International Journal of Pharmaceutics 402 (2010) 213220

Contents lists available at ScienceDirect

International Journal of Pharmaceutics

journa l homepage: www.e lsev ier .com

Pharmaceutical Nanotechnology

Optimization of formulation and process variableindustr

Dhananj ajanRajesh Va Custom Pharm , Andhb Department oc Department o , Indiad Department o henna

a r t i c l

Article history:Received 8 JulReceived in re23 SeptemberAccepted 28 SAvailable onlin

Keywords:NanosuspensionZeta potentialBead millBoxBehnken designParticle size diAnalysis of var

entifpensis useant tn thiio of p

potential whereas milling time and milling speed has signicant effect on the particle size distributionof nanosuspension. The X-RD pattern of drug milled at high and low speed reveals no form conversionwhen compared to unmilled drug. The BoxBehnken design used in this study helped in identifying thefactors affecting the particle size distribution d(90), zeta potential and, scalability of nanosuspension.The derived polynomial equation and contour graph aid in predicting the values of selected independent

1. Introdu

As per FDto an approconsideredquivalent toadministeretage to phathe existingalso benemay providaction (LiveRapamune

Nanosusnanoparticltant. Nanosactive moie

CorresponServices, FTO-lapur, RR Dist.Tel.: +91 4044

E-mail ad(D.S. Singare).

0378-5173/$ doi:10.1016/j.stribution d(90)iance (ANOVA)

variables for preparation of optimum nanosuspension formulations with desired properties. 2010 Elsevier B.V. All rights reserved.

ction

A a nanoparticulate drug is not considered as genericved product and therefore can be patented; and areas newdrug, becausenanoparticulatedrug isnotbioe-amicrocrystalline or solubilized formof the samedrug,d at the same dosage. It also offers a unique advan-rmaceutical companies of product line extension fordrug formulations. Such product line extension can

t consumers by giving new drug dosage form whiche fewer side effects, lower doses and faster onset ofrsidge and Ruddy, 2009). The best example of this is.pensions are liquid dispersion consisting of solid druges which are stabilized by polymer and or surfac-izing has been proven to be an effective tool for anty considered as brick dust candidate. There are

ding author at: Dr. Reddys Laboratories Ltd., Custom PharmaceuticalII, Admin. Building, Survey Nos. 42, 45 and 46, Bachupally, Qutubul-Hyderabad, Andhra Pradesh, 500090, India.642379/2378; fax: +91 4044645212.dresses: dhananjays@drreddys.com, dhananjaysingare@yahoo.co.in

two main approach for formulating a nanosuspension i.e. topdown and bottom up technology. The bottom up technologyinvolves dissolving drug in a solvent which is then added tonon-solvent to precipitate the crystals. The top down approachrelies on mechanical attrition to render large crystalline par-ticles into nanoparticles. The Top Down Technologies includeMedia Milling (Nanocrystals), High Pressure Homogenizationin water (Dissocubes), High Pressure Homogenization in non-aqueous media (Nanopure) and combination of Precipitationand High-Pressure Homogenization (Nanoedege). Table 1 listssome of the FDA approved products relying on nanotechnol-ogy.

Nanosuspensions for oral route aremainly characterized bypar-ticle size distribution (PSD), zeta potential, crystalline status, anddissolution velocity and saturation solubility. A particle of less than400nm is considered to be acceptable for a nanosuspension to beadministered intravenously (Raval et al., 2006). For a physically sta-ble nanosuspension solely stabilized by electrostatic repulsion, azeta potential of 30mV is required as a minimum. In the case of acombinedelectrostatic and steric stabilization, as a roughguide lineof 20mV is sufcient (Muller et al., 2001). The crystalline struc-ture of nanosuspension is important for drugs existing in differentpolymorphic form. This is mainly conrmed by DSC, X-RD or wideangle X-ray analysis (WAXS). Dissolution velocity and saturation

see front matter 2010 Elsevier B.V. All rights reserved.ijpharm.2010.09.041ial perspective

ay S. Singarea,d,, Seshasai Marellaa, K. Gowthamrooturia, Parchuri Srinivasa Raoa

aceutical Services-Formulations, Dr. Reddys Laboratories Ltd., Bachupally, Hyderabadf Pharmaceutics, JSS College of Pharmacy, Ooty, Tamilnadu, 643001, Indiaf Pharmaceutics, Maratha Mandals College of Pharmacy, Belgaum, Karnataka, 590010f Pharmaceutics, Karpagam College of Pharmacy, Karpagam University, Coimbatore, C

