Transcription of PARTICLE SIZE
1 PARTICLE SIZEDISTRIBUTIONDrugDeliveryTechnologyMa y2009 Vol9No5xxINTRODUCTIO NThe need for par ticl e size control in the manufactureofpharmaceuticalsis becomingincreas ingl y apparen t asthe pharmaceuticalindust ry at tem pts to capi talize onsome APIs with less -than-i deal sol ubili ty ,signif icant advancesin drug del ivery have been made inwhich a finely di vi ded API, wit h the concom itantincr ease in specific surface area, has res ulted inincr eased bioav ailabi lit y. Preci se parti cle size controltechnolo gies hav e al so ass is ted in the de velopm ent ofdrug delivery platfor ms for the del ivery of amedi cament to the lung. As thes e trends have occu rred,the need for highly reproducible part icle sizeassessmenttechniqueshas grown si gnificantly in thepast decade.
2 The int eres t in part icle size meas urementswil l remain high, part icul arly in view of FDA trendstoward recommending more thoroughdescript ions ofpart icle size distributions in subm is si ons in which theemphasisof a drug product clai m is bas ed in a tightlycontrolledparticle si PA RIS ON OF ME THOD S TOME AS URE PARTICLESI ZEDI STRI BUTIO NParti cle sizing of di spers ion can be accomplishedusing laser scatteringor diffract ion techniquesor bydisc centrifugetechni ques if high res olution of the sizedistr ib ution is er scatt ering requires verylow PARTICLE concen trati ons, usual ly requiringsignif icant sample dilut ion. The par ticl es in the samplemust be below 1 microm eter in si ze and free to undergoBro wnian motion. For las er dif fraction met hods,dil ut ion is again often requiredto opti mize the intensityof di ffract ed li ght at the det ect ors, though di lution requi re ment sare not as stri ngent as for scat teri ng met hodsgive wei ght -averagepart icle si ze, and al though these can beFIG URE 1 Placeboat 400 XMagnificationLightMicroscopicDeterminat ionofParticleSizeDistributioninanAqueous GelBy.
3 Phil o Morse , MS , and An dre w Lox le y, PhDFIGU RE 2 ActiveGel at 400 XMagnificationPARTICLE SIZEDISTRIBUTION mathematically con verted to number-weighteddi strib utions, the con versionscan producemisleadingar sc cen trifuge methodsrely on the abi lit yof the par ticles to move through the sampl eunder the influenceof a centripetalforcege nerated in a spinningdisc cont aining thesa mple; so the sample viscos ity mus t be lowenough that the force overcomesvi scousre sistance to PARTICLE movementin the fi some drug product formu lat ions can bedi luted without significantchange to the part icl esi ze distribut ion (allowingappropri ate sam pleconcentrationsand viscositiesfor theafo rementionedmethods)for the developmentof highl y vis cous gel-basedproducts , whoseAPI PARTICLE size distributionma y be affect ed bysi gnificantsample dilution,standardmet hodsmay not be not a pharmaceuticalcontract researchorganization(C RO), Part icleSci ences routinelydev elops such vi scoussystems for clients (especially fo r topicalpro ducts)
4 , and to enable useful par ticl e sizi ng ofsuch products,has dev eloped two met hods tode termine th e par ticle size distribut ion ofsuspendedAPI in a viscous aqueous gel thatinv olve a min imum of sample preparat ion andcan anal yze samples with broad part icl e sizedi strib methodsare based on las er di ffracti onusing a specif ically designedcell for vi scouspa ste analysi s, and image analys is of opticalphotomicrographsusing image anal ysi s soft wareto identify particlesand numerically bin themaccordingto shape and method of PARTICLE size di st rib uti on det ermi nation by opticalmicroscopyand image analys is is a technolo gy-intensivemethodre quiring the capacityto automat icall y acquire and analyze a largenumber of icl e Sciences uses a powerfulopticalmicroscopefitted with a dedicate d di gital cam era and autom ated stageand focusingmov ement, controlledby soft ware that also handl es theanalysis of th e images s enables aut omat ic col lect ion ofthe large number of image object s requiredfor stati st ical ly relev antanalysis, which includesmeasurementof lengt h, widt h, area, circledi ameter.
5 Roughness, careful selectionof obj ecti ves and cam era, the techniqueal so has a broad dynam ic range in whi ch the upper limit is severalmillimet ers at low magni fi cat ion, and the low er limit that is cl ose to 1microm et er , whi ch is correl at ed to the resolutioninherent in the us e ofwhi te light illum inat ion. A signi fi cant advantageof microscopyoverlas er di ffract ion is veri fi abl e and cal ibrat ed accuracy , as cal ibrat ion ofthe ins trument may be carried out wi th the use of NIST traceabl e st agemicrom et ers and verifi ed by the use of the monodisperselat exmicros pheres .Thi s is opposedto that av ailable for laser diffract ion,whi ch is bas ed on firs t pri nci ples , and the measurementmay only beveri fied with the us e of monodi spers e latex microspheresbut nocorrect ions or cal ibrati ons may be performedto modify theSampleMagnification# Objectsd10d50d90 Placebo Spiked w/ Spiked w/ Spiked w/ Spiked w/ LE 1 SpikedPlaceboGel LinearityResultsAdded PARTICLE Size (m)Determined Size (d50 in m)% LE 2 Accuracyof MethodAdded PARTICLE Size (m)Determined Size (d50 in m)%RSD(n=4) LE 3 Precisionof mil Thickness (25mm) mil Thickness (50mm) mil Thickness (125mm)
6 LE 4 Gel ThicknessStudyxxDrugDeliveryTechnologyMa y2009 Vol9No5 PARTICLE SIZEDISTRIBUTION instrumentresult if an inaccurateres ult is calmicroscopy, how ever, suffers in the need for a very large num ber ofobserv atio ns. ISO 13322-1containsa guide for the num ber of part iclesre quired at a 95% conf idence based on the wi dth of the parti cle sizedi strib ution. The shortcomingin thi s of course is the analyst has nokno wledge of this fact a priori, and the range of di st ributi on wi dthsde scribed by ISO 13322-1are to o few to descri be man y real-worldpa rticle size applicationseven though the largest of the distributionsre quires an extraordinarilyhigh number of preparationin the case of microscopyis ver y simp le,re quiring onl y the sandwichingof ~100 mi crolit ers of sampl e betweena slide and a cover slip with gentle pressureto achi eve a samplethicknessof ~25 size standardsare as sort ed monodispersepolystyrenelat ex standardsranging from microm eters to microm eters, andcommerciallyavailable polydisperse gl as s microbeadstandardsof 1 to10 micrometer, 3 to 30 micrometer.
