Green DOE based RP-HPLC method for the simultaneous determination of Azelastine and Losartan in spiked human plasma samples

TL;DR

A green HPLC method was developed to simultaneously determine Azelastine and Losartan in human plasma samples.

Contribution

A novel green RP-HPLC method using DOE optimization for drug analysis in plasma is introduced.

Findings

The method achieved rapid and accurate quantification of Azelastine and Losartan in spiked plasma.

The developed approach was validated for linearity and precision according to ICH guidelines.

Greenness assessment confirmed the environmentally friendly nature of the method.

Abstract

A green RP-HPLC approach coupled with spectrofluorometric detection was established for the simultaneous estimation of the co-administered drugs Azelastine (AZL) and Losartan (LOS) in spiked human plasma samples. Preliminary trials were carried out for determination of the Critical Method Parameters (CMPs) and Critical Quality Attributes (CQAs). Design of Experiment (DOE) was developed relying on the use of Central Composite Design (CCD) for the optimization of conditions to establish a simple, rapid, cost-effective and environmentally benign approach. The chromatographic separation was based on using a mobile phase composed of methanol: acetonitrile: 0.02M phosphate buffer of pH 3.25 with 0.05% tri-ethylamine (60: 1.5: 38.5, v/v/v) at a flow rate of 1.2 mL/min and injection volume of 10µL. The fluorescence detection was carried out at 245 nm/400 nm and 290 nm/360 nm for estimation of LOS and AZL, respectively. The retention times of AZL and LOS were 8.9 ± 0.2 min and 13.4 ± 0.1 min, respectively. The developed method was validated according to ICH Guidelines with linear relationship in concentration ranges of 0.10–50.0µg/mL for AZL and 0.30–40.0µg/mL for LOS, while in spiked plasma samples the concentration ranges were 0.025–0.07µg/mL for AZL and 0.025–0.10 µg/mL for LOS. The developed approach was successfully applied for the quantitation of both drugs in spiked human plasma samples. Complex MoGAPI and AGREE methodologies were used for greenness assessment.

The online version contains supplementary material available at 10.1038/s41598-026-39426-2.

Figures

Typical chromatogram for the separation of AZL (7.0µg/mL) and LOS (5.0µg/mL) under the described chromatographic conditions.

Solution ramps for optimal levels of each CMP using Derringer Desirability Function.

Chromatograms showing the separation of a: AZL(0.07µg/mL) and LOS(0.10µg/mL) in spiked human plasma, b: blank plasma sample and c: diluent.

Results for greenness assessment using Complex MoGAPI (a) AGREE (b) methods for the proposed method.

Full text

Introduction

Design of Experiments (DOE) employs multivariate statistical techniques that offer several advantages over univariate ones. It reduces the total number of experiments required for analytical method development. It also minimizes analysis time and reagents consumption giving a more environmentally benign analytical method. DOE also, enables the development of mathematical models to be used for assessment of the different effects between several factors in order to identify the crucial few ones^1^. The present DOE was used to develop a green, sensitive and time saving RP-HPLC method based on determining the Critical Method Parameters (CMPs) that affect the Critical Quality Attributes (CQAs)^2^.

Chronic low-grade inflammation, endothelial dysfunction, and platelet activation are major contributors to the pathogenesis of cardio metabolic and allergic diseases. Current therapies typically target individual aspects of these processes, yet evidence suggests that multi-target strategies may be more effective in modifying diseases progression. Histamine, a product originated from mast cells, can play several physiological functions in the pathophysiology of atherosclerosis^3^. Serum levels of histamine usually show significant increase in case of atherosclerosis, diabetes and hypertension diseases.

Azelastine (AZL) shown in Figure S1a^4^, is an anti-allergic mast cell stabilizer acting by preventing mast cells from histamine releasing, that alleviates conjunctivitis caused by allergies^5^. Numerous analytical approaches were reported for AZL quantitation including spectrophotometry^6–11^, spectrofluorimetry^12,13^, LC-ESI/MS/MS^14^ and RP-HPLC^15–20^ methods.

