Differential Scanning Calorimetry

DSC thermograms of LV, RVS, PVA and their physical mixture are displayed in Fig. 1. Thermogram of LV presented a broad endothermic peak at about 80 °C and another sharp endothermic peak at about 230 °C corresponding to its melting [34]. RVS presented two peaks observed at 150 and 220 °C due to its semi-crystalline nature [35, 36]. These peaks appeared almost unchanged in the thermogram of the physical mixture. The additional peak at 195 °C could be attributed to PVA [37].

Fig. 1

DSC thermograms of RVS, LV, PVA and their physical mixtures

UV spectrophotometric method

UV spectra of both drugs showed severe overlap; a usual drawback in spectrophotometric analysis (Supplementary Fig. 1). LV spectrum had more sensitive and strong absorption bands which made separation of RVS more difficult. Several methods were tried to overcome this problem.

Derivative ratio method (DR)

This method involves the division of the mixture spectrum by the spectrum of one component (divisor) to create the derivative spectrum of the other component. The calculated derivative spectrum will be independent on the divisor, so it can be determined without interference [38]. The selection of divisors is of particular importance to get low noise and high sensitivity. Different concentrations of RVS (2, 15 and 30 µg/mL) and LV (2, 10 and 25 µg/mL) divisors were tested before developing the method. The concentrations of 30 µg/mL of RVS and 25 µg/mL of LV were the selected divisors regarding low noise and maximum sensitivity.

The derivative orders of the ratio spectra (1st, 2nd, 3rd and 4th) were also tried to attain the best wavelength interval. Besides, different Delta λ values were tested (2, 4, 8 and 16) as they affect the shape, intensity and position of peaks of the analyzed drug. In addition, a scaling factor was attempted. The first derivative ratio (1DR) was the most effective for RVS determination with scaling factor 10 and Δλ = 8 nm, while the third derivative ratio (3DR) was created with scaling factor 100 and Δλ = 16 to measure LV. The wavelength choice was tested as well, where the amplitudes at 250 nm for RVS (Fig. 2a), and 268 nm and 295 nm for LV (Fig. 2b) showed optimum recovery percentages in laboratory-prepared mixture.

Fig. 2

Proposed spectrophotometric methods for simultaneous RVS and LV quantification in phosphate buffer. a First DR spectra of RVS (5–30 µg/mL) using LV 25 µg/mL as a devisor at 250 nm. b Third DR spectra of LV (2–25 µg/mL) using RVS 30 µg/mL as a divisor at 268 and 295 nm. c First D spectra of LV (2–25 µg/mL) at 351 nm in phosphate buffer. d Mean-centered ratio spectra of LV using RVS 30 µg/mL as a divisor at 364 nm

First derivative

The first derivative (1D) spectrum (Fig. 2c) showed zero crossing points of the RVS only at 351 nm at which LV was determined. Different Δλ values were tried (2, 4, 8 and 16), where Δλ = 8 nm and scaling factor = 10 provide maximum peak heights with minimum noise.

Mean centering method (MNCN)

Basically, it is a simple method involving dividing the target analyte by the interfering components, then mean centering the ratio spectra resulted using MATLAB software [39]. The zero-order absorption spectra of LV were divided by RVS (30 µg/mL). The attained ratio spectra were mean-centered where the peak amplitudes were recorded at 346 nm (Fig. 2d).

Validation of method as per ICH recommendation

Linearity: The linearity range was 5–30 µg/mL for RVS and 2–25 µg/mL for LV. Regression parameters were calculated (Table 3).

Table 3 Regression, assay validation and statistical parameters for RVS and LV spectrophotometric determination

Accuracy: The standard addition technique was used to check the accuracy. Both drug concentrations were obtained from the corresponding regression equations, and percentage recoveries revealed good accuracy of the established methods (Table 3).

Precision: It was calculated using 3 concentrations of each of RVS (5, 20, 30 µg/mL) and LV (2, 10, 25 µg/mL). Samples were analyzed three times on the same day (Repeatability) and three successive days (interday) by the proposed methods. Table 3 shows that the values of the RSD are acceptable.

Selectivity: The proposed methods’ selectivity was achieved by the analysis of different laboratory-prepared mixtures of RVS and LV. Simultaneous determination of the two drugs was achieved without any interference (Supplementary Figs. 2 and 3). Thus, the established methods were proved to be selective as indicated by acceptable recovery results ranging from 98.55 to 101.25 (Supplementary Table 1).

Based on the previous findings, the developed methods were successfully carried out for the simultaneous determination of RVS and LV in their new combined topical preparation without preliminary separation. The derivative ratio (DR) method was the only selected technique for monitoring the two studied drugs in the dissolution study because 1DR was the only effective one for RVS determination.