e i n f o

y 2010vised form2010eptember 2010e 7 October 2010

a b s t r a c t

The objective of this study was to idcharacteristic and scale up of nanosusperspective. BoxBehnken design wapolymer to drug and ratio of surfactmilling speed. Responses measured id(90). TheANOVA test reveals that rat/ locate / i jpharm

of nanosuspension: An

b, Giriraj T. Kulkarnic,

ra Pradesh, 500090, India

i, 641021, India, India

y and optimize formulation and process variables affectingonmanufacturing process on beadmill considering industriald for this study. Formulation factors evaluated were ratio ofo drug, whereas process parameters were milling time ands study include zeta potential and, particle size distributionolymer to drug andmilling speedhas signicant effect on zeta

guruHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

214 D.S. Singare et al. / International Journal of Pharmaceutics 402 (2010) 213220

Table 1FDA approved products based on nanoparticles technology.

Product, active ingredient,company

Manufacturing approach

Rapamune

Emend , ApTricor , FenMegace , ESPar PharmacTriglide , FePharma

solubility amethods.

The stabgoverned btion of propquantity plCommonlycellulose (HK-30), andlizer includand docusaratio (steric2007). At hof the adsodecrease intial and a dthe gastrointhese electrreduced inTherefore itbilizer. Procvary based

As seen fbased on mhigh pressupension us250mL to 4lation andto increasedecrease thtime while

In ordersion, the revariablemuthis relationkeeping othexpensive,unpredictabDesignof exmethod of otionship beresources, bmodel whicand or proc

Sudhir Vand procession usingwas used instabilizer cidization prsignicant oA similar apied the formCyclosporin

that emulsication rate is highly dependent on drug concentra-tion, surfactant and co-surfactant added. Apart from this severalresearcharticles are available forbetterunderstandingofnanocrys-

rmulated by other than media milling technology. Thoughf thediamblic dnanof manh-prn thchnocan bg susrodup (Dspenlecteobjd prosusThe

er tosusptorsevelsearetersloxics parthis(BBDehnk

teria

ateri

loxics Lablcelluals Cseddy.

anosu

osusunifande u) andmixe) oper ofe sizytt

rform, Sirolimus, Wyeth Top-down, media millingrepitant, Merck Top-down, media millingobrate, Abbott Top-down, media milling,Megestrol acetate,euticals

Top-down, media milling

nobrate, Skye Top-down, high-pressurehomogenization

re generally performed using ofcial pharmacopoeial

ility and robustness of a nanosuspension is mainlyy various formulation and process variables. Selec-er steric and electrostatic stabilizer and its optimumays a major role in formulating a nanosuspension.used steric stabilizer includes hydroxypropylmethylPMC), hydroxypropyl cellulose (HPC), povidone (PVPpluronics (F68 and F127) whereas electrostatic stabi-es polysorbate (Tween-80), sodium laurylsulphate (SLS)te sodium (DOSS). A suitable working polymer to drugstabilizer) is from 0.05:1 to 0.5:1 (Kesisoglou et al.,

igh stabilizer concentrations, well above of the plateaurption isotherm, electrostatic stabilizers can cause athe diffuse layer leading to a decreased zeta poten-ecreased physical stability. Electrolytes are present intestinal tract and the contact of the nanocrystals witholytes cannot be avoided. Electrostatic stabilization isits efciency in an electrolyte containing environment.is important to nd the optimal concentration for a sta-essing factors for formulating a stable nanosuspensionon the equipment selected for manufacturing.rom Table 1majority of products existing inmarket areedia milling technique due its many advantages overre homogenization. Therefore, scale-up of a nanosus-ing bead mill with varying operating capacity fromL and more requires a better understanding of formu-process variables. A major emphasis should be giventhe solid content (API) in nanosuspension which wille processing time and in turn also decrease the dryingconverting it to a solid dosage form.to obtain best formulation and process of nanosuspen-lationship between controllable variable and qualityst be understood. The traditional method used to studyship involves changing one variable at a time, whileers as constant. This approach has been proved to belaborious and also unfavorable to x errors that arele and at times even unsuccessful. On the other handperiment (DoE) canserveasanefcient andeconomicalbtaining the necessary information to understand rela-tween variables. DoE provides not only efcient use ofut also provides a method of obtaining a mathematicalh can be used to characterize and optimize formulationess.erma et al. (2009) evaluated the effect of formulations variables affecting the preparation of nanosuspen-