7 And 10 to 100 mi crometersizera method describedherein was devel oped for the analys is of age l based on hydro xyethyl cellulos e (HEC) and cont aining (w /w) micronizedAPI. The placebo was preparedin the same fas hionas the active, but without API. Th e monodispersest andards werepre pared at appro ximatel y (w/w) by adding approximately 1microliter of 1% (w/w) dispersionsof latex mi crospheresto 10 g ofpl acebo, mixed thoroughlyby hand, centrifugedat 150 g for 30minut es to remove entrai ned ai r bubbl es. Polydispersestandardswerepreparedat (w /w) by the addi ti on of g standardbeads to 20g of placebo gel , mi xed thoroughly by hand , and centrifugedat 150 ALLE NG ES TO VA LI DA TIO N OF PH YSI CALCH ARA CTE RI ZATI ON METH ODSIf part icl e size is to be a qual ity control criteria for a givenproduct or cl ai ms of product stabil ity are to be made based on aspeci fic part icl e size, the met hod of PARTICLE size distributiondet erm inati on will have to be val idat ed.
8 The US FD A cG MP sect (e)reques ts met hods to be val idated. The accuracy,sens itivit y,speci fici ty, and reproducibi lit y of tes t methodsemployedby the fi rmshal l be es tabl ished and docum ent ed. Such validationanddocum ent at ion may be accom pli shed in accordancewith sect (a).Thes e requi rem ents incl ude a statementof each met hodus ed in tes ting the sam pl e to meet proper standardsof accuracyandrel iabi lity, as appl ied to the tes ted met hods are incapabl e at thi s ti me of API specifici ty. Anopt ical sys tem is current ly avai labl e that incorporatesnear infraredspect ros cop y as a det ecti on met hod (SyN IRgi, MalvernInstr um ent s,Ltd.). Thi s sys tem shoul d be abl e to di scern between API andexci pi ent s or impuri ties . However , for this discus sion, all part icl eswi ll be incl uded in the micros copi c and laser-sizingtechni quesregardl es s of the ident ity of the part icl en ignored in part icl e sizi ng met hod validationare theparam et ers of detect ion limi t and quant itation limit of whi ch onlythe det ect ion limit may be address ed by referenceto the inst rum entmanufacturerscl ai ms.
9 Range and linearity can be examinedin thesame way as in al l ot her techni ques in which placebo or vehi cl e isspiked wi th standard mat eri al , and measurementis carried out in theintended fashi on. In the case of aut om ated image analysistechni ques , the lower bound of the range will be defi ned by the areainscri bed by a mini mum num ber of pi xels that are capable ofcarryi ng any size/ shape inform at thi s case, the lower limi t was defined in the image analysi srout ine as a 5 x 5 pi xel square or a diameterof approximately1microm et er .Accuracyand preci sion can be assessedby thepro xi mity of the experi ment al val ues to those publishedfor thestandard mat eri al , and the coeffi ci ent of variationcalculat ed fromrepeat ed measurem ent s of the spi ked sam ple, ermedi at e preci sion or anal yst -t o-analystvariationcan be seen byDrugDeliveryTechnologyMay2009 Vol9No5xxFIGU RE 3 PlaceboWithLatexStandardsPARTICLE SIZEDISTRIBUTIONthe examinationof multiplepreparat ions of the materi al by multipleanalysts and calculationof the res ult ing RS D betw een analyst s andthe RSD of all samples pooled.
10 Res olut ion of the method, defi ned asthe ability of a techniqueto diffe rentiat e bet ween di scret emonodisperse par ticle sizes, can be addressedin the linearit the definitionis bas ed on an ins trument al abili ty tore solve monodispersepar ticle siz es that are mixed, the as sessmentbe comes more diff MIN ARY VALIDATION OF AMI CR OSCOPYTE CHNIQ UEThe deter minationof particl e size by microscopyrequires theuse of define d routines that rigorous ly control the coll ecti on ofphotomicrographsand the binary processingof the im age result gures 1 and 2 are examplepho tomi crographsof the pl acebo andactive HEC gels, respecti vely , at 400X magni ficat the counting/im age analysi s techni que forpa rticle sizing was accomplishedby mixi ng pl acebo gel wit h latexbe ads of known size, performingthe proposedsam ple preparation,and anal yzing the slide via the count ing/ acquisiti on rout ine.