The platelet activation and endothelial dysfunction are the main risk factors for atherosclerosis and its consequent complications^21^. Losartan (LOS) shown in Figure S1b^4^, is an angiotensin II type 1 receptor blocker. It -not only lowers blood pressure-but also reduces oxidative stress, vascular inflammations, and fibrosis through inhibition of renin–angiotensin signaling^22^. ARB members act by blocking angiotensin II type 1 (AT1) receptors; thus, improve endothelial dysfunction and prevent platelets activation, adhesion and aggregation exerted by nitric oxide^23^. LOS is used for treatment of high blood pressure and heart failure^24^. Numerous analytical methods were reported for its quantitation , via spectrofluorimetric^25–28^ spectrophotometric^29–31^ and RP-HPLC^32–35^ methods.

The combination of AZL and LOS produced superior improvement in endothelial functions, platelets activation, biochemical parameters and reduction in serum glucose levels compared to drug alone, suggesting synergistic benefits^21,36,37^. Given their favorable safety profiles and absence of major reported interactions, investigating this combination in humans represents a promising repurposing strategy to address overlapping inflammatory and vascular dysfunction in cardio metabolic and allergic disease^21^. This fact supports the rationale for further investigation of this combination as a novel therapeutic strategy for conditions driven by metabolic, inflammatory, and vascular abnormalities. Therefore, this study aims to establish the first DOE RP-HPLC approach for the assay of the frequently co-administrated drugs, namely AZL and LOS in spiked human plasma sample sand for their therapeutic drug monitoring for in-patients.

Materials and methods

Materials and reagents

AZL (99.30% pure) and LOS (99.50% pure) were obtained as gifts from Sigma Pharmaceutical Company (Quesna, Egypt). Human plasma samples were purchased from Blood Bank Center at Mansoura University Hospital, Mansoura, Egypt. HPLC grade acetonitrile (ACN) and methanol (MeOH) were obtained from Fisher, UK. Analytical grade orthophosphoric acid and potassium dihydrogen phosphate were obtained from Sigma-Aldrich, Germany. Tri-ethylamine of HPLC grade was obtained from Oxford Laboratory, UK.

Apparatus and software

The analysis was performed using Shimadzu Prominence-i series LC-2030C 3D Plus system (Shimadzu, Kyoto, Japan) equipped with a quaternary RS pump, an RS auto sampler injector, a thermostated RS column compartment, and Prominence fluorescence detector (RF-20A). Chromatographic separation was achieved using Knauer ODS column (150 × 4.6 mm, 5 µm). The data acquisition was carried out using Lab solutions software (Shimadzu, Japan). Cooling centrifuge (Sigma, Germany), vortex mixer (VELP, Scientifica, China), pH-Meter 3510(Jenway, UK) and Vacuum pump Rocker 811 Lab (Lingya Dist. Kaohsiung City ,Taiwan) were also used. Nylon membrane filter 0.22 μm (Millipore, Ireland) was used for the filtration of buffer solution before use.

The software used for Design of Experiments (DOE) was Design-Expert version11.Analytical GREEnness Metric (AGREE) tool is freely available and downloadable from the open-source https://mostwiedzy.pl/AGREEBeta version of software was applied for the greenness assessment. The software used for the Modified Green Analytical Procedure Index (MoGAPI) method is also available freely at https://fotouhmansour.github.io/MoGAPI/ that enables applications.

Chromatographic conditions

Separation was performed using a mobile phase consisted of MeOH:ACN:0.02M potassium dihydrogen phosphate containing 0.05% tri-ethylamine (60:1.5:38.5, v/v/v). The buffer pH was adjusted to 3.25 using ortho-phosphoric acid. The fluorescence detection was accomplished at 245/400nm and 290/360nm for LOS and AZL, respectively. 10µL injection volume and 1.2 mL/min flow rate were adopted all over the runs.

Preparation of standard solutions

For preparation of 1000.0µg/mL stock solutions of each of AZL and LOS,100.0 mg of each drug was weighed and transferred into two 100-mL volumetric flasks then dissolved using 50 mL of MeOH. The resulting solutions were diluted and completed to the volume using the same solvent. The resulting stock solutions were stored in refrigerator at 4 °C. For working standard solutions, from each stock solution 5.0 mL were transferred into two separate 50-mL volumetric flasks and completed with the mobile phase to obtain 100.0µg/mL working solutions.

Construction of calibration curves

Increasing aliquot volumes were transferred from each working standard solution into separate series of 10-mL volumetric flasks then completed to the mark using the mobile phase to attain solutions of concentration ranges of 0.10–50.0µg/mL for AZL and 0.30–40.0µg/mL for LOS. 10µL Aliquots were injected from each solution under the described chromatographic conditions. The average peak areas were plotted against their corresponding concentrations in order to generate calibration curves and the regression equations were then derived.