tals fomost othemethe putive oftime o

Higpensiothis tehand,startinsion pscale-unanosufore se

Thetion anof nan(RSM).polyma nanoing facscale dfor a reparam

Mesuch aated indesignBoxB

2. Ma

2.1. M

MeReddymethychemicpurchathis stu

2.2. N

Nanone ismediaber. Th5.0mdolphCanadachambparticl0.2-mmwas pemicrouidization technique. A 2(51) factorial designthis study. Formulation factors such as stabilizer type,

oncentration and processing factors such as microu-essure,milling time and temperaturewere identied asr critical factors affecting stability of nanosuspension.proachwas used by Zidan et al. (2007)where they stud-ulation variables affecting nanoemulsied particles ofe A, using a Box Behnken design. The study highlights

a rate of 10and time aswas controlouter jacke

2.3. Design

Initial sclation andproducts existing in market are developed based onilling technique; there is no researchwork available inomain on the same considering the industrial perspec-suspension development; which includes the cost andufacturing.

essure homogenization relies on the forcing of a sus-rough a small gap which makes miniaturization oflogy less straightforward. Media milling, on the othere per formed by agitation of devices containing thepension and milling media. Furthermore, nanosuspen-ction by media milling is characterized by its ease ofate and Patravale, 2004), making results generated onsions in downscaled designs valuable and was there-d for this study.ective of the present study was to identify formula-ocess variables affecting the characteristic and scale uppension, on bead mill using response surface methodobjective of this study also includes optimization ofdrug ratio and surfactant to drug ratio that could giveension with ideal characteristic, to optimize process-that could affect the manufacturing process on largeopment. The objective also includes to set a guidelinech scientist to eliminate the initial screening of processwhile formulating a nanosuspension on bead mill.am was used as a model drug for this study. Responsesticle size distribution and zeta potential were evalu-study. Based on several advantages of BoxBehnken) over central composite design we decided to applyen experimental design.

ls and methods

als

am USP used in this study was manufactured by Dr.oratories Limited, Hyderabad, India. Hydroxypropyllose 6cps (HPMC) was purchased from Samsung neo., LTD., Korea and sodium lauryl sulphate (SLS) wasfrom Stepan Co., USA. Puried water USP was used in

spension preparation

pension preparation involves two main steps; the rstorm dispersion of drug and stabilizers in dispersionsecond one is particle size reduction in milling cham-niform dispersion of drug (initial particle size d(90)stabilizer in dispersionmediawas prepared using Hei-

r (Model: RZR2051 Control, Rose Scientic Ltd., Alberta,eratedat 500 rpm.This suspensionwas loaded inmillingbeadmill (Model: Lab Star 1, NetzschMill, Germany) fore reduction. The milling media used for this study wasrium-stabilized zirconium beads. Themilling operationed in a re-circulationmode with the suspension fed at

0mL/min. The bead mill was operated at specic speeddesigned by the DoE. The temperature of suspensionled duringmilling by circulating coldwater through thet.

of experiment

reening trialswere carriedout for evaluating the formu-processing aspects of nanosuspension. Various factors

guruHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

D.S. Singare et al. / International Journal of Pharmaceutics 402 (2010) 213220 215

Table 2Variables for BoxBehnken study.

Independent factors Design level

Uncoded Coded Uncoded level Coded level

Ratio of polymer to drug A 0.1:1.0 10.3:1.0 00.5:1.0 +1

Ratio of surfactant to drug B 0.02:1.0 10.04:1.0 00.06:1.0 +1

Milling time (h) C 3.0 14.5 06.0 +1

Milling speed (rpm) D 2500.0 1

like concenfactant to dof milling mcritical to gThe resultspolymer tospeed are thpotential of

Based ondesign wasingparametThe four indpolymer tospeed. All t1). The conmilling medwere kept sware was uwere designwere run imodel. A batal trials. Talevels usedtion of nanoDoE.

