Development of foaming vaginal tablets with clotrimazole. Part I. Optimization of the formulation using design of experiments approach

Objectives. The aim of this study was to optimize the formulation of certain immediate release foaming vaginal tablets, with fast vaginal disintegration (part I), fallowing their time stability (part II). Material and methods. Modde optimizing software and a Box-Bhenken experimental design was used to establish an optimal formula that would provide maximum “in vitro” behaviour advantages, including stability during preservation; and the improvement of the foaming effect. There were prepared 15 formulations of foaming vaginal tablets in order to choose the optimal formula. Outcomes. The main factor, with influence over all the studied parameters, is the percentage of the effervescent mixture in the formulation. The excess of citric acid has a main role in the assurance of local pH; in the improvement of the flowing properties of the powder mixtures before compression; and in the improvement of the effervescent process, having, on the other hand, an effect of decreasing the stability of tablets and increasing their humidity adsorption. Sodium lauryl sulphate, at the chosen concentration levels (0.5-1.5%), did not influence the dependent variables, it only produced a strong-er resistance of the foam. The optimal formula of a foaming vaginal tablet with an optimal stability during preservation must contain 0.5% sodium lauryl sulphate, while the effervescent mixture must contain a percentage of 24% excess of citric acid, and it must represent 44% of the total amount of excipients. Conclusions. The experimental determinations of the optimal formula were close to the theoretical values predicted by the program, this certifying the validity of the optimization and the conclusion draw after analysis the experimental design.


INTRODUCTION
Foaming vaginal pharmaceutical dosage forms allow the uniform spread of the doses on the whole mucosal surface of the vagina, and assures a rapid disintegration of the pharmaceutical form, associated by a lack of flowing from the site of application (1,2). Being conceived for a mucosal route, these tablets must carry out some standards (2,3) concerning especially their form which has to be conformed to an easily applicable device without producing vaginal lesions; and concerning their dissolution capacity, which has to assure the dissolution in a very low liquid volume (~2 ml), associated with a diffusion and a spread between all the vaginal mucosal velocities. Another problem is how to assure their stability in time.
The aim of this study was to optimize the formulation of immediate release foaming vaginal tablet formulation with fast vaginal disintegration using a Box-Bhenken experimental design and Modde optimizing software. There were analysed 15 responses, concerning the properties of the powder mixtures, the properties of the tablets, considering that the studied responses are the most important characteristics of a foaming vaginal tablet.

Software and experimental design
To perform the study, a Box-Bhenken experimental design 3 3 with three factors and three levels was used. Experimental design, coefficient calculation, statistic parameters calculation and evaluation of the quality of fitting were performed with Modde 12.0 (Sartorius Stedim Data Analytics AB, Sweden). The independent variables (formulation factors considered the most influents in the case of foaming vaginal tablets) and the levels of variation are presented in table 1.
The dependent variables are presented in table 2.

Methods
Formulation and preparation of the foaming vaginal tablets After consulting the literature (16)(17)(18)(19)(20) we used the following ingredients to prepare foaming vaginal tablets: effervescent mixture (15-65%) with alkaline component sodium hydrogen carbonate associated with anhydrous sodium carbonate (10% from the alkaline amount for a better stability) and as acid component citric acid (in excess to obtain an optimum pH -between 4.2 and 6 after foaming), clotrimazole 10% for a 100 mg dose in an 1g weight tablet; foaming agent sodium laurylsulphate (0.5-1.5%); Ludipress as diluent; Methocel E15 2% as binder, magnesium stearate 1% as lubricant and Aerosil 1% as foam resistance enhancer.
The foaming vaginal tablets were obtained by direct compression in normal thermo-hygrometrical conditions, using a Korsch EK 0 excentric press, equipped with droplet shaped die set punch. The powders were individually dried until a 0.5% humidity before    Flowing time of the powder mixtures (min) Y 3 4 Mechanical resistance of the tablets (kg) Y4 5

RESULTS AND DISCUSSIONS
Friability of the tablets (%) Y 5 6 Foaming time (min) Y 6 7 Maximal volume of the foam (ml) Y 7 8 Resistance of the foam (min) Y 8 9 Weight of CO2 loosed by 20 minute foaming (g) Y 9 10 pH generated in 30 ml distilled water Y 10 11 pH generated in 30 ml (FVS) pH 4.2 Y 11 12 pH generated in 30 ml acetate buffer, pH 5.5 Y 12  (2). The volume and resistance of the foam are responses which shoes a great dependence on the formulation factors (independent variables).

