactan LC batch No. B009801 (Intervet International, Munich,

12_150185_Yehia.indd 73 20/02/16 10:25 AM

74 Yehia et al.: Journal of aoaC international Vol. 99, no. 1, 2016

Germany), where each syringe was labeled to contain 75 mg CFQ per 8 g ointment.

Chemicals.—Sodium tetraphenylborate (TPB) and 2-hydroxy propyl β-cyclodextrin were purchased from Sigma-Aldrich (Steinheim, Germany). Polyvinyl chloride (PVC) and dibutyl sebacate (DBS) were obtained from Fluka (Steinheim, Germany). Tetrahydrofuran (THF) was obtained from BDH (Poole, England). The Britton-Robinson buffers in the pH range of 2.0–12.0 were prepared by mixing equal volumes of 0.04 mol/L acetic acid, 0.04 mol/L boric acid, and 0.04 mol/L phosphoric acid. Then the required pH values were adjusted using 0.2 mol/L NaOH standard solution. Sulfuric acid 98%, sodium hydroxide and hydrogen peroxide 30% were purchased from Adwic, Qalyubia, Egypt. Disodium hydrogen phosphate dihydrate, phosphoric acid and acetone were purchased from E. Merck, Darmstadt, Germany.

Degraded samples.—CFQ hydrolytic degradation products were prepared by way of acidic or alkaline media and the dissolution of 100 mg CFQ in either 50 mL 0.05 N sulfuric acid or 0.1 N sodium hydroxide, respectively. The solutions were separately refluxed for 2 h and then neutralized to pH 7.0. The neutral solutions were evaporated, reconstituted in methanol, and filtered. The alcoholic extracts were evaporated to obtain degradation products in solid form. For photodegraded CFQ samples, 50 mg CFQ were dissolved in 50 mL water and subjected to a UV light source at 254 nm for 1 h. The oxidative degradation products were prepared by refluxing 50 mg CFQ with 50 mL 2.5% hydrogen peroxide aqueous solution for 1 h; then excess hydrogen peroxide was decomposed by heating. Complete degradation of CFQ samples was confirmed by RP-TLC using disodium hydrogen phosphate dihydrate (2.0 g %, w/v), and adjusted to pH 3.5 by phosphoric acid–acetone (15:10, v/v) as a developing system. Spots were detected under a UV lamp at 254 nm.

Fabrication of Membrane Sensors

The ion association complex was prepared by mixing ca 10 mL CFQ and TPB aqueous solutions (each 1 × 10−2 mol/L), then filtered, washed with cold water, and dried at room temperature. In two separate Petri dishes (5 cm diameter), 10 mg of the previously prepared ion association complexes were mixed with DBS (0.35 mL) and PVC (0.19 g) to prepare sensor 1. In the second Petri dish, the same previously mentioned reagents were again mixed but this time with the addition of 2-hydroxy propyl β-cyclodextrin (10 mg) to prepare sensor 2. The two mixtures were separately dissolved in THF (5 mL),

covered with a filter paper and left to stand overnight for slow solvent evaporation at room temperature. Master membranes of 0.1 mm in thickness were obtained from which disks were cut (about 8.0 mm in diameter) and cemented to flat end of PVC tubing with the aid of THF. A mixed solution consisting of equal volumes of 1 × 10−4 mol/L CFQ and 1 × 10−4 mol/L sodium chloride was used as an internal reference solution for the two sensors. An Ag/AgCl electrode (3 mm diameter) was used as an internal reference electrode. The sensors were conditioned by soaking in a solution of 1 × 10−4 mol/L CFQ for 24 h and then stored in the same solution when not in use.

Potentiometric Measurements

The potential of a cell comprising the proposed ion-selective electrode and Ag/AgCl double-junction-type external reference electrode [Orion 900200 (Thermo Scientific, Waltham, MA), 3.0 M KCl saturated with AgCl as an inner filling solution, and 10% KNO3 as a bridge electrolyte) was measured with a Jenway digital ion analyzer (Model No. 3330; Essex, United Kingdom). pH adjustments were carried out using a Jenway pH glass electrode.