2.4. Particle

The parMalvern Zetimes. Thethe respons

2.5. Zeta po

The zetaMalvern Zeat least thrcalculations

Table 3Formula comp

Ingredients

MeloxicamHPMC 6cpsSLSPuried wat

qs, quantity su

2.6. Solid state characterization

2.6.1. Powder X-ray diffraction (PXRD)Initial unmilled suspension of meloxicam, nanosuspension

milled at hdried usingMumbai, In85 C; outletion pressuusing a X-rline as theage, a 40-mrange of 3

ults

fect osize

tal oulatrticlexpeted iresp

functmial

+ 1D rereprvalu

strib, responstranon ofquenary Son, hSquathe be 5.VAof thgressor inof thes thosuspillingterm2950.0 03400.0 +1

tration of drug, ratio of polymer to drug, ratio of sur-rug, solvent for nanosuspension, millingmedia, volumeedia, milling time andmilling speed were identied asive a product in nano range and with required stability.from the initial screening trials suggested that ratio ofdrug, ratio of surfactant todrug,milling timeandmillinge main factors which affects the particle size and zetathe nanosuspension.the number of factors and their level, a BoxBehnken

used to evaluate the effect of formulation and process-ersaffecting thephysicalpropertiesofnanosuspension.ependent factors identied for this study were ratio ofdrug, ratio of surfactant todrug,milling timeandmillinghese factors were operated at three levels (+1, 0 andcentration of drug, type of polymer, type of surfactant,ia, volume of milling media, solvent i.e. puried waterame for all the experiments. Design-Expert 7b1.1 soft-sed to conduct the study. A total of 26 experimentsed by the software with 2 center points. Experiments

n random order to increase the predictability of thetch size of 250g was kept constant for all experimen-ble 2 shows the independent factors and their designin this study. Table 3 list out the formula composi-suspension with respect to the ratios as mentioned in

size measurement

ticle size of nanosuspension was measured usingtasizer ZS200. Each sample wasmeasured at least threeaverage values were employed for the calculations ofe surfaces.

tential

potential of nanosuspension was measured using

3. Res

3.1. Efparticle

A toof formand pafor allpresen

Thelinearpolyno

y = 0whereable, y

Thesize di443nmthe respowerselectithe SeSummdeviatiSumoflyzingin Tabl

ANOnitudeThe replots fquacyidentiof nanand mmodeltasizer ZS200 at 250.5 C. Each sample was measuredee times. The average values were employed for theof the response surfaces.

osition of nanosuspension batches.

Quantity/batch (g)

16.01.68.00.320.96

er qs

fcient to 250g.

milling speterms. Thein Table 6.details of thvalue for th(0.20) withas predict

Thenamined by D(3) for resp

y1 (zeta potigh speed (3500 rpm) and low speed (2500 rpm) werespray dryer (Spray Dryer Model: LU 222, Labultima,

dia) under following set of condition: inlet temperature,t temperature, 60 C; feed rate, 2.5mL/min; atomiza-re 2kg/cm2. The PXRD pattern of samples was recordeday diffractometer (Bruker axs, D8 Advance) with a Cusource of radiation. Standard runs using a 40-kV volt-A current, and a scanning rate of 0.013 min1 over a 245 were used.

and discussion

f independent factors on zeta potential (y1) anddistribution (y2)

f 26 experiments were carried out to study the effection and processing factors affecting the zeta potentiale size distribution of nanosuspension. Response datarimental runs of BoxBehnken experimental design isn Table 4.onses obtained for this study are well modeled by aion of the independent variables; hence the rst orderwas used for approximating the function.

1 + 22 + 33 + 44 + D (1)

presents noise or error, represents independent vari-esents response and represents coefcient.es of response y1 (zeta potential) and y2 (particle

ution d(90)) ranges from 11.3 to 33mV and 211 topectively. The ratio of maximum to minimum for bothes y1 and y2 is 2.92 and 2.09, respectively; thereforesformation was not applied to the obtained values. Themodel for analyzing the response was done based ontial Model Sum of Squares, lack of t test and Modeltatistics. The Prob> F value of P F is less than 0.05). It alsoe signicant factors that affect the responses y1 and y2ension. For zeta potential the ratio of polymer to drugspeed (factors A and D) were identied as signicantswhereas for particle size distributionmilling time anded (factors C and D) were identied as signicant modeldetails of ANOVA for response y1 and y2 are mentionedThe multiple regression term were also analyzed. Theere vaules ismentioned inTable7. The PredR-squarede responses was found to be in reasonable agreementh the Adj R-squared value which indicates that modeled the responses values well.lmathematicalmodel in terms of coded factors as deter-esign-Expert software is shown below in Eqs. (2) andonse y1 and y2, respectively.

ential) = +21.02 5.93 A + 0.083 B + 0.80 C+2.60 D (2)

guruHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

HpHighlight

216 D.S. Singare et al. / International Journal of Pharmaceutics 402 (2010) 213220

Table 4Factor level and response data for BBD study.