Calibration in spiked human plasma

The frozen human plasma samples were left to thaw and equilibrate to room temperature before analysis. Then at 2000 rpm, Multipulse vortex was used for 30s. to vortex the thawed plasma samples to ensure uniformity of the contents. Various aliquots from working standard solutions (25.0µg/mL) of each of drug with an aliquot of 200µL of blank plasma sample were transferred into different centrifuge tubes. The obtained solutions were diluted to4mL with methanol and vortexed twice at 2000 rpm for 30 s. for well mixing and to confirm the protein precipitation. The resulting solutions were centrifuged for 30min at 6000rpm. Then 1mL aliquots were taken from each supernatant into a series of 5-mL volumetric flasks and the solutions were completed with mobile phase. The solutions were filtered using cellulose acetate syringe filter (0.22 μm) then 10µL aliquots were injected from each solution under the previously mentioned chromatographic conditions. The average peak area values were plotted against the corresponding concentrations of each drug for construction of calibration graphs and the corresponding regression equations were then derived.

Preliminary studies and method optimization with central composite design (CCD)

Preliminary studies were performed to select the most important parameters (CMPs) that have the most obvious effects on the RP-HPLC method performance or (CQAs) and that will be involved in the optimization step. Many factors such as: column length, column temperature, flow rate, injection volume, type of buffer, type of organic solvent, detector, buffer concentration, percentage of organic solvent and buffer pH, were considered. Only three parameters were considered as the most significant factors or (CMPs) that are affecting the (CQAs) or method performance (%MeOH: percentage of MeOH, %ACN: percentage of ACN and pH of buffer). The CMPs and CQAs were determined from the preliminary trials. During the optimization, the CQAs values were used to choose the most optimum separation parameters by mathematical technique based on Derringer’s Desirability Algorithm, while the non-significant parameters will be ignored and kept constant. Then further optimization was achieved by means of response surface methodology for identification of the optimum levels of each CMP.

The most significant three CMPs (%MeOH, % ACN and buffer pH) were, moreover, optimized by means of Central Composite Design (CCD) with three-levels to determine the optimum level of each CMP. The design composed of total of 20 experimental runs: 14 non-center points (8 factorial runs + 6 axial points with α = 1.68) and 6 center points to consider the experimental errors. The optimization procedure based on five CQAs named: AZL tailing factor (t_f_-AZL), resolution (Rs), LOS tailing factor (t_f_-LOS), LOS AUC and run time.

Results and discussion

An HPLC approach was established for the separation and quantitation of LOS and AZL in human plasma samples upon co-administration, based on the use of DOE. Figure 1 shows typical chromatogram for the separation of AZL and LOS under the optimized chromatographic conditions.Fig. 1. Typical chromatogram for the separation of AZL (7.0µg/mL) and LOS (5.0µg/mL) under the described chromatographic conditions.

Preliminary studies

Preliminary studies based on scientific knowledge were conducted through several trials using different organic modifiers, mobile phase compositions, columns, flow rates, and aqueous phases. However, some problems were encountered, including inadequate resolution between the two drugs and pronounced peak tailing for each. Tailing factor and resolution are mainly affected by buffer pH and percent of organic modifiers. Depending on the results of preliminary studies the most important parameters used were: CMPs for the optimization (A: %MeOH, B: %ACN and C: buffer pH). Five responses or (CQAs): resolution (Rs), AZL tailing factor (t_f_-AZL), LOS tailing factor (t_f_-LOS), LOS AUC and run time. However, other parameters such as room temperature, buffer concentration (0.02M) and flow rate (1.2 mL/min) were hold constant .The temperature of the column was the least significant factor and had no added effect on the chromatographic separation. The need to carry out an optimization strategy appeared due to the variability of factor settings required to improve each measured response separately that suggested non-linearity, which is best explained by three level response surface optimization designs.