Sensor Calibration

Calibration of the conditioned CFQ sensors was performed by separately immersing the sensors in conjunction with a double-junction Ag/AgCl reference electrode in different CFQ standard solutions (1 × 10−6 to 1 × 10− 2 mol/L) prepared in buffer solution pH 7.0. The sensors were allowed to equilibrate while stirring and recording the potential difference (Electromotive force) in readings within ±1 mV. Membrane sensors were washed between measurements with the buffer. The recorded potentials were plotted as a function of logarithm CFQ concentrations in buffer pH 7.0 at 25°C.

Effect of pH

The effect of pH on the response of the proposed sensors (potential) was investigated using 1 × 10−2 mol/L and 1 × 10−3 mol/L aqueous CFQ standard solutions at different pH values, ranging from 2.0 to 12.0. The obtained potential at each pH value was recorded.

Selectivity Measurements

The potentiometric selectivity coefficients ( ,KA Bpot ) were

evaluated according to International Union of Pure and Applied Chemistry (IUPAC) guidelines using the separate solutions method (16) in which the potentials were measured for CFQ cation (A) and interfering ion (B) standard solutions, both having the same activity ( )a aA B= . Different interfering ions with concentrations of 1 × 10−3 mol/L in buffer pH 7.0 were used.

Analysis of Synthetic Mixtures

Potentiometric measurements were carried out in different synthetic mixtures containing a fixed concentration of CFQ (1 × 10−3 mol/L) while varying concentrations of acidic degradation products in buffer pH 7.0. The potential of each

Figure 1. Chemical structure of CFQ (molecular weight 626.68).

12_150185_Yehia.indd 74 20/02/16 10:25 AM

Yehia et al.: Journal of aoaC international Vol. 99, no. 1, 2016 75

mixture was recorded with the two sensors and the concentration of CFQ was calculated from the corresponding regression equations.

Applications

Pharmaceutical formulations.—Two amounts of suspension and ointment formulations equivalent to 157 mg CFQ each were accurately transferred into two separate 25 mL volumetric flasks, and the volumes were filled to the mark with methanol. The solutions were sonicated for 10 min to enhance drug extraction. Ten-fold dilutions of the prepared solutions were applied using buffer pH 7.0, then the potentials of the prepared buffered solutions were measured. The concentrations of CFQ were calculated from the corresponding regression equations.

Spiked plasma samples.—Spiked plasma samples were prepared by transferring 2.5 mL aliquots from CFQ standard solution (1 × 10−4 mol/L) into a 25 mL volumetric flask, where the volume filled to the mark with bovine plasma. Twenty milliliters of spiked plasma sample were placed into a 25 mL beaker and the initial potential measured using sensor 2. While potentiometric measurement proceeded, 0.5 mL aliquot from 1 × 10−4 mol/L CFQ standard solution was added; then a second potential was measured after stabilization. The two measured potentials, before and after standard addition, were used to calculate the concentration of CFQ in the spiked plasma samples.

Spiked milk samples.—Spiked milk samples were prepared by transferring 2.5 mL aliquots from CFQ standard solution (1 × 10−4 mol/L) into a 25 mL volumetric flask and the volume filled to the mark with cow milk. Twenty milliliters of spiked milk sample were placed into a 25 mL beaker and the initial potential measured using sensor 2. While potentiometric measurement proceeded, successive additions of 0.5 mL aliquot from 1 × 10−4 mol/L CFQ standard solution were applied. Stable potentials were measured after each addition. A graphical plot of the standard addition curve was used to calculate the concentration of CFQ in the spiked milk samples.

Kinetic Study

Determination of reaction order.—Sensor 2 was immersed in conjunction with a double-junction Ag/AgCl reference electrode into 22.5 mL buffer pH 10.0. The buffer solution was thermostatically controlled at 25°C. Once the potential stabilized, 2.5 mL aliquots from 1 × 10−2 mol/L CFQ standard solution were added. The potential of the solution having initial concentration [C0 = 6.27 × 10−4 g/mL (1 × 10−3 mol/L)] was measured at zero time, then potential measurements were carried out in 15 min intervals. The reduced concentrations of CFQ were calculated using the corresponding regression equation, performed for CFQ in pH 10.0 at 25°C. The log% of the remaining concentration was plotted against time to determine the reaction order and to calculate rate constant (k), half-life (t1/2), and t90%.