Run Block Factor: A Factor: B Factor: C Factor: D Response 1 Response 2Ratio of polymer to drug (g) Ratio of surfactant to drug (g) Milling time (h) Milling speed (rpm) Zeta potential (mV) PSD d(90) nm

1 Block 1 4.8 (0) 0.64 (0) 4.5 (0) 2950 (0) 21.8 2892 Block 1 4.8 (0) 0.96 (+1) 6.0 (+1) 2950 (0) 22.7 2383 Block 1 1.6 (1) 0.64 (0) 3.0 (1) 2950 (0) 33.0 2654 Block 1 8.0 (+1) 0.64 (0) 4.5 (0) 2500 (1) 12.0 3265 Block 1 1.6 (1) 0.64 (0) 4.5 (0) 2500 (1) 25.4 4436 Block 1 4.8 (0) 0.64 (0) 4.5 (0) 2950 (0) 13.5 3347 Block 1 4.8 (0) 0.64 (0) 3.0 (1) 3400(+1) 5.63 2498 Block 1 4.8 (0) 0.64 (0) 6.0 (+1) 3400(+1) 22.6 2199 Block 1 1.6 (1) 0.96 (+1) 4.5 (0) 2950 (0) 25.0 311

10 Block 1 1.6 (1) 0.64 (0) 4.5 (0) 3400(+1) 28.3 24011 Block 1 8.0 (+1) 0.96 (+1) 4.5 (0) 2950 (0) 16.5 31112 Block 1 4.8 (0) 0.96 (+1) 4.5 (0) 2500 (1) 23.7 30913 Block 1 4.8 (0) 0.96 (+1) 3.0 (1) 2950 (0) 18.8 25014 Block 1 8.0 (+1) 0.64 (0) 6.0 (+1) 2950 (0) 19.1 28815 Block 1 8.0 (+1) 0.32 (1) 4.5 (0) 2950 (0) 15.9 35816 Block 1 8.0 (+1) 0.64 (0) 3.0 (1) 2950 (0) 17.0 41817 Block 1 4.8 (0) 0.64 (0) 3.0 (1) 2500 (1) 11.3 42418 Block 1 8.0 (+1) 0.64 (0) 4.5 (0) 3400(+1) 16.8 22519 Block 1 4.8 (0) 0.32 (1) 4.5 (0) 2500 (1) 17.8 35020 Block 1 1.6 (1) 0.32 (1) 4.5 (0) 2950 (0) 27.7 35521 Block 1 4.8 (0) 0.32 (1) 4.5 (0) 3400(+1) 23.8 23822 Block 1 4.8 (0) 0.96 (+1) 4.5 (0) 3400(+1) 24.0 24623 Block 1 4.8 (0) 0.32 (1) 6.0 (+1) 2950 (0) 23.3 23624 Block 1 4.8 (0) 0.32 (1) 3.0 (1) 2950 (0) 21.2 27725 Block 1 4.8 (0) 0.64 (0) 6.0 (+1) 2500 (1) 15.1 25126 Block 1 1.6 (1) 0.64 (0) 6.0 (+1) 2950 (0) 29.1 211

g, gram; h, hours; rpm, rotation per minute; nm, nanometer; mV, millivolts; PSD, particle size distribution; BBD, Box Behnken design.

Table 5Fit summary for zeta potential and particle size distribution.

Source Sum of squares df F value Prob> F P-value

y1 y2 y1 y2 y1 y2 y1 y2

Sequential sum of model squareLinear 511.34 58049.83 4 4 11.57 6.10 F P-value

y1 y2 y1 y2 y1 y2 y1 y2

Model 511.34 58049.83 4 4 11.57 6.10

D.S. Singare et al. / International Journal of Pharmaceutics 402 (2010) 213220 217

Fig. 1. Perturbtial).

y2 (PSD d(9

A positivsign indicatcients of A ilevel of polyB, C and Drate of respB,C andD incentration oSimilarly, thd(90) with i

The theovalues of reand were fo

A perturaffects thethat the resshows inseresponse y1a slight slopWhereas incompared tresponse y1A and D arey1. But to h

Table 7Values for regr

Sr. No.