Evaluation of greenness of UV spectrophotometric method

Green analytical chemistry is a concept of promoting the utilization of energy-efficient instruments, reducing the consumption of toxic compounds and minimizing waste production in analytical procedures. Hence, this approach is well-regarded and commonly applied in analytical chemistry laboratories [40].

In this work, we used two common and complementary methods for evaluating the greenness of analytical techniques to assess the proposed spectrophotometric method, namely The Analytical Eco-Scale and GAPI. The Analytical Eco-Scale is a quantitative method that subtracts penalty points out of 100 points. It evaluates various factors, including the quantity and hazard level of reagents, energy consumption, occupational hazards and the amount of waste generated [41]. As shown in Table 4, the proposed technique scored 92 and outperforms the two previously reported spectrophotometric methods [15, 16]. The second method for the assessment of the greenness is GAPI. It is a color code evaluation of each stage. The red color symbolizes indicates high environmental impact. Yellow and green colors represent medium and low impacts [42]. GAPI tool shows that our proposed method has a higher level of greenness in comparison to the two reported methods (Fig. 3).

Table 4 Penalty points of the proposed methods according to the Analytical Eco-ScaleFig. 3

GAPI approach for greenness assessment of: a the developed UV spectrophotometric methods, b the reported UV spectrophotometric method for RVS [15], c the reported UV spectrophotometric method for LV [16]

Tensile strength

The developed dressings had TS values (4.18 ± 0.1–14.1 ± 0.23 MPa) (Table 2). The quadratic model was the best as indicated by the highest regression coefficient value. The model’s significance was assessed by P < 0.001. The quadratic model had a predicted R2 (0.9781) close to the adjusted R2 (0.9952) (Supplementary Tables 2 and 6) indicating a good correlation. The developed equation’s fitting degree is considered fine as indicated by a non-significant lack of Fit; F value of 2.10. Figure 4a–c shows the response surface plots drawn by Design-Expert. The final equation developed based on data analysis for prediction of the response was:

$$\begin }\;} = }.58 + 3.35A - 1.64B + 1.21C - 0.5025AB \hfill \\ \quad \quad \quad \quad \quad \quad + 0.1375AC - 0.4475BC + 1.80A^ + 0.2622B^ - 0.2073C^ \hfill \\ \end$$

Fig. 4

Response surface plot illustrating the effect of variables on TS (MPa), %EB and in vitro drugs release a % PVA and % PG b % PVA and number of FT cycles and c % BG and number of FT cycles on TS (MPa) (R1) of the prepared dressings, d %PVA and %PG e % PVA and number of FT cycles and f % BG and number of FT cycles on EB (R2) of the prepared dressings, g % PVA and % PG h % PVA and number of FT cycles and i % PG and number of FT cycles on LV release (R3) of the prepared dressings, j % PVA and % PG k % PVA and no of FT cycles and l % PG and number of FT cycles on RVS release (R4) of the prepared dressings

Elongation to break (%EB)

The developed dressings %EB values ranged from 44.18 ± 1.0 to 110.05 ± 2.8% (Table 2). The high coefficient value indicates that the quadratic model was the best model with high significance (P < 0.001). Its predicted R2 (0.9022) was close to the adjusted R2 (0.9832) (Supplementary Tables 3 and 6) indicating a good correlation. Non-significant lack of fit F value of 5.76 reveals the fine degree of fitting for the developed equation. Figure 4d–f shows the response surface plots drawn by Design-Expert. The final equation developed was:

$$\begin }\;} = 74.03 + 2.32A + 26.49B - 8.05C - 0.2200AB \hfill \\ \quad \quad \quad \quad - 9.02AC + 0.3700BC - 3.30A^ + 7.41B^ - 3.95C^ \hfill \\ \end$$

In vitro drug release

Results for LV and RVS release (Table 2 and Fig. 5) showed the lowest release (37.81, 26.66% for LV and RVS, respectively) at 12th h for F15, while F17 showed the highest release at the same time (73.31 and 65.36% for LV and RVS, respectively).

Fig. 5

Cumulative LV (––) and RVS (---) release from wound dressings. a F1–F6, b F7–F12, c F13–F17

Box–Behnken statistical analysis confirmed the suitability of the quadratic model as illustrated by regression coefficient value. The model was significant (P < 0.001) (Supplementary Tables 3 and 4). The predicted R2 (0.9286 and 0.9590 for LV and RVS, respectively) were close to adjusted R2 (0.9623, 0.9837 for LV and RVS, respectively) (Supplementary Tables 5 and 6) indicating good correlation. Non-significant lack of fit F values (0.3164 and 0.5452 for LV and RVS, respectively) suggest a fine degree of fitting for the developed equations. Figure 4h–i shows the response surface plots drawn by Design-Expert. The final equations for predicting the responses were:

$$\begin }\;}\;}\;}\;} = 52.85 - 12.44A + 4.37B - 4.54C + 0.2325AB \hfill \\ \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \; + 0.9500AC + 0.3050BC + 0.8392A^ + 1.85B^ + 2.82C^ \hfill \\ \end$$

$$\begin }\;}\;}\;}\;} = 43.36 - 13.26A + 3.59B - 4.53C + 2.10AB \hfill \\ \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \; + 2.02AC - 0.2150BC + 0.4562A^ + 1.94B^ + 1.36C^ \hfill \\ \end$$