Effect of pH on the reaction rate.—From 1 × 10−2 mol/L CFQ standard solution, 2.5 mL were transferred into three 50 mL beakers, each containing 22.5 mL buffer with different pH values (9.0, 9.5, and 10.0) and thermostatically controlled at 25°C. The potentials were continuously monitored as previously described. The log% of the remaining concentration was plotted

against time for all three buffers. The rate constant values, their corresponding t1/2 and t90% were calculated.

Effect of temperature on the reaction rate.—The previously described procedure for continuous potential monitoring for the three buffered solutions (pH 9.0, 9.5, and 10.0) was followed, while thermostatically heated at 25, 30, and 35°C. Likewise, the log% of the remaining concentration was plotted against time for different temperatures. In addition, an Arrhenius plot was drawn to determine the effect of temperature on the hydrolysis rate.

Results and Discussion

Ion-selective electrode theory and characterization methodology for chemical application were adopted over the last few decades (17). Stability indicating electrochemical ion-selective electrodes (18) along with application on plasma samples (19) followed. The introduction of ionophores enhanced selectivity and improved electrode response toward electroactive analytes (20). More recently, monitoring kinetic study of the degradation process of electroactive drugs was carried out using membrane electrodes either off- (21) or online (22) during hydrolytic degradation.

In this work, two membrane sensors for CFQ were used for its determination in the presence of degradation products. Applications of membrane-selective electrodes were further used to determine CFQ in dosage form, and sensor 2 was used to determine CFQ in different biological samples without the need for pretreatment or purification steps. Furthermore, kinetic behavior of the degradation process was monitored using sensor 2. Upon review of the literature, we found electroanalytical methodology has not been previously used for the determination of CFQ and that online kinetic study had also never been applied specifically to observe the degradation process. Only offline HPLC kinetic monitoring has been previously reported (23). The studied drug possessed a positive quaternary amine of 5,6,7,8-tetrahydroquinolinium ion which suggests the use of TPB as counter anion for the cationic CFQ in the two proposed PVC membrane sensors. The formation of stable water-insoluble ion association complexes between CFQ and sodium TPB in a 1:1 ratio was confirmed by Nernstian responses close to +60 mV, which is the typical value for monovalent cationic drugs.

Concerning the stability of CFQ, the intact drug was subjected to acidic and alkaline hydrolysis, UV light, and oxidation. Degradation processes were confirmed by RP-TLC: two bands at Rf 0.40 and 0.82 for the two degradation products of hydrolysis and photodegradation, respectively, as well as a third band at Rf 0.87 observed for the product of oxidative degradation (Figure 2). Intact CFQ at the Rf 0.62 band was monitored to confirm complete degradation. Based on previous studies of structurally related cephalosporin (Cefpirome), the common degradation pathway involved the cleavage of the carbon nitrogen bond attached to position three in the cephalosporin nucleus (12–14). Therefore 5,6,7,8-tetrahydroquinoline (Rf 0.4) was suggested. This degradation product is liable to air oxidation, which can occur during preparation steps, and converts into 5,6,7,8-tetrahydroquinoline 1-oxide, which is stabilized by resonance (24). Mass spectrometry revealed a

12_150185_Yehia.indd 75 20/02/16 10:25 AM

76 Yehia et al.: Journal of aoaC international Vol. 99, no. 1, 2016

molecular ion peak m/z at 148, which confirms the proposed common degradation product (Figure 3).

Characteristics of the Proposed Ionophore-Based Sensor

Molecular recognition between host molecules (ionophore) and guest molecules (drug) improves the response of ionophore-based potentiometric sensors (sensor 2). Cyclodextrins (CDs) are the most popular ionophores in use. In either their natural or synthetic form, CDs act as molecular receptors. In the case of natural CD, cooperative binding with certain guest molecules is mostly attributed to intermolecular hydrogen bonding between CD molecules. Intermolecular interactions between host and guest molecules (hydrogen bonds, hydrophobic interactions, and van der Waals forces) contribute to cooperative binding processes when synthetic CDs are used (25). Although the size and geometry of the guest mainly govern binding strength, it is possible to influence host–guest interactions by modifying the three hydroxyl groups on each glucose unit. Indeed, the use of 2-hydroxy propyl β-cyclodextrin enhances the interaction properties between host and guest molecules (26). Positive CFQ ions prefer the high-donation sites (OH groups) of the 2-hydroxy propyl β-cyclodextrin structure. Therefore, sensor 1, where the ionophore was absent, has a lower slope than sensor 2 (Table 1).

