Rights / License: Research Collection In Copyright - Non ...6563/eth... · nanosuspension),...

Research Collection

Doctoral Thesis

Production and stabilization of nanosuspensions of poorlysoluble drug substances

Author(s): Mendes Cerdeira, Ana Maria

Publication Date: 2012

Permanent Link: https://doi.org/10.3929/ethz-a-007613986

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.

ETH Library

https://doi.org/10.3929/ethz-a-007613986

http://rightsstatements.org/page/InC-NC/1.0/

https://www.research-collection.ethz.ch

https://www.research-collection.ethz.ch/terms-of-use

DISS. ETH NO. 20759

Production and Stabilization of Nanosuspensions of

Poorly Soluble Drug Substances

A dissertation submitted to

ETH ZURICH

For the degree of

Doctor of Sciences

Presented by

Ana Maria Mendes Cerdeira

Licenciatura em Ciências Farmacêuticas,

Universidade de Lisboa

Born on

09th October 1971

Citizen of Portugal

Accepted on the recommendation of

Professor Marco Mazzotti (ETH Zurich), examiner

Professor Bruno Gander (ETH Zurich), co-examiner

Zurich, 2012

Para ti mãe que me proporcionastes a possibilidade de ser quem sou com todas as dificuldades e

sacrificio que isso implicou desde que o pai nos teve de deixar tão cedo nas nossas vidas e na vida dele.

A ti devo hoje estar a escrever estas palavras nesta tese com muito amor para ti.

Para o meu pai, que descanse em paz.

V

Acknowledgments

Above all and foremost my sincere thank you to Professor Bruno Gander; he is the best supervisor one

can wish to have. He supported me in all the moments, at all the levels, from the beginning until the

very end of this work. He was more than a supervisor, he was and he is also a friend!

I am deeply greatful to Professor Marco Mazzotti for accepting me as his doctor student and for his

sharpness of thinking and straithforward way of getting things done. He has never given up of

supporting and motivating me to conclude my studies!

I want to express my gratitude to all who were involved at a certain moment of my long journey.

Without everyone of you, it would not have been possible:

- Stefan Merkle (Cilag, Schaffhausen, Switzerland) who told me that it was possible, that others had

worked and simultaneously done their thesis. His words were - go for it!

- Els Poupeye (Janssen, Beerse, Belgium) supported my thesis work and she made sure that I had

financial support for the equipment and materials.

- Professor António Almeida (School of Pharmacy, Lisbon, Portugal) who always motivated me to do

my PhD studies helped me in the first contacts with ETH, Zurich.

- Filippo Rendina (Janssen Latina, Italy), Colleen Michaels (Janssen US) and Dirk de Smaele

(Jansseen Beerse, Belgium) never stopped to motivate and support me.

- Cédric Gysel and Wolfgang Zimmermann (Cilag, Schaffhausen, Switzerland) supported me with the

equipment installation and maintenance.

- Yavor Kamdzilov, Floriano Monnico (Cilag, Schaffhausen, Switzerland) and Sarah Barthold (ETH,

Zurich) for their experimental contibutions.

- Cilag board who allowed me to to use their facilities and equipment.

- Pierre-Alain and Harry Tiemessen my good friends that were always on my side!

- Elke Walter for the invaluable scientific imput in the first years of my thesis and for being my

dearest friend always full of energy and motivation spirit.

Gil, Lu, mom and friends, I am very greatful for your patience and continuous overlasting support,

mostly in the stressful moments when I had to combine my workload at Cilag, where I have always

worked 100%, with the workload of my thesis that had to be done on my free time, including holidays!

To all of you … in all difficult moments of your life remember …. The greater the difficulty, the more

the glory in surmounting it …. Epicurus (BC 341-270) and above everything, we always learn!

Zurich, 2012, Ana Cerdeira

VII

Summary A substantial portion of new drug substances fails full development because of poor bioavailability,

which is partly caused by very low solubility and dissolution rate. Nanosuspensions are one of the

approaches to increase dissolution rate of practically insoluble drug substances, which is mediated by

the newly created high surface area. Nanosuspensions are already used in marketed products and in

products under development for different administration routes. Nonetheless, nanosuspension

technology still faces numerous formulation, processing, and stability challenges. Main problems that

often occur during nanosuspension manufacturing and storage are particle agglomeration and crystal

growth, both of which are caused by the increased surface energy. To minimize such particle size

changes, primary particles need to be adequately stabilized by ionic or non-ionic surfactants and/or

polymers, which provide electrostatic repulsion or steric hindrance against particle-particle attraction

(Chapter I).

For enhancing the efficiency of nanosuspension development it is crucial to define the most influential

factors affecting the critical quality attributes (CQAs) of such formulations. In this thesis, we studied

the importance of physico-chemical drug substance characteristics, as well as formulation and

manufacturing process parameters that affect nanosuspensions CQAs, focusing mainly on particle size

and dissolution. With this objective, three practically water-insoluble drug substances were selected: the

anti-fungal miconazole (MIC) and itraconazole (ITR) and the anti-viral etravirine (ETR). The drug

substances were nanoground by wet media milling. With miconazole, we studied the effect of type and

amount of surface active and polymeric excipients on achievable nanoparticle size and particle size

stability (Chapter II). Among the different types of stabilizers tested (povidone, poloxamers, cellulose

ethers; cationic and anionic surfactants), the mixture of hydroxypropylcellulose (HPC) and sodium

dodecyl sulfate (SDS) stabilized best the MIC nanosuspensions. High amounts of drug substance (up to

20%) and HPC (up to 5%) provided more efficient nanogrinding for MIC. Optimal SDS-concentration

(0.0125 - 0.05%, w/w) was a compromise between adequate MIC surface charge and wetting, on the

one hand, and crystal growth, on the other hand as both MIC solubility and zeta-potential increased

with the increase of SDS. We also found that not only the interaction drug substance / stabilizers was

important, but also the interaction between HPC and SDS as there is competitive displacement of

adsorbed HPC by increasing SDS concentration above the critical aggregation concentration (CACSDS-

HPC 0.05%). Storage of optimized nanosuspensions at 5 °C for up to 6 months caused only minor

changes in particle sizes, whereas storage at 25 °C resulted in particle agglomeration and crystal

growth.

In Chapter III, we studied nanogrinding process parameters (e.g., stress number; stress energy of the

grinding beads; specific energy input) for nanogrinding MIC, ITR, and ETR. While the three drugs

VIII

possess differing physico-chemical properties, their elastic and plastic properties were similar.

Therefore, the millability and particle size stability depended primarily on adequate electrostatic and

steric stabilization of the nanoparticles by SDS and HPC. The extent of particle size reduction of all

three drug substances depended mainly on the specific energy input of the grinding process (defined by:

grinding time; grinding bead size; stirrer tip speed), which was found optimal at ~ 15 MJ/kg. Under the

provision of proper particle stabilization the three drug substances could be milled to a mean size of ~

130 nm, with 90% of all particles being smaller than ~ 250 nm.

To gain deeper insight into the stabilizing mechanisms of the most efficient excipient mixture used

(SDS and HPC), we quantified the adsorption of both excipients on the drug nanoparticles. For this, a

near-infrared (NIR) method was established to quantify simultaneously SDS and HPC in the suspension

media; by calculation, the amounts of adsorbed excipients were then determined (Chapter IV). Second

derivative of NIR signals was used to establish calibration curves in the concentration ranges of interest

for SDS (0.03 - 0.3%) and HPC (0.75 - 7.5%). The suitability of the NIR method was verified by

evaluating the linearity, accuracy, precision, and specificity. While SDS adsorbed to similar extent on

the three drug substances, i.e., up to 122 µg/m2 (4.2 x 10-7 mol/m2) with increasing SDS concentration

(0.05 - 0.2%, w/w), HPC adsorption was approximately threefold higher on ITR (2100 - 2300 µg/m2)

and four to five times higher on ETR (3100 - 3500 µg/m2) than on MIC (700 - 800 µg/m2). This data

confirmed the higher surface hydrophobicity of ITR and ETR (Chapter V).