Optimization using response surface design

Response surface methodology was used for optimization of CMPs that resulted from the preliminary studies in order to identify the optimal levels of each CMP taking potential quadratic effects into account. Central Composite design (CCD) was carefully chosen to assess the effects of the three CMPs (%MeOH, % ACN and buffer pH). The design composed of 20 experimental runs: 6 center points to consider the experimental errors and 14 non-center points (8 factorial runs + 6 axial points with α = 1.68). These experimental runs assisted in developing a model which might optimize the chosen CMPs based on the results of five CQAs namely: resolution (Rs), AZL tailing factor (t_f_-AZL), LOS tailing factor (t_f_-LOS), LOS AUC and run time (Table S1). Examining the 3D response surfaces (Fig. 2) and model coefficients (Table 1) that were obtained, all the models were quadratic and all models showed high R^2^ and adjusted R^2^ values more than 0.9 as displayed in Table S1. Also, values of insignificant Lack-of-fit relative to pure error and higher-order terms (x^2^) showed that, the variables behaved non-linearly, where all confirm good model fitting.

• Analysis of the optimization experimental outcomes: Fig. 2(a-e): Response surfaces CCD for factor interaction for optimization of the proposed DOE method; a: resolution between AZL & LOS peaks, b: AZL-t_f_, c: LOS-t_f_, d: LOS- AUC and e: run time.Table 1. Coefficients and ANOVA statistical analysis for the three studied CMPs of the optimization design.A(%MeOH)B(%ACN)C(pH of buffer)ABACBCA^2^B^2^C^2^Rs**-0.764****-0.4791.177**-0.070**1.9770.8380.332**-0.0040.797p-values**< 0.0001< 0.0001****< 0.00010.104< 0.0001****< 0.0001****< 0.00010.877< 0.0001T_f_ AZ-0.0250.2150.102****-0.0510.000-0.1630.0290.1550.081p-values < 0.0001 < 0.0001**** < 0.0001**** < 0.00010.930 < 0.0001**** < 0.0001**** < 0.0001**** < 0.0001T_f_ LOS-0.019**-0.0020.049****-0.0360.0040.042****-0.081****-0.118****-0.068p-values0.0040.580** < 0.00010.0010.4680.000 < 0.0001**** < 0.0001**** < 0.0001Area (LOS)-1,163,630****-1,104,380****-1,699,76059,2551,110,0201,105,980-764,638****-603,926****-174,454p-values < 0.0001**** < 0.0001**** < 0.00010.181 < 0.0001**** < 0.0001**** < 0.0001**** < 0.00010.000run time-4.157-0.5721.5410.9240.0760.9241.406**-0.1850.718p-values** < 0.00010.001**** < 0.00010.0010.6570.001 < 0.00010.1070.001**

The procedures that were followed to identify the effect of each CMP on each response (CQA) include:

1. Examination of fitting statistics (R^2^andadjusted R^2^).
2. ANOVA interpretation and examination of factor significance and model residuals.
3. Interpretation of prediction equation coefficients relying on magnitude and sign.
4. Examination of the integrity of ANOVA diagnostics plots.
5. Characterization of significant factor performance relying on the obtained model graphs.

The confirmation was based on coefficients’ magnitude and sign from prediction equation and p-value magnitude from ANOVA results. Each parameter coefficients with positive sign showed that the response was positively affected by the parameter, while a negative sign indicated that the parameter and the response had an inverse relationship. Table S1 shows the 20 experiments and their experimental results. Interaction plots were examined as very helpful tools for additional characterization of CMPs impacts on each CQA to determine the critical parameters and chose the optimum level of each. The following parameters were determined:

• Resolution between AZL & LOS peaks (Rs):

Resolution factor was influenced by three CMPs (A, B&C); buffer pH (C) had the most significant and important impact while ACN% (B) had the least impact. Increasing factor (C) increased (Rs),while, increasing factors (A) and (B) decreased (Rs) as shown in coefficient table (Table 1) and Perturbation plot (Figure S2, a).The Rs value of 4.5 could be obtained using 2.42% ACN, 57.45% MeOH and at pH 3.5 as shown in 3D Surface plot (Fig. 2, a). There was a factor-to-factor interaction between factors (A&C) positively affected (Rs) as shown in (Table 1) and Interaction Plot (Figure S3, a).

• Tailing factor (AZL-t_f_)

The tailing factor of AZL (AZL-t_f_) was strongly affected by ACN % (B) and buffer pH (C) with positive effect, MeOH% (A) slightly affected (AZL-t_f_) with negative effect as shown in coefficient table (Table 1) and Perturbation plot (Figure S2, b).Therefore, decreasing buffer pH (C) to 3.25 and 2% ACN led to decrease in (AZL-tf) value to be 1.4 as shown in 3D Surface Plot (Fig. 2 b). There was factors-interaction between (B&C) with negative effect as shown in Interaction plot (Figure S3, b).