The electrochemical performance characteristics of the proposed CFQ sensors were evaluated according to IUPAC standards (16). Two different assemblies for each sensor were used over a 1 month period and the measurements carried out at 25°C. The results are provided in Table 1. Typical calibration plots for both sensors are shown in Figure 4. Sensor 1 has a

relatively lower slope with narrow linearity range. The slopes obtained from the calibration plots are 54.9 and 57.1 mV per decade for sensors 1 and 2, respectively. The electrodes’ response to the CFQ cation is related to its activity rather than its concentration, therefore deviation from an ideal Nernstian slope at 25°C (60 mV) occurs. However less significant deviation was observed for sensor 2. Constant potential readings were displayed for both sensors during day-to-day measurements and the change in calibration slopes were within ±2 mV per decade over a period of 21 and 35 days for sensors 1 and 2, respectively. Detection limits for the two sensors were calculated according to the IUPAC definition; indubitably sensor 2 had a relatively lower LOD, as can be seen in Table 1. The regression equation of the linear response between measured potential (E) in millivolts and the corresponding concentration of CFQ (C ) in molar is represented by the following equation:

logE K S C= + (1)

where (K) represents the standard electrode potential in buffer pH 7.0 at 25°C and (S) is the Nernstian response.

Effect of pH

The potential stability of the proposed sensors over various pH ranges was also examined. The results revealed a fairly

Figure 2. TLC chromatogram of degraded CFQ samples using disodium hydrogen phosphate dihydrate (2.0 g %, w/v) and adjusted to pH 3.5 by phosphoric acid–acetone (15:10, v/v) as a developing system. Detection at 254 nm.

Figure 3. Mass spectra of the suggested CFQ degradation product.

Table 1. Electrochemical response characteristics of the two investigated CFQ sensors

Parameter Sensor 1 Sensor 2

Slope, mV per decadea 54.9 57.1

Intercept, mV 201.6 199.1

LOD, mol/Lb 3.1 × 10−5 6.0 × 10−6

Correlation coefficient 0.9994 0.9998

Response time, sec 40 30

Working pH range 6.0–8.5 5.5 – 8.5

Concentration range, mol/L 3.2 × 10−5 to 10−2 10−5 to 10−2

Stability, days 21 35

Repeatabilityc 0.94 0.75

Intermediate precisiond 1.24 0.89

Accuracy ± SD 99.97 ± 0.72 99.99 ± 0.57a Average of four determinations.b LOD calculated at the interception of extrapolated arms in the

potential profile.c Intraday (n = 9) RSD, % of concentrations 10−2, 10−3, and

10−4 mol/L CFQ.d Interday (n = 9) RSD, % of concentrations 10−2, 10−3, and

10−4 mol/L CFQ.

12_150185_Yehia.indd 76 20/02/16 10:25 AM

Yehia et al.: Journal of aoaC international Vol. 99, no. 1, 2016 77

stable potential over pH ranges 6.0–8.5 and 5.5–8.5 for sensor 1 and sensor 2, respectively (Figure 5). Beyond this pH range, the potential was unstable, which may be attributable to the degradation of CFQ in acidic and alkaline media, forming 5,6,7,8-tetrahydroquinoline with tertiary amine group of pKa 6.0. Therefore potential increment in acidic pHs (below 5.0) was observed, due to protonation of the proposed degradation product (ionized) or interference of hydronium ions. Whereas the potential decreased in alkaline pHs (above 8.5) since the degradation product becomes no longer protonated (unionized).

Selectivity of the Proposed Sensors

The results obtained through a separate solution method (Table 2) show reasonable selectivity by both sensors for CFQ determination compared with degradation products and

several other interfering ions commonly used as pharmaceutical additives or that can be found in biological fluid. ,KA B

pot was calculated using the following equation:

log 1 log,KE ES

z z aA Bpot B A

A B A( )=−

+ − (2)

where S is the slope of the calibration plot, aA is the activity of CFQ, zA and zB are the charge numbers, and EA and EB are the measured potentials for CFQ and the interfering ion, respectively.

