INTRODUCTION:

Response surface methodology (RSM) was a collection of statistical and mathematical techniques that has been successfully used to determine the effects of several variables and optimize processes. Optimization of formulation design can be used in formulation and development of pharmaceutical products due to the wide array of parameters and variables that must be controlled to achieve desire release pattern and meet other performance criteria. Box–Behnken designs do not have axial points, thus all design points fall within the safe operating zone. These designs also ensure that all factors are never set at their high levels, simultaneously.^1,2

Furthermore; Box–Behnken designs have fewer design points. Also, each factor requires only three levels instead of the five required for central composite designs (unless alpha is equal to one), which may be experimentally more convenient and less expensive to run than central composite designs with the same number of factors.

Oral sustained release formulations have drawbacks in respect to variation of gastric emptying time results in variable drug absorption. Too rapid gastrointestinal transit can lead to inadequate drug release from the dosage form above the absorption zone, resulting in diminished effectiveness of the given dose when the drug presents an absorption window. Prolongation of gastric residence time of a rate controlled oral drug delivery system can rectify these problems by minimizing the inter-subject variability known as ‘peak and trough’ effect, and also improve the bioavailability, specially for drugs having a narrow absorption window in the upper part of the GI tract.³ One of the most feasible approaches for achieving a prolonged and predictable drug delivery in the GI tract is to control the gastric residence time, i.e. gastro retentive dosage form through bioadhesion. Over the last two decades, there has been considerable interest in mucoadhesive drug delivery systems for their potential to optimize localized drug delivery, by retaining a dosage form at the site of action, or systemic delivery, by retaining a formulation in intimate contact with the absorption site by adhere to mucosal surfaces; devices that rapidly increase in size once they are in the stomach to retard the passage through the pylorus.^{4, 5}

Hibiscus esculentus is planted in many areas of the world and is an indigenous of India. Hibiscus esculentus mucilages are used as thickeners, flavoring for different foods and as egg white substitute, fat substitute in chocolate bars and frozen dairy desserts.⁶ Hibiscus esculentus polysaccharides are found to be random coil polysaccharides consisting of galactose, rhamnose, and galacturonic acid. The repeating unit was reported to be α-(1-2)-rhamnose and α-(1- 4)-galacturonic acid residues including disaccharide side chains and degree of acetylation up to 58 (DA = 58).⁷ Diltiazem is a calcium channel blocker, with an elimination half-life of 3 to 4.5 hours, an absorption zone from the upper intestinal tract and requires multiple daily drug dosage to maintain adequate plasma concentrations in management of angina pectoris and hypertension. Efficacy of the administered dose may get reduced due to incomplete drug release from the device above the absorption zone.⁸ These disadvantages can overcome by developing gastroretentive mucoadhesive drug delivery systems that can be retained in the stomach and assist in improving the oral sustained delivery of drugs that have an absorption window in a particular region of the gastrointestinal tract. These systems help in continuously releasing the drug before it reaches the absorption window and thus ensuring optimal bioavailability. The present study was aimed to design and optimize the intragastric sustained release mucoadhesive oral tablets by using diltiazem hydrochloride as a model drug to increase the residence time of the drug in to the stomach and release for extended period of time.

MATERIALS AND METHODS:

Materials:

Diltiazem Hydrochloride was obtained as a gift by Zim Laboratories, Kalmeshwar, Nagpur (India). Okra mucilage was isolated by simple maceration from Hibiscus esculantus unripe fruits which were purchased from the local market, Amravati district Maharashtra, in the month of September 2012. All excipients were of USP/NF grades and all other chemicals used were of analytical grades.

Preparation of diltiazem mucoadhesive sustained release matrices:

The experiments consisted of 17-run Box Behnken design points generated with the aid of Design-Expert® V7 software. The actual and coded values are given in table 1. All tablet formulations were prepared by using wet granulation technique. Polyvinyl pyrrolidone K30 was accurately weighed and dissolved in isopropyl alcohol.

Table 1: Composition of GRMDDS of DLT HCL for preliminary optimization by runs designed in coded values according to Box-Behnken design

Formulation Code	Independent Variables (coded)						Ingredients (mg)
	Coded level			Actual value			Ingredients (mg)
	X1	X2	X3	HEC	XNG	Load	DLT	DCP	PVP	TLC	MGST
DF01	-1	-1	0	40	35	3	90	124	5	3	3
DF02	1	-1	0	60	35	3	90	104	5	3	3
DF03	-1	1	0	40	45	3	90	114	5	3	3
DF04	1	1	0	60	45	3	90	94	5	3	3
DF05	-1	0	-1	40	40	1	90	119	5	3	3
DF06	1	0	-1	60	40	1	90	99	5	3	3
DF07	-1	0	1	40	40	5	90	119	5	3	3
DF08	1	0	1	60	40	5	90	99	5	3	3
DF09	0	-1	-1	50	35	1	90	114	5	3	3
DF10	0	1	-1	50	45	1	90	104	5	3	3
DF11	0	-1	1	50	35	5	90	114	5	3	3
DF12	0	1	1	50	45	5	90	104	5	3	3
DF13	0	0	0	50	40	3	90	109	5	3	3
DF14	0	0	0	50	40	3	90	109	5	3	3
DF15	0	0	0	50	40	3	90	109	5	3	3
DF16	0	0	0	50	40	3	90	109	5	3	3
DF17	0	0	0	50	40	3	90	109	5	3	3