TABLE 4. Matrix of the responses obtained for the determinations immediate after preparation
As we can see, there are little differences between the humidity of powder mixtures before compression and the humidity of the tablets resulted, but the initial humidity of the powders was in all cases below 0,5% after drying. The conclusion is that during the compression process a humidity absorption phenomenon takes place. The results showed that the amount of humidity absorbed is independent on the formulation and depends on the environmental termo-hygrometrical conditions (22).

The influence of formulation factors on powder flow properties
Graphs representing the influence of formulation factors on Carr's (Y 1 ) are illustrated in Fig. 2.a. Carr's index increases with increasing the percentage of effervescent mixture (X 2 ) and decreases with the increase of the excess citric acid (X 3 ), being independent of the percentage of sodium lauryl sulphate. Graphs representing the influence of formulation factors on Hausner ratio (Y 2 ) are illustrated in Fig. 2.b. Hausner ratio increases with increasing the percentage of effervescent mixture (X 2 ) and decreases with the increase of the excess citric acid (X 3 ), being independent of the percentage of sodium lauryl sulphate, the results being consistent with those obtained for Carr's index. Graphs representing the influence of formulation factors on flowing time of powders are illustrated in Figure  2.c. Flowing time increases with increasing the percentage of effervescent mixture (X 2 ), so with The results obtained for the study of the flowing time show very good flow for all formulations studied (Hausner ratio between 11.11 and 21.21; Carr's index between 1.11 and 1.26), and in terms of its variation depending on formulation factors, the results show that the effervescent mixture makes the flow difficult, probably due to interactions between mixture components, which can lead to an enhancement of powder humidity during processing. Citric acid has a very good flow, so the excess of citric acid facilitates the flow. So, a good flow is achieved with a small amount of effervescent mixture (which corresponds to a higher ratio of diluent Ludipress), and a higher percentage of excess citric acid.

