1. INTRODUCTION:

Ciprofloxacin (Figure 1A) is a second-generation fluoroquinolone, which kills the bacteria by inhibiting enzymes (DNA gyrase and topoisomerase IV) necessary to separate bacterial DNA.

Its ophthalmic solution is commonly used to treat bacterial infection of the eye including conjunctivitis and corneal ulcers. It can be also co-formulated with naphazoline (Figure 1B), a sympathomimetic agent with marked alpha adrenergic activity, to relieve redness, puffiness, and itchy/watering eyes due to colds, allergies, or eye irritations.

The literature review has highlighted that ciprofloxacin could be analyzed by UV-Vis spectrophotometry ^[1-3]; while naphazoline by derivative UV spectroscopy and chemometrics-assisted UV spectrophotometry ^[4-7]. No official procedure is given in well-known pharmacopeias for the simultaneous determination of ciprofloxacin and naphazoline.

(A)

1-cyclopropyl-6-fluoro-4-oxo-7-(piperazin-1-yl)-quinoline-3-carboxylic acid

(B)

2-(naphthalen-1-ylmethyl)-4,5-dihydro-1H-imidazole

Figure 1. Chemical structure of ciprofloxacin (A) and naphazoline (B)

Up till now, chemometric quantitative calibration techniques in spectral analysis has been gaining importance in the quality control of active compounds in mixtures and pharmaceutical formulations ^[8,9].

The aim of this study was to propose the combination of UV spectrophotometry and chemometrics (CLS, ILS and PCA) for the simultaneous determination of ciprofloxacin and naphazoline in their binary mixtures and eye drops.

2. EXPERIMENTAL:

2.1. Apparatus and software

Absorption spectra were registered and treated by using a UNICAM UV 300 double beam spectrophotometer (Thermo Spectronic, USA) with a fixed slit width (1.5 nm) connected to an IBM computer loaded with Thermo Spectronic VISION32 software and 1-cm quartz cells. The zero-order spectra were recorded in the wavelength range of 190 – 400 nm against a blank (water) at Intelliscan mode to enhance the signal-to-noise ratio of absorbance peaks without extended scan duration with a Dl = 0.1 nm (i.e. 30-120 nm/min). The spectra were differentiated and smoothed by using Savitzky – Golay filter. The chemometric data treatment was done using MATLAB R2013a software (The MathWorks, Natick, MA, USA).

2.2. Reagents and standard solutions

Ciprofloxacin hydrochloride (CIP) and naphazoline nitrate (NAP) were kindly provided by the National Institute of Drug Quality Control (Vietnam). De-ionized doubly distilled water was used throughout. Stock solutions of CIP (3000 mg/L) and NAP (500 mg/L) were freshly made in water. A concentration set of standard solutions were prepared in 50-mL volumetric flasks by using the same stock solutions.

2.3. Sample solution

Two domestic commercial formulations containing CIP 15 mg + NAP 2.5 mg per 5-ml vial i.e. Diflox (Minh Hai Pharmaceutical Joint Stock Company, Vietnam) and Hadolmax (Ha Tay Pharmaceutical Joint Stock Company, Vietnam) were studied. For each formulation, the content of five vials was mixed well. Appropriate dilution was then made in a 50-mL volumetric flask to obtain the test solution ca. CIP 30 mg/L + NAP 5 mg/L.

3. THEORETICAL BACKGROUND:

3.1. Classical Least Squares (CLS)

CLS is also known as direct least squares, the Beer's law method, and sometimes as the K-matrix calibration algorithm. It involves the application of Multiple Linear Regression (MLR) to the classical expression of the Beer–Lambert law of spectroscopy:

A=K×C

This equation is a matrix equation and it can be written as a linear equation system:

A₁ = K₁₁C₁ + K₁₂C₂ + …+ K_1CC_C

A₂ = K₂₁C₁ + K₂₂C₂ + …+ K_2CC_C

A₃ = K₃₁C₁ + K₃₂C₂ + …+ K_3CC_C

… … …

A_W = K_W1C₁ + K_W2C₂ + …+ K_WCC_C

Where, A_w is the absorbance at the w^th wavelength; K_wc is the calibration coefficient for the c^th component at the w^th wavelength; C_c is the concentration of the c^th component.

K-matrix and unknown concentration (C₀) are determined as follows.

K = (C^TC)^-1 C^TA

C₀ = A₀ K^T (KK^T)^-1

Where, C^T and K^T are the transposes of C-matrix and K-matrix, respectively.

3.2. Inverse Least Squares (ILS)

ILS is also known as inverse Beer's law method, and sometimes as the P-matrix calibration algorithm. It also involves the application of MLR to the inverse expression of the Beer–Lambert law of spectroscopy:

C=P×A

This equation can be written as a linear equation system:

C₁ = P₁₁A₁ + P₁₂A₂ + …+ P_1WA_W

C₂ = P₂₁A₁ + P₂₂A₂ + …+ P_2WA_W

C₃ = P₃₁A₁ + P₃₂A₂ + …+ P_3WA_W

… … …

C_C = P_C1A₁ + P_C2A₂ + …+ P_CWA_W

Where, A_w is the absorbance at the w_th wavelength; P_cw is the calibration coefficient for the c^th component at the w^th wavelength; C_c is the concentration of the c^th component.