1234

erturbation graph for effect of individual factor on response y1 (particle sizetion d(90).

ponse y1. Factors C and D were focused more for generatingtour plot for response y2.relationship between the dependent and independent

eswthe ethe eC anxed ln resotenfrompendation graph for effect of individual factor on response y1 (zeta poten-

0)) = +294.65 + 8.42 A 12.42 B 36.67 C57.17 D (3)

e sign represents a synergistic effect, while a negativees an antagonistic effect. In case of y1 negative coef-n the model refer to decreased zeta potential at highermer to drug ratio. Similarly, the positive coefcients ofindicate the increase in zeta potential with increasingective factor. Whereas for y2 the negative coefcient ofthemodel refers todecreasing thePSDd(90) as the con-f surfactant, milling time and milling speed increases.e positive coefcients of A indicate the increase in PSDncreasing concentration of respective factor.retical values were compared with the experimental

Fig. 2. Pdistribu

for res2D con

Thevariablresentshowsfactorsy1 at and C ois the pminedwill desponses using diagnostic case statistic reports (Table 8)und to be in reasonably close agreement.bation graphwas plotted to nd those factors thatmostresponse. A steep slope or curvature in a factor showsponse is sensitive to that factor. A relatively at linensitivity to change in that particular factor. In case offactor A shows a steep slope, factors C and D exhibite or a noticeable bend and factor B shows a at line.case of response y2 factors C and D shows a steep slopeo A and B. Figs. 1 and 2 represent perturbation plot forand y2. Based on ANOVA and perturbation plot factorsmost ideal for generating 2D contour plot for responseave a better understanding factor B was also evaluated

ession term.

Terms Values

y1 y2

R-squared 0.6878 0.5376Adj R-squared 0.6283 0.4495Pred R-squared 0.5444 0.2755Adeq precision 11.706 8.777

Fig. 3. Contouto drug (B) onere further elucidatedusing contour plots. Figs. 35 rep-ffect of factors A, B, C and D on the response y1. Fig. 3ffect of factors A and B on response y1 at xed level ofdD. Fig. 4 shows the effect of factorsA andDon responseevel of factors B and C. Fig. 5 shows the effect of factors Aponse y1 at xed level of factors B and D. Zeta potentialtial at the hydrodynamic shear plane and can be deter-particle mobility under an electric eld. The mobilityon surface charge and electrolyte concentration. As canr plot for effect of ratio of polymer to drug (A) and ratio of surfactantzeta potential (y1).

218 D.S. Singare et al. / International Journal of Pharmaceutics 402 (2010) 213220

Table 8Actual and predicted values of the responses.

Std order Actual value Predicted value Residual

y1 (mV) y2 (nm) y1 (mV) y2 (nm) y1 (mV) y2 (nm)

1 27.7 355 26.87 298.65 0.83 56.352 15.9 358 15.00 315.49 0.90 42.513 25.0 311 27.03 273.82 2.03 37.184 16.5 311 15.17 290.65 1.33 20.355 11.3 424 17.62 388.49 6.32 35.516 15.1 251 19.22 315.15 4.12 64.157 21.0 249 22.82 274.15 1.82 25.158 22.6 219 24.42 200.82 1.82 18.189 25.4 443 24.35 343.40 1.05 99.60

10 12.0 326 12.48 360.24 0.48 34.2411 28.3 240 29.55 229.07 1.25 10.9312 16.8 225 17.68 245.90 0.88 20.9013 21.2 277 20.13 343.74 1.07 66.7414 18.8 250 20.30 318.90 1.50 68.9015 23.3 236 21.73 270.40 1.57 34.4016 22.7 238 21.90 245.57 0.80 7.5717 33.0 265 26.15 322.90 6.85 57.9018 17.0 418 14.28 339.74 2.72 78.2619 29.1 211 27.75 249.57 1.35 38.5720 19.1 288 15.88 266.40 3.22 21.6021 17.8 350 18.33 364.24 0.53 14.2422 23.7 309 18.50 339.40 5.20 30.4023 23.8 238 23.53 249.90 0.27 11.9024 24.0 246 23.70 225.07 0.30 20.9325 21.8 289 21.02 294.65 0.78 5.6526 13.5 334 21.02 294.65 7.52 39.35

be seen from Fig. 3 the zeta potential values increases at low levelof factor A where the level of factor B is high. HPMC is a nonionicpolymer anthe particledue towhicexcellent disurface leadof nanosuspfactor levelin nanosuspleads to a relayer of the

