Multi-factor optimization method for regulating and controlling in-vitro release performance of O / W type cream

Abstract

The invention discloses a multi-factor optimization method for regulating and controlling the in-vitro release performance of O / W type cream, and belongs to the technical field of pharmaceutical preparations. According to the method, a quantitative relation model between the drug release rate and the process parameters is established through statistical experimental design in combination with parameter screening, response surface modeling and prediction verification. According to the method, firstly, Plackett-Burman design is adopted to identify key influence factors, then, a steepest climbing method is adopted to clearly optimize a central point of an area, and finally, a Box-Behnken response surface method is utilized to establish a quadratic polynomial regression model for predicting, regulating and controlling drug in-vitro release behaviors. The typical model cream verifies that the method can effectively improve the process regulation and control efficiency, and the obtained model has good prediction precision and stability.

Description

Technical Field

[0001] The present invention belongs to the field of pharmaceutical preparation process optimization, and specifically relates to a modeling and prediction method for regulating the in vitro release performance of an oil-in-water (O / W) cream preparation based on a multi-factor statistical design method. Background Art

[0002] Creams are typical semisolid preparations. The drug release rate from the matrix directly affects their efficacy and safety, and release behavior is synergistically regulated by emulsification, homogenization, and cooling process parameters. In vitro release testing (IVRT) is a core method for evaluating cream performance. The variability in drug release from O / W creams prepared with different process parameters further complicates in vitro model construction, making it difficult to optimize release profiles using existing methods.

[0003] Traditional process optimization for creams relies heavily on single-factor experiments or empirical adjustments, making it difficult to analyze the interactions and nonlinear effects between multiple parameters. For example, while increasing homogenization pressure can reduce droplet size (increasing surface area and promoting drug release), it can also lead to drug degradation or excipient structural damage due to shear heat, resulting in insufficient process robustness and significant batch-to-batch release variability.

[0004] While there have been occasional attempts to employ response surface methods in the prior art, these have primarily focused on microstructure and rheological properties, lacking a systematic approach to optimizing the release kinetics of O / W creams. Furthermore, most existing studies rely on empirical parameter adjustments without incorporating statistical experimental designs, resulting in a fuzzy process window and poor reproducibility. Therefore, a data-driven approach to optimizing process parameters is urgently needed to balance release rate and batch stability. Summary of the Invention

[0005] The present invention aims to provide an in vitro release control method suitable for an O / W cream preparation. The method screens key process factors based on the Plackett-Burman design, quickly locates the optimization region through the steepest climbing method, establishes the multi-factor interaction relationship with the help of the Box-Behnken response surface model, and realizes the prediction of process parameters and the control of release behavior through the regression model.

[0006] In order to achieve the above object, the technical solution proposed by the present invention is:

[0007] A multi-factor optimization method for regulating the in vitro release performance of an O / W cream, comprising:

[0008] S1 Plackett-Burman design was used to screen the process parameters that significantly affected the release behavior during the preparation of the cream;

[0009] S2 Based on the main effect direction and intensity of key parameters, the steepest climbing test is designed to determine the center point of the response surface and the optimization area;

[0010] S3 was based on the Box-Behnken response surface methodology, with the screened process parameters as factor variables and the in vitro release rate as the response variable, to construct a quadratic polynomial model;

[0011] S4 inputs the target release parameters and infers one or more feasible process parameter combinations through the regression model;

[0012] S5 compares the deviation between the predicted release parameters and the measured values, and iteratively updates the prediction model when the deviation exceeds the tolerance range.

[0013] Furthermore, in step S1, the screening method is: using Design-Expert software to analyze the experimental data, calculate the difference in response values ​​of each process parameter at high and low levels, perform variance analysis on the experimental data, and identify significant effects in the Plackett-Burman experiment.

[0014] Furthermore, in step S1, the screened process parameters include at least one of homogenization temperature, homogenization pressure, number of homogenization cycles, and cooling rate.