All excipients except magnesium stearate and talc were accurately weighed and passed through # 80 meshes. Calculated amount of the drug, polymer (HEC, XNG) and filler dicalcium phosphate were mixed thoroughly. A sufficient volume of the specified granulating agents were added slowly to achieve the granulation endpoint. After the enough cohesiveness obtained, granules were dried at room temperature to evaporate the isopropyl alcohol and then were dried at 50°C for 30 minutes. The semi-dried granules were passed through # 10 mesh and drying was continued for another one hour. The granules were collected and passed through # 12 meshes. Magnesium stearate ad talc was passed through # 80 meshes to lubricate the granules. The granules were compressed using 10 mm round standard concave punches at different compression load. The evaluation of the main effects, interaction, and quadratic effects of the input and response variables was performed using a Box–Behnken statistical screening design that had three center points. The mathematical relationship between the input and output variables was generated using Design-Expert® V7 software (Stat–Ease, Inc., Minneapolis, Minnesota, USA). The dependent variables studied the percent cumulative drug release (% CDR) at 8th hr (Y₁), time to release 80 % (t₈₀) of drug (Y₂), mucoadhesive strength (Y₃) and mucoadhesive time (Y₄).

FTIR spectral and DSC analysis⁹:

FTIR spectrums were recorded on samples prepared in potassium bromide (KBr) disks using FTIR spectrophotometer (Model-1601 PC, Shimadzu Corporation, Japan). The scanning range was 400 to 4000 cm^-1. The FTIR spectrums of pure diltiazem, 1:1:1:1 ratio of diltiazem: Hibiscus esculentus mucilage: xanthan gum: povidone and formulation blend were taken. The DSC analysis was carried out using differential thermal analyzer (Shimadzu DSC-60, Shimadzu Limited, and Japan). A 1:1:1 ratio of diltiazem: Hibiscus esculentus mucilage: xanthan gum: povidone were weighed into aluminum crucible and the DSC thermo grams were recorded at a heating rate of 10⁰C/min in the range 20⁰C to 280⁰C, at a nitrogen flow of 20 ml/min.

Pre and post-compression parameters:

Properties of the gastro retentive mucoadhesive granules were evaluated for bulk density, tapped density, angle of repose and compressibility as in-process characterization. Properties of the gastro retentive mucoadhesive tablets, such as appearance, thickness and diameter, surface pH, hardness, friability, weight variation and content uniformity were determined. The hardness was determined by using a Monsanto tablet hardness tester. Friability was determined using Roche friability testing apparatus (EF-1W, Electrolab Mumbai, India) as per procedures described in the Indian Pharmacopoeia (IP). Weight variation was also performed according to the IP procedure. Thickness of tablets was determined using a digital vernier caliper (Mitutoyo Corporation, Kawasaki, Japan).^{10, 11}

In-vitro drug release and kinetics study:

In-vitro drug release study was carried out using USP II (paddle method) apparatus in 900 ml of 0.1 N HCl (pH 1.2) for 12h. The temperature of the dissolution medium was kept at 37± 0.5ºC and the paddle was set at 100 rpm. 10 ml of sample solution was withdrawn at specified interval of time and filtered through Whattman filter paper. The sample was replaced with fresh dissolution medium. The sample diluted to a suitable concentration

with 0.1 N HCL. The absorbance of the withdrawn samples was measured at 237 nm using a Shimadzu UV-

Visible spectrophotometer. The dissolution studies of all the batch formulations were performed in six replicates.¹² In order to describe the kinetics of drug release from controlled release formulation; various mathematical equations have been used like zero order, first order, higuchi model, hixson–crowell cube root law and korsmeyer peppas equation. An Excel based program DD Solver was used to analyze model dependent kinetics. Following formulae were used for this purpose.

Zero order kinetics: Q=K_ot ------------------------------------(1)

First order kinetics: In Q=In Q_o-K_t----------------------------(2)

Higuchi kinetics: Q=k_Ht^1/2-------------------------------------- (3)

Hixson and Crowell cube root model: Q_o^1/3-Q_t^1/3 =k_H C (4)

Baker and Lonsdale model: 3/2[1-(1-F)^2/3]-F=k_BLt ----- (5)

Weibull model: m =1- exp[-(t-T_t)^β/α ------------------ --(6)

Hopfenberg Model: F = 100[1-(1-k_HB. t)ⁿ] -------- -- -- (7)

Korsemeyer and peppas model: M_t/M_∞ =ktⁿ-- ------- -(8)

Where

Q= the drug release at t, Ko= zero order rate constant, Q_o =initial drug release, k_H =Higuchi dissolution constant, Q_t =remaining amount of drug, t= time, F=the fraction of drug release at t=time and k_BL=release rate constant, Mt =drug released at time t, M∞ =quantity of drug released at infinite time; k = the kinetic constant and n = the release exponent. Again, the Korsmeyer–Peppas model was employed in the in vitro drug release behavior analysis of these formulations to distinguish between competing release mechanisms: Fickian release (diffusion-controlled release), non-Fickian release (anomalous transport), and case-II transport (relaxation-controlled release). When n is ≤0.43, it is Fickian release. The n-value between 0.43 and 0.85 is defined as non-Fickian release. When, n is ≥0.85, it is case-II transport.^13-15

Swelling index:

Swelling study of individual batch was carried out using USP dissolution apparatus-II (rotating paddle), in 900 ml of 0.1NHCl which is maintained at 37±0.5°C, rotated at 100 rpm. Weight of individual tablet was taken prior to the swelling study (W1).The tablet was kept in a basket. The tablet was removed every one hour interval up to 12 hour and excess water removed carefully using filter papa.¹⁶The swollen tablets were reweighed (W2); Percent hydration (swelling index) was calculated using following formula,

% Swelling Index = (W₂) – (W₁)/ (W₁) x 100 -------- (9)