Fig. 4. Contouon zeta potent

potential ismeasured, to a larger distance from the particle surface.Consequently the measured zeta potential is lower. In such cases

tentials of about 20mV are still sufcient to fully stabilizetem4 shed frspethisnd eobilid SLS is an anionic surfactant. At low level of factor Asurface of drug is not covered so densely with HPMC,h SLS diffuses faster to the particle surfaces; since it hasspersion properties. Adsorption of SLS onto the particles to high zeta potential value. The zeta potential valueensiondecreases at high level of factorA, irrespective ofB. At this stage due to increased concentration of HPMCension its adsorption on drug particles increaseswhichduction of themeasured zeta potential. The adsorptionstabilizer shifts the plain of shear, at which the zeta

zeta pothe sys

Fig.observmillingson forsteric aticle mr plot for effect of ratio of polymer to drug (A) and milling speed (D)ial (y1).

Fig. 5. Contouzeta potentialin combination with steric stabilization.ows the effect of factors A and D on response y1. It wasom the gure that zeta potential value increases at highed and lower polymer concentration. The probable rea-may be that due to highmilling speed the adsorption oflectrostatic stabilizer is more which increases the par-ty and zeta potential value. The zeta potential valuesr plot for effect of ratio of polymer to drug (A) andmilling time (C) on(y1).

D.S. Singare et al. / International Journal of Pharmaceutics 402 (2010) 213220 219

Fig. 6. Contour plot showing effect of milling speed (D) andmilling time (C) on PSDd(90) (y2).

increases at low level of factor A where the level of factor C is high(Fig. 5).

The effect of factors D, C, A and B on y2 is shown in Figs. 68.The effectsshown in FiA and millieffects of rais shown inlinear patted(90) of nanspeed increhigher millimay be, hig

Fig. 7. Contou(A) on PSD d(9

Fig. 8. Contou(C) on PSD d(9

impaction oenergy inpu

es.can bses as maes dudecpolyally,bea

creasnufagh molymer to drug ratio to be milled at higher milling speed areof milling speed, D and Milling time, C on PSD d(90) isg. 6. Fig. 7 represents effect of ratio of polymer to drug,ng speed, D on the PSD d(90) of nanosuspension.Thetio of polymer to drug and Milling time, C on PSD d(90)Fig. 8. In Fig. 6, the PSD d(90) response surface varies inrn with increasing milling speed and milling time. PSDosuspension decreases as the milling time and millingases. Theminimumvalue of PSD d(90) 211.0nm is at theng time andmilling speed. The probable reason for thish energy and shear forces generated as a result of the

particlAs

decreafor thiparticlin turnhigherdrasticsion onand iningmaand hihigh pr plot showing effect ofmilling speed (D) and ratio of polymer to drug0) (y2).

unsuitableried out to ospeed for iThe millingnanosuspen

3.2. Effect opattern of A

The powpension opshowed nowith initiadrug was oslight decrenanosuspenXRD patter

3.3. Optimiusing graph

Optimizels of factor plot showing effect of ratio of polymer to drug (A) and milling time0) (y2).

f the milling media with the drug which provides thet tobreak the themicroparticulatedrug intonano-sized

e seen from Fig. 7 the PSD d(90) of nanosuspensiont the lower polymer concentration. The probable reasony be, at lower polymer concentration collision of druge to high impaction ofmillingmedia is increasedwhichreases the PSD d(90) of nanosuspension. Moreover atmer concentration the viscosity of polymer increaseswhich alters or hinders the processing of nanosuspen-d mill. Issues such as increase in product temperaturee in pressure on the bead mill were also observed dur-cturingnanosuspensionwithhighpolymer to drug ratioilling speed. These states that nanosuspension withformanufacturing and scale up. But an attemptwas car-vercome this issue by operating the bead mill at lower

nitial 10min and further increasing the speed slowly.time also shows a prominent effect on the PSD d(90) ofsion even at high polymer concentration (Fig. 8).

f milling speed on the powder X-ray diffractionPI

der X-ray diffraction study of spray dried nanosus-erated at high milling speed and low milling speedsignicant shift in the main peaks when compared

l unmilled drug. The characteristic peaks for milledbserved at same 2 value as that of unmilled drug. Aase in intensity of peaks was observed with spray driedsion operated at higher milling speed. Fig. 9 shows then of spray dried nanosuspensions.