Synthetic Mixture Analysis

In practical applications, primary and interfering ions are present at the same time and the electrode may behave quite differently in these cases as opposed when either type of ion is solely present. However, results obtained upon analysis of synthetic mixtures containing different ratios of intact drug and acidic degradation products varying from 1:1 to 1:6 showed that both sensors can be successfully used for selective determination of CFQ in the presence of its acidic degradation products (Table 3). Therefore both sensors are recommended in stability studies.

Determination of CFQ in Pharmaceutical and Biological Samples

Pharmaceutical formulation.—Potentials (E) of the buffered dosage form solutions were measured using sensors 1 and 2, and the CFQ concentrations (Cunk) were calculated as follows:

10CunkE KS=−

(3)

Bovine plasma.—To eliminate potential drift during measurement in plasma samples, standard addition was applied, where the potential (E1) of a 20 mL spiked plasma sample is measured, then 0.5 mL 1 × 10−4 mol/L CFQ standard solution is added to the first volume, and the potential (E2) remeasured. Applying both potentials, the following two equations are produced:

101

CunkE KS=−

(4)

Figure 4. Profile of the potential in millivolts plotted against the CFQ log concentrations obtained using sensors 1 and 2.

Figure 5. Effect of pH on the performance of sensors 1 and 2.

Table 2. Potentiometric selectivity coefficient (KA,Bpot )

for the two investigated CFQ sensors using the separate-solutions method

Interferent 1 × 10−3, mol/L

Selectivity coefficienta

Sensor 1 Sensor 2

Na+ 9.1 × 10−3 8.4 × 10−3

K+ 1.3 × 10−2 1.1 × 10−2

NH4+ 1.4 × 10−2 1.1 × 10−2

Mg2+ 2.1 × 10−2 1.5 × 10−2

Ca2+ 1.9 × 10−2 1.4 × 10−2

Acidic degradation products 6.5 × 10−3 4.1 × 10−3

Alkaline degradation products 6.5 × 10−3 4.4 × 10−3

Oxidative degradation products 6.5 × 10−3 2.6 × 10−3

a Average of three separate determinations.

12_150185_Yehia.indd 77 20/02/16 10:25 AM

78 Yehia et al.: Journal of aoaC international Vol. 99, no. 1, 2016

102C V C V

V Vunk o std std

o std

E KS

++

=−

(5)

where Cunk and Vo are the concentration and volume of plasma samples, respectively, whereas Cstd and Vstd are the concentration and volume of the added CFQ standard solution, respectively.

The response with standard addition (Equation 5) is divided by the response without a standard addition (Equation 4) to obtain the following equation:

10

10

2

1

C V C VC V

V V

V

unk o std std

unk o

o std

E KS

o

E KS

( )( )

+=

+−

− (6)

Finally, the unknown concentration in the plasma samples (Table 4) is calculated as follows:

1 10 12 1

C

C VV

VV

unk

std std

o

std

o

E ES

=

+

−− (7)

Cow milk.—Cow milk is a complex matrix containing approximately 87.7% water, 4.9% lactose, 3.4% fat, 3.3% protein, and 0.7% minerals. Highly interfering milk ions may affect CFQ cation response. As a result, successive standard additions to spiked milk samples are required to overcome any matrix interference, where the initial potential for the sample is measured (E1). Then potential measurements are taken after each added standard (E2). Equation 6 was transformed into linear expressions as follows:

10 12 1V V

VC VC V

o std

o

E ES std std

unk o

+= +

− (8)

10 12 1V V V

CC Vo std

E ES o

unkstd std( )+ = +

− (9)

Equation 9 is a linear expression, where the graphical plot

of ( )102 1

V Vo std

E ES+−

versus C Vstd std has a straight line with

slope 1Cunk

. Moreover, as with traditional standard addition

calibrations, linear extrapolation with an x-intercept, where

y = 0, results in C Vunk o− from which Cunk can be calculated. Figure 6 shows the standard addition curve for the milk sample.

Table 4 contains the results of the pharmaceutical and biological applications. Both sensors were used for CFQ determination in the pharmaceutical samples with high accuracy and precision. On the other hand, sensor 2 is specifically recommended for accurate CFQ determination in biological samples, with no plasma protein or milk fat interference.