The influence of formulation factors on tablets properties immediate after preparation
Graphs representing the influence of formulation factors on mechanical resistance of the tablets are illustrated in Fig. 3a. The results show that the mechanical resistance of effervescent tablets does not depend on the formulation variables chosen. This is understandable because the mechanical resistance depends mainly on the compression force, which has changed from one formulation to another. However, it was not introduced in the experimental design, because it was not monitored, So, in this study the mechanical resistance and the compression force were variables that the experimental design was unable to control. However, there is a tendency of mechanical resistance decrease with increasing the percentage of all the variables chosen, so with the decrease of Ludipress percentage, and the graph has a minimum at 65% effervescent mixture and 1.5% sodium lauryl sulphate. We can thus draw the conclusion that an increase of Ludipress percentage may increase mechanical resistance. This result is understandable, given the good compressibility and compactibility properties of this excipient (16).
Graphs representing the influence of formulation factors on the tablets friability (Y 5 ) are illustrated in Fig. 3.b. Although for friability the results are poorly fit with the model and there is a great variability of data (from improper values 5-6% to rather small values 0.45%), the friability depending generally on other factors than those chosen for the study, yet it may be observed that friability decreases with increasing the percentage of effervescent mixture (X 2 ), the chart registering a minimum at 65% effervescent mixture. Also, at 25% effervescent mixture, most formulations were inadequate in terms of friability, although they had a high mechanical resistance (from 9.5 to 14). These results suggest that the effervescent mixture reduces the friability, probably by the interactions of its components, and that there is probably an optimal ratio between the diluent and the effervescent mixture that allows these interactions and thereby obtaining tablets with appropriate friability, but this phenomenon may have negative effects on foaming and stability in time.
Graphs representing the influence of formulation factors on foaming time (Y 6 ) are illustrated in Fig. 3.c. Foaming time increases with increasing the percentage of sodium lauryl sulphate (X 1 ) and decreases with increasing the percentage of effervescent mixture (X 2 ), but for this formulation variable (X 2 ), a nonlinear effect is observed. Regarding the concentration of sodium lauryl sulphate, some data in the literature state that the foaming slows down the effervescence speed, due to changes in fluid properties in which the disintegration takes place (16,17,18), so the higher the concentration of sodium lauryl sulphate, the higher the duration of disaggregation time will be, but this effect may be due to the fact that with increasing the percentage of sodium lauryl sulphate a higher compression force is needed, leading to tablets that have a slow disaggregation through foaming. Regarding X 2 variable (percentage of effervescent mixture), its growth decreases foaming time, because if the amount of effervescent mixture is greater, the decomposition will be faster, with a broader interaction between the effervescent mixture components, a more abundant effervescence and a faster consumption of these components. For values greater than 45% effervescent mixture (value for which the chart has a minimum), the foaming time does not decrease with increasing the percentage of effervescent mixture, probably because over this threshold value the amount of effervescent mixture is so great that time between decomposition and consumption of the effervescent reaction is longer than for the concentration of 45%, and in this case we must also take into account the dissolution time of the effervescent mixture components in the small volume of vaginal fluid. Another explanation is that a high percentage of effervescent mixture needs a too high compression force, leading to an interaction between components during compression and a strengthening of the tablet core, slowing down the final stage of disaggregation (1,17,19).
Graphs representing the influence of formulation factors on maximal volume of foam (Y 7 ) are illustrated in Fig. 3.d. From the graphical representations it may be concluded that the volume of the foam formed in the disaggregation process depends mostly on the percentage of effervescent mixture (X 2 ) and increases with its amount increase. The volume of foam depends very little on the excess citric acid (X 3 ), decreasing with its increase and does not depend on the percentage of sodium lauryl sulphate (X 1 ). This behaviour may be explained by the fact that to achieve the maximal volume of foam, several competing processes take place simultaneously: foam formation, foam maintenance and foam breaking. Depending on the speed of these processes, the final volume of foam will result. If the foam formation speed and its maintenance is greater than its

