DEVELOPMENT AND VALIDATION OF A STABILITY-INDICATING HPLC-DAD METHOD WITH ANN OPTIMIZATION FOR THE DETERMINATION OF DIFLUNISAL AND NAPROXEN IN PHARMACEUTICAL TABLETS

Abstract

An artificial neural network (ANN) was employed to model the chromatographic response surface for the isocratic separation of binary mixtures of diflunisal (DFL) and naproxen (NAP). The two drugs were separated using RP-HPLC and detected with a photodiode array detector (HPLC-DAD). The ANN was trained using the composition of the mobile phase including the percentage of acetonitrile in the mobile phase and the pH of the aqueous phase as input variables while the capacity factors were used as the output variables. ANN was trained to accurately describe the retention factors of each of DFL and NAP within a defined experimental space of aqueous phase pH 2.7–5.8 and 38–60% acetonitrile in the mobile phase. Using the best performing ANN, the optimum conditions predicted were 0.05 M sodium dihydrogenphosphate aqueous solution adjusted to pH 3.4 and acetonitrile in a ratio of (50: 50, v/v). The optimized conditions were used to develop a stability-indicating HPLC method for the quantitative determination of DFL and NAP in tablet dosage forms. The drugs were subjected to oxidation, hydrolysis, photolysis, and heat to apply stress conditions. Complete separation was achieved for the parent compounds and all degradation products. The method was validated according to ICH guidelines.

Keywords:

• Artificial Neural Network optimization
• diflunisal
• HPLC-DAD
• method development
• naproxen
• stability indicating

ACKNOWLEDGMENT

This research project was supported by a grant from the “Research Center of the Center for Female Scientific and Medical Colleges,” Deanship of Scientific Research, King Saud University.

Notes

^a Training data.

^b Verification data.

^c Testing data.

R²: Coefficient of multiple determination; MSE: Mean square error; RMS: Root mean square error; ERR%: Percentage relative error.

LOD: Limit of detection; LOQ: Limit of quantitation; S_y/x: Standard deviation of residuals; S_a: Standard deviation of intercept; S_b: Standard deviation of slope; F: Variance ratio, equals the mean of squares due to regression divided by the mean of squares about regression (due to residuals).

t _R : Retention time, in minutes; N: Number of theoretical plates;

k: Capacity factor; α: Selectivity, between each two successive peaks;

R _s : Resolution, between each two successive peaks; T _f :Tailing factor.

*Mixtures were prepared in the working range for DFL and NAP determination as mentioned in Table 3.

E _r %: Percentage relative error; SD: Standard deviation;

RSD (%): Percentage relative standard deviation.

*Average of three concentrations 0.5, 2.5, 5.0 µg/mL for both DFL and NAP, respectively (n = 6).

Commercial tablets are labeled to contain 200 mg DFL and 200 mg NAP per tablet.

^a Mean ± RSD% for five determinations.

^b Theoretical values of F and t for p = 0.05 and n = 5 are 6.39 and 2.31, respectively.

^c For standard addition of 50% at the nominal content (n = 5).

Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/ljlc.