Degradation Kinetics

The offline monitoring of degradation reaction is a time-consuming procedure. On the contrary, to accurately estimate constants for the reaction rate, online monitoring is recommended, given the real-time observation of CFQ hydrolysis. Sensor 2 was used for this study thanks to a shorter response time and longer-lasting stability. The linear relationship obtained from plotting log% of the remaining concentration against time (Figure 7) indicated a first-order degradation process for CFQ. However, as hydrolysis took place in large excesses of buffer (pH 10.0), pseudo–first-order reaction rate was suggested to be at play, which is the term used when two reactants are involved in the reaction but the change in concentration of one of the reactants (buffer) is negligible as a result of its presence in large excess compared with the change in concentration of the other reactant (drug). The Nernstian responses (S ) of the calibration curves at

Table 3. Determination of CFQ in synthetic mixtures by the proposed sensors

concentration, mol/L CFQ recovery %, mean ± SDa

CFQ

Acidic degradation

products Sensor 1 Sensor 2

1 × 10−3 1 × 10−3 101.41 ± 1.43 101.44 ± 1.39

1 × 10−3 4 × 10−3 101.85 ± 1.80 101.38 ± 1.32

1 × 10−3 5 × 10−3 100.92 ± 2.16 100.56 ± 1.53

1 × 10−3 6 × 10−3 103.74 ± 2.16 102.23 ± 1.58a Average of three determinations.

Table 4. Determination of CFQ by the proposed sensors in its pharmaceutical and biological samples

Application

Recovery %, mean ± SD

Sensor 1 Sensor 2

Market samples

Cobactan suspension 99.17 ± 0.87a 99.93 ± 0.44a

Cobactan LC ointment 98.88 ± 0.61a 99.75 ± 0.35a

Biological samples

Bovine plasma N/A 101.58 ± 4.44b

Cow milk N/A 101.28 ± 4.75b

a Average of three determinations.b Average of six determinations.

Figure 6. Standard addition plot on 20 mL spiked milk sample (1 × 10–5 mol/L), using a successive 0.5 mL addition of 1 × 10–4 mol/L CFQ standard solution.

12_150185_Yehia.indd 78 20/02/16 10:25 AM

Yehia et al.: Journal of aoaC international Vol. 99, no. 1, 2016 79

these pH values were checked and confirmed as not having significant differences compared with the calibration plot for pH 7.0 (data not provided). Therefore C0 and concentrations at t-time (Ct) can be calculated as follows:

CEc KS=−

1000

(10)

CtEct KS=−

10 (11)

where Ec0 and Ect are the initial and t-time potentials, respectively. To overcome different intercepts (K ) in each calibration plot, the following equation was concluded to calculate log% of the remaining concentration at each of the three pH values:

2 log 20

0CC

E ES

t ct c+ = +− (12)

Equation 12 is a newly designed mathematical equation for the kinetic calculations for log% of the remaining concentration in potentiometric applications.

Different parameters that affect the reaction rate were studied. The effect of temperature and pH on CFQ degradation was studied by conducting reactions at different temperatures using different buffer pH values as shown in Figure 8. K and t1/2 were calculated for each temperature. In addition, log k was plotted against the reciprocal of the temperature in Kelvin (Arrhenius plot), as shown in Figure 9, to demonstrate the effect of temperature on the rate constant. It was concluded that as the temperature increased, the rate of hydrolysis increased with a decrease in t1/2 (Table 5). The pH of the buffer used was another factor that affected the reaction rate. Therefore, buffers with different pH values were used to study the hydrolysis reaction. Table 5 shows that the rate of hydrolysis increased as pH increased. Because the degradation process was accelerated within a narrow temperature range, the effect of pH was more obvious than the effect of temperature on the CFQ hydrolysis rate, especially at pH 10.0.

Finally, the proposed electrochemical method was statistically compared with the manufacturer’s RP-HPLC method. The calculated values show that there is no significant difference between the two methods in terms of accuracy and precision (Table 6).

Conclusions

In conclusion, the sensors described were sufficiently simple and selective for the quantitative determination of CFQ in presence of its degradation products in its pure form and in pharmaceutical formulations. The better response characteristics and selectivity coefficients of sensor 2, along with its wider linearity and lower detection limit, made accurate determination of CFQ in low concentrations in bovine plasma and cow milk spiked samples feasible without pretreatment or separation steps that have been required in previously reported methods. Initial plasma concentration of CFQ was almost determined using sensor 2. However the CFQ MRL for milk was lower than the LOD. In fact, research to decrease the LOD in sensors will be an area of future exploration. Moreover, online kinetic monitoring of the CFQ hydrolysis process is a convenient way to estimate reaction rate and half-life, in addition to the accurate real-time determination of any remaining drug. Although temperature changes were conducted within a narrow range, the

Figure 7. First-order plot for CFQ hydrolysis (1 × 10–3 mol/L) with buffer pH 10.0 at 25°C.