[0015] Furthermore, the parameter ranges are: homogenization temperature 55-75°C, homogenization pressure 30-200 bar, homogenization cycle number 1-3 times, and cooling rate 1-3°C / min.

[0016] Furthermore, in step S2, the release rate variation range after single-step adjustment in the steepest climbing test is 10% to 30%.

[0017] Furthermore, in step S3, the number of center points included in the Box-Behnken design should not be less than 3.

[0018] Furthermore, in step S3, the quadratic polynomial model is:

[0019] Y is the release rate of triamcinolone acetonide, are the coefficients of the polynomial, X5 is the homogenization pressure, X6 is the number of homogenization cycles, and X4 is the cooling rate.

[0020] Furthermore, the in vitro release rate values ​​are all measured by calibrated vertical diffusion cells.

[0021] Furthermore, in step S5, when the relative deviation RD between the predicted release parameter and the measured value is greater than 5%, the model is updated and the prediction result is recalculated.

[0022] The beneficial effects of the present invention are:

[0023] The present invention has the advantage of using two statistical experimental designs to establish a multi-factor optimization method suitable for regulating the release performance of creams. This method process can quickly identify significant variables that affect the release performance in the cream preparation process, significantly reducing the experimental workload. The constructed quadratic polynomial model has clear data standards and can be used to analyze the interaction effects between factors. The model can be iteratively optimized dynamically, and the prediction model is updated by expanding the experimental data matrix to avoid extreme risk conditions. With the cream release performance as the response value, the stability and release rate are balanced by a satisfaction function, and the expected drug release rate is obtained by changing the significant variables within the experimental parameter range. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] FIG1 shows the change in droplet size after changing the number of homogenization cycles in Example 1.

[0025] FIG2 shows the change in droplet size after changing the homogenization pressure in Example 1.

[0026] FIG3 shows the changes in drug recovery and cream viscosity after changing the homogenization temperature in Example 1.

[0027] FIG4 is a half-normal probability plot of the effects of the Plackett-Burman design process factors on the release rate of triamcinolone acetonide in Example 2.

[0028] FIG5 is a Pareto diagram showing the effects of the Plackett-Burman design process factors on the release rate of triamcinolone acetonide in Example 2.

[0029] FIG6 is a response surface diagram of the Box-Behnken design homogenization pressure and the number of homogenization cycles in Example 5.

[0030] FIG7 is a response surface diagram of the Box-Behnken design of homogenization pressure and cooling rate in Example 5.

[0031] FIG8 is a response surface diagram of the Box-Behnken design homogenization cycle number and cooling rate in Example 5.

[0032] FIG9 is a graph showing the release curve of triamcinolone acetonide from the two extreme point verification experiments in Example 6.

[0033] FIG10 is a graph showing the release curves of triamcinolone acetonide before and after the model is updated in Example 7. DETAILED DESCRIPTION

[0034] To more clearly illustrate the technical solution of the present invention, the method of the present invention is described in detail below in conjunction with specific embodiments. Based on the embodiments disclosed in the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection disclosed in the present invention. However, the scope of protection of the present invention is not limited to the following embodiments.

[0035] A multi-factor optimization method for regulating the in vitro release performance of an O / W cream comprises the following steps:

[0036] S1 Plackett-Burman design was used to screen the process parameters that significantly affected the release behavior during the preparation of the cream;

[0037] S2 Based on the main effect direction and intensity of key parameters, the steepest climbing test is designed to determine the center point of the response surface and the optimization area;

[0038] S3 Based on the Box-Behnken response surface methodology, a quadratic polynomial model was constructed with the key parameters as factor variables and the in vitro release rate as the response variable;

[0039] S4 inputs the target release parameters and uses the regression model to infer one or more feasible process parameter combinations.

[0040] S5 compares the deviation between the predicted release parameters and the measured values, and iteratively updates the prediction model when the deviation exceeds the tolerance range.

[0041] Furthermore, in step S1, the screened process parameters include at least one of the following: homogenization temperature, homogenization pressure, number of homogenization cycles, and cooling rate.