Where W₁-initial weight of tablet, W₂- weight of the swollen tablet

In-vitro mucoadhesion strength and wash off test for mucoadhesion time:

Mucoadhesive strength of the tablets was measured on the modified physical balance¹⁷. Goat stomach mucosa was used as a model membrane and buffer media 0.1N HCl was used as moistening fluid. It was then tied to a Teflon‐coated glass slide and this slide was fixed over the protrusion in the Teflon block using a thread. The block was then kept in glass beaker. The beaker was filled with 0.1N HCl up to the upper surface of the goat stomach mucosa to maintain stomach mucosa viability during the experiments. The one side of the tablet was attached to the glass slide of the right arm of the balance and then the beaker was raised slowly until contact between goat mucosa and mucoadhesive tablet was established. A preload of 5 g was placed on the slide for 5 min (preload time) to established adhesion bonding between mucoadhesive tablet and goat stomach mucosa. After the completion of preload time, preload was removed from the glass slide and water was then added in the plastic container in left side arm by peristaltic pump at a constant rate of 100 drops per min. The weight of water required to detach mucoadhesive tablet from stomach mucosa was noted as mucoadhesive strength in grams. Each measurement was carried out in triplicate and the results were averaged. The mucoadhesive strength was determined using following formulas;

Mucoadhesive Strength (g) = (wt. of beaker + wt. of water) – wt. of empty beaker---------------------------- (10)

Force of adhesion (N) = Mucoadhesive strength× 9.81/1000 ------- (11)

The mucosa was fixed on a glass slide using double sided adhesive and one side of glass slide was fixed to thread whose another end was fixed with the arm of tablet disintegration test apparatus.¹⁸ A side of each tablet was wetted with dissolution medium and was attached to the mucosa by applying a light force with a fingertip for 20 seconds. The beaker was filled with 900 ml of simulated gastric fluid and kept at 37˚C; after 2 minutes the slide was placed in a beaker and the apparatus was started. Care was taken that while up and down motion of the arm tablet should remain in medium. The behaviour and mucoadhesive time of tablet were monitored until complete detachment or dissolution occurred.

Statistical Analysis by Design Expert software:

Statistical analysis and optimization data obtained from all DLT HCL formulations were analyzed using Design Expert software version 7.0.10 (Stat-Ease, USA) and used to generate the study design and the response surface plots. Polynomial models, including linear, interaction and quadratic terms were generated for all the response variables using the software. The polynomial equation for quadratic model or second order regression model generated by this experimental design is as follows:

Yi = b₀ + b₁X₁ + b₂X₂ + b₃X₃ + b₁₂X₁X₂ + b₁₃X₁X₃ + b₂₃X₂X₃ + b₁₁X₁² + b₂₂X₂² + b₃₃X₃²…………… (12)

Where, Yi is the dependent variable; b₀ is the intercept; b₁ to b₃₃ are the regression coefficients; and X₁, X₂ and X₃ are the independent variables that were selected from the trial trials. The main effects of X₁, X₂ and X₃ represent the average result of changing one variable at a time from its low level to its high level. The interaction terms (X₁X₂, X₁X₃, X₂X₃, X₁², X₂², and X₃²) represent changes when two variables are simultaneously changed. The best fitting model was selected based on comparisons of several statistical parameters, including the coefficient of variation(CV), coefficient of determination (R²) and adjusted coefficient of determination (adjusted R²) provided by Design Expert software. In addition, analysis of variance (ANOVA) was used to identify significant effects of factors on response regression coefficients. The F test and P values were also calculated using the software. The relationship between the dependent and independent variables was further elucidated using response surface plots.¹⁹ These plots are useful to study the effects of various factors on the response at a given time and to predict the responses of dependent variables at intermediate levels of independent variables.

Evaluation of Optimized Formulation:

A simple model independent approach uses a difference factor (f1) and a similarity factor (f2) to compare dissolution profiles. The difference factor calculates the percent difference between the two curves at each time point and is a measurement of the relative error between the two curves. It is expressed as:

f₁ = {[Σⁿ _t=1(Rt - Tt)]/ [[Σⁿ _t=1Rt]} × 100 ---------------(13)

where n is the number of time points, R is the dissolution value of the reference (pre change) batch at time t, and Tt is the dissolution value of the test (post change) batch t at time t.

The closeness of drug release profile to that of target profile (market product) was calculated using FDA recommended similarity factor (f2 value), that must be within 50-100 for similarity was calculated as follows:

ƒ₂= 50 log {[1+1/n Σⁿ_n=1(Rt-Tt)²]^-0.5×100} ---- ------(14)

Where, Rt and Tt are percent of drug which was dissolved at each time point for the test and reference products respectively, n is the number of time points considered²⁰.

X-ray imaging:

The selected formulation for in-vivo bioadhesion will reformulate for small (8 mm) excluding diltiazem hydrochloride (placebo) and containing 30% barium sulphate, as opaqueing agent by same method used in formulation. X-ray photographs were performed after the approval of the protocol (1370/ac/10/CPCSEA/IAEC/04 on dated 17/01/2015) from Institutional Animal Ethical Committee (IAEC) of Government College of Pharmacy, Amravati, India. The formulations are administered to rabbits through plastic tubing followed by flushing of 25 to 30 ml of water. The radiographs were made just before the administration of the formulation to ensure the absence of any radio-opaque material in the stomach. X-ray photography of the abdominal region was taken at 2 h time intervals for a period of 12 hours after administration of the tablets using X-ray machine (MXD-1000, Medford, and Jaipur, India).