zation of formulation and processing parametersical optimization method

ation of nanosuspension was performed to nd the lev-rs AD which gives y1 in 20 to 25mV range and y2

220 D.S. Singare et al. / International Journal of Pharmaceutics 402 (2010) 213220

nmill

Fig. 10

d(90) of 350required raand 2563.0these valuewere prepaagreementmization prplot showinto get the re

4. Conclus

Optimizis a compleaffects the

sucor naer coFig. 9. XRD patterns for milled and u

appliedeters fpolym. Overlay plot for optimized parameters of nanosuspension.

400nm. Under this the model predicted y1 and y2 innge at A, B, C and D values of 3.9 (g), 0.40(g), 3.8 (h)(rpm), respectively for a batch size of 250g. By usings of factors three different batches of nanosuspensionsred. The values of observed y1 and y2 were in very closeto the predicted one. By this the reliability of the opti-ocedure was established. Fig. 10 represents an overlayg the optimized parameters suggested by DoE softwaresponses in required range.

ion

ation of nanosuspension using media milling approachx process since it involves large number of factors thatcharacteristic of nanosuspension. The RSM was been

plays a signpension. Msignicantsion. The gsweet spotphysicochecertain formmer concenmanufacturof the nanomer concenof nanosuspshow any fo

Acknowled

The authable reviewthankDr. Raunderstand

References

Date, A.A., Patcles. Curr.

Kesisoglou, F.,and bioph

Liversidge G.,com/Artic

Muller, R.H., Jmulationsfuture. Ad

Raval, J.A., PateSystems, v

Verma, S., Lanto unders185198.

Zidan, A.S., SM.A., 2007cyclosporidesign anded drug.

cessfully to optimize formulation and process param-nosuspension. From this study it was concluded thatncentration (ratio of polymer to drug) andmilling speedicant role in controlling the zeta potential of nanosus-illing time and milling speed were considered to befactors which affected the PSD d(90) of nanosuspen-raphical optimization method helped in nding theor design space to get nanosuspension with desiredmical properties. The study also helped in identifyingulation and processing parameters, such as high poly-tration and high milling speed which may affects theing of nanosuspension at higher scale. Zeta potentialsuspension was found to dependent more on the poly-tration compared to surfactant concentration. Millingension at high speed and for long period of time didntrm conversion of drug.

gements

or wishes to thank Dr. Ganesan M for providing valu-comments on the researchwork.We alsowould like tovi Pillaiwhohas always stressed for enhanced scienticing and been a rich source of inspiration.

ravale, V.B., 2004. Current strategies for engineering drug nanoparti-Opin. Colloid Interface Sci. 9, 222235.Panmai, S., Wu, Y., 2007. Nanosizingoral formulation development

armaceutical evaluation. Adv. Drug Deliv. Rev. 59, 631644.Ruddy S.B., 2009. Will nanoparticles deliver? http://www.dddmag.-le-Nanoparticles-For-Drug-Delivery-050509.aspx.acobs, C., Kayser, O., 2001. Nanosuspensions as particulate drug for-in therapy Rationale for development andwhatwe can expect for thev. Drug Deliv. Rev. 47, 319.l, J.K., Patel,M.M., 2006.NanosuspensionsasParticulateDrugDeliveryol. 4, http://www.pharmainfo.net/reviews/.b, Y., Gokhale, R., Burgess, D.J., 2009. Quality by design approachtand the process of nanosuspension preparation. Int. J. Pharm. 377,

ammour, O.A., Hammad, M.A., Megrab, N.A., Habib, M.J., Khan,. Quality by design: Understanding the formulation variables of ane A self-nanoemulsied drug delivery systems by BoxBehnkendesirability function. Int. J. Pharm. 332, 5563.

Optimization of formulation and process variable of nanosuspension: An industrial perspectiveIntroductionMaterials and methodsMaterialsNanosuspension preparationDesign of experimentParticle size measurementZeta potentialSolid state characterizationPowder X-ray diffraction (PXRD)

Results and discussionEffect of independent factors on zeta potential (y1) and particle size distribution (y2)Effect of milling speed on the powder X-ray diffraction pattern of APIOptimization of formulation and processing parameters using graphical optimization method

ConclusionAcknowledgementsReferences