[0042] Preferably, the parameter ranges are: homogenization temperature 55-75°C, homogenization pressure 30-200 bar, homogenization cycle number 1-3 times, and cooling rate 1-3°C / min.

[0043] Furthermore, in step S2, the release rate variation range after single-step adjustment in the steepest climbing test is 10% to 30%.

[0044] Furthermore, in step S3, the number of center points included in the Box-Behnken design should not be less than 3.

[0045] Furthermore, the in vitro release performance values ​​described in the present invention are all measured by calibrated vertical diffusion cells.

[0046] Furthermore, in step S4, the predicted process parameter value is adjusted to an integer value according to the actual process feasibility.

[0047] Furthermore, in step S5, when the RD between the predicted value and the actual measured value is greater than 5%, the response surface model is updated and the prediction result is recalculated.

[0048] Example 1 Determination of process parameter range

[0049] Using triamcinolone acetonide cream as a model cream, a single-factor experiment was used to determine the numerical range of the process parameters of the O / W cream, as shown in Table 1.

[0050] Table 1 Single factor experimental design of key process parameters of cream

[0051]

[0052] The experimental results are as follows Figure 1 、 2 As shown in Figures 3 and 4, the droplet size of O / W creams should be between 1 and 100 μm. Too little pressure and too short a time will not evenly disperse large particles. Too long a time will break up already dispersed small particles, forming even smaller droplets that gradually transform into micelles under the action of surfactants, affecting the accuracy of the measurement results. Too low a homogenization temperature increases the viscosity of the cream and hinders the release of the drug from the matrix. Temperatures exceeding 75°C increase drug instability and lead to degradation. Therefore, the homogenization temperature should be between 55 and 75°C, the homogenization pressure should be between 30 and 200 bar, and the number of homogenization cycles should be between 1 and 3.

[0053] Example 2 Plackett-Burman Design (PBD) Screening Key Parameters

[0054] In this example, triamcinolone acetonide cream was used as a model cream to verify the prediction of drug release rate by process parameters in the polynomial model.

[0055] Step 1: Experimental Design

[0056] A PBD (Physical Design of Batch) screening of key process parameters was designed using Design Expert 13 software. Factors included emulsification temperature (X1), emulsification time (X2), emulsification speed (X3), cooling rate (X4), homogenization pressure (X5), number of homogenization cycles (X6), and homogenization temperature (X7). This experimental design included a total of 15 factor points, with three replicate center points to assess error. The release rate of triamcinolone acetonide from the cream was used as the response. A detailed design is shown in Table 2.

[0057] Table 2 Experimental design and results of significant factor screening

[0058]

[0059] Step 2: Preparation of triamcinolone acetonide cream

[0060] The parameters for each process step in the cream preparation were set according to the experimental parameters in Table 1. During the emulsification process, the mixture was stirred at speeds of 100 and 1200 rpm, heated to 50 and 70°C, and stirred for 10 and 40 min. During the homogenization process, the colostrum was poured into a high-pressure homogenizer and homogenized at pressures of 30 and 150 bar for 1 and 5 passes, respectively. The homogenization temperature was set at 55 and 65°C, and the emulsion was then stirred at a rate of 1.3°C / min and cooled to room temperature.

[0061] Step 3: Calculation of release rate

[0062] The in vitro release experiment of the cream was conducted using the fully automatic transdermal diffusion system KX-10VPC with a release area of ​​1.767 cm 2 The receiving cell volume was 12 mL, and the receiving solution consisted of 0.05% sodium lauryl sulfate (20% ethanol) in saline. Sink conditions were maintained throughout the experiment. Before the experiment, the nylon membrane was soaked in the receiving solution for 30 minutes. Afterward, 300 mg of cream was evenly applied to the release membrane. The release membrane, containing the sample, was secured between the receiving and supply cells. A magnetic stirring bar was placed in the receiving cell, and 12 mL of receiving solution was added. A magnetic stirring system was installed on the bottom of the system table, wrapped in an insulating interlayer. The speed was set to 600 rpm and the temperature to 32°C. Sampling was performed at set intervals of 30, 60, 120, 180, 240, and 360 minutes. Each sample consisted of 200 μL of receiving solution, which was then supplemented with the same volume of fresh blank receiving solution.