P-matrix and unknown concentration (C₀) are determined as follows.

P = (A^TA)^-1 A^TC

C₀ = A₀.P

Where, A^T is the transpose of A-matrix.

3.3. Principal Component Analysis (PCA)

This model-building procedure has two steps: (i) the eigenvectors of the centered absorbance data matrix are calculated and (ii) the concentration data matrix is regressed by MLR. This procedure can be mathematically expressed as follows.

A_proj = V_c^TA

In this equation, A_projindicates the matrix containing the new coordinates (the projections); V_c^T represents the matrix containing the basis vectors, one column for each factor retained; whilst A denotes the original training set absorbance matrix. When the matrix A_proj is known, the unknown concentration matrix can be calculated as below.

C = FA_proj

Here, F represents the calibration coefficient for the obtained linear equation system.

4. RESULTS AND DISCUSSION:

In this study, three chemometrics-assisted UV spectrophotometric methods i.e. CLS, ILS and PCA were proposed for the simultaneous determination of CIP and NAP in binary mixture without prior treatment. This approach overcomes the obstacle of severe spectral overlapping as shown in Figure 2, which precludes the possibility of using direct UV spectrophotometric method for the simultaneous determination of NAP in the presence of CIP.

At first, the construction of the training set concentration matrix was done with nine binary mixtures containing 24 – 36 mg/L of CIP and 4 – 6 mg/L NAP (Table 1). The composition of the mixtures was randomly designed so as to obtain non-correlated concentration profiles.

In order to have good quantitative results with a mixture, the additivity of absorbances must be obeyed for spectrophotometric methods used. Also taking into account that NAP’s absorbances markedly reduced, the wavelength region of 225 – 265 nm was deliberately chosen. As a result, the absorbance data matrix for the training set concentration matrix was obtained by measuring zero-order absorbances in this region with Δλ = 1 nm for CLS and PCA techniques. The major drawback of these techniques is that all of the interfering species must be known and their concentrations are included in the model. This need can be eliminated by using ILS technique, for which Δλ = 5 nm was fixed. The reason is that ILS’s inverted matrix has dimensions equal to the number of wavelengths in the spectrum and this number should not exceed the number of calibration samples.

Figure 2. UV spectra of CIP 30 mg/L, NAP, 5 mg/L, binary mixture of CIP 30 mg/L + NAP 5 mg/L, and their corresponding addition spectrum

In CLS and ILS techniques, the absorbance data matrix and concentration data matrix were used for calibration and regression to predict the unknown concentrations of CIP and NAP in their binary mixtures and pharmaceutical formulations. On the other hand, in PCA technique the variance and covariance matrices corresponding to the absorbance data matrix were calculated for the basis vectors and matrix containing the new coordinates.

The predictive ability of chemometric models is evaluated in terms of the Standard Error of Prediction (SEP) and the Standard Error of Calibration (SEC), which are given by the following equations.

Where, C_i^Added is the added concentration of drug; C_i^Found is the predicted concentration of drug; n is the total number of synthetic mixtures; and p is the number of components in the mixtures.

The results obtained in the application of CLS, ILS and PCA techniques to the same binary mixtures of the training set concentration matrix are indicated in Tables 1 and 2. The SEP and SEC values were reasonably acceptable (< 0.8 for both compounds). The correlation between added and found constituent concentrations is defined by r values, which were close to 1 meaning that no interference was seen in this set of synthetic mixtures.

Table 1. Results obtained for the determination of CIP and NAP in synthetic mixtures by chemometrics-assisted UV spectrophotometry

No.	Mixture		Recovery (%)
No.	Mixture		CLS		ILS		PCA
	CIP	NAP	CIP	NAP	CIP	NAP	CIP	NAP
1	30	5	101.5	98.6	102.2	98.3	101.9	99.2
2	24	6	99.3	97.9	100.0	97.5	100.5	98.3
3	36	6	98.5	99.3	98.9	99.7	97.9	100.1
4	36	5	98.3	101.9	97.9	102.1	99.1	102.6
5	30	6	102.3	101.5	101.8	101.6	103.0	100.9
6	36	4	102.5	98.9	103.1	99.5	103.3	99.4
7	24	4	99.4	99.1	100.2	99.6	99.9	98.9
8	24	5	98.8	99.3	98.3	98.8	99.1	98.4
9	30	4	100.9	102.6	100.5	101.7	99.8	103.0
Mean			100.2	99.9	100.3	99.9	100.5	100.1
SD			1.7	1.7	1.8	1.6	1.9	1.7

Table 2. Statistical summary of chemometrics-assisted UV spectrophotometric methods