Stability Studies:

The stability studies of all the formulations were studied at different temperatures using the reported standard procedure. The tablets were wrapped in aluminum foil and placed in Petri dishes. These containers were stored at ambient humid conditions, at long term (25 ±2 °C), intermediate temperature (30 ± 2 °C) and accelerated condition (40 ± 2 °C) for a period of 6 months. The samples were analyzed for physical changes such as color, texture, drug content, hardness, friability, mucoadhesive strength and drug release characteristics and compared with the initial values before storage.²¹

RESULTS:

IR spectral assignments for diltiazem HCl reveals that it gives characteristic peaks at 3056 cm^-1, 3035 cm^-1, 2966 cm^-1, 2837 cm^-1, 2393 cm^-1, 1745 cm^-1, 1680 cm^-1, 839 cm^-1, and 781 cm^-1 frequencies in the region of 400 to 4000 cm^-1. Frequencies of functional groups of pure drug remained intact in physical mixture containing different polymers; these results confirmed that, no interaction occurred between DTZ HCl and polymers on mixing. A sharp melting transition of Diltiazem hydrochloride pure drug was observed at 217.49°C. In formulation and physical mixture melting endotherm at 208.44 and 211°C was observed respectively. The precompression parameters of granules for all the formulations were shown in table 2. The bulk and tapped density give an insight on the packaging and arrangement of the particles and the compaction profiles of the material. Compressibility values up to 15% usually results in good to excellent flow properties and indicate desirable packing characteristics.

Table 2: Data for precompression studies of DLT HCL factorial preliminary batches

Formulation Code	Derived Properties		Flow Properties
Formulation Code	Bulk density (g/cm3) (mean ± SD)	Tapped density (g/cm3) (mean ± SD)	Carr's index (%) (mean ± SD)	Hausner’s ratio (mean ± SD)	Angle of Repose (°) (mean ± SD)
DF01	0.56 ±0.023	0.68 ±0.025	9.68±0.14	1.11±0.004	28.6±0.37
DF02	0.50 ±0.020	0.61 ±0.044	13.79±0.23	1.16±0.013	26±0.0014
DF03	0.53 ±0.023	0.62±0.031	11.67±0.24	1.13±0.004	25±0.46
DF04	0.51 ±0.010	0.60±0.025	8.93±0.24	1.10±0.009	28±0.83
DF05	0.58±0.015	0.68±0.024	12.12±0.71	1.14±0.0014	25±0.37
DF06	0.56±0.014	0.66±0.056	12.5±0.20	1.14±0.007	25±0.20
DF07	0.56±0.016	0.65±0.024	12.5±0.58	1.14±0.007	28±0.007
DF08	0.55±0.010	0.64±0.031	11.29±0.22	1.13±0.014	28±0.06
DF09	0.55 ±0.020	0.63±0.024	12.7±0.46	1.15±0.004	27±0.83
DF10	0.54±0.016	0.62±0.018	12.9±1.46	1.15±0.013	28.6±0.06
DF11	0.53±0.015	0.64±0.056	13.11±0.22	1.15±0.002	26.3 ±0.20
DF12	0.54±0.024	0.63±0.018	12.9±0.71	1.15±0.009	30±0.020
DF13	0.55±0.018	0.65±0.044	11.29±0.58	1.13±0.011	25.5±0.17
DF14	0.54±0.018	0.63±0.044	11.48±0.58	1.14±0.011	26.4±0.17
DF15	0.55±0.018	0.66±0.044	11.29±0.58	1.17±0.011	25.5±0.17
DF16	0.55±0.018	0.64±0.044	12.7±0.58	1.14±0.011	26.3 ±0.17
DF17	0.55±0.018	0.66±0.044	11.29±0.58	1.17±0.011	26.4±0.17

Each value represents the mean ± standard deviation of 3 trails

Table 3: Post-compression parameters of DLT HCL formulations (IPQC) test for factorial batches

F. Code	Diameter (Mean±SD)(mm)	Thickness (Mean±SD) (mm)	Hardness (Mean±SD)(kg/cm²)	Friability (Mean±SD) (%)	Average weight (Mean±SD) (mg)	Drug content (Mean±SD) (%)
DF01	10±0.08	3.12 ± 0.16	6.9±0.14	0.60±0.22	297.92±0.15	101.04 ± 2.33
DF02	10±0.02	3.88 ± 0.04	6.7±0.09	0.32±0.14	300.01±0.08	99.88±0.87
DF03	10±0.09	3.92 ± 0.18	7.6±0.22	0.22±0.14	300.15±0.17	100.38±0.67
DF04	10±0.03	3.88 ± 0.08	6.3±0.16	0.26±0.26	299.87±0.04	99.14±1.59
DF05	10±0.01	3.16 ± 0.06	7.5±0.21	0.21±0.26	297.8±0.42	100.46±0.77
DF06	10±0.09	3.28 ± 0.03	7.6±0.24	0.38±0.26	299.98±0.13	97.04 ± 2.33
DF07	10±0.05	3.18 ± 0.07	6.2±0.12	0.60±0.26	298.65±0.17	101.17±1.09
DF08	10±0.04	3.88 ± 0.08	6.2±0.15	0.45±0.26	299.96±0.11	99.47±0.93
DF09	10±0.01	3.19 ± 0.04	6.8±0.18	0.51±0.14	299.85±0.10	101.86±1.30
DF10	10±0.01	3.88 ± 0.08	6.7 ±0.05	0.65±0.14	297.5±0.25	100.87 ± 1.31
DF11	10±0.02	3.98 ± 0.06	7.3±0.22	0.51±0.14	299.92±0.15	99.83±1.35
DF12	10±0.09	3.16 ± 0.16	6.8±0.14	0.50±0.35	300.01±0.15	99.69±0.53
DF13	10±0.09	3.90 ± 0.03	6.7±0.05	0.56±0.15	298.80±0.14	98.04 ± 1.06
DF14	10±0.04	3.45 ± 0.03	6.7±0.05	0.55±0.35	299.85±0.15	98.60 ± 0.60
DF15	10±0.06	3.59 ± 0.25	6.7±0.73	0.55±0.25	298.90±0.16	98.76 ± 3.32
DF16	10±0.09	3.61 ± 0.23	6.7±0.05	0.57±0.35	299±0.05	98.60 ± 3.20
DF17	10±0.06	3.60 ± 0.13	6.6±0.70	0.56±0.05	298±0.12	98.50 ± 3.35