[0063] The supernatant was analyzed by HPLC, and the triamcinolone acetonide content in the sample was calculated using the external standard method. The cumulative release per unit area was plotted as the dependent variable against the square root of time, and the slope of the linear relationship was defined as the release rate.

[0064] In this embodiment, 6 parallel groups of in vitro release experiments of the cream were set up for each experimental point, and the slope of the linear fit between the average value of the 6 release rates and the square root of time was finally taken as the release rate.

[0065] In this embodiment, the cumulative release amount of triamcinolone acetonide in the cream is calculated as follows:

[0066]

[0067] Where: C n is the concentration at the nth sampling point, μg mL-1; V1 is the volume of the receiving pool, mL; V2 is the volume of a single sampling, mL; A is the area of ​​the receiving pool, 1.767 cm 2 ; Q is the cumulative permeation amount of the drug per unit area, μg·cm -2 .

[0068] Step 4: Result Analysis

[0069] Design-Expert 13 software was used to analyze the data in Table 1, calculate the difference in response values ​​of each factor at high and low levels, perform variance analysis on the experimental data, identify significant effects in the Plackett-Burman experiment, and draw a Pareto chart of the standardized effects of factors.

[0070] Perform multivariate fitting on the data and obtain the equation:

[0071] Y=-0.1538+0.006083X1-0.003167X2-0.000833X3+0.04583X4+0.02125X5+0.08417X6+0.01917X7.

[0072] Figure 4 and Figure 5 It is a half-normal probability distribution diagram and Pareto diagram of the effects of each factor, showing the influence of each process factor on the release rate of triamcinolone acetonide. Figure 4 In the figure, the farther the effect point is from the distribution fitting line, the more significant the influence of the factor on the effect, which means that the homogenization pressure, homogenization temperature, homogenization cycle number and cooling rate have the most significant influence on the effect. Figure 5 The Pareto chart shows the upper and lower lines representing the significance thresholds of P = 0.05 and P = 0.01, respectively. Bars exceeding these lines represent factors with a significant effect on the effect. Homogenization pressure, homogenization temperature, and number of homogenization cycles all significantly affect the release rate of triamcinolone acetonide. However, because the matrix components in the model cream formulation are prone to crystallization during the cooling process, affecting drug release, the inclusion of cooling rate as a factor makes subsequent results more reliable. Furthermore, the homogenization process in the validation experiment used an ATS AH-NANOSOP high-pressure homogenizer. Although the equipment was connected to an external temperature-controlled circulating water bath system, the physical material of the feed cup caused the emulsion to cool rapidly and partially adhere to the cup wall, resulting in temperature unevenness when the colostrum entered the homogenization chamber. To avoid inevitable errors introduced by the equipment, the homogenization temperature variable was eliminated from the Box-Behnken design. In subsequent experiments, this parameter was set to the geometric mean of the range verified in Example 1, i.e., 65°C.

[0073] Example 3 Steepest climbing test of significant factors

[0074] The results of the Plackett-Burman experiment show that homogenization pressure, number of homogenization cycles, and cooling rate all have a positive effect on the release rate, and the homogenization pressure has the most significant impact. Therefore, X5 (homogenization pressure) is used as the climbing unit.

[0075]

[0076] The ramping experiment took into account both the workload and experimental conditions. The release rate variability after a single ramping step ranged from 10% to 30%. To avoid missing the optimal point or inaccurate results, the homogenization pressure step was set at 15 bar and the cooling rate at 0.5°C / min. To avoid Ostwald ripening caused by over-homogenization, the homogenization cycle step was set at 0.5 (i.e., half the sample volume was homogenized and then mixed with the unhomogenized sample). The experimental design and results are shown in Table 3. The release rate variability after a single ramping step ranged from 10% to 30%, demonstrating that the triamcinolone acetonide release rate in Treatment 4 reached its maximum. Therefore, Treatment 4 was selected as the center point of the response surface.