Evaluation of wound dressing physical and water-related characters

The formulated wound dressings had average weight and pH values of 320 ± 15–440 ± 33 mg and from 6.2 ± 0.25 to 6.7 ± 0.14, respectively (Table 5). Drug content was in the range of 96.50 ± 0.93–98.45 ± 0.22% for LV and 96.98 ± 0.98–99.21 ± 0.36% for RVS (Table 5). Folding endurance ranged from 264 ± 19 to > 300 (Table 5).

Table 5 Evaluation of wound dressing physical and water-related characteristics

A swelling study showed increasing SI up to the maximum, followed by an equilibrium (no further water absorption) (Supplementary Fig. 4).

WVTR values of the prepared dressings ranged from 1967.39 ± 30.6 to 2629.15 ± 36.5 g/m2/day (Table 5).

Optimization of wound dressing formulations

The mechanical property is important to ensure optimum physical protection of the wound by the applied dressing. Sufficient mechanical properties are also needed to support cellular processes including proliferation, tissue remodeling and angiogenesis. For %EB, dressings also needed to be flexible, soft and elastic to adapt to different body parts. Percent elongation to break was measured as an indicator of the dressing’s extendibility from its initial length up to its break (extendibility) [29, 43]. Hence, optimum values for TS (MPa) and %EB were set to maximize TS and %EB to ensure dressing durability and flexibility in the applied area.

Dressings are generally designed to be replaced every 1–3 days [10, 44]. The current study aimed to ensure dressings provide controlled drug release for 24 h while maintaining dressing integrity. So, an optimum value for LV and RVS release at 12th h was set to 50% to ensure controlled drug release for 24 h.

Optimum formulation with 0.741 desirability was identified using Design-Expert software and was formulated using 8% PVA, and 9%PG and subjected to one FT cycle. Optimized formulation showed predicted values of 9.542 MPa for TS, 113.975 for %EB with about 50.00, 39.84% LV and RVS release, respectively. Experimental values were 9.45 ± 0.67 MPa for TS, 112.6 ± 3.8 for %EB and 52.3 ± 1.4, 38.99 ± 1.6% for LV and RVS release, respectively.

Wound healing evaluation

Time-dependent progression in wound contraction percent was observed in tested groups (Fig. 6a and b). The normal healing process was observed in the control group which exhibited only slight wound contraction percent (11.57 ± 1.13, 25.93 ± 4.54and 56.94 ± 2.32% after 4, 7 and 12 days, respectively). Group 1 (treated with LV-wound dressing) showed wound contraction percentages of 17.12 ± 2.73, 50.46 ± 1.13 and 71.75 ± 3.24% after 4, 7 and 12 days, respectively. For group 2 (RVS wound dressing), contraction percentages were 21.75 ± 3.69, 43.05 ± 3.4 and 68. ± 2.32% after 4, 7 and 12 days, respectively. Group 3 (LV/RVS wound dressing) showed significantly (p < 0.001) higher wound contraction percent compared to other groups (32.4 ± 2.27, 61.5 ± 1.85 and 84.3 ± 1.98% after 4, 7 and 12 days, respectively).

Fig. 6

Wound healing study in rat model. a Percentage wound contraction, b photographs wound healing phases and c histopathology in control group and in groups receiving LV, RVS, LV + RVS wound dressings (inflammation (LI), edema (*) and congestion (**), regenerated epithelium (E), reticular layer (R), scab development (S) and hair follicle (H)

Histopathological examination (Fig. 6c) provides evidence for wound healing progression. Histopathological evaluation of the control group showed that the wound area had significant inflammation (LI), considerable edema (*) and marked congestion (**). The epithelium was only partially regenerated (E) in some wound areas with a partial development of the reticular layer (R). Scab development (S) was highly observed in the wound area.

In animals treated with LV-wound dressing, tissue sections showed a high degree of inflammation (LI), moderate edema (*) and minor congestion. Epithelium was partially restored (E). Animals treated with RVS wound dressing showed wound areas with moderate inflammation (LI), considerable edema (*) and congestion (**) in the newly formed tissues. There were areas for fully regenerated epithelium (E) and some areas of organized reticular layer (R) indicating RVS wound healing ability. In animals treated with RVS/LV-wound dressing, tissue sections showed a lower degree of inflammation (LI) and congestion (**). Granulation tissue formation was observed indicating a successful re-epithelialization process. The epithelium was well developed with some areas containing hair follicles (H).