Parameter	CLS		ILS		PCA
	CIP	NAP	CIP	NAP	CIP	NAP
SEP	0.63	0.10	0.72	0.10	0.75	0.10
SEC	0.52	0.08	0.59	0.08	0.61	0.08
r	0.9947	0.9954	0.9937	0.9951	0.9931	0.9954
Intercept	- 0.2280	0.0779	- 0.1640	0.0778	- 0.0347	0.0806
Slope	1.0097	0.9830	1.0090	0.9827	1.0063	0.9843

The chemometrics-assisted spectrophotometric methods (CLS, ILS and PCA) were successfully applied to the analysis of CIP and NAP in eye drops (Table 3). The results were obtained with good repeatability (RSD < 2%). To assess the accuracy of these methods, standard addition technique was carried out i.e. the pre-analyzed sample was spiked with known quantities of CIP and NAP equivalent to 20% of their label claim in eye drops. The recoveries were in the range of 98 – 102% for both compounds indicating that the problem of interference of the mixture and/or additives (e.g. polysorbate 20, benzalkonium chloride) could be successfully solved. Moreover, the results were statistically compared with each other. It is seen that at 95% confidence level, there was no significant difference between the accuracy (evaluated by one-way ANOVA test, calculated F value < tabulated F value) and precision (evaluated by Bartlett test, calculated χ²value < tabulated χ²value) among all proposed methods.

Table 3. Assay results for the determination of CIP and NAP in eye drops

	% Label claim (mean ± SD, n = 6) F_0,05;2;15 = 3.682; χ²_0,95;2 = 5.991
Method	Diflox		Hadolmax
Method	CIP	NAP	CIP	NAP
CLS	101.1 ± 1.1	98.5 ± 1.9	102.3 ± 1.9	97.8 ± 1.7
ILS	100.8 ± 1.5	98.9 ± 1.9	101.9 ± 1.8	98.5 ± 1.8
PCA	101.6 ± 1.8	99.0 ± 1.4	101.5 ± 2.2	98.9 ± 1.6
F	0.439	0.137	0.203	0.642
χ²	1.072	0.528	0.246	0.064

5. CONCLUSION:

The proposed methods i.e. chemometrics-assisted UV spectrophotometry (CLS, ILS, PCA) could be applied with great success for the resolution of the binary mixture of CIP and CAF, whose spectra are strongly overlapped. For the same commercial formulations, these methods were statistically comparable with regard to good accuracy and precision suggesting that they could be used for the routine analysis of CIP and NAP in eye drops. This study stresses on the use of chemometrics to develop simple and reliable spectrophotometric methods without any prior steps of separation and extraction common to classical determination processes.

6. REFERENCES:

1. Sultan SM, Suliman FO. Flow injection spectrophotometric determination of the antibiotic ciprofloxacin in drug formulations. Analyst. 117(9); 1992: 1523-1526.

2. Fratini L, Schapoval EES. Ciprofloxacin determination by visible light spectrophotometry using iron(III) nitrate. International Journal of Pharmaceutics. 127(2); 1996: 279-282.

3. Walily AFME, Belal SF, Balery R. Spectrophotometric and spectrofluorimetric estimation of ciprofloxacin and norfloxacin by terary complex formation with eosin and palladium (II). Journal of Pharmaceutical and Biomedical Analysis. 14(5); 1996: 561-569.

4. Nour el deen S, Maha H, Mohammed A, Rehab A. Spectrophotometric, chemometric and chromatographic determination of naphazoline hydrochloride and chlorpheniramine maleate in the presence of naphazoline hydrochloride alkaline degradation product. Bulletin of Faculty of Pharmacy, Cairo University. 51(1); 2013: 57-68.

5. Joseph-Charles J, Bertucat M. Simultaneous determination of naphazoline nitrate and tetramethylthionine base in eye drops by first-derivative UV spectrophotometry. Analytical Letters. 32(2); 1999: 373-382.

6. Hajian R, Shams N, Kaedi, I. Application of ratio derivative spectrophotometry for simultaneous determination of naphazoline and antazoline in eye drops. E-Journal of Chemistry. 7(4); 2010: 1530-1538.

7. Effat S, Massoud A, Hassan F, Alma A. A rapid derivative spectrophotometric method for simultaneous determination of naphazoline and antazoline in eye drops. Chemical and Pharmaceutical Bulletin. 54(1); 2006: 119-122.

8. Forina M, Casolino MC, de la Pezuela Martinez C. Multivariate calibration: applications to pharmaceutical analysis. Journal of Pharmaceutical and Biomedical Analysis. 18(1-2); 1998: 21-33.

9. Graciela ME, Nicholaas (Klaas) MF, Héctor CG, Arsenio M de la P, Alejandro CO, Ronei JP. Second- and third-order multivariate calibration: data, algorithms and applications. Trends in Analytical Chemistry. 26(7); 2007: 752-765.

Received on 16.05.2014 Modified on 25.05.2014

Asian J. Research Chem. 7(5): May 2014; Page 461-465