[0077] Table 3 Steepest climbing test design and results

[0078]

[0079] Example 4 Box-Behnken response surface design

[0080] Based on the experimental parameter ranges and center points of significant factors derived from the steepest ramp test, a three-factor, three-level response surface analysis scheme was designed using Design-Expert 13 software, with homogenization pressure, number of homogenization cycles, and cooling rate selected as optimization variables for the Box-Behnken experiment. Significant factors were assigned numbers of -1, 0, and 1, representing low, medium, and high experimental values. The specific response surface design factors and levels are shown in Table 4.

[0081] In this experimental design, a total of 15 experimental points were included, which were divided into 12 extreme points and 3 replicated central points to estimate experimental error. The release rate of triamcinolone acetonide from the cream was used as the response value. The detailed plan is shown in Table 5.

[0082] Table 4 Factors and levels of BBD response surface design

[0083]

[0084] Table 5 BBD response surface experimental design and results

[0085]

[0086] Example 5 Data Processing in Box-Behnken Experimental Design

[0087] Regression analysis was performed on the experimental variables and reaction indicators in the Box-Behnken design, and the regression formula of process variables on triamcinolone acetonide release rate was obtained:

[0088] Y is the release rate of triamcinolone acetonide, are the coefficients of the polynomials, are process variables: X5 (homogenization pressure), X6 (number of homogenization cycles), X4 (cooling rate), is the interaction term between process variables, is the quadratic term of the process variable. The results of the variance analysis of the regression equation are shown in Table 6. The regression significance of the model (P = 0.002) indicates that the model relationship between the independent variable and the response value is significant, while the lack of fit term (p = 0.0687) is not significant, which indicates that the regression equation fits the actual situation well. In addition, the determination coefficient R 2 It is 0.9736, which proves that the model can explain 97.36% of the response value variation. Therefore, this model can be used to analyze and predict the effect of process parameters on the release rate of triamcinolone acetonide.

[0089] The F value can reveal the degree of influence of various factors on the experimental indicators. According to the F value, the contribution rate of each factor is ranked as A>B>C, specifically homogenization pressure>number of homogenization cycles>cooling time. By performing variance analysis on the various parameters of the model, we found that the linear term A (homogenization pressure) in the model has a very significant effect on the release rate of triamcinolone acetonide (p<0.01), the interaction term BC (number of homogenization cycles and cooling rate) has a significant effect on the release rate of triamcinolone acetonide (p<0.01), and the quadratic term A 2 、B 2、 C 2 There was also a significant effect (p<0.01). However, the effects of the interaction terms AB and AC on the release rate of triamcinolone acetonide were relatively small (p>0.05).

[0090] Table 6 Box-Behnken analysis of variance

[0091]

[0092] Figures 6 to 8 The results of the response surface analysis of the effects of homogenization pressure, number of homogenization cycles, and cooling rate on the release rate of triamcinolone acetonide are presented. The interaction between these two factors becomes more significant when the curvature and steepness of the response surface plot increase. Figure 6 This figure demonstrates how homogenization pressure and the number of homogenization cycles affect the release rate of triamcinolone acetonide when the cooling rate is zero. As the homogenization pressure and the number of homogenization cycles gradually increase, the release rate of triamcinolone acetonide also increases. The contour lines are roughly circular and relatively sparsely distributed. This pattern indicates that the interaction between homogenization pressure and the release rate of triamcinolone acetonide has no significant effect. The circular contour lines reveal the balanced interaction between the various factors, while the sparse distribution suggests that the interaction does not significantly affect the response variable.