Each value represents the mean ± standard deviation (n=3)

A 3-factor, 3-level design was used to optimize the tablets by studying the effect of independent variables (concentration of hibiscus esculentus (X₁), concentration of xanthan gum(X₂) and compression load (X₃) on the dependent variable (cumulative percentage drug released at 8 h (Q8h), time required to release drug (t_80%), and mucoadhesive strength and mucoadhesion time. Response Surface Quadratic Model, F-value of 3.82, 7.91, 3.71 and 4.05 implies the models are significant for Y₁, Y₂ Y₃ and Y₄ respectively. In Y₁ response, X₁, X₂++2+-, X₃++2+, Y₂ response X₂X₃, X₂++2+-, X₃++2+-, Y₃ response X₂++2+- and Y₄ response X₁ are significant model terms. The "Lack of Fit F-value" of 264.09, 9.64, 988.86 and 10.89 implies the Lack of Fit is significant. Adeq Precision the ratio of 5.843, 9.690, 6.151 and 6.188 for Y₃ and Y₄ respectively indicates an adequate signal. This model can be used to navigate the design space.²⁶ The application of response surface methodology yielded the following regression equations:

Y₁ (% drug release at 8 hrs) Q8 =98.74-3.7675 X₁-0.49375 X₂-1.90125 X₃-2.8875 X₁ X₂+2.3025 X₁ X₃+3.56 X₂ X₃+0.55 X₁²-5.4075 X₂²-7.8525 X₃²----(15)

Y₂ (80% Drug release at time) T80% =7.1+0.1375 X₁-0.065 X₂+0.115 X₃+0.075 X₁ X₂-0.07 X₁ X₃-0.29 X₂ X₃-0.16 X₁²+0.385 X₂²+0.57 X₃²---------------------------(16)

Y₃ (Mucoadhesive strength) =14.202+3.38 X₁-0.17125 X₂-2.80375 X₃-0.09 X₁ X₂-0.01 X₁ X₃+1.5075 X₂

X₃+2.53275 X₁²+9.36525 X₂²-0.33475 X₃²-----------(17)

Y₄ (Detachment time) =6.174118+1.1175 X₁+0.73875 X₂-0.72625 X_{3 ---------------------------------------------------------------------} (18)

Summary of ANOVA results in analyzing lack of fit (LOF) and pure error for the responses of DLT formulations is given in the table 4.

Table 4: Statistical for DLT HCL response surface models

Statistic	Y1	Y2	Y3	Y4
Std. Dev.	4.358504	0.1988	4.136486	1.236808
Mean	92.75882	7.474118	19.64353	6.174118
C.V. %	4.698749	2.659845	21.05775	20.03215
PRESS	2117.971	3.9413	1914.049	38.34609
R-Squared	0.830707	0.910458	0.82687	0.482969
Adj R-Squared	0.613045	0.795333	0.604275	0.363655
Pred R-Squared	-1.69641	-0.27566	-1.76671	0.003014
Adeq Precision	5.842746	9.690265	6.151223	6.188117

The impact of independent factors (i.e., HECM, XNG and load) at different levels on the in-vitro release rate of DLTH, mucoadhesive strength and mucoadhesive time is shown on contour and three-dimensional response surface plots in Fig.7 to 10 respectively. The three-dimensional response surface plots depict a curvilinear relationship between the factors and the response.

Table 5: Composition of checkpoint formulations, the predicted and experimental values of response variables

Code	HEC	XNG	LOAD	Response variable	Measured value	Predicted value	% prediction error	% RSD
DLTF1	49.11	44.91	4.79	Q8	88.78	88.58	0.226	0.159
				T80%	7.65	7.70	0.649	0.461
				Mucoadhesive Strength	21.05	21.35	1.405	1.001
				Detachment time	6.12	6.15	-0.488	0.346
DLTF2	49.85	44.96	4.89	Q8	87.56	87.55	0.011	0.008
				T80%	7.8	7.76	0.515	0.364
				Mucoadhesive Strength	21.8	21.65	0.693	0.488
				Detachment time	6.5	6.20	4.839	3.341
DLTF3	49.93	44.88	4.86	Q8	87.95	87.80	0.171	0.121
				T80%	7.86	7.74	1.550	1.088
				Mucoadhesive Strength	21.25	21.41	0.747	0.530
				Detachment time	6.1	6.21	1.771	1.264

Figure 11: Overlay plot indicating the region of optimal process variable setting for DLT HCL mucoadhesive tablets

