Effect of Granules Properties on the In-vitro and In-vivo Performance of Ibuprofen Sustained Release Matrix Tablets

The impact of variations in the wet granulation step during the manufacturing process on the in-vitro and in-vivo performance of ibuprofen sustained release matrix tablets was investigated. Two batches were produced under different wet granulation conditions. The granules of the first batch (TI) were characterized by having a lower bulk density (0.56 glml), a higher percentage of fines (56.7% wlw) and a smaller geometric mean diameter (dg), 600 pm. While the granules of batch (T2) were characterized by having a more coherent properties, a higher bulk density (0.66 glml), a lower percentage of fines (36.9% wlw) and a larger dg, 720 pm. Three large scale production batches (BI, B2, B3) were manufactured similarly to T2 and found to have granules possessing similar properties. In-vitro tests showed that tablets of T I had a statistically significant higher release rate constant than tablets of either T2, B1 ,B2 or B3. In-vivo tests were done using T I and T2 tablets. Although T I and T2 were bioequivalent with respect to Cmax and AUC, T2 exhibited a statistically significant longer sustained release characteristics than T I (P<0.05).


Introduction
Ibuprofen (IB) is a propionic acid derivative of a non-steroidal antiinflammatory drug (NSAID) widely used for the treatment of rheumatoid arthritis and as an analgesic and antipyretic agent [I].For a short biological half life NSAIDs, a sustained release dosage forms are desirable in order to allow twice or once daily administration of the drug to reduce side effects due to high plasma concentration and to improve patient compliance.Different sustained release dosage forms for ibuprofen were proposed [2-71.One of the designs, which was successful in the formulation of sustained release IB dosage form, was the bimodal drug release pattern [8, 91.The bimodal design aimed at delivering a proportion of IB at colon, and consequently a high morning drug plasma concentration might be achieved [8].The benefit of a high early morning levels of drug in plasma is to overcome morning stiffness symptom of rheumatoid disease.It was found that the choice of matrix material, amount of drug incorporated in matrix, additives type and amounts, the hardness of the tablet, density variation and tablet shape could affect the release rate and the mechanism of release of the drug [lo].
The aim of this study was to investigate the effect of variation in granules properties on the in-vitro and in-vivo performance of IB sustained release matrix tablets.

Results and Discussion
All tablet products were found complying with the required specifications of assay, weight variation, thickness and hardness.The results of the physical properties of T I , T2, 61, 62 and 63 tablets are shown in tab.1 which indicated reproducibility of the tabletting process.

Granules properties Micoroscopical and Density Propetties of granules
Microscopical examination showed that T I granules were more irregular in shape than T2 granules.Visual inspection of granules indicated that T2 granules were more coherent than T I granules and indicated also from the bulk density of 0.56 and 0.66 glml for T I and T2 respectively.The bulk density of the granules of B1, B2, or B3 (production batches) was similar to T2 (tab.1).T I granules exhibited the highest percentage of fines (< 600pm).B1, B2 and B3 granules had similar physical properties to T2 which indicated a reproducible granulation process was attained with acceptable batch to batch variation.This was not the case with T I as the wet granulation process was different which resulted in production of granules possessing different physical properties as shown in tab.1.

Dissolution properties of granules
The dissolution results of T I and T2 granules showed T I releasing the drug faster than T2 due to the presence of higher percentage of fines, where 71.7% (21 .96%)and 55.2% (50.25%) of IB were released after 1 5 minutes of dissolution for T I and T2 respectively.After 60 minutes of dissolution T I released 94.2% (20.54%) of IB, while T2 released 86.1 % (50.35%) of IB.

Dissolution of the dosage forms
In-vitro dissolution profiles of IB from 6 different products are shown in fig. 1.

Dissolution Models
The reference product showed the highest dissolution rate.Tab. 2 shows the mathematical modeling parameters and regression data of the dissolution results.It was noticed that the dissolution data did not fit neither the zero-order nor the firstorder kinetics (? ~0.99).Furthermore, it was notable that the dissolution profiles fitted the Higuchi model (12 >0.99) indicating that within the limitation of the model, the dissolution data were consistent with a diffusional mechanism of release.However, matrix dissolution or erosion, which is an important characteristic of swellable and erodible systems is not considered in Higuchi model kinetics.Therefore, additional analysis was done using Korsmeyer-Peppas and Hixson-Crowell models to make more definitive conclusions.According to the Korsmeyer-Peppas semi-emperical exponential equation, the best overall function was an anomalous non-Fickian transport mechanism (0. Tab. 2. Mathematical model parameters and regression data of the dissolution results.