Transforming nanosuspensions into powders is one way to maintain stability during storage, although

drying can also affect negatively nanoparticle size and redispersibility. Upon spray- or freeze-drying the

nanosuspensions, the itraconazole particles agglomerated to a larger extent than did the miconazole

particles, resulting in lower dissolution rate. In presence of mannitol or microcrystalline cellulose

(MCC) (drug : excipient = 1:1, w/w), the dissolution rate of both nanoground drug substances was

enhanced (Chapter V). While for MIC both spray- and freeze-drying yielded fast dissolving powders

(twice of the amount of MIC was dissolved after 10 to 20 min as compared to the coarse

nanosuspension), spray-drying was more efficient for ITR as only very little agglomerates were

generated, which resulted in fast dissolution (60% of ITR dissolved in less than 10 min, as compared to

30 to 45 min for the same amount of coarse suspension). Freeze-drying of the itraconazole

nanosuspension revealed to be more complex. The increased amount of MCC (drug : excipient = 1:2,

w/w) to limit the amount of aggregate resulted in an increased dissolution rate in the first 10 - 20 min

with 55% of ITR dissolved in 10 min versus 28% for the coarse suspension.

In conclusion, this thesis provides useful new insight into nanosuspension formulation and processing.

The experiments provide new tools for developmental work, and the results improve our understanding

of critical parameters affecting nanosuspensions CQAs. As an outlook, during the course of this work, I

have always felt that there is a lack of fundamental understanding of the key mechanisms governing

IX

nanosuspensions to allow for the development of reliable and accurate first-principles models of

nanosuspensions and related processes. It is obvious that such models would be extremely useful in

research and in industrial practice. I believe that their development is one of the important next research

steps in this field.

XI

Zusammenfassung

Ein beachtlicher Anteil neuer Arzneistoffmoleküle besitzen eine sehr geringe Wasserlöslichkeit und

Lösungsgeschwindigkeit, was zu ungenügender Bioverfügbarkeit und somit zum Scheitern des

Arzneistoffs im Laufe der Produkteentwicklung führen kann. Die Formulierung von Nanosuspensionen

stellt eine effiziente Möglichkeit dar, die Lösungsgeschwindigkeit von praktisch wasserunlöslichen

Arzneistoffen substanziell zu erhöhen. Die erhöhte Lösungsgeschwindigkeit wird durch die stark

vergrösserte Oberfläche der nanovermahlenen Arzneistoffpartikel erzielt, die ca. 100 – 200 nm gross

sind. Als Nanosuspensionen formulierte Medikamente für verschiedene Anwendungswege (peroral,

parenteral) gibt es bereits seit mehreren Jahren auf dem Markt und neue Produkte sind in Entwicklung.

Nichtsdestotrotz sind in der Nanosuspensionstechnologie immer wieder neue Herausforderungen der

Formulierung, Prozesse und Stabilität der Endzubereitung zu bewältigen. Die Hauptschwierigkeiten

liegen dabei meist in Partikelagglomeration und –aggregation sowie im Kristallwachstum, welche

aufgrund der hohen Oberflächenenergie der Nanopartikel während der Herstellung und Lagerung

auftreten. Um solches Grössenwachstum der Primärpartikel zu minimieren, müssen diese durch

geeignete oberflächenaktive und/oder polymere Hilfsstoffe stabilisiert werden; geeignete Hilfsstoffe

vermitteln ionische Abstossung zwischen den Partikeln und/oder eine sterische Barriere, um

übermässige Partikelannäherung oder das Auskristallisieren von gelösten Molekülen auf der

Partikeloberfläche zu verhindern (Kapitel I).

Die effiziente Entwicklung von Nanosuspensionen erfordert die Kenntnis der Einflussgrössen, welche

kritische Qualitätsattribute solcher Formulierungen bestimmen. Das Ziel dieser Dissertation bestand

darin, Zusammenhänge zwischen physikalisch-chemischen Arzneistoff- und Hilfsstoffeigenschaften

sowie Formulierungs- und Prozessparametern einerseits und kritischen Qualitätsattributen von

Nanosuspensionen andererseits, insbesondere Vermahlbarkeit der Arzneistoffe, erzielbare

Partikelfeinheit und Auflösungsgeschwindigkeit, zu untersuchen und besser zu verstehen zu. Für die

Untersuchungen wählten wir drei praktisch wasserunlösliche, hydrophobe Arzneistoffe: die

Antimykotika Miconazol und Itraconazol, sowie das Virostatikum Etravirin. Die Arzneistoffe wurden

mittels Nassmahlung bis zu einer mittleren Partikelgrösse von 120 – 150 nm verkleinert. Am Beispiel

von Miconazol untersuchten wir zuerst den Einfluss verschiedener oberflächenaktiver und polymerer

Hilfsstoffe auf dessen Vermahlbarkeit, erzielbare Partikelfeinheit und Partikelgrössenstabilität (Kapitel

II). Von den untersuchten Hilfsstoffen (Povidon, Poloxamere, Celluloseether; kationische und

anionische Tenside) stabilisierte eine Mischung von Hydroxypropylcellulose (HPC) und Natrium-

dodecylsulfat (SDS) die Miconazol Nanosuspensionen am besten. Hohe Konzentrationen an Arzneistoff

(bis 20%) und HPC (bis 5%) förderten die Nassvermahlungseffizienz. Der optimale SDS-

Konzentrationsbereich (0.0125 – 0.05%, m/m) war ein Kompromiss zwischen Vermittlung von

XII

genügender Benetzung und Oberflächenladung der Miconazol Partikel einerseits und Beschränkung des

Kristallwachstums andererseits; höhere SDS-Konzentrationen vermittelten nicht nur erhöhte

Oberflächenladung, sondern begünstigen auch die Löslichkeit und damit das Kristallwachstum von

Miconazol. Für die nanopartikel-stabilisierenden Effekte waren nicht nur die Interaktionen zwischen

Arzneistoffpartikel und Hilfsstoffen von Wichtigkeit, sondern auch jene zwischen SDS und HPC; wir

konnten zeigen, dass die Erhöhung der SDS-Konzentration über die sogenannte kritische

Aggregationskonzentration hinaus (CACSDS-HPC 0.05%) zu einer kompetitiven Verdrängung des

adsorbierten HPC führt, was die Stabilität der Nanopartikel beeinträchtigte. Optimierte

Nanosuspensionen konnten ohne signifikante Partikelgrössenveränderungen während 6 Monaten bei 5

°C gelagert werden, während Lagerung bei 25 °C zu Partikelagglomerationen und Kristallwachstum

führte.

Die für Miconazol optimierten Hilfsstoffe und Hilfsstoffkonzentrationen erwiesen sich auch für

Itraconazol und Etravirin genügend geeignet, um mit vergleichbaren Nanosuspensionsformulierungen

der drei Arzneistoffe Prozessparameter der Nassvermahlung (Stresszahl; Stressenergie der Mahlperlen;

Spezifischer Energieeintrag) zu untersuchen (Kapitel III). Die vergleichbaren mechanischen

Eigenschaften der drei Arzneistoffe widerspiegelten sich in deren ähnlichen Vermahlbarkeit. Diese hing

fast ausschliesslich von der elektrostatischen und sterischen Stabilisierung der Partikel durch SDS und

HPC ab, d.h. von den physikalisch-chemischen Eigenschaften der Partikel, und nur in geringem Masse

von Prozessparametern. Andererseits wurde das Ausmass der Vermahlung der drei Arzneistoffe

hauptsächlich durch den spezifischen Energieeintrag definiert, der durch die Mahlzeit, die Grösse der

Mahlperlen sowie die Mahlstiftgeschwindigkeit bestimmt war. Der optimale Energieeintrag betrug ca.