FIGURE 3. The influence of formulation factors on tablets properties
a -mechanical resistance, Y4; b -tablets friability, Y5; c -flowing time, Y6; d -foam volume, Y7 e -foam resistance, Y8; f -amount of CO2 released, Y9; g -pH generated in water, Y10; h -pH generated in SVF, Y11; i -pH generated in acetate buffer, Y12 X1 -percentage of sodium lauryl sulphate, X2 -percentage of effervescent mixture, X3 -percentage of excess of citric acid in the effervescent mixture breaking, the volume increases; if the foam formation speed is smaller than the foam breaking speed, the foam does not increase in time, but remains at a minimum volume, or does not form at all. The results suggest that when the foam formation process is predominant, it depends only on the effervescent mixture. Perhaps increasing the percentage of effervescent mixture increases formation speed and maintenance of foam; foam breaking has a constant speed that depends on the studied environment (SVF) and is independent of the percentage of effervescent mixture and the percentage of sodium lauryl sulphate. We can say that there is no need to increase the percentage of sodium lauryl sulphate to increase the foam volume; there is probably a threshold concentration of sodium lauryl sulphate above which the foam breaking has a constant speed, and this concentration is below the minimum concentration used in the experience design, taking into consideration the small volume of the environment used for the experiment.
Graphs representing the influence of formulation factors on foam resistance (Y 8 ) are illustrated in Fig.  3.e. Although for foam volume we obtained an independent behaviour of the sodium lauryl sulphate percentage, foam resistance increases with the increase of sodium lauryl sulphate percentage (X 1 ), but also with the increase of effervescent mixture percentage (X 2 ); for the latter, the dependence is nonlinear. These results may be explained by the fact that a higher percentage of sodium lauryl sulphate produces a higher density foam which, because of the low surface tension of the foam spheres (16,17), withstands the mechanical action of external forces, this effect being greater with increasing the percentage of effervescent mixture that generates the foam spheres.
The three-dimensional graph shows a maximal concentration at 45% effervescent mixture, suggesting that an increase of the percentage of effervescent mixture over 45% will not increase the foam resistance, but will decrease it, probably due to the fact that above this threshold value the release of carbone dioxide resulted in the effervescence is high, leading to larger foam spheres that break faster (20,21).
Graphs representing the influence of formulation factors on the amount of CO 2 released after foaming (Y 9 ) are illustrated in Fig.  3.f. The amount of carbon dioxide released after effervescence increases with the increase of effervescence mixture percentage (X 2 ) and citric acid excess (X 3 ). This behaviour is understandable (19,22), because if effervescent mixture percentage is higher, the stoichiometric quantity of carbone dioxide resulting from the reaction of effervescence will be higher. Citric acid excess leads to the acidification of the environment and to the decrease of carbon dioxide solubility, which promotes its release.
Graphs representing the influence of formulation factors on pH values generated in water (Y 10 ) after tablet disintegration in water are shown in Fig. 3.f. Regarding the pH produced by effervescent tablets disintegration, it may be observed that it decreases with the increase of the percentage of effervescent mixture (X 2 ) and with the increase of citric acid excess (X 3 ). Since all formulations contain a certain excess of citric acid, the higher the effervescent mixture percentage, the higher the amount of citric acid excess and the amount of citric acid in each tablet is, leading to a lower pH after effervescence. Another factor involved is carbon dioxide released after the effervescence reaction, which may also acidify the environment studied (19,21).
Graphs representing the influence of formulation factors on pH values recorded after tablet disintegration in simulated vaginal fluid (SVF) (Y 11 ) are illustrated in Fig. 3 Fig. 3.h.The graphs representing the variation of pH in acetate buffer pH 5.5 show that its change, depending on the formulation, is due only to citric acid excess (X 3 ) and is independent of the percentage of effervescent mixture (X 2 ). We can drag the conclusion that an excess of citric acid in the effervescent couple lowers the possible buffering capacity, while the effervescent mixture itself cannot cause such an effect, which can be demonstrated by the fact that in most cases the pH decreases and only in one case it increases slightly (pH = 5.77 for N6 formulation).

Optimal formula determination
The domains and theoretical levels chosen for the dependent variables and the experimental responses obtained for the optimal formula are represented in table 5.
The optimal domains for the levels of dependent variables were chosen considering that we need an equilibrium between an easy preparation; quality assurance; time stability; easy disintegration; energetic foaming with keeping of the local mucosal environmental conditions. Based on the chosen levels for the dependent variables, the optimizing program recommended the following levels for the formulation factors, of an intravaginal foaming tablet: 0.5% sodium laurylsulphate (X 1 ); 44% total amount of effervescent mixture in the formulation (X 2 ); with 24% excess of citric acid in the effervescent mixture (X 3  in order to choose the optimal formula with the Box Bhenken optimization design based on the Modde optimization software. The main factor involved in the behaviour of tablets immediately after preparation is the percentage of effervescent mixture. There is a very small influence of sodium lauryl sulphate on the dependent variables. The concentration of sodium lauryl sulphate was considered to be decisive only for two properties: foaming time and foam resistance. An interesting result is that the volume of foam does not depend on the percentage of sodium lauryl sulphate (for the chosen concentration levels). The optimal formula of a foaming vaginal tablet, obtained with the Box-Bhenken design, must contain 0.5% sodium lauryl sulphate, while the effervescent mixture must contain a percentage of 24% citric acid excess, and it must represent 44% of the total amount of excipients. The experimental determinations on the optimal formula were close to the theoretical values predicted by the program, this attesting the validity of the optimization, based on the Box-Bhenken experimental design. generated in pH = 5.5 acetate buffer The theoretical values predicted by the optimizing program for the optimal formula, were very close to the practical values obtained by preparation and analysis of the recommended formula; these attesting the validity of optimization based on Modde optimizing software.

CONCLUSIONS
Fifteen formulations of foaming vaginal fast release tablets for local antimycotic action were prepared,

Conflict of interest: none declared
Financial support: none declared