In-vitro Evaluation
The similarity factor (f,) showed significant deviations from the acceptance limits for the comparison of T I and T2 products with the reference products (R).These results indicated T I and T2 products were not similar to R. The dissolution profiles of B1, B2 and 63 were found to be similar to that of T2.f2 values were 65.1, 60.3 and 80.3 for B1, B2 and B3 respectively which indicated an acceptable batch to batch variation.However, this was not the case with T I as the metric value was found outside the recommended limits of similarity, 47.1 % (acceptance criteria is 50% or more).The difference between T I and other test products (T2, B1, B2 and B3) with regard to the time required for 100% release of IB was found to be significant (Pc0.05).The rate constant of Higuchi's model for T I was higher than that of eitherT2, B1, 82 or B3 as shown in tab. 2 indicating a faster diffusion rate for T I tablets.Furthermore, it was observed visually that upon dissolution T I tablets eroded faster than T2.The rate constants of Hixson-Crowell model showed T I had Kp value 1.5 times of eitherT2, B1, B2 or B3, an indication of faster erosion.A faster diffusion and erosion of T I tablets could be related to the physical properties of their granules as shown previously.

In-vivo study: Bioavailability of R, T I and T2
Fig. 2 shows the mean ( S D ) plasma concentration-time profiles of IB after 600mg single-dose administration of R, T I and T2 products.: time to maximum concentration; Ke : elimination rate constant; t I l 2 : elimination half life; MRTin-vivo : mean residence time in-vivo; MRTin-Vitro : mean residence time in-vitro; MDT: the mean dissolution time in-vivo = MRTin-vivo (tablets) -2.63 (MRTin-vivo for ibuprofen solution according to reference 17); HVDtSO% Cmax : half value duration, the time range which 50% of the observed maximum plasma concentration is attained; AUC12-24 : area under the plasma concentration-time curve from 12-24 hours after drug administration.
Tab. 3. Pharmacokinetic parameters (mean + SD, n=10) obtained from concentrationtime data for reference (R) and two test Products (TI and T2) after a single dose of 600 mg ibuprofen as a sustained release dosage form.

Test products versus Reference product
Peak plasma concentration attained by the products T I and T2 are significantly lower than the reference R (Pc0.05).No significant difference was observed for the values of t , , , within the three products (P>0.05).The extent of absorptions (AUCO-24 and AUCo-,, pg.h./ml) for R when compared with T I and T2 were found not statistically significantly different (P>0.05).Elimination rate constant (ke) and elimination half life (tql2) values for R and T I showed a difference that was not statistically significant.Ke and t l I 2 were significantly different for T2 which clearly reflected product-related differences in drug release as shown by its dissimilar in-vivo profiles (fig.2).The 90% confidence intervals based on parametric testing of the log-transformed data of the ratio T/R for the Cmax were 0.67-0.84for T I and 0.60-0.82for T2.They were outside the generally used acceptance criteria for bioequivalence (0.70 -1.43).Furthermore, the corresponding values of the extent of absorption (bioavailability) represented by AUCO-24 were 0.85-1.02for T I and 0.83-1.03for T2 which were within the generally used acceptance criteria for bioequivalence (0.80 -1.25).Evaluation of AUCo-, for T I and T2 showed similar results.

In-vivo study: Sustained release characteristics of R, T I and T2
In-vitro data showed T2 releasing the drug in a rate significantly slower than T I (tab.2).Despite the slower release rate of T2, the bioavailability data indicated that its extent of absorption represented by AUCO-24 and AUCo, were not significantly different from those values of T I or R and were within the acceptance range of bioequivalence requirements as described in the previous section.In order to describe the sustained release characteristics of the products MRTin-vitrol MRTinviva, MDTin-vivo, HVDtso% Cmaxl C12, % CiZ/Cmax and AUC12-24 were determined as shown in tab. 3. It is known that the higher values of these parameters represent greater sustained release performance.The mean dissolution time (MDTin-vivo) for T2 was greater than MDTin-vivo of either R or T I and showed a difference that was statistically significant (P<0.05).Furthermore, MRTin-Vitro and MRTin-vivo values of T2 were significantly higher than the corresponding values of either R or T I products (W0.05).C12, %C12/Cmax and HVDtJO% Cmax values were significantly higher (pc0.05) for T2 than R or T I .The C12 value of T I was not significantly higher than the C12 value of R (P> 0.05) although the values of HVDtSO% Cmax and % Cl2ICmax were marginally significant (P~0.05).The difference in these parameters between T2 and R or T I explained the differences in the residual AUC represented by AUC12-z4.The value of AUC12-24 in case of T2 was significantly higher than the corresponding value of R (P< 0.05) or T I (Pc0.025).The higher values of Cq2, % C121 Cmax, HVDtso% Cmax and AUC12-24 could be due to the arrival of a portion of the tablet in the colon, where it was then disintegrated and ibuprofen was dissolved and absorbed as previously pointed out [9].However, this characteristic was lost in the case of T I tablets.T I product was matrix tablet prepared from smaller particle size and less coherent granules which produced higher dissolution rate (tab. 2)and smaller MRTinwvitr0 whereas T2 matrix tablets were made from larger particle size and more coherent granules , showed slower dissolution rate (tab.2) and higher MRTin-vitro .Thus T I tablets were expected to erode or disintegrate faster in the gastrointestinal tract with larger proportion in the small intestinal region and allowing minimum proportion to reach the ascending colon.From the results described above, T2 appeared to have a more prolonged in-vivo delivery of IB than R and T I .