• ooTailing factor (LOS-t_f_)

The tailing factor of LOS (LOS-t_f_) was mostly affected by negative nonlinear quadratic effect of ACN % (B^2^) and MeOH% (A^2^) as shown in coefficient table (Table 1) and Perturbation plot (Figure S2, c). There was factors-interaction between (B&C) positively affected (LOS-t_f_) and between (A&B) negatively affected (LOS-t_f_) as shown in Table 1. LOS-t_f_ value of 1.4 could be obtained using pH 3.5, 4.3% ACN and 64.7% MeOH as shown in3D Surface Plot (Fig. 2, c).

• Area 2 (LOS- AUC):

Buffer pH factor (C) had the most important and significant effect on (LOS- AUC) as shown in coefficient table (Table 1) and Perturbation plot (Figure S2, d). The LOS- AUC value of 9.276*10^6^was obtained using 2% ACN, 56.077% MeOH at pH 3.5 as shown in 3D surface plot (Fig. 2, d). There was factor-to-factor interaction between (A&C) and (B&C) positively affected the (LOS-AUC) as shown in (Table 1) and in Interaction plot (Figure S3, c & d).

• Run time:

Run time was affected mostly by %MeOH (A) with negative effect. %ACN (B) and Buffer pH (C) had positive effect on run time as shown in coefficient table (Table 1) and Perturbation plot (Figure S2, e). There was factor-to-factor interaction between (A&B) and (B&C) positively affected run time as shown in (Table 1) and in interaction plot (Figure S3, e & f). The run time can be obtained 13 min using 2% ACN, 60.05% MeOH at pH 4.9 as shown 3D surface plot (Fig. 2, e).

Optimization criteria aid to select the optimum level of each CMP in order to enhance the method performance and efficiency, the next criteria were described:

1. To maximize R_s_ between (AZL and LOS) in range (3–4).
2. To minimize T_f_ values of both (AZL and LOS).
3. To minimize run time in range (10–14 min).
4. To maximize the Area of LOS.

Using Derringer’s Desirability Algorithm,38 solutions were obtained, the proposed optimum level of each CMP from the most suitable solution were(A) 60% MeOH%, (B) 1.5% ACN% and (C) 3.25 buffer pH as shown in Fig. 3.Fig. 3. Solution ramps for optimal levels of each CMP using Derringer Desirability Function.

Finally, the optimum mobile phase was composed of (60:1.5:38.5) % MeOH: %ACN:0.02M Phosphate buffer (with 0.05 v/v % tri-ethylamine) and buffer pH of 3.25,with predicted attribute values of Rs value of 4.48, AZL t_f_ (1.41), LOS t_f _ (1.44), run time (13.99 min.) and LOS-AUC (6.96910^6^) as shown in Fig. 3, with a desirability value of 0.954.These proposed optimum separation chromatographic levels were tested and verified three times, and the calculated mean values for each response were Rs (4.45), AZL t_f_ (1.42), LOS t_f _(1.45), run time (13.99 min.) and LOS AUC (6.92210^6^). To demonstrate the model predictability, the observed and predicted values were compared and all results were satisfactory with low prediction error.

Method validation

The established DOE approach was validated according to ICH guidelines^38^. System suitability parameters values were presented in Table S2 at optimum chromatographic conditions.

Linearity and range

The developed RP-HPLC approach was linear covering ranges of 0.1-50µg/mL for AZL and 0.3-40µg/mL for LOS in pure form, while in spiked plasma samples ,it was0.025–0.07µg/mL for AZL and 0.025–0.1 µg/mL for LOS as shown in Table2 and Table 3.Table 2. Results of regression parameters of the proposed method.DrugAZLLOSConcentration range (µg/mL)0.10–50.00.30–40.0r0.99990.9999a46,5317153b397,898102,390S_a_34783076S_b_195.4198.8S_(y/x)89407911LOD (µg/mL)0.0290.099LOQ (µg/mL)0.0870.30a, intercept; b, slope; r, correlation coefficient.S_a, standard deviation of intercept; S_b_, standard deviation of slopeS_y/x_, residual standard deviation.LOD, limit of detection; LOQ, limit of quantitation.Table 3. Results of analysis of AZL and LOS in spiked human plasma.Taken Conc. (µg/ mL)AZLLOSAZLLOSObtained Conc(µg/ mL)Recovery %Obtained Conc(µg/ mL)Recovery %0.0250.0250.02598.8860.02599.4050.050.070.050100.6060.06998.7030.070.100.071101.7320.101101.2480.100.300.09999.1710.30099.936Mean %100.09999.823S.D1.3241.076% RSD1.3221.078S.D: standard deviation, % RSD: percent relative standard deviation.