[0093] Figure 7This figure demonstrates how homogenization pressure and the number of homogenization cycles affect the release rate of triamcinolone acetonide when the cooling rate is zero. As the homogenization pressure and the number of homogenization cycles gradually increase, the release rate of triamcinolone acetonide also increases. A thorough analysis of the response surface contours reveals that the contours are roughly elliptical, suggesting a significant interaction between these two variables.

[0094] Figure 8 The figure shows how the homogenization pressure and the number of homogenization cycles affect the release rate of triamcinolone acetonide when the cooling rate is 0. As the homogenization pressure and the number of homogenization cycles gradually increase, the release rate of triamcinolone acetonide also increases.

[0095] Example 6 Verification of the Predictive Ability of the Quadratic Regression Equation

[0096] Triamcinolone acetonide cream was used as a model cream. Response surface methodology was used to calculate a quadratic polynomial regression equation for the drug release rate as a function of homogenization pressure, number of homogenization cycles, and cooling time. This regression model was used to predict the peak release rate of triamcinolone acetonide under these process conditions. Three batches of cream were prepared at each predicted peak. In actual experiments, process parameters were appropriately adjusted, and validation experiments (n = 6) were conducted at two peaks using the same release rate measurement method. The results are shown in Table 7.

[0097] Table 7 Verification test results

[0098]

[0099] Note: The experimental values ​​are the release rates of triamcinolone acetonide in three batches of cream, expressed as mean ± standard deviation.

[0100] like Figure 9 As shown in the figure, the release curve of the cream prepared under the process parameters at the two extreme points is linear, with a cumulative drug release rate of less than 30% over 6 hours, which meets the requirements of the Higuchii model assumption. The RSD of the release rate of the six groups within the batch is less than 15%. The average cumulative release amount of the six groups in each batch is linearly fitted with the square root of time. The linear coefficient is the release rate of triamcinolone acetonide. The maximum release rate of triamcinolone acetonide after adjustment is the average release rate of three experiments of 1.42±0.04μg / cm 2 / t 1 / 2 The minimum release rate of triamcinolone acetonide was 2.14±0.06μg / cm 2 / t 1 / 2 , which were close to the predicted values, with relative deviation (RD) < 5%.

[0101] This shows that the established model and preparation process have high operability and stability, and can be reliably used to predict the release rate of triamcinolone acetonide.

[0102] Example 7 Dynamic Optimization of Quadratic Regression Model

[0103] Parameter prediction was performed using the quadratic polynomial model from Example 5. Due to the spatial symmetry of the response surface, an input value of 1.80 revealed five optimal solutions, as shown in Table 8. Process parameters were selected based on experimental conditions and equipment performance. Because high homogenization pressure can cause Ostwald ripening of emulsion droplets, Solution 1 was selected for experimental verification.

[0104] Table 8 A series of parameter combinations corresponding to the predicted values

[0105]

[0106] The verification results are shown in Table 9. Experiments were conducted on the prediction parameter values, and the results are as follows Figure 10 As shown, the cumulative drug release rate over 6 hours was less than 30%, meeting the assumptions of the Higuchii model. The RSD of the release rates among the six groups within the batch was less than 15%. While ensuring consistency between manual operation and experimental instrumentation, the relative deviation between the experimental value and the input value was greater than 5%. This experimental data was added to the original Box-Behnken data matrix to optimize the model.

[0107] The regression formula after updating the model:

[0108] Y' is the release rate of triamcinolone acetonide, are the coefficients of the polynomials, are process variables: X5 (homogenization pressure), X6 (number of homogenization cycles), X4 (cooling rate), is the interaction term between process variables, is the quadratic term of the process variable.

[0109] Table 9 Predicted values ​​and experimental values ​​before and after model update

[0110]

[0111] The new model was used to input the target release rate value, and 6 parameter combinations were obtained. The parameter scheme was selected based on the experimental cost and instrument performance, as shown in Table 9, 2. Figure 10 As shown in the results, the new model predicted the release curve of the cream prepared under the process parameters with good linearity. The cumulative drug release rate in 6 hours was less than 30%, which met the requirements of the Higuchii model assumption. The RSD of the release rate of the six groups within the batch was less than 15%. The average cumulative release amount of the six groups in each batch was linearly fitted with the square root of time. The linear coefficient was the release rate of triamcinolone acetonide. The maximum release rate of triamcinolone acetonide after adjustment was the average release rate of 1.884±0.064μg / cm3 for three experiments. 2 / t 1 / 2, which is close to the predicted value, with a relative deviation (RD) of <5%.