In-vivo study: Steady state performance
The advantage of T2 over R was shown by reporting that the morning steady-state mean plasma concentrations after administration of T2 tablets was significantly higher than that for R capsules, being 18.0 and 10.5 pglml, respectively.This was interpreted as being a result of higher C12 for T2 tablets [8].
A high morning drug plasma concentration is considered as an advantage because it is useful to overcome morning stiffness, which is characteristic symptoms of many rheumatic conditions [ I I].R and T2 products are administered twice daily, so C12 as IB plasma concentration is an important pharmacokinetic parameter to be determined.In this investigation as shown in tab.3, the C12 of T2 was significantly higher (P < 0.05) than C12 of either R or T I .However, R and T I products did not show significant difference for their C12 values (P > 0.05).When MDTin-vivo or MRTinviva was plotted versus C12 values of R, T I and T2 a linear relationship was obtained (0.90 < ?< 0.99).Thus, it could be predicted that T I would produce C12 at steady-state conditions significantly lower than Cq2 of T2.Consequently, the advantages of lower fluctuation of the steady state concentrations and higher CI 2 value attained by T2 tablets will not be achieved by the administration of T I tablets.
In conclusion, variations in the wet granulation process were reflected on the in-vitro and in-vivo sustained release characteristics of ibuprofen sustained release matrix tablets.In-vitro, T I showed a significant higher diffusion and erosion rates than T2 (Pc0.05).Although T I and T2 were bioequivalent with respect to Cmax , A u c 0 -2 ~~" ~ AUC o-, , T2 exhibited a statistically significant longer sustained release characteristics than T I (Pc0.05) as represented by the parameters CI2, MRTin-vivo, MRTin-vitro , MDTin-vivo HVDtsoo/, Cmax C121Cmax and AUCI 2-24 .

Materials
Ibuprofen; lactose monohydrate; maize starch; magnesium stearate; titanium dioxide; polysorbate 80; and talc were materials of pharmaceutical grade and supplied by the Arab Pharmaceutical Manufacturing Co., Sult, Jordan.Ammonio methacrylate copolymer, Type B, NF (Eudragit RS) granules were supplied by Rohm Pharma Polymers, Degussa, Darmstadt, Germany.Hypromellose 291 0 (hydroxypropyl methylcellulose), Methocel E l 5 Premium-29% methoxyl and 8.5% hydroxypropoxyl content and viscosity grade 15 cP was supplied by Colorcon, Kent, UK.Colloidal silicon dioxide (Aerosil200, Degussa, Germany) and Sodium starch glycolate (Primojel, Avebe America Inc., NJ, USA) were used.All chemicals and solvents were of analytical grade and supplied by E.Merck, Germany.Distilled water was used to prepare aqueous solutions and granulating agents.enb bid 300 mg ibuprofen sustained release (SR) spansules in capsules (Smithkline Beecham, UK) were purchased locally and used as reference product (R).Production of tablets 600 mg If3 sustained release tablets were produced according to the general formula as reported in tab. 4. Five different batches were prepared, T I and T2 as pilot batches; and B1, B2 and B3 as large scale production batches.The powder mixture was prepared by mixing together 16, half amount of lactose powder and hypromellose in Gral mixer and granulator (Collette, Belgium) for 10 minutes with the mixer adjusted at low speed and the chopper at high speed.required amount of the dispersion gradually in five portions to the powder mixture and mixed for 3 minutes after each addition.The speed of either mixer or chopper was adjusted at low speed setting.The other batches T2, B1, B2 and 83 were wet granulated similarly using a more diluted Eudragit RS dispersion (1 5% wlw) allowing more water to be incorporated into the wet mass.Addition was performed in 6 portions allowing mixing for 5 minutes after each addition.For all batches wet granules were then dried and milled using standard processes.Particle size analysis was done in duplicate on 100 gm of screened granules through sieves 2000, 1250, 1000, 800, 600 pm and receiver.Shaking was conducted for 15 minutes.Diluent granules made from lactose powder and starch and granulated with hypromellose 5% aqueous solution were prepared using the same equipment.
Final powder mix was then prepared by mixing drug granules with diluent granules, Primojel and talc for 15 minutes.Aerosil was then added and mixing was performed for another 5 minutes.At the end of mixing operation magnesium stearate was added and mixed for 5 minutes.The final powder mix was compressed into tablets (oblong, 19 mm x 9 mm) using 22 station rotary tabletting machine (Perfects -5, Fette, Germany) under controlled hardness (20 Kp + 10%).Film coating was done using Accela cota (Manesty, UK).Physical characteristics like tablet weight, thickness and hardness (Core) were controlled.