15 MJ/kg für alle drei Arzneistoffe. Unter der Voraussetzung geeigneter Partikelstabilisierung konnten

alle drei Arzneistoffe auf volumenbezogene mittlere Partikeldurchmesser von ca. 130 nm gemahlen

werden (volumenbezogen besassen 90% aller Partikel Durchmesser von < 250 nm; d90 = 250 nm).

Im Bemühen, die partikel-stabilisierenden Mechanismen der effizientesten Hilfsstoffe (Mischung von

SDS und HPC) besser zu verstehen, quantifizierten wir deren Adsorption an die Arzneistoffpartikel.

Dazu haben wir eine Nahinfrarot-Methode (NIR) entwickelt, welche die gleichzeitige Quantifizierung

von SDS und HPC in den Nanosuspensionen ohne weitere Probenaufbereitung ermöglichte. Aus der

Differenz von Hilfsstoffkonzentrationen in den Suspensionsmedien mit und ohne

Arzneistoffnanopartikel wurde die Menge adsorbierter Hilfsstoffe ermittelt (Kapitel IV). Mit Hilfe der

zweiten Ableitung der NIR-Signale wurden Kalibriergeraden in den relevanten

Konzentrationsbereichen von SDS (0.03–0.3%) und HPC (0.75–7.5%) erstellt. Die Validität der NIR

Methode wurde anhand von Linearität, Genauigkeit, Präzision und Spezifizität der Messungen

aufgezeigt. Während SDS mit steigender Konzentration (0.05 - 0.2%, m/m) in ähnlichem Masse an die

drei Arzneistoffe adsorbierte, nämlich bis ca. 122 µg /m2 (4.2 x 10-7 mol/m2), adsorbierte dreimal mehr

XIII

HPC an Itraconazol (2100-2300 µg/m2) und vier- bis fünfmal mehr an Etravirin (3100-3500 µg/m2) als

an Miconazol (700-800 µg/m2). Diese Ergebnisse bestätigten die höhere Oberflächenhydrophobie von

Itraconazol und Etravirin gegenüber jener von Miconazol (Kapitel V).

Eine sehr effektive Möglichkeit die Lagerstabilität von Nanosuspensionen zu gewährleisten liegt in der

Umwandlung in eine Trockenform, auch Trockensuspension genannt. Jedoch kann eine solche

Umwandlung Partikelagglomeration und –aggregation verursachen, was die Redispergierbarkeit der

Zubereitung beeinträchtigt. Mit den verwendeten Trocknungsverfahren (Sprühtrocknung und

Gefriertrocknung) agglomerierte Itraconazol in stärkerem Masse als Miconazol, was die

Auflösungsgeschwindigkeit von Itraconazol verlangsamte. Unter Zugabe von Mannitol oder

mikrokristalliner Cellulose (MCC) (Arzneistoff : Hilfsstoff = 1:1, m/m) resultierte für beide

Arzneistoffe eine erhöhte Auflösungsgeschwindigkeit (Kapitel V). Sowohl die sprühgetrockneten wie

auch die gefriergetrockneten Miconazol Formulierung (mit Mannitol oder MCC) lösten sich innerhalb

von ca. 20-30 min vollständig auf, wogegen die analog getrocknete grobdisperse Formulierung (ohne

Hilfsstoffe) nur ca. 20-40% dieser Menge freisetzen konnten. Für Itraconazol erwies sich die

Sprühtrocknung als geeigneter als die Gefriertrocknung, da weniger Agglomerate gebildet wurden;

sprühgetrocknete Itraconazol-Formulierung löste sich in 10-20 min vollständig auf (für 75% Auflösung

der grobdispersen Suspension waren 60 min notwendig). Die erfolgreiche Gefriertrocknung der

Itraconazol-Nanosuspension erforderte eine Erhöhung des Mannitol- oder MCC-Anteils auf ein

Arzneistoff : Hilfsstoff Verhältnis von 1:2 (m/m), um die Bildung von Aggregaten zu vermeiden und

eine hohe Auflösungsgeschwindigkeit zu gewährleisten (55% aufgelöst in 10 min gegenüber 28% mit

der grobdispersen Suspension).

Zusammenfassend eröffnet diese Dissertation neue und für die pharmazeutische Entwicklung relevante

Erkenntnisse über Formulierung und Prozesse von Arzneistoff-Nanosuspensionen. Die Arbeit

beschreibt auch neue Methoden, um die Entwicklung von Nanosuspensionen zu unterstützen und

entscheidende Parameter und deren Bedeutung für kritische Qualitätsattribute der Nanosuspensionen

besser zu verstehen. Im Sinne eines Ausblicks sei jedoch festgehalten, dass die Schlüsselmechanismen

der Nanosuspensionsformulierung nach wie vor zu wenig gut verstanden werden, um verlässliche und

genaue Grundprinzip-Modelle von Nanosuspensionen und zugehörigen Prozessen zu entwickeln.

Solche Modelle wären sowohl für die Forschung wie auch industrielle Entwicklung extrem hilfreich.

Ich bin überzeugt, dass die Erarbeitung solcher Modelle ein wichtiger nächster Schritt in diesem Gebiet

sein sollte.

XV

Table of Contents Abbreviations ................................................................................................. XIX Chapter I ............................................................................................................. 23 Introduction ........................................................................................................ 23

1. Rational of developing drug nanosuspensions .......................................................................... 23 2. Nanosuspensions: Formulation and manufacturing .................................................................. 25 3. Nanosuspension applications .................................................................................................... 27 4. Structure of the thesis ................................................................................................................ 31

Chapter II ........................................................................................................... 37 Miconazole nanosuspensions: Influence of formulation variables on particle size reduction and physical stability ................................................................ 37

1. Introduction ............................................................................................................................... 37 2. Materials and Methods .............................................................................................................. 39

2.1. Materials .......................................................................................................................... 39 2.2. Production of nanosuspensions ....................................................................................... 39 2.3. Design of experiments (DOE) ......................................................................................... 40 2.4. Particle size distribution by laser light diffraction ........................................................... 40 2.5. Optical microscopy .......................................................................................................... 41 2.6. Determination of contact angle........................................................................................ 41 2.7. Quantification of miconazole solubility in the nanosuspensions .................................... 41 2.8. Viscosity measurement of the nanosuspensions .............................................................. 42 2.9. Zeta-potential measurement of the nanosuspensions ...................................................... 42 2.10. Particle size stability during storage ................................................................................ 42

3. Results ....................................................................................................................................... 42 3.1. Excipient screening for miconazole nanogrinding .......................................................... 42 3.2. Optimization of miconazole nanogrinding ...................................................................... 46 3.3. Stability of miconazole nanosuspensions during storage ................................................ 50

4. Discussion ................................................................................................................................. 52 5. Conclusions ............................................................................................................................... 55

Chapter III .......................................................................................................... 59 Role of milling parameters on nanogrinding of drug substances of similar mechanical properties ........................................................................................ 59

1. Introduction ............................................................................................................................... 59 2. Materials and Methods .............................................................................................................. 61

2.1. Materials .......................................................................................................................... 61 2.2. Surface Morphology of Unprocessed Drug Particles ...................................................... 61 2.3. Mechanical Properties of the Drug Substances ............................................................... 62 2.4. Miconazole, Itraconazole, and Etravirine Formulations for Nanogrinding ..................... 63 2.5. Mill and Grinding Media ................................................................................................. 63 2.6. Particle Size Measurement .............................................................................................. 64 2.7. Viscosity of the Nanosuspensions ................................................................................... 64