Figure 8. First-order plot for CFQ hydrolysis (1 × 10–3 mol/L) with buffer pH 9.0, 9.5, and 10.0 at different temperatures.

12_150185_Yehia.indd 79 20/02/16 10:26 AM

80 Yehia et al.: Journal of aoaC international Vol. 99, no. 1, 2016

research revealed how little effect temperature has on reaction rates compared with increases in pH. Accordingly the proposed sensors can be used to routine analyze CFQ in quality-control laboratories.

Acknowledgments

We thank and gratefully acknowledge the Analytical Chemistry Department, Faculty of Pharmacy, Cairo University for its financial support. We also express sincere gratitude to Mohamed Khaled Abd El-Rahman, Analytical Chemistry, Faculty of Pharmacy, Cairo University (Cairo, Egypt), for his guidance and beneficial discussions throughout this work. His expertise in electrochemical science improved this research and nourished our scientific and technical knowledge.

References

(1) Becker, M., Zittlau, E., & Petz, M. (2004) Anal. Chim. Acta. 520, 19–32. doi:10.1016j.aca.2004.04.022.

(2) Limbert, M., Isert, D., Klesel, N., Markus, A., Seeger, K., Seibert, G., & Schrinner, E. (1991) Antimicrob. Agents Chemotherapy 35, 14–19. doi:10.1128/AAC.35.1.14

(3) Schlegelová, J., Dendis, M., Benedík, J., Babák, V., & Rysánek, D. (2003) Vet. Microbiol. 92, 327–334. doi:10.1016/ S0378-1135(02)00409-1

(4) Shpigel, Y., Levin, D., Winkler, M., Saran, A., Ziv, G., & Böttner, A. (1997) J. Dairy Science 80, 318–323. doi:10.3168/jds.S0022-0302(97)75941-1

(5) Ehinger, A. M., Schmidt, H., & Kietzmann, M. (2006) Vet. J.172, 147–153. doi:10.1016/j.tvjl.2005.02.029

(6) Swinkels, J.M., Lam, T J.G.M., Green, M.J., & Bradley, A.J. (2013) Vet. J. 197, 682–687. doi:10.1016/j.tvjl.2013.03.010

(7) Junza, A., Amatya, R., Barrón, D., & Barbosa, J. (2011) Chromatogr. B. 879, 2601–2610. doi:10.1016/j .jchromb.2011.07.018

(8) Pérez-Burgos, R., Grzelak, E.M., Gokce, G., Saurina, J., Barbosa, J., & Barrón, D. (2012) Chromatogr. B. 899, 57–65. doi:10.1016/j.jchromb.2012.05.002

(9) Quesada-Molina, C., Claude, B., Garcia-Campaña, A. M., del Olmo-Iruela, M., & Morin, P. (2012) Food Chem. 135, 775–779. doi:10.1016/j.foodchem.2012.05.014

(10) Quesada-Molina, C., García-Campaña, A.M., & del Olmo-Iruela, M. (2013) Talanta 115, 943–949. doi:10.1016/j .talanta.2013.07.008

(11) Council Regulation (EEC) No. 2377/90 (1990) Off. J. Eur. Communities, L224, 1–8

(12) Elghobashy, M.R., Bebawy, L.I., Abbas, S.S., & Shokry, R.F. (2013) Chromatographia 76, 1141–1151. doi:10.1007/ s10337-013-2516-5

(13) Sugioka, T., Asano, T., Chikaraishi, Y., Suzuki, E., Sano, A., Kuriki, T., Shirotsuka, M., & Saito, K. (1990) Chem. Pharm. Bull. 38, 1998–2002. doi:10.1248/cpb.38.1998

(14) Zalewski, P., Skibiński, R., & Cielecka-Piontek, J. (2014) J. Pharmaceut. Biomed. Anal. 92, 22–25. doi:10.1016/j.jpba.2013.12.039

(15) Doĺhań, A., Jelińska, A., & Manuszewska, M. (2014) Acta Pol. Pharm. 71, 249–64.