Quality control tests
Assay of IB tablets and capsules was done according to the USP 24 monograph.The dissolution profile of each dosage form was obtained using Erweka dissolution apparatus (Hensenstamm, Germany) with paddles rotating at 75 rpm and 900 ml of USP phosphate buffer solution (pH 7.2) heated at 37C0.For dissolution of granules, paddles were allowed to rotate at 50 rpm instead of 75 rpm and the dissolution medium was diluted with distilled water (2:l) while the pH was kept at 7.2.Such conditions allowed 100% of IB to be released in more than 60 minutes..The drug concentration was determined spectrophotometrically versus a standard solution (DU-7 Spectrophotometer, Beckman, USA) at 275 nm.

In-vivo tests
Randomized, single dose crossover studies were performed on 10 healthy male volunteers aged between 18 and 40 years over 3 treatment periods with one week washout phase after each period according to a previously published method and followed ICH guidelines [8].The dose of 600 mg IB (one tablet of either T I or T2, 2 capsules of en bid' as a reference) was given orally with 250 ml of orange juice after having a light breakfast.Withdrawal of blood samples (7ml) via a cannula inserted into a forearm vein was done immediately before the dose (0 time) and then at 0.5, 1, 1. 5, 2, 3, 4, 6, 8, 10, 12, 14, 16, 20 and 24 hours post administration.Blood samples were drawn into heparinised tubes, centrifuged at 3000 rpm for 5 minutes, and the plasma was stored frozen at -20°C until the day of analysis.Plasma analysis for IB concentrations was carried out using a validated HPLC method [8].

In-vitro data analysis
Drug dissolution from solid dosage forms has been described by different kinetic models like zero order , first order, Higuchi square root of time model, Korsmeyer-Peppas semi-empirical exponential equation model and Hixson-Crowell cubic root of the unreleased fraction of drug versus time model [ I 2,131.

Evaluation of similarity
Comparison of dissolution profiles were done using the similarity factor (f2) as adopted by FDA and EMEA (European Agency for the Evaluation of Medicinal Products) as a criterion for the assessment of the similarity between two in-vitro dissolution profiles [14, 151.FDA and EMEA suggest that two dissolution profiles are considered similar if f2 value is between 50 and 100.The test is sensitive to measurements obtained after either test or reference batch are dissolved more than 85%.Shah et al [I41 recommended that, the number of sample points be limited not more than one, once any of the product reaches 85% dissolution.

In-vivo data analysis
The bioavailability parameters of the three products were determined by a standard non-compartmental method and ANALYSIS OF VARIANCE (ANOVA) statistics were used for bioequivalence evaluations.Pharmacokinetic parameters were calculated.The maximum IB plasma concentration (Cmax, pglml) and the corresponding peak time (tmax, h) were determined by the examination of the individual drug plasma concentration-time profiles.The area under the curve to the last measurable concentration (AUCO-24, pg.hlml) and the area under the curve from 0 to 12 hours (AUCO-121 pg-hlml) or from 12 to 24 hours after administration (AUCI2-24, pg.hlml) were calculated by the linear trapezoidal rule.AUCo-, (pg.hlml) was calculated as: [(AUCO-24) + (Ct I Ke) ] where Ct is the last detectable plasma concentration and Ke, is the elimination rate constant (h-I) .Half Value Duration (HVDtSo% Cmax, h) analysis was used to evaluate the sustained release nature of the product [5].It is the time range which 50% of the observed maximum plasma level concentration is attained.% Cl2I Cmax , the percentage of the ratio of CI 2 (the plasma concentration at the end of the intended dose interval) and Cmax.This ratio provides an indicator of the peak-trough fluctuation to be expected after steadystate administration.A higher percentage also indicates a better performance as a sustained release dosage forms that are given twice daily.In-vivo mean residence time (MRTin-vivo, h), in-vitro mean residence time (MRTin-vitm, h) and in-vivo mean dissolution time of the product (MDT in-vivo) were calculated according to Banaker [I61 and Shargel and Yu [ I 71.The pharmacokinetic parameters AUC and CmaX were assumed to be log-normally distributed.Log-transformed values of these pharmacokinetic parameters were analyzed by performing ANOVA analyses using SAS statistical program.A 5% level of significance was used for all comparisons.The two one-sided tests for bioequivalence and 90% confidence intervals for the ratios of the geometric means were calculated.The recommended range of bioequivalence was 80-125% for AUC and 70-143% for C , , , .