Table 6: Kinetic models for optimized formulation of DLT HCL

Formula	Model	Parameters
Formula	Model	R_obs-pre	R²	R²_adj	SS	AIC	MSC	k	n
DLTH01	Zero-order	0.940	0.632	0.632	3773.94	109.066	0.223	10.214
	First-order	0.979	0.955	0.955	464.38	81.829	2.319	0.262
	Higuchi	0.992	0.981	0.981	192.67	70.393	3.198	30.241
	Korsmeyer –Peppas	0.994	0.987	0.986	129.37	67.216	3.443	33.941	0.443
	Hixson-Crowell	0.979	0.938	0.938	636.37	85.925	2.003	0.068
	Hopfenberg	0.979	0.955	0.951	464.62	83.836	2.164	0.001	595.650
	Quadratic	0.975	0.923	0.916	786.72	90.682	1.638	-0.009	0.185
Dilzem-SR	Zero-order	0.908	0.554	0.554	5226.21	113.299	0.042	10.803
	First-order	0.994	0.986	0.986	158.45	67.851	3.538	0.319
	Higuchi	0.981	0.959	0.959	476.69	82.169	2.436	32.137
	Korsmeyer-Peppas	0.986	0.971	0.969	336.35	79.636	2.631	37.648	0.422
	Hixson-Crowell	0.994	0.984	0.984	185.13	69.874	3.382	0.084
	Hopfenberg	0.994	0.988	0.987	135.66	67.832	3.539	0.041	7.118
	Quadratic	0.988	0.966	0.963	394.91	81.723	2.471	-0.011	0.214

DISCUSSION:

In the development of dosage form, crucial issue is to design an optimized pharmaceutical formulation in short time period with marginal trials. The peaks obtained in the spectra of all mixture correlates with the peaks of drug spectrum and were identical with the peaks of pure drug and physical mixture of drug and HEC mucilage and pure drug with all other excipients mixture which ensured that there was no any chemical interaction between them. DSC study confirmed that the presence of other excipients did not affect the drug nature and it was well maintained in the selected formulation. But not significant change in the position of this peak or broadening of peak in the thermogram of drug and excipients mixture was observed with respect to the thermogram of pure drug. From the powder characteristic i.e. angle of repose, compressibility index and hausners ratio it was concluded that the powder possesses good to excellent flow characteristics. The values of precompression parameters evaluated were within prescribed limits and indicated good free flowing property. The evaluation studies on granules for all the formulations were proved to be within limits and shown good derived and flow properties. All the post compression parameters are evaluated were prescribed limits and results were within IP acceptable limits. The results shows that, as the proportion of mucilage in the tablets were increased, the percent swelling increased, and the percent erosion decreased. Similar results were earlier reported for mucilage of Hibiscus Rosa Sinensis; matrix tablets formulated using pure mucilage showed greater swelling and lesser erosion as compared with the matrix tablets containing mucilage and drug.²² In the present investigation, dibasic calcium phosphate, a water-insoluble additive, was employed, with the formulation DF01 containing the highest proportion of dibasic calcium phosphate. The formulation DF01 showed the least percent swelling and highest percent erosion, while the formulation DF10 showed the least percent erosion and highest percent swelling. Thus, the matrix formed by the higher proportion of mucilage will provide more gel strength and longer diffusion path length as compared with the matrix formed with smaller proportion of mucilage. This may because as the dry polymer becomes hydrated, the mobility of the polymer chains increase, thereby increasing the hydrodynamic volume of the polymer compact, which allows the compact to swell²³. As polymer chains become more hydrated and the gel becomes more dilute, the disentanglement concentration may be reached, i.e. the critical polymer concentration below which the polymer chains disentangle and detach from a gelled matrix. These events result in simultaneous swelling, dissolution, and erosion.²⁴ From the above results it was found that polymers having high molecular weight and high viscosity exhibited higher adhesion. By this it was found that polymers having high molecular weight & high viscosity exhibited higher adhesion. The higher mucoadhesive potential observed with formulations containing only HEC as compare to XNG. This might have prevented the tablet from quick over-hydration and formation of slippery and weak mucilage that are easily removable from the mucosal surface. On the other hand, XNG might have formed weaker, more easily fractured gels due to its comparatively low molecular weight in addition to very low rate of swelling, resulting in low mucoadhesive strength. It could be due to the glass-rubbery transition which provides plasticization to hydrogel resulting in a large adhesive surface for maximum contact with mucin and flexibility to the chains for interpenetration with mucin²⁵ resulting, consequently, in the augmentation of bioadhesive strength.

DLT HCL is a highly water soluble drug with pK_aof 4.2; as a result, it is difficult to control dissolution early in acidic solution. Thus, the higher solubility of DLT HCL in 0.1 N HCl accounted for > 20 % release of the drug. The results show that more than 20% of the drug dissolved during the first 1 h in 0.1 N HCl. Formulation DF10 and DF12 tablets gave release profile close to the theoretical sustained release needed for DLT HCL. Formulation DF01 and DF05, composed of polymer 25%, failed to sustain release beyond 8 hr. Between formulation DF02 to DF13, formulated employing polymer between 28 to 33%, formulation DF02 with higher tablet hardness gave slower (t₅₀ is 3.1 hr) and complete release of DLT HCL over a period of 12 h. Overall curve fitting showed that the drug release from sustained release matrix tablets followed first order model for burst release and Korsmeyer-Peppas model for dissolution (the critical value of n = 0.765 and 0.383 suggesting both Fickian and non-Fickian diffusion). The most of the formulations showed peppas model as best fit model, also DF01 and DF05 batches shows first order and higuchi models as best fit model. The n values thus obtained ranged from 0.252 to 0.459. Formulations DF12 to DF17 exhibited anomalous (non-fickian transport) diffusion mechanism with n value ranging 0.5384 to 0.679. It can be observed from the results that the release rate data of all the batches of DLT HCL matrix tablets formulated using mucilage as the matrix fitted to the higuchi’s square root release kinetics, as indicated by highest value of r². Hence, diffusion coupled with erosion may be the mechanism of diltiazem hydrochloride release from DF04, DF06 DF09 and DF12 while other formulations release mechanism were fickian transport.