3. Results ....................................................................................................................................... 64 3.1. Particle Size and Surface Morphology of the Drug Particles .......................................... 64 3.2. Young’s Modulus and Plastic Deformation of Miconazole, Itraconazole, and Etravirine Drug Particles ............................................................................................................................... 65 3.3. Effect of Milling Parameters on Miconazole Particle Size Reduction and Nanosuspension Viscosity ....................................................................................................................................... 66 3.4. Importance of the SDS Concentration in the Formulations for Particle Size Reduction and Nanosuspension Viscosity ............................................................................................................ 72

XVI

3.5. Effect of Milling Parameters on Particle Size Reduction of Miconazole, Itraconazole, and Etravirine ...................................................................................................................................... 76

4. Discussion ................................................................................................................................ 78 5. Conclusions .............................................................................................................................. 82

Chapter IV ......................................................................................................... 87

Simultaneous quantification of polymeric and surface active stabilizers of nanosuspensions by using near-infrared spectroscopy .................................. 87

1. Introduction .............................................................................................................................. 87 2. Materials and Methods ............................................................................................................. 88

2.1. Preparation of miconazole nanosuspensions ................................................................... 89 2.2. Nanosuspensions dispersed phase characterization: particle size and surface area, particle morphology, and ζ-potential ......................................................................................................... 89 2.3. Nanosuspensions continuous phase characterization: solubility of drug substance and surface tension .............................................................................................................................. 90 2.4. Quantification of HPC and SDS adsorbed to nanoparticles in nanosuspensions............. 94

3. Results ...................................................................................................................................... 95 3.1. Miconazole nanoparticle size, ζ-potential, and solubility ................................................ 95 3.2. Critical micellar concentration of SDS and SDS–HPC aggregation concentration ......... 95 3.3. NIR spectrometry for quantification of SDS and HPC in miconazole nanosuspensions 96 3.4. Quantification of SDS and HPC adsorbed onto miconazole nanoparticles in nanosuspension ........................................................................................................................... 101

4. Discussion .............................................................................................................................. 103 5. Conclusions ............................................................................................................................ 106

Chapter V ......................................................................................................... 111

Formulation and drying of miconazole and itraconazole nanosuspensions111 1. Introduction ............................................................................................................................ 111 2. Materials and Methods ........................................................................................................... 113

2.1. Materials ........................................................................................................................ 113 2.2. Nanogrinding of the drug substances ............................................................................. 114 2.3. Spray-drying and freeze-drying of coarse drug suspensions and nanosuspensions ....... 114 2.4. Particle size characterization ......................................................................................... 115 2.5. Zeta-potential of the nanosuspensions ........................................................................... 116 2.6. HPC and SDS adsorption onto miconazole and itraconazole nanoparticles .................. 116 2.7. Determination of drug substance content of the dried products .................................... 116 2.8. Determination of discriminative dissolution media ....................................................... 117 2.9. Dissolution testing of the dried powders ....................................................................... 118 2.10. X-ray powder diffraction ............................................................................................... 118 2.11. Statistics ......................................................................................................................... 118

3. Results .................................................................................................................................... 119 3.1. Excipient screening for efficient nanogrinding and stabilization of miconazole and itraconazole nanoparticles........................................................................................................... 119 3.2. Effect of the amounts of drug substance, HPC, and SDS on miconazole and itraconazole particle size ................................................................................................................................. 122 3.3. HPC and SDS adsorption onto the drug nanoparticles .................................................. 123 3.4. Spray-dried and freeze-dried drug suspension formulations ......................................... 125

3.4.1. Drug substance content.............................................................................................. 125 3.4.3. X-ray diffraction ........................................................................................................... 128

3.5. Dissolution behaviour of the spray-dried and freeze-dried products ............................. 129 4. Discussion .............................................................................................................................. 132 5. Conclusions ............................................................................................................................ 137

Chapter VI ....................................................................................................... 143

XVII

Conclusions and Outlook ................................................................................ 143 1. Conclusions ............................................................................................................................. 143 2. Outlook .................................................................................................................................... 150

Appendix I ........................................................................................................ 157 Supplementary information for chapter V ................................................... 157 Appendix II ....................................................................................................... 167 Supplementary information for chapter VI .................................................. 167

Abbreviations

XIX

Abbreviations

AFM Atomic Force Microscopy

ANOVA Analysis of Variance

AUC Area under the plasma concentration-time curve

BALB/c mice The most widely used inbred strains used in animal experimentation –

immunodeficient (lacks thymus - not able to produce T-Cells)

BCS Biopharmaceutical classification

BCS class II Low solubility and high permeability

BCS class IV Low solubility and low permeability

BK Benzalkonium chloride

C Concentration of drug substance in suspension

CAC Critical Aggregation Concentration

CI Confidence interval

CMA Critical Material Attribute

Cmax Maximum (peak) plasma drug concentration

CMC Critical micellar concentration

CPP Critical Process Parameter

CQA Critical Quality Attribute

dB Diameter of grinding beads

DmB Wear of the grinding beads at the given time of milling

DMSO Dimethyl sulfoxide

DOE Design of Experiments

DS Sodium docusate

D50 Mean particle diameter

D90 90% undersize particle diameter

D [4,3] Volume weighted mean

D [3,2] Surface area weighted mean

F (chapter IV) Force acting on the balance

FT-NIR Fouriertransform near Infrared

HPC Hydroxypropylcellulose

HPC-LF Hydroxypropylcellulose Mw 95,000

HPC-EF Hydroxypropylcellulose Mw 80,000

HPMC Hydroxypropylmethylcellulose

Abbreviations

XX

ICH International Conference Harmonization

ITR Itraconazole

E Young’s modulus or modulus of elasticity

EMA (EMEA) European Medicine Agency

EPR Enhanced permeability and retention effect

ETR Etravirine

i.m. Intra-muscular

i.v. Intravenous

Ka Acid dissociation constant

L (chapter III) Applied load in atomic force microscopy

L (chapter IV) Wetted length

m Mass of drug substance

n Stirrer speed

N Stress number of grinding beads

N0 Power consumption of the unloaded mill

Nt Power consumption at time t of milling

MIC Miconazole

MCF-7 Human breast adenocarcinoma cell line

MCC Microcrystalline cellulose

MPS Mononuclear Phagocyte System

NIR Near Infrared

PASG Pharmaceutical Analytical Sciences Group

PBS Phosphate buffered solution

PEG Polyethylenoglycol

PI3-K Phosphatidylinositol 3-kinase

PK Pharmacokinetics

pKa −log Ka

PEO Polyethylene oxide

PPO Polypropylene oxide

PVDF Polyvinylidene difluoride

PVP Povidone

R Radius of atomic force microscope indenter

RMSECV Root Mean Square of Cross Validation

RSD Relative standard deviation

s.c. Subcutaneous

Abbreviations

XXI

SDS Sodium dodecyl sulphate

SD Standard deviation

SEB Stress energy of grinding beads

SEI Total energy input relative to the total mass

SEM Scanning electronic microscopy

SGF Simulated gastric fluid

SSA Specific surface area

t Milling time

TPGS d-α-tocopherol polyethylene glycol 1000 succinate

w/w Weight fraction

v/v Volume fraction

Greek symbols

δ Indentation depth in atomic force microscopy measurements under

elastic deformation

ε Porosity of the bulk of beads

ϕB Volume fraction of the grinding beads relative to the volume of the

grinding chamber

ρB Density of grinding beads

σ Surface or interfacial tension

θ Contact angle

ν Rotational speed of the stirrer tip

νs Poisson’s ratio (for atomic force microscopy measurements)

Abbreviations

XXII

Chapter I

Page 23 of 182

Chapter I

Introduction

The aim of this introduction is to place the investigations of this Ph.D. study in the context of current

pharmaceutical industry strategy to develop and manufacture new medicinal products, particularly, drug

nanosuspension formulations

1. Rational of developing drug nanosuspensions

The current processes of developing new medicinal products are insufficiently efficient. For every

5,000–10,000 molecules that enter the research and development pipeline, approx. 250 molecules enter

the preclinical program, five are tested in clinical trials, and only one receives approval for market

introduction (Shah et al, 2011). Estimates for the average cost of bringing a new drug substance to

market range between 0.8 and 2 billion $; main expenditures are caused when a new molecule fails at

late pre-clinical stage and by the increasing costs of phases 1 to 3 of clinical studies (Orloff et al.,

2009).

