Rifampicin (99.6%) was kindly gifted by Novartis (Bangladesh) Limited, Bangladesh. Ethyl acetate was of analytical grade and used without further purification.

A UV spectrophotometer (UV- 1800 240V, SHIMADZU) equipped with a desktop (Intel® Core™ i3 processor) having UVprobe multifunctional UV control software was used to develop and validate the suggested method.

A simple UV spectroscopic method was developed and validated for the analysis of rifampicin in a combined solid dosage form. The development and validation procedure is as follows.

Ten (10) mg of standard rifampicin was measured using electronic balance (Model-ATY224, SHIMADZU) and diluted to 100 ml with distilled water to prepare a standard solution of concentration of 100 μg/mL.

Ten (10) mL of ethyl acetate and 10 mL of water was taken in a separating funnel and shaken well. The ethyl acetate layer was then collected in a volumetric flask. This solution was used as blank.

Two (2.0) mL of stock solution was diluted to 10 mL using distilled water. This solution was then extracted by 10 mL of ethyl acetate. The diluted ethyl acetate solution was scanned from 200 to 400 nm in UV spectrophotometer to determine the λ_max.

The water and ethyl acetate were mixed at a ratio of (water: ethyl acetate) 1: 0.25, 1: 0.5, 1: 1, 1: 1.5, and 1: 2 to get the optimum extraction ratio.

The mixtures (water: ethyl acetate = 1: 1 (v/v)) were allowed to stand for different time interval such as 1.0, 1.5, 2.0, and 3.0 h to find the best separation time.

The ICH guidelines (International Conference on Harmonisation (ICH), 2005) were followed to validate the analytical method. The newly developed UV spectrophotometric method was validated for parameters like specificity, linearity, accuracy, precision, limit of detection (LOD) and limit of quantitation (LOQ).

The specificity of the method was confirmed through establishing the identity of rifampicin by Thin Layer Chromatographic (TLC) technique. The standard rifampicin solution and the extracted sample (tablet) solution were spotted on Thin Layer Chromatography (TLC) plates (Silica gel F₂₅₄) and the plates were developed using mobile phase comprising of 10 and 15% methanol in chloroform.

The linearity graphs were plotted for the concentration range of 2.5–35.0 μg/ml. Least square analysis was used to obtain the slope, intercept, correlation coefficient (R²) and regression equation of the method.

Accuracy represents the bias of the method. Five different concentrations such as 10 μg/mL, 15 μg/mL, 20 μg/mL, 25 μg/mL, 30 μg/mL, and 35 μg/mL of standard rifampicin solution were used to find the accuracy of the method. Three replicates of each concentration were used and then their average concentration was used to calculate the accuracy by applying the following equation: Accuracy (%)= Observed concentrationActual concentration×100 The accuracy should be within 15% of the actual value, except at the lower limit of quantitation (LOQ), where it should be ≤20% (Xiong et al., 2009).

Precision of the assay was assessed with respect to repeatability (intraday) and intermediate precision (inter-day). The precision of the current method was determined by analyzing three different concentrations. Three replicates of each concentration were used to calculate the relative standard deviation (RSD) by following the equation below: RSD (%)= Standard deviationMean×100 The calculated RSD at each concentration level should not exceed 15%, except for LOQ, where it should not exceed 20% (Mostafavi et al., 2009).

The LOD was calculated according to following equation: LOD= 3.3×SDσ Where, SD = standard deviation of response

σ = slope of regression line

The LOQ is the measure of an analytical method that can quantify an analyte with adequate accuracy and precision and it can be calculated as follows: LOQ= 10×SDσ Where, SD = standard deviation of response

σ = slope of regression line

Two brands of combined solid dosage form of rifampicin, isoniazid and pyrazinamide are available in Bangladesh. So, the two brands (coded S1 and S2) were assayed. The powdered form of each drug (equivalent to 10 mg of rifampicin) was taken in a 100 mL volumetric flask and dissolved in distilled water. Sonication was done for about 20 min and the solution was then diluted up to the mark with distilled water. The solution was then filtered through Whatman No. 1 filter paper. The filtered solution was further diluted according to the need and extracted by using 10 ml of ethylacetate. Ethyl acetate layer was collected in a volumetric flask. Absorbances of these ethyl acetate solutions were determined.