The main effects of X₁ and X₂ represent the average result of changing one variable at a time from its low to high value. From the polynomial equation of Y₁ (Eq 15), it is clearly evident that as the % of mucilage (-19.48X₁), % xanthan gum (-8.12X₂) increases, Y₁ (Drug release at 8 hrs) decreases. Moreover, coefficient of X₁ is larger than X₂. Thus, effect of independent variable X₁ is more dominant factor than X₂ at Y1. From the polynomial equation of Y2 (Eq 16), HECM and load has a synergistic effect on time required to 80% release of DLTH, whereas XNG have an antagonistic effect as described in equation 16. Furthermore, HECM and XNG exhibit a synergistic interaction on the release of DLT as can be seen from the positive value of the coefficient for this factor. Similarly, XNGM have a synergistic effect on mucoadhesive strength where as XNG and load has an antagonistic effect as described in equation 16. And HECM and XNG have synergistic effect on mucoadhesive retain time and load has inversely effect on same. The highest Q8 for DLTH ≥80% is achieved when the levels of HECM are high and when those for XNG are between a low and intermediate level. In addition, ≥80% DLTH is released when the amounts of HECM and XNG per tablet are >50 mg and <45 mg, respectively. It is evident that the antagonistic effect of XNG in the attainment of Q8 ≥80% for DLTH is achieved when the content of XNG per tablet is >45 mg, at which point an increase in the content of HECM to maximum levels of 60 mg per tablet resulted in <80% drug release. It can be concluded that a Q8 ≥80% for DLTH may be achieved when the amount of HEC and XNG per tablet is >50 mg and <45 mg, respectively, and when a high level of load is included in the formulation. From the results of regression analysis it was concluded that all the three independent factors significantly contributed to Q8 (P < 0.01). As stated earlier, increase in the concentration of HEC resulted in increased mucoadhesive strength. HEC is a long chained, anionic polymer and so its mucoadhesion is attributable to the formation of physical (including hydrogen) bonds with the mucus components.

It possesses a large number of hydroxyl groups that are responsible for adhesion. Formation of hydrogen bonds between the hydrophilic functional groups of mucoadhesive polymers and the mucus layer or the mucosal surface is a prerequisite for extensive and longer mucoadhesion. The increased sites for bond formation may explain the increase in bioadhesion with increase in its concentration. It is capable of developing additional molecular attractive forces by electrostatic interactions with negatively charged mucosal surfaces or negatively-charged sialic acid groups of the mucus network. The higher mucoadhesive potential observed with formulations containing only HEC as compare to XNG. This might have prevented the tablet from quick over-hydration and formation of slippery and weak mucilage that are easily removable from the mucosal surface. On the other hand, XNG might have formed weaker, more easily fractured gels due to its comparatively low molecular weight in addition to very low rate of swelling, resulting in low mucoadhesive strength. The optimum values of selected variables obtained using Design Expert software were 49.99 mg hibiscus esculentus mucilage, 44.97 mg xanthan gum and 4.48 ton compression load fulfilled the optimal criteria of best regulation of the release rate and bioadhesive characteristics with t80% of 7.6 h, Q8h of 89.83 % and bioadhesive strength of 22.14 g. Thus, besides controlling drug release, the formulation has definite gastro retentive potential to retain the drug in the gastric environment and upper part of intestine. The optimized formulation was found to release about 98.82% drug in sustained release manner for 12 h. In-vivo mucoadhesive study in rabbit confirmed the retention of intact tablet in stomach up to 10 hrs, which implied that the tablet may remain in upper GI tract for much longer time, indicating its suitability as GRMDDS. Stability studies of the optimized formulation under accelerated storage conditions as per ICH guidelines did not reveal any degradation of the drug and changes in the in-vitro release profiles of the optimized formulation after storage for 6 months were statistically insignificant as compared to the control sample (ANOVA, p > 0.05).

In conclusion, the results of the present study indicated that a use of Box-Behnken design, global desirability function in optimization of gastro retentive mucoadhesive dosage form containing diltiazem hydrochloride. The amount of hibiscus esculentus (X₁), the amount of xanthan gum (X₂) and compression load (X₃) has significant influence on mucoadhesion and in-vitro drug release (i.e. M and Q8). The statistical model generated using the multiple linear regression and global desirability function showed good predicative power for the experimental value. Based on the results, it can be concluded that the statistical tools like global desirability and Box-Behnken design are quite helpful in understanding the interaction(s) between different independent variables and for rapid formulation development. The DLT mucoadhesive matrices were 49.99 mg hibiscus esculentus mucilage, 44.97 mg xanthan gum and 4.8 ton compression load fulfilled the optimal criteria of best regulation of the release rate and bioadhesive characteristics with t_80%of 7.6 h, Q8h of 89.83 % and bioadhesive strength of 22.14 g.

ACKNOWLEDGEMENTS:

The authors gratefully acknowledge the faculty of Government College of Pharmacy, Amravati for providing technical assistance and facilities to carry out the research. Furthermore, this work was supported by providing gift samples of diltiazem hydrochloride from Zim Laboratories Kalmeshwar, India.

REFERENCES:

1. Box G.EP and Behnken DW. Technometrics. 2(4); 1960:455-475.

2. Minitab, Inc. Minitab user's guide 2: Data analysis and quality tools. Release 12 for Windows. State College, Penn.: Minitab. Moran, R. C. 1997.

