Only one year after the approval of SOF and SMV individually, their combination was also approved, omitting the need for poorly tolerated interferon and achieving high cure rates in patients with and without cirrhosis [15, 16]. Although SMV can be easily determined in the presence of SOF, the opposite is not true. This is evident from their spectra which overlap throughout SOF absorption spectrum between 200–290 nm (Fig. 2). In this work, three chemometric methods were developed; isosbestic point, ratio subtraction and dual wavelength methods to address the problem of overlapping. The choice of chemometric methods was to offer simplicity, rapidity, affordability as well as reliability to analysts faced with the challenge of simultaneous determination of SOF and SMV in their bulk powders and pharmaceutical dosage form. No prior separation was required, just simple mathematical manipulations of the investigated drugs’ spectra that doesn’t require special equipment or extensive training and could be reliably applied for routine analysis.
4.1 Analytical methods
Isosbestic point method: Upon examining the absorption spectra of equal concentration of SOF and SMV, spectral overlap is observed between 200–290 nm and the two spectra intersect at two wavelengths (isosbestic points: 258 and 273 nm). In this method, 273 nm was chosen after careful consideration since it provided more accurate results judged by % recoveries obtained. An experimental confirmation of the isosbestic point was attained by examining the absorbance of 20 µg mL− 1 of SOF, 20 µg mL− 1 of SMV and a mixture of 10 µg mL− 1 of SOF and 10 µg mL− 1 of SMV (Fig. 3). In all three cases, the absorbance value was the same at the isosbestic point. SMV was directly determined using the mixture’s absorbance at 335 nm (SMV λmax) where SOF doesn’t interfere.
Ratio subtraction method
After recording the mixture’s spectrum (200–400 nm), it was divided by the spectrum of 20 µg mL− 1 SMV (B`). The resulting spectrum (Spec 1) represents. The constant is the absorbance plateau between 325–345 nm and after subtracting the ratio spectrum of SOF/SMV (Spec 2) was obtained. A simple multiplication of the ratio spectrum (Spec 2) by B` resolved SOF original spectrum that was part of the mixture. SOF can now be determined from the resolved spectrum at its λmax of 260 nm using previously constructed calibration graphs. SMV can be directly determined from the mixture’s spectrum at its λmax of 335 nm at which SOF had no absorbance (Fig. 4).
Dual wavelength method: After thorough inspection of SOF and SMV spectra, two wavelengths emerged as best candidates for this method: 261 nm and 294 nm. At both wavelengths SMV absorbance was the same (ΔASMV= zero) while SOF absorbance was different, and this difference was also directly and strongly correlated to SOF concentration. The next step was to construct a linear calibration graph using pure SOF solutions and their corresponding difference in absorbance at both wavelengths (A261 nm – A294 nm). This calibration graph was used to directly find SOF concentration in the mixture. SMV concentration in the mixture could be directly found from its absorbance at 335 nm, where SOF doesn’t interfere (Fig. 5).
4.2 Validation of the proposed methods
ICH guidelines regarding linearity, accuracy, precision, limit of detection and limit of quantitation [17] were followed to validate the methods presented in this work.
4.2.1 Linearity:
Absorbance values at 335 nm were plotted against corresponding SMV concentration to construct its calibration curve. The linear regression equation’s terms were calculated where correlation coefficient was 0.9999 (Table 1). The linearity range was 3–50 µg mL− 1 with LOD as low as 0.47 µg mL− 1 Fig. (S1).
Method I
The same manipulation was done for SOF but at 273 nm (isosbestic point). The linear regression equation’s terms listed in Table 1 show a correlation coefficient of 0.9998 over a linear range between 2–50 µg mL− 1 with LOD of 0.60 µg mL− 1 Fig. (S2).
Method II
SOF absorbance values at 260 nm were used to construct its calibration graph and compute its linear regression equation. A correlation coefficient of 0.9998 over a concentration range of 2–50 µg mL− 1was attained (Table 1) with a similar LOD of 0.53 µg mL− 1 Fig. (S3).
Method III
SOF ΔA values (A261 nm – A294 nm) were found to be strongly correlated (r = 0.9998) to their corresponding concentrations over a range of 4–50 µg mL− 1 through linear regression analysis (Table 1). LOD was calculated as 0.54 µg mL− 1 which is similar to the other two methods Fig. (S4).
4.2.2 Accuracy
Laboratory prepared mixtures of SMV and SOF at different known concentrations were used to assess the developed methods’ accuracy. For each mixture the new methods were applied, the absorbance values were recorded and employed into the corresponding linear regression equation (Table 1) to calculate the relevant drug’s concentration. The percent of the calculated concentrations to their true known counterparts (% recoveries) were calculated (Table 2) and found to have a mean close to 100% with RSD around 1%, which indicate the method’s high level of accuracy.
4.2.3 Precision
Two levels of precision were assessed: intra- and inter-day. This was achieved by applying the proposed methods in three replicates of three different laboratory prepared mixtures of SMV and SOF. The analysis was performed in the same day at three different times (intra-day) and over three different days (inter-day). The percent recoveries and their RSD were calculated and found to be ranging between 98.5 and 102.8% with RSD almost always ≤ 2% (Table 3) proving the high level of precision of the proposed methods.
4.3 Analysis of the pharmaceutical dosage form:
The methods presented in this work were utilized for quantitation of SMV and SOF in their pharmaceutical formulation (Merospevir® capsules and Sofolanork® tablets) laboratory made mixture. The results obtained were statistically compared to those of reported methods [18] using t- and F-tests. The comparison revealed no significant differences at 95% confidence level (Table 4). Selectivity of the method was examined by studying the effect of the possible interference due to the presence of the common tablet excipients which used as coating and core for tablet such as, titanium dioxide, lactose monohydrate, magnesium stearate, and talc. Different mixtures containing different excipients in ratios similar to those present in the pharmaceutical formulations were prepared and analyzed by the proposed procedure. Results presented in (Table 5) show that the presence of either of these excipients did not significantly the results of the method as the %recovery values are close to 100%.
4.4 Greenness evaluation of the proposed system
Analysts have a lot of responsibility when it comes to protecting the environment and people from harmful chemicals and organic waste that are produced as a result of chemical and pharmaceutical activities [19, 20]. Green chemistry must be created and upgraded on a regular basis. To assess an analytical method's 'ecological worth,' recent considerations such as the analytical eco scale score and the Environmental Quality Methods Index marking have been utilized [21], [22]. In the present work, Eco-Scale Score was utilized to determine the greenness of the proposed system. An analytical eco-scale assessment result is a number that represents a penalty point deducted from a total of 100; it is a result obtained for 'ultimate green analysis.' These points highlight the risks that researchers face during the study process. The greener the analysis, the higher the score (indicated by a large number) [23]. The eco-scale score for the developed technique was 95 because there was no extraction step, no heating, and the energy-consuming procedure was less than 0.1 kW h per sample. Results in Table 6 indicate that the present method was environmentally friendly.