From discovery to marketing and post-marketing stages, the development of a medicinal product must

fulfil the current requirements of health regulatory authorities. To ensure safety, efficacy and quality,

the formulation and manufacturing process must be stable and robust, and provide medicinal products

with consistent critical quality attributes such as consistent dose, drug dissolution and storage stability.

The increasing use of combinatorial chemistry, phage display libraries, and high-throughput screening

has facilitated and speeded up greatly the synthesis and in vitro bioactivity testing of new molecules

(Chaubal, 2004). Introduction of new molecules with appropriate biological activity into pre-clinical

formulation development often causes major challenges (Shah et al., 2011), because of incomplete

physico-chemical characterization, unfavourable properties (e.g., poor solubility and poor

permeability), limited availability, and short development timelines. Early drug candidates often vary

considerably in particle size, crystallinity and polymorphic form, as their manufacturing process and

physical properties have not yet been optimized (Maas et al., 2007).

Chapter I

Page 24 of 182

Two important factors in the development of medicinal products are drug solubility and permeability, as

they determine to a large extent the bioavailability of a drug substance (Shah et al, 2011). Actually, the

number of relatively large-sized and/or less soluble molecules displaying permeability-limited and/or

solubility-limited absorption has greatly increased over the last years (Lipinski, 2001; Lipinski, 2002;

Merisko-Liversidge et al., 2003; Di , et al., 2012; Shah et al., 2011). It has been reported that 75% of

recent and current drug development candidates show low solubility in water (Di et al., 2012; Kawabata

et al., 2011) and thus belong to the classes II (high permeability) and IV (low permeability) of the

Biopharmaceutical Classification System (BCS) (Amidon et al., 1995).

Several measures are currently used to increase the aqueous solubility of very slightly water-soluble

drug candidates, including synthesis of an optimal salt form or a pro-drug, pH adjustment of the

aqueous medium in case of weakly acidic or basic substances, addition or use of organic solvents,

inclusion into cyclodextrins, micelles, microemulsions or liposomes, and, last but not least,

micronization and nanonization. While such measures can significantly improve aqueous solubility,

they may also affect (positively or negatively) the permeability and stability of the candidate drug

substances. Further, some of the measures involving the use of organic solvents or surfactants may not

provide sufficient solubilisation effect at safe concentrations for low potency drug substances requiring

high doses (Merisko-Liversidge et al., 2003); indeed, for solubilizing important amounts of drug

substance, high concentrations of excipients may be needed, which may cause safety issues, especially

during long-term use in humans (Chaubal, 2004). In this respect, the optimal salt form, optimal pH of

the aqueous medium and micro- and nanosuspension formulations are preferable.

Early formulation development of very slightly water-soluble or practically water-insoluble compounds

is crucial to enable pre-clinical and clinical pharmacokinetic (PK), efficacy, and toxicity studies.

Importantly, processes and dosage forms developed for pre-clinical studies should be transferable to

clinical formulations and their intended administration route (Chaubal, 2004). Thus, developability of

an early formulation is crucial. In this respect, nanosuspension formulations are again highly beneficial,

particularly for low potency compounds requiring high doses of drug substance with little excipients

and when the formulation should be administrable by the oral and parenteral routes (Gao et al, 2012).

Chapter I

Page 25 of 182

2. Nanosuspensions: Formulation and manufacturing

Nanosuspensions are aqueous suspensions containing one or several submicron-sized drug substances

and appropriate stabilizers. Stabilizers include excipients that enable nanogrinding of the drug particles,

prevent crystal growth or nanoparticle aggregation during storage, pH-buffering substances,

preservatives, and other components that may be needed for further processing (e.g., transforming into a

solid form) or administration to patients (e.g., sweeteners, colorants) (Merisko-Liversidge et al., 2003).

The term nanosizing, as used in this work, describes the reduction of suspended drug particles down to

the submicron size range.

The main challenge in nanosuspension technology is prevention of particle agglomeration or

aggregation and crystal growth. (Merisko-Liversidge et al., 2003). At the nanometer scale, attractive

van der Waals and dispersive forces between particles come into play. Such attractive forces increase

dramatically as particles approach each other, which ultimately results in irreversible aggregation

(Peukert et al., 2005; Kesisoglou et al., 2007). Key to drug nanoparticle technology is the successful

compensation of the extra free energy of freshly exposed surfaces (Choi et al., 2005). The tendency of

the smaller particles in a suspension to dissolve and re-crystallize on the larger particles represents a

mode of instability, termed Ostwald ripening. Ostwald ripening becomes important with particles

smaller than 0.5 µm. In general, the speed of Ostwald ripening is governed by molecular diffusion or

surface reaction. Diffusion-controlled growth predominates if the particle size distribution in the

suspension is large (i.e., in presence of an important fraction of smaller particles) and the solubility is

high. In this situation, Ostwald ripening can be lowered by narrowing the particle size distribution. The

alternative mechanism of surface reaction predominates under very low supersaturation, i.e., when the

solubility of the smaller and larger particles is similar. Ostwald ripening via reaction-controlled

mechanism can be prevented by the addition of polymeric stabilizers to the suspension.

Irrespective of the mechanism, particle growth can be prevented or at least minimized by steric

hindrance and/or electrostatic repulsion. Steric hindrance is primarily achieved by adsorbing polymers

onto particles, while for electrostatic repulsion, ionic surfactants or polymers are used. Use of

surfactants or surface active polymers additionally promotes wetting and dispersion of the drug

particles, which are usually very hydrophobic.

Commonly used polymeric stabilizers for nanosuspensions include cellulose ethers, such as

hydroxypropylcellulose (HPC) and hydroxypropylmethylcellulose (HPMC), povidone, and poloxamers

(types 188, 407 and 338). Commonly used surfactant stabilizers are either non-ionic, such as the

Chapter I

Page 26 of 182

polysorbate types, or anionic, such as sodium dodecyl sulfate (SDS) and docusate sodium (SD). For

effective nanosuspension stabilization, the drug substance: stabilizer ratio may vary between 20:1 to

2:1, (w/w). While insufficient amounts of stabilizers remain ineffective for preventing particle

agglomeration, excessive quantities may promote crystal growth by Ostwald ripening (Merisko-

Liversidge et al., 2003). Naturally, only excipients with established safety profiles should be used for

the stabilization of nanosuspensions (Kesisoglou et al., 2007).

Particle size reduction can be achieved mainly by bottom-up and top-down processes (Verma et al.,

2009; Kawabata et al., 2011): (1) Particle formation through micro-precipitation, chemical synthesis, or

complexation; (2) particle comminution (nanosizing) through high energy homogenization (Liversidge

et al., 1995; Müller et al., 2006). Processes can also be combined to achieve synergistic effects as

adopted in the NANOEDGE® technology platform (Müller et al., 2006).

A widely used process in nanosuspension technology is wet media milling (NanoCrystals® technology)

in high-shear energy mills (Liversidge et al., 1995). In this process, drug substance powder is suspended

in an appropriate medium (mostly an aqueous or aqueous-organic solution containing appropriate

stabilizers). To the drug dispersion, milling medium is added under continuous stirring; typical milling

media comprise beads of ceramics (cerium- or yttrium-stabilized zirconium dioxide), stainless steel,

glass or highly cross-linked polystyrene resin, all of which can be obtained in different sizes (typically

between 0.1 and 1 mm). The slurry of drug suspension and milling beads is then introduced into the

milling chamber of an appropriate mill. Shear forces generated by the movement of the milling medium

lead to particle size reduction. Required milling time mainly depends on the hardness of the drug

particles, viscosity of the drug suspension, temperature, size and density of the milling medium, and

total energy input during milling (Stenger et al., 2005). Milling time can last from about 30 min to

several hours or even days (Shegokar, et al., 2010).

Nanomilling can be done at lab scale with as little as 100 mg in a few millilitres of medium by using the

Nanomill® system (élan Drug Discovery, PA, USA). Production can be readily scaled up to several

litres, e.g., by using Dynomill® (chamber volumes of 300 and 600 ml; Glen Mills, Cliffton, NJ, USA)

in flow-through mode or Netzsch® mills (chamber volumes of 2, 10 and 60 L; Netzsch, Exton, PA,

USA). To produce large-size batches, mills are preferably configured in circulation mode. The

NanoCrystal® technology has successfully expanded the use of nanosuspensions for oral, inhalation,

intravenous, subcutaneous, intramuscular and ocular administration (Merisko-Liversidge et al., 2008)

(Shegokar, et al., 2010).

Chapter I

Page 27 of 182

Nanosuspensions are typically converted to a solid dosage form for clinical formulations. Prior to

drying, redispersants need to be added to the nanosuspension to ensure complete redispersion of

nanoparticles into their primary, pre-drying state. Sugars, such as sucrose, lactose, and mannitol, are

commonly used as redispersants in oral formulations. The sugar molecules serve as “protectants” and

prevent nanoparticles from aggregating as they are concentrated during drying.

When intended for oral use, nanosuspension-derived solids have to self-re-disperse into nanoparticles in

the gastro-intestinal medium in order to maintain the intended benefit of fast dissolution.

Nanosuspensions intended for i.v. administration have to fulfil further quality criteria such as limited

contaminants from wear of the grinding beads, appropriate syringeability by the intended syringe and

needle (Wong et al., 2008), and sterility. Due to potential stability impairment of aqueous

nanosuspensions by terminal sterilization (autoclaving, irradiation), nanosuspensions are preferably

produced under aseptic conditions. If particle sizes of an injectable nanosuspension are much smaller

than 200 nm, terminal sterile filtration might even be considered.

3. Nanosuspension applications

Nanosuspensions can be applied to all phases of pharmaceutical development from pre-clinical, clinical

phase 1 and 2 to phase 3 and commercial, depending on the drug substance characteristics and

application. Nanosuspensions can be formulated as a liquid, semi-solid or a solid formulation depending

on its application and the development phase.

Their major advantage remains on the nanoparticle size and the consequent increase in surface area and

low toxicity. Particle size engineering can be used to achieve the intended dissolution rate in case the

drug substance adsorption is dependent on the solubility or both solubility and permeability. The

nanoparticle nanosize and increase in the specific surface area constitute both their major advantage and

manufacturability and stability challenges. Therefore, one of their obvious critical to quality attributes is

the particle size distribution as well as re-dispersion both in the gastric intestinal fluids or plasma.

Nanosized drug particles in liquid or dried nanosuspension formulations offer several therapeutic

benefits. Firstly, drug nanoparticles exhibit enhanced dissolution rate, which is due to the greatly

increased specific surface area, as described by the Nernst–Brunner and Noyes–Whitney equations.

Secondly, drug nanoparticles also show increased solubility, which is due to the vapour and dissolution

pressure of solid particles that increases when particles sizes decrease below 1000 nm, and more

particularly below 100 nm), as described by the Freundlich–Ostwald equation (Kesisoglou et al., 2007).

From an industrial perspective, nanosuspensions are advantageous, because they can be applied by

Chapter I

Page 28 of 182

many different routes, most importantly the oral and parenteral routes, including intravenous

administration. Upon parenteral administration, drug nanoparticles offer the potential for passive or

active targeting. When administered intravenously, drug nanoparticles can be taken by the mononuclear

phagocyte system (MPS) providing targeting for macrophage related disorders and accumulation in

liver and spleen (Shegokar et al., 2010). To lower and retard uptake by the MPS and maintain the drug

nanoparticles longer in circulation, particles can be decorated with hydrophilic polymers, such as

polyethylene glycol (PEG). Such PEGylated drug nanoparticles would then passively accumulate in

tumour tissue by the so-called enhanced permeation and retention (EPR) effect (Merisko-Liversidge et

al., 2003; Chaubal, 2004). For active targeting, drug nanoparticles can be coated with specific

functional groups. For example, thiamine coated nanoparticles demonstrated preferential uptake in the

brain via the blood–brain–barrier thiamine transporter (Chaubal, 2004). In oral administration, drug

nanoparticles show enhanced bioavailability and more balanced drug absorption between fed and fasted

states. Taken together, all the beneficial properties of drug nanoparticles may offer means for

accelerating lead selection and drug development processes (Chaubal, 2004).

Nanoparticles have found wide use in recent years for various oral, injectable, inhalational and even

intradermal applications (Chaubal, 2004). The first nanosuspension-based product launched on the

market using the NanoCrystal® technology was the immune suppressant sirolimus (Rapamune®,

Wyeth) in 2000, followed by aprepitant (Emend®, MERCK) in 2003, fenofibrate (TriCor®, ABBOTT)

in 2003, and megestrol acetate (Megace, ES PAR PHARM) in 2005. All these products are

administered orally. One of the first demonstrations of the biopharmaceutical benefits of

nanosuspensions in oral administration was with danazol, a poorly soluble gonadotropin inhibitor for

treatment of, e.g., menorrhagia. The absolute bioavailability of conventional danazol microsuspensions

in beagle dogs was only 5.2%, but 82.3% when administered as an aqueous nanosuspension (at same

dose of 200 mg); at the same time, Tmax was reduced and Cmax increased approx. 15-fold (Liversidge et

al., 1995). Another example is aprepitant, a low solubility drug substance used on the drug product

Emend® which has been approved for the prevention of chemotherapy-induced nausea and vomiting.

Aprepitant nanosuspensions are prepared by nanogrinding using hydroxypropyl cellulose and sodium

dodecyl sulfate as stabilizers (Wu et al., 2004).

The NanoCrystal® dispersion provided 3.5-fold increase in exposure in humans at a dose of 100 mg

compared with a tablet formulation made of micronized particles and also eliminated food effects on

absorption in a dog model (Wu et al., 2004).

In various instances, intravenous drug injection is required to meet the clinical needs. However, when a

formulation is administered intravenously, particle size must be smaller than 5 µm to avoid capillary

blockade and embolism. Nanosuspensions offer a good alternative as they can be injected intravenously

Chapter I

Page 29 of 182

using relatively high doses in low volumes and achieve 100% bioavailability (Gao et al., 2012).

Merisko-Liversidge (2011) compared the performance of marketed paclitaxel solution (Taxol®), which

is formulated as microemulsion with Cremophor® EL and ethanol to solubilise the drug substance, with

paclitaxel nanosuspension. The results demonstrated that the nanosuspension was better tolerated and

more efficacious.

More recently, nanosuspensions have also been introduced as long-acting injectable formulations.

Tailoring the particle size distribution at the nanosize level may afford control of dissolution rates,

while offering the advantages of high drug dose and stability (Baert et al., 2009). Injectable long-acting

formulations are currently available in several therapeutic areas requiring long-term treatment or

prophylaxis, such as for psychiatric disorders or contraception (Van Klooster et al., 2010). A successful

long-acting nanosuspension is Invega® Sustenna® (Jansen), a once-a-month extended release

nanosuspension of paliperidone palmitate for intramuscular (deltoid or gluteal muscle) injection.

Paliperidone palmitate nanosuspension is obtained by nanogrinding in presence of a surfactant and

marketed as a single-use prefilled syringe (Samtani et al., 2009; Merisko-Liversidge et al., 2011).

Another example of a long-acting (once-a-month) injectable nanosuspension is with the very poorly

water soluble antiretroviral anti-HIV rilpivirine. Rilpivirine is currently in development as a long acting

nanosuspension formulation, for treatment and prophylaxis against HIV (Baert et al., 2009). Rilpivirine

nanosuspension with 200 nm sized particles was injected intramuscularly (i.m.) or subcutaneously (s.c.)

in dogs as a single-dose and achieved stable plasma concentration profiles detectable up to 3 months.

The 200 nm sized particles provided improved early release (higher Cmax) as compared with 400 or 800

nm sized particles. Pharmacokinetic simulations demonstrated that once-a-month s.c. or i.m.

administrations of rilpivirine in humans will maintain plasma concentrations above the minimal

therapeutic concentrations of 73 to 95 ng/ml as observed after daily oral dosing with 25 mg rilpivirine

(Baert et al., 2009; Van Klooster et al., 2010).

The following, are further examples of applications of nanosuspensions.

Docetaxel (Duopafei®) is a practically water-insoluble drug substance with low bioavailability and high

toxicity; it is used in the treatment of different types of cancer (ovarian, breast and other type of

tumours). The presently marketed product contains high concentration of the non-ionic surfactant

polysorbate 80. Docetaxel-loaded nanosized lipid carrier was prepared by high pressure

homogenization using soya lecithin as surfactant followed by freeze-drying in presence of mannitol.

The mean particle size of the freshly prepared and the freeze-dried drug-loaded lipid particles were

200.0 ± 3.4 nm and 223.3 ± 4.3 nm, respectively (Wang et al., 2011). At two weeks after i.v. injection

into B16 melanoma-bearing mice, the group treated with the nanosuspension showed smaller tumor

volumes (P < 0.05) and lower tumour weights (P < 0.01) than the Duopafei® group; this suggests that

the nanosuspension inhibited effectively tumour growth, while no antitumor effect was observed in the

Chapter I

Page 30 of 182

blank group. Besides the higher antitumor efficacy, the nanosuspension –treated group showed also

increased survival rate and strongly reduced drug toxicity (Wang et al., 2011).

In Sporanox® i.v. for intravenous administration, itraconazole is solubilized by hydroxypropyl-β-

cyclodextrin and propyleneglycol. Because of the cyclodextrin, Sporanox® i.v. must not be

administered to renally impaired patients. To overcome this limitation and to accommodate maximal

dose, an itraconazole nanosuspension with mean particle size of 581 ± 18 nm (measured after 24

months storage at 5 °C) was developed (Rabinow et al., 2007). When administered i.v. to rats, the

nanosuspension was better tolerated than itraconazole solution, especially at higher doses. The higher

tolerated drug levels achieved with the nanosuspension in the target organs reduced candida albicans

colony counts and increased survival rates (Rabinow et al., 2007).

The investigational anticancer compound SN 30191 is practically water-insoluble inhibitor of

phosphatidylinositol 3-kinase (PI3-K). The activation and/or over-expression of PI3-K has been

associated with a number of human cancers; PI3-K has been shown to disturb the fine equilibrium

between cell division, growth and apoptosis, which leads to abnormal cell growth and possibly

resistance to therapy. SN 30191 was formulated by high pressure homogenization as a nanosuspension

with a mean particle size of less than 150 nm, (Sharma et al., 2011). A toxicity study in mice (i.v.

administration of 2.5 – 20 mg/kg body weight) revealed that the tolerated dose with the nanosuspension

was at least four times higher (10 mg/kg versus 2.5 mg/kg) than with the solution formulation

(consisting of 10% (v/v) DMSO, 40% (v/v) PEG 300, 50% (v/v) PBS and 0.5% (w/v) Cremophor®

ELP) (Sharma et al., 2011).

Another case where nanosuspension technology has proven to be highly beneficial is with

camptothecin, a natural alkaloid and topoisomerase I inhibitor with strong antitumor activity. The

compound had, however, to be abandoned in early clinical trials, because of its poor solubility and

toxicity. A camptothecin nanosuspension was developed using a solvent precipitation method, which

yielded particle sizes in the range of 200 nm to 700 nm (Zhang et al., 2011). When tested in MCF-7

xenografted BALB/c mice, the nanosuspension suppressed significantly tumor growth with the drug

concentration in the tumor being five times higher at 24 h as compared to a drug solution in

propylenglycol-water mixture (Zhang et al., 2011).

A final example concerns the topoisomerase II inhibitor asulacrine with activity against breast and lung

cancers (Ganta et al., 2009). Asulacrin nanosuspension was prepared by high pressure homogenization

using poloxamer 188 as stabilizer for the drug nanoparticles (d50 0.133 ± 0.02 µm; d90 0.702 ± 0.02

µm). In vitro dissolution testing revealed a sustained profile reaching 42% of the dose (5 mg in 500 ml

Chapter I

Page 31 of 182

medium) within 5 h (as compared to 6% when a micronized powder was used). A pharmacokinetic

study in mice showed that the nanosuspension produced reduced Cmax (12.2 ± 1.3 µg ml−1 versus

18.3 ± 1.0 µg ml−1) and AUC-values (18.7 ± 0.5 µg ml−1 h versus 46.4 ± 2.6 µg ml−1 h) as compared to

the drug solution in a 1:1 mixture of dimethylacetamide and propylenglycol. The plasma profiles

indicated that the nanosuspension formulation maintained low plasma drug for a longer period of time,

which can potentially help to overcome the dose-dependent toxicity of asulacrin. The authors of the

study that asucraline nanosuspensions may give added value by allowing a reduction in either the dose

or its frequency of administration (Ganta et al., 2009).

4. Structure of the thesis

Nanosuspensions represent a very promising, universal formulation approach from pre-clinical to

commercialization for poorly soluble drug substances, as they can be administered by various routes

(parenteral, oral, ophthalmic and nasal), and exhibit substantially increased solubility and dissolution

rate, hence improved bioavailability (Müller et al., 2006; Kipp, 2004). Nonetheless, major challenges of

nanosuspension technology and formulations remain unsolved or only partly solved.

The present work focused on the nanosuspension technology of stirred media mills and formulation and

manufacturing process challenges and its influence on some of the most important nanosuspensions

critical quality attributes, which are major challenges in nanosuspension formulation; particle size after

wet-nanogrinding and storage of the liquid formulations, crystallinity, drug substance content, re-

dispersion and dissolution rate of dried powders.

First the current chapter (Chapter I) provides an overview on the nanosuspension technology

applications and formulation and manufacturing process strategies to evaluate how the current work

could contribute to improve the current knowledge and help to understand and solve some of the current

issues in nanosuspension formulation.

The first experimental chapter (Chapter II) provides a systematic approach to select the most

appropriate stabilizers and to investigate the best quantitative composition to achieve the intended

particle size distribution and stability using miconazole as the model drug substance. Understand the

importance of the excipient selection to provide adequate stabilization during milling and stability as

well as some of the characteristics of those excipients that can influence their suitability for

nanosuspension stabilization was also an objective. After the selection of the excipients the

Chapter I

Page 32 of 182

investigation focused on the quantitative composition and its influence in the particles size and particle

size stability

Next, Chapter III focuses on the study of the nanogrinding process parameters and their influence on

the particle size distribution. Miconazole, itraconazole and etravirine drug substance mechanical

properties as plastic and elastic deformation were measured to access the importance of these in the

nanogrinding energy profiles to achieve the intended particle size.

In Chapter IV a rapid assay method to quantify the stabilizers hydroxypropylcellulose and sodium

dodecyl sulfate applicable to different drug substances was developed. This rapid assay used

miconazole as the model drug substance.

The use of such quantitative method can be used to compare the amounts of steric and electrostatic

stabilizers adsorbed to the nanoparticle surface and understand how this attribute can influence

nanosuspension formulation and further processing in solid formulations. This was one of the objectives

of Chapter V, where miconazole and itraconazole were used. Further objectives of this chapter were

the comparison of the adequacy of spray drying and freeze drying to dry nanosuspensions and the

influence of adding the so called matrix formers as mannitol and microcrystalline cellulose.

Finally, conclusions are drawn in Chapter VI, and an outlook is presented.

Chapter I

Page 33 of 182

References

Amidon, G.L., Lennernas, H., Shah, V.P., Crison, J.R., 1995. A theoretical basis for a biopharmaceutic

drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability.

Pharm. Res. 12, 413–420.

Baert, L., Van‘t Klooster, G., Dries, W., François, M., Wouters, A., Basstanie, E., Iterbeke, K.,

Stappers, F., Stevens, P., Schueller, L., Van Remoortere, P., Kraus, G., Wigerinck, P., Rosier, J.,

2009. Development of a long-acting injectable formulation with nanoparticles of rilpivirine

(TMC278) for HIV treatment. Eur. J. Pharm. Biopharm. 72, 502–508.

Ganta, S., Paxton, J.W., Baguley, B.C., Garg, S., 2009. Formulation and pharmacokinetic evaluation of

an asulacrine nanocrystalline suspension for intravenous delivery. Int. J. Pharm. 367, 179–186.

Gao, L., Liu, G., Ma, J., Wang, X., Zhou, L., Li, X., 2012. Drug nanocrystals: In vivo performances. J.

Controlled Release 160, 418–430.

Chaubal, M.V., 2004. Application of formulation technologies in lead candidate selection and

optimization. Drug Discov. Today, 9, 603–609.

Choi, J-Y, Yoo, J.Y., Kwak, H-S., Nam, B.U., Lee, J., 2005. Role of polymeric stabilizers for drug

nanocrystal dispersions. Curr. App. Phys. 5, 472–474.

Di, L., Fish, P.V., Mano, T., 2012. Bridging solubility between drug discovery and development. Drug

Discov. Today, 17, 486–95.

Kawabata, Y., Wada, K., Nakatani, M., Yamada, S., Onoue, S., 2011. Formulation design for poorly

water-soluble drugs based on biopharmaceutics classification system: Basic approaches and practical

applications. Int. J. Pharm. 420, 1– 10.

Kesisoglou, F., Panmai, S., Wu, Y., 2007. Nanosizing – oral formulation development and

biopharmaceutical evaluation. Adv. Drug Del. Rev. 59, 631–644.

Kipp, J.E., 2004. The role of solid nanoparticle technology in the parenteral delivery of poorly water-

soluble drugs. Int. J. Pharm. 284, 109–122.

Lipinski, C.A., 2001. Avoiding investment in doomed drugs, is poor solubility an industry wide

problem? Curr. Drug Disc. 17–19.

Lipinski, C., 2002. Poor aqueous solubility-an industry wide problem in drug discovery. Am. Pharm.

Rev. 5, 82–85.

Liversidge, G., Cundy, K., 1995. Particle size reduction for the improvement of oral bioavailability of

hydrophobic drugs: I. Absolute oral bioavailability of nanocrystalline danazol in Beagle dogs, Int. J.

Pharm. 125, 91–97.

Maas, J., Kamm, W., Hauck G., 2007. An integrated early formulation strategy – From hit evaluation to

preclinical candidate profiling. Eur. J. Pharm. Biopharm. 66, 1–10.

Chapter I

Page 34 of 182

Merisko-Liversidge E., Liversidge, G.G., Cooper, E.R., 2003. Nanosizing: a formulation approach for

poorly-water-soluble compounds. Eur. J. Pharm. Sci. 18, 113–120.

Merisko-Liversidge, E.M., Liversidge, G.G., 2008. Drug nanoparticles: formulating poorly water-

soluble compounds. Toxicol. Pathol. 36, 43–48.

Merisko-Liversidge, E., Liversidge, G.G., 2011. Nanosizing for oral and parenteral drug delivery: A

perspective on formulating poorly-water soluble compounds using wet media milling technology.

Adv. Drug Deliver. Rev. 63, 427–440.

McKee, J., Rabinow, B., Cook, C., Gass, J., 2010. Nanosuspension Formulation of Itraconazole

Eliminates the Negative Inotropic Effect of SPORANOX® in Dogs. J. Med. Toxicol. 6, 331-336.

Müller, R.H., Keck, C.M., 2006. Drug nanocrystals of poorly soluble drugs produced by high pressure

homogenization. Eur. J. Pharm. Biopharm. 62, 3–16.

Orloff, J., Douglas, F., Pinheiro, J., Levinson, S., Branson, M., Chaturvedi, P., Ette, E., Gallo, P.,

Hirsch, G., Mehta, C., Patel, N., Sabir, S., Springs, S., Stanski, D., Evers, M. R., Fleming, E., Singh,

N., Tramontin, T., Golub, H., 2009. The future of drug development: advancing clinical trial design.

Nat. Rev. Drug. Discov. 8, 949–957.

Peukert, W., Schwarzer, H-C., Stenger, F., 2005. Control of aggregation in production and handling of

nanoparticles. Chem. Eng. Proc. 44, 245–252.

Rabinow, B.E., 2004. Nanosuspensions in drug delivery. Nat. Rev. Drug Discov. 3, 785–796.

Samtani, M. N., Vermeulen, A., Stuyckens, K., 2009. Population Pharmacokinetics of Intramuscular

Paliperidone Palmitate in Patients with Schizophrenia: A Novel Once-Monthly, Long-Acting

Formulation of an Atypical Antipsychotic. Clin. Pharmacokinet. 48, 585–600.

Shah, A.K., Agnihotri, S.A., 2011. Recent advances and novel strategies in pre-clinical formulation

development: An overview J. Controlled Release 156, 281–296.

Sharma, P., Zoran, D., Zujovic, Z.D., Bowmaker, G.A., Denny, W.A., Garg, S., 2011. Evaluation of a

crystalline nanosuspension: Polymorphism, process induced transformation and in vivo studies. Int.

J. Pharm. 408, 138–151.

Shegokar, R., Müller, R.H., 2010. Nanocrystals: Industrially feasible multifunctional formulation

technology for poorly soluble actives. Int. J. Pharm. 399, 129–139.

Stenger, F., Mende, S., Schwedes, J., Peukert, W., 2005. Nanomilling in stirred media mills. Chem.

Eng. Sci. 60, 4557–4565.

The FDA's Drug Review Process: Ensuring Drugs Are Safe and Effective

http://www.fda.gov/drugs/resourcesforyou/consumers/ucm143534.htm

Tonkens, R., 2005. An Overview of the Drug Development Process. The Physician Executive, May,

48–52.

Chapter I

Page 35 of 182

Van ¢t Klooster, G., Hoeben, E., Borghys, H., Adriana Looszova, A., Bouche, M.P., Van Velsen, F.,

2010. Pharmacokinetics and Disposition of Rilpivirine (TMC278) Nanosuspension as a Long-Acting

Injectable Antiretroviral Formulation. Antimicrob. Agents Chemother. 54, 2042–2050.

Verma, S., Gokhale, R., Burgess, J.D., 2009. A comparative study of top-down and bottom-up

approaches for the preparation of micro/nanosuspensions. Int. J. Pharm. 380, 216–222.

Wang., L., Liu, Z., Liu, D., Liu, C., Juan, Z., Zhang, N., 2011. Docetaxel-loaded-lipid-based-

nanosuspensions (DTX-LNS): Preparation, pharmacokinetics, tissue distribution and antitumor

activity. Int. J. Pharm. 413, 194– 201.

Wong, J., Brugger, A., Khare, A