The newly developed method was compared with the reported Q-absorbance ratio spectrophotometric method (Khamar and Patel, 2012b). The two methods are employed to assay rifampicin following the method described above and the difference between the results of two methods was statistically evaluated by applying t-test and ANOVA (single factor) analysis. The statistical calculations were performed using Microsoft® Excel 2007.

Five microspecies with four ionized (I2, I3, I4, and I5) and one unionized form (I1) of isoniazid were found and only the unionized form (I1) predominates (100%) in the pH range of 6.8–10.2 (Figure 2).

The pyrazinamide exhibits three species one of which is unionized (P1) and the remaining two were ionized forms (P2 & P3) (Figure 3). The abundance of unionized form was 100% in the pH range of 2.6–9.6.

The calculation of pKa over the pH of 0.0–14.0 revealed that rifampicin has 14 ionized species (R2–R15) along with the unionized form (R1) (Figure 4). It is clear from Figure 4 that from pH 6.8–7.6, three species of rifampicin (R1, R2, and R6) predominates (87.2–98.5% in abundance) in the solution.

The partition coefficient (logP) was calculated at 0.1M electrolyte concentration using consensus method. The logP value (Table 1) demonstrates that isoniazid and pyrazinamide are hydrophilic whereas rifampicin is lipophilic i.e., has higher affinity to n-octanol.

The logD was calculated at electrolyte concentration of 0.1M using consensus method implemented in MarvinSketch. The plot of logD vs. pH for isoniazid, pyrazinamide and rifampicin has been presented in Figure 5. From pH 0.0 to 14.0, the logD for isoniazid and pyrazinamide is negative in quantity, indicating both drugs have higher affinity and hence, higher solubility in water. However, rifampicin has logD value over 2 from pH 6.0 to 8.0 suggesting it has higher solubility in n-octanol over this pH range.

The aqueous solubility, in terms of logS, has been presented in Figure 6. The figure indicates that isoniazid, pyrazinamide and rifampicin over the pH of 5.2 to 10.8 were soluble (42.38 mg/mL), freely soluble (363.3 mg/mL) and practically insoluble (<0.1 mg/mL) in water, respectively. However, Table 2 represents intrinsic solubility of these drugs and suggests similar findings as described above.

The isoelectric point is the pH where the molecule has no net charge. The pI value of isoniazid, pyrazinamide and rifampicin is shown in Table 3 and Figure 7.

Figure 4 suggests that rifampicin carries little or no net charges at pH near to 7.0 and hence, is suitable for extraction at this pH. Although isoniazid and pyrazinamide also has no net charge at this pH, but their higher solubility in aqueous media will retard their extraction.

The solubility of isoniazid, pyrazinamide and rifampicin were calculated in 80 different solvents using Abraham descriptors (Abraham et al., 2010). From the virtual screening it was found that 1,2-dichloroethane, 1-butanol, 1-decanol, 1-heptanol, 1-hexanol, 1-octanol, 1-pentanol, 2-methyl-2-propanol, butyl acetate, diethyl ether, ethyl acetate, fluorobenzene, isopropyl myristate, and octadecanol are suitable for the extraction of rifampicin from the water mixture. However, considering the availability, cost and hazard we choose ethyl acetate as an extracting solvent.

The solubility of isoniazid, pyrazinamide and rifampicin were calculated in ethyl acetate using Abraham descriptors (Abraham et al., 2010). The calculated results (Table 4) revealed that isoniazid, pyrazinamide and rifampicin were slightly soluble (1.37 mg/mL), very slightly soluble (0.37 mg/mL) and freely soluble (107.81 mg/mL), respectively. Hence, ethyl acetate was chosen to extract rifampicin from the water mixture of isoniazid and pyrazinamide.