Limits of detection and quantitation

The LOD and LOQ were determined according to Eq. (1) and Eq. (2).The results were shown in (Table2).

\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{LOQ}\hspace{0.17em}=\hspace{0.17em}10\text{ Sa}/\mathrm{b}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{LOD}\hspace{0.17em}=\hspace{0.17em}3.3\text{ Sa }/\mathrm{b}$$\end{document}

Where b is the slope of calibration curve and S_a_ is the standard deviation of y-intercept of regression lines.

Accuracy

Accuracy of the developed method was evaluated by determining the mean %recoveries for both drugs using different three concentration levels, for AZL :7.5, 15.0 and 45.0µg/mL and for LOS :5.0, 10.0 and 30.0µg/mL as shown in Table S3.

Precision

The precision was evaluated using three concentration levels and three replicate analyses of both drugs on the same day for intraday precision and at different three successive days for interday precision. The calculated values of S.D and % RSD were less than 2% as presented in Table S4 indicating the good precision of developed approach.

Robustness

The performance of the developed approach remained unaffected by the deliberate, but minor changes of various experimental factors, confirming its robustness. The S.D, %RSD values and %recoveries are shown in Table S5 and all values were satisfactory.

Selectivity

The optimized method’s selectivity was found to be unaffected by the presence of biological matrices of plasma as shown in (Fig. 4, a &b) and unaffected by the diluents used throughout the procedure (Fig. 4 c).Fig. 4. Chromatograms showing the separation of a: AZL(0.07µg/mL) and LOS(0.10µg/mL) in spiked human plasma, b: blank plasma sample and c: diluent.

Results of analysis in spiked plasma samples

As displayed in (Fig. 4, a) the developed approach was successfully used for the simultaneous determination of AZL and LOS in spiked plasma samples and the calibration curves covered concentration ranges of 0.025–0.07µg/mL for AZL and 0.025–0.10 µg/mL for LOS as shown in Table 3.

Greenness evaluation method

Complex Modified Green Analytical Procedure Index (Complex MoGAPI) and Analytical GREEnness (AGREE) methods were applied to assess the greenness of proposed method^39–41^. While the (Complex MoGAPI)^42^ based on a pictogram estimation of greenness of many analytical stages, with stress on energy requirements and reagent toxicity, the AGREE approach provides a comprehensive framework for assessing sustainability from a variety of perspectives, including as economic and safety factors^43^. The 12 principles of green analytical chemistry were assessed using the AGREE software, producing a circular pictogram that identifies areas that are green and those that could want improvement^43^. (Fig. 5 a & b) display the results from the two evaluation methods for the developed approach. In comparison with conventionally developed RP-HPLC methods, the DOE-assisted HPLC approach is greener due to fewer optimization experiments, reduced solvent consumption, and lower waste generation, which is reflected in its improved AGREE and Complex MoGAPI greenness scores.Fig. 5. Results for greenness assessment using Complex MoGAPI (a) AGREE (b) methods for the proposed method.

Green methods are not necessarily practical, so the practicality of the established approach was assessed applying Blue Applicability Grade Index (BAGI)^44^ and Click Analytical Chemistry Index (CACI)^45^ methodologies as displayed in (Figure S4 a & b).The developed approach had low cost and complexity using CACI^45^ method and had high greenness and bench suitability regarding to BAGI^44^.

Conclusion

A highly sensitive, fast and green DOE RP-HPLC approach was developed and optimized based on DOE for the simultaneous separation of AZL and LOS and quantification in spiked plasma samples as co-administered drugs. The proposed DOERP-HPLC method is qualified as candidate for bioanalytical analyses. It can be adopted for therapeutic drug monitoring for in-patients. AGREE and Complex MoGAPI methodologies were used for evaluation of developed approach greenness.

Supplementary Information

Supplementary Information.