(16) Umezawa, Y., Buhlmann, P., Umezawa, K., Tohda, K.,& Amemiya, S. (2000) Pure Appl. Chem. 72, 1851–2082. doi:10.1351/pac200072101851

(17) Buck R. P. (1978) Theory and principles of membrane electrodes, H. Freiser (Ed),Vol. 1, Chap. 1, Ion-Selective Electrodes in Analytical Chemistry, Plenum Press, New York,NY.

(18) Hassan, S.S., Amer, M.M., Abd-El-Fatah, S.A., & El-Kosasy, A.M. (1998) Talanta 46, 1395–1403. doi:10.1016/ S0039-9140(98)00008-3

(19) El-Ghobashy, M.R., Yehia, A.M., & Mostafa, A.A. (2009) Anal. Lett. 42, 123–140. doi:10.1080/00032710802568606

Table 5. Kinetic data of CFQ alkaline degradation

pH Temperature, °C k, 1/min t1/2, min t90%, min

9.0 25 0.011 62.690 9.318

30 0.019 36.697 5.454

35 0.033 20.753 3.084

9.5 25 0.019 35.823 5.324

30 0.031 22.290 3.313

35 0.053 12.970 1.928

10.0 25 0.025 27.862 4.141

30 0.048 14.467 2.150

35 0.070 9.931 1.476

Figure 9. Arrhenius plot for CFQ hydrolysis (1 × 10–3 mol/L) with buffer pH 9.0, 9.5, and 10.0.

Table 6. Statistical analysis of the results obtained by the proposed sensors and by the manufacturer’s method for the determination of CFQ in pure form

Value Sensor 1 Sensor 2 Manufacturer’s methoda

Mean 99.98 99.99 100.32

SD 0.72 0.57 0.69

RSD, % 0.72 0.57 0.69

n 6 4 6

Variance 0.52 0.32 0.48

Student’s t-test 0.835 (2.228)b 0.789 (2.306)b N/A

F-value 1.08 (5.05)b 1.50 (9.01)b N/Aa RP-HPLC method using a C18 stainless steel column (250 × 4.6 mm,

5 μm); a mobile-phase mixture of 90 mL acetonitrile, 12 mL phosphoric acid, and 3.45 g sodium perchlorate monohydrate in 1 L water; pH 3.6 adjusted with diethylamide, with a flow rate of 1.0 mL/min, and UV detection at 270 nm.

b The numbers in parentheses are the corresponding tabulated values of t and F at P = 0.05.

12_150185_Yehia.indd 80 20/02/16 10:26 AM

Yehia et al.: Journal of aoaC international Vol. 99, no. 1, 2016 81

(20) El-Kosasy, A. M., Nebsen, M., Abd El-Rahman, M. K., Salem, M. Y., & El-Bardicy, M. G. (2011) Talanta 85, 913–918. doi:10.1016/j.talanta.2011.04.071

(21) El-Sayed, M.A. (2014) Sensor. Actuat. B-Chem. 190, 101–110. doi:10.1016/j.snb.2013.08.065

(22) Abd El-Rahman, M.K., Zaazaa, H.E., Badr El-Din, N., & Moustafa, A.A. (2015) Talanta 132, 52–58. doi:10.1016/j.talanta.2014.08.068

(23) Shantier, S.W., Gadkariem, E.A., Adam, M.O., & Mohamed, M.A. (2013) Intl. J. Biomed. Sci. 9, 162–167

(24) Sainsbury, M. (2001) Heterocyclic Chemistry, E.W. Abel (Ed), The Royal Society of Chemistry, London, United Kingdom, 18–41. doi:10.1039/ 9781847551061-00018.

(25) Sideris, E.E., Valsami, G.N., Koupparis, M.A., & Macheras, P.E. (1999) Eur. J. Pharm. Sci. 7, 271–278. doi:10.1016/ S0928-0987(98)00035-9

(26) Flink, S., van Veggel, F.C.J.M., & Reinhoudt, D.N. (2000) Adv. Mat. 12, 1315–1328. doi:10.1002/ 1521-4095(200009)

12_150185_Yehia.indd 81 20/02/16 10:26 AM