Theoretical investigation of acid dissociation constant (pKa), partition coefficient (logP), distribution coefficient (logD), isoelectric point (pI), and solubility in water (logS) of isoniazid, pyrazinamide and rifampicin suggests that pH range of 6.0–8.0 will favor the extraction of rifampicin using organic solvent (ethyl acetate) as distribution coefficient is higher at this pH range. So, we extracted rifampicin using ethyl acetate at different pH such as 2.6, 4.1, 7.3, 8.3, and 9.7.

The results (Table 5) indicate that pH 7.3 would be suitable to extract rifampicin which is in accordance with our theoretical findings. So, the extraction of rifampicin was performed at pH 7.4 ± 0.1.

The computed enthalpy, Gibbs free energy and entropy are presented in Table 6. The results suggest that rifampicin possesses higher entropy in ethyl acetate (−28.0 J/K/mol) as compared to water (−53.7 J/K/mol) whereas isoniazid and pyrazinamide have higher entropy in water (3.8 and −1.0 J/K/mol, respectively) than ethyl acetate (2.4 and −3.3 J/K/mol, respectively).

The mixture of water and ethyl acetate at different ratio were investigated to get the optimum volume of water and ethyl acetate. Table 7 suggests that solvent ratio of (water: ethyl acetate) 1: 1, 1: 1.5, and 1: 2 are suitable to extract rifampicin from water. So, we select 1: 1 solvent ratio (lowest optimum volume of ethyl acetate) for extracting rifampicin.

The result of extraction is presented in Table 8. From the table it is clear that separation time for 2 h yielded good recovery with less variability. However, the room temperature was maintained to 25°C ± 0.5°C and the cap of the extracting apparatus were kept closed during the whole extraction time to minimize evaporation of ethyl acetate.

The specificity test was performed using Thin Layer Chromatographic (TLC) technique. The standard and extracted sample solution of rifampicin were spotted on Thin Layer Chromatography (TLC) plates (Silica gel F₂₅₄) and the plates were developed using mobile phase consisting of 10 and 15% methanol in chloroform. A single brown spot was found for both solutions at R_f (Retardation factor) of approximately 0.54 when the developed plates were sprayed with vanillin/H₂SO₄ spray reagent.

The λ_max of rifampicin in ethyl acetate solution was found to be 344.0 nm. The absorbances of the sample solution were plotted against concentration range of 2.5–35.0 μg/mL and a good correlation coefficient was obtained. The correlation coefficient (R²) for the best fitting line was 0.990 which indicates good linear relationship of the newly developed method. The slope (m) and intercept (c) of the calibration curve were found to be 0.027 and 0.032, respectively (Figure 8 and Table 9).

The accuracy of the proposed method was in the range of 96.7 ± 0.9% to 101.1 ± 0.4% which is within the acceptable limit. The result of accuracy is presented in Table 10.

The precision of the method has been shown in Table 11. The intraday precision and inter-day precision in terms of % RSD was in the range of 1.09–1.70% and 1.63–2.99%, respectively.

Since all the values of accuracy and % RSD of precision study were within the acceptable range, the results indicate that the method is reliable, reproducible and accurate.

The LOD and LOQ of the developed method were found 0.83 μg/mL and 2.52 μg/mL, respectively.

The assay result of the two marketed preparation S1 and S2, available in Bangladesh were 108.8 ± 1.21% and 115.5 ± 9.37%, respectively (Table 12).

The results of the proposed and the reported method (Khamar and Patel, 2012b) are presented in Table 13. At 95% confidence interval the calculated F_(6.323) and t_(2.515) value were less than their respective critical value (7.71 and 2.776, respectively). This indicates that the difference between the two methods is statistically insignificant. The comparison with the reported method shows that the developed method is accurate and precise.