[0112] It is proved that the prediction model calculated by this method can be refitted by merging the validation data. The model after iterative updating has good predictive performance and the process combination has good operability.

[0113] As can be seen from the above examples, the present invention provides a multi-factor method for regulating the in vitro release performance of an O / W cream. In this example, using triamcinolone acetonide cream as a model, a Plackett-Burman design experiment was used to screen for process factors with significant effects on the drug release rate from the cream. A steepest ramp experiment was used, combined with specific operating conditions, to determine the parameter ranges and response surface centers of key influencing factors. Within the specified parameter ranges, a Box-Behnken design experiment was conducted, and a quadratic polynomial regression equation was used to predict the effects of process parameter variables on the drug release rate from the cream. Validation of the model cream revealed that the extreme points were within the experimental parameter range, indicating a reliable model. The results demonstrated that the model predicted release rates using this method was highly accurate, with the relative deviation (RD) between the predicted and experimental values ​​at the model extreme points being no greater than 5%. When data deviations occurred, the original data matrix was expanded to update the iterative prediction model. This method can be used to flexibly and accurately screen significant factors and effectively regulate various factors influencing the drug release rate from O / W creams.

[0114] The above embodiments are only used to illustrate the present invention. Any equivalent transformations and improvements based on the technical solution of the present invention should not be excluded from the protection scope of the present invention.

Claims

1. A multi-factor optimization method for regulating the in vitro release performance of an O / W cream, characterized in that: The method comprises: S1 Plackett-Burman design was used to screen the process parameters that significantly affected the release behavior during the preparation of the cream; S2 Based on the main effect direction and intensity of key parameters, the steepest climbing test is designed to determine the center point of the response surface and the optimization area; S3 was based on the Box-Behnken response surface methodology, with the screened process parameters as factor variables and the in vitro release rate as the response variable, to construct a quadratic polynomial model; S4 inputs the target release parameters and infers one or more feasible process parameter combinations through the regression model; S5 compares the deviation between the predicted release parameters and the measured values, and iteratively updates the prediction model when the deviation exceeds the tolerance range.

2. The method according to claim 1, wherein: In step S1, the screening method is: using Design-Expert software to analyze the experimental data, calculate the difference in response values ​​of each process parameter at high and low levels, perform variance analysis on the experimental data, and identify significant effects in the Plackett-Burman experiment.

3. The method according to claim 2, characterized in that In step S1, the screened process parameters include at least one of homogenization temperature, homogenization pressure, homogenization cycle times, and cooling rate.

4. The method according to claim 3, characterized in that The parameter ranges are: homogenization temperature 55~75°C, homogenization pressure 30~200 bar, homogenization cycle number 1~3 times, and cooling rate 1-3°C / min.

5. The method according to claim 4, characterized in that In step S2, the release rate after single-step adjustment in the steepest climbing test varies in a range of 10% to 30%.

6. The method according to claim 5, characterized in that In step S3, the number of center points included in the Box-Behnken design should not be less than 3.

7. The method according to claim 6, characterized in that In step S3, the quadratic polynomial model is: Y is the release rate of triamcinolone acetonide, are the coefficients of the polynomial, X5 is the homogenization pressure, X6 is the number of homogenization cycles, and X4 is the cooling rate.

8. The method according to claim 7, characterized in that The in vitro release rate values ​​were all measured using a calibrated vertical diffusion cell.

9. The method according to claim 7, characterized in that In step S5, when the relative deviation RD between the predicted release parameter and the measured value is greater than 5%, the model is updated and the prediction result is recalculated.

Images