3. Hoffman A, Stepensk D, Lavy E, Eyal S, Klausner E, Friedman M. Pharmacokinetic and pharmacodynamic aspects of gastroretentive dosage forms. Int. J. Pharm. 277; 2004:141-153.

4. Bruschi ML and Freitas O. Oral bioadhesive drug delivery systems. Drug. Dev. Ind. Pharm. 3; 2005: 293-310.

5. Ahuja A., Khar R.K. and Ali J.: Mucoadhesive drug delivery systems, Drug. Dev. Ind. Pharm. 23; 1997: 489-515.

6. Alamri MS., Mohamed AA and Hussain S. Effects of alkaline-soluble okra gum on rheological and thermal properties of systems with wheat or corn starch. Food Hydrocolloids. 30; 2013: 541-551.

7. Tomada M, Shimada K, Saito Y and Sugi M. Okra polysaccharides structure. Chem. Pharm. Bull. 28; 1980: 2933-2940.

8. Kathleen Parfit, Martindale, The complete drug reference. Pharmaceutical Press, London, Edition 32, 1999, 854-857.

9. Verma RK and Garg S. Selection of excipients for extended release formulations of glipizide through drug–excipient compatibility testing. J. Pharma. Biomed. Ana. 38; 2005: 633-644.

10. Indian Pharmacopoeia. Controller of Publications, Ministry of health and family welfare. Government of India, Delhi, Vol 1. 1996, 256-257.

11. Srivastava AK, Wadhwa S, Ridhurkar D and Mishra B. Oral sustained delivery of atenolol from floating matrix tablets-formulation and in-vitro evaluation. Drug Dev. Ind. Pharm. 31; 2005: 367-374.

12. Lamoudi L, Chaumeil JC and Daoud K. Development of gastro intestinal sustained release tablet formulation containing acryl-EZE and pH-dependent swelling HPMC K 15 M. Drug Dev. Ind. Pharm. 38; 2012: 515-520.

13. Kanagale P. Pharmaceutical Development of Solid Dispersion Based Osmotic Drug Delivery System for Nifedipine. Curr. Drug Del. 5; 2008: 306-11.

14. Korsmeyer RW. Mechanisms of solute release from porous hydrophilic polymers. Int. J. Pharm. 15(1); 1983: 25-35.

15. Wise Donald L. Handbook of Pharmaceutical Controlled Release, New York Marcel Dekker, Inc., 2000:155-179, 183-205, 255-267.

16. Lalla JK.and Gurnancy RA. Polymers for mucosal delivery swelling and mucoadhesive delivery. Ind. Drugs. 39; 2002: 270-276.

17. Patel B, Patel P, Bhosale A, Hardikar S, Mutha S.and Chaulang G. Evaluation of tamarind seed polysaccharide (TSP) as a mucoadhesive and sustained release component of Nifedipine buccoadhesive tablet & comparison with HPMC and Sodium CMC. Int. PharmTech Res. 1(3); 2009: 404-410.

18. Rajput GC, Majumdar FD, Patel JK, Patel KN, Thakor RS, Patel BP and Rajgor NB. Stomach specific mucoadhesive tablets as controlled drug delivery system-a review work. Int. J. Pharma & Bio. Res. 1(1); 2010: 30-41.

19. Lewis GA, Mathieu D and Phan‑Tann‑Luu R., Pharmaceutical Experimental Design, New York, Marcel Dekker, 1999, 265‑76.

20. Freitag G. Guidelines on Dissolution Profile comparison, Drug Inform. J. 35; 2001: 865-874.

21. ICH Topic Q1A (R2): “Note for Guidance on Stability Testing: Stability Testing of New Drug Substances and Products”, 2003, CPMP/ICH/2736/99.

22. Jani GK and Shah DP. Evaluation of mucilage of Hibiscus Rosa sinensis Linn as rate controlling matrix for sustained release of diclofenac. Drug Dev. Ind. Pharm. 34; 2008: 807-16.

23. Wan LS, Heng WS and Wong LF. Matrix swelling: A simple model describing extent of swelling of HPMC matrices. Int. J. Pharm. 116; 1995:159-168.

24. Nixon PR and Patel MV. Diffusion coefficients of polymer chains in the diffusion layer adjustments to a swollen hydrophilic matrix. J. Pharm. Sci. 86; 1997: 1293-1298.

25. Shah V and Patel R. Studies on mucilage from Hibiscus Rosa Sinensis linn as oral disintegrant. Int. J. Appl. Pharm. 2(1); 2010: 18-21.

26. Armstrong AN and James KC., Pharmaceutical experimental design and interpretation, Edition 2, Bristol, PA, Taylor and Francis Publishers, 1996,131-192.

Received on 13.04.2016 Modified on 12.05.2016

Research J. Pharm. and Tech. 2016; 9(7):817-830

DOI: 10.5958/0974-360X.2016.00156.6

Check points		Formulation			% RSD
		DLTH01	Dilzem SR
t₅₀		2.650	2.171		14.05
t₈₀		6.153	5.041		14.05
Q₈		89.25	93.46		3.25
Similarity factor (f2)		55.15
Dissimilarity factor (f1)		7.72
Statistics of similarity factors for DLTH Formulation
Overall statistics	Mean_R vs Individual_T				Mean_R vs Mean_T
	Mean			SE
f2	55.08			0.24	55.15
Is f2 ∈[50,100] between Mean_R and Mean_T					Yes
Similarity of R and T					Accept
f1	7.72			0.14	7.72
Is f1 ∈[ 0,15 ] between Mean_R and Mean_T					Yes
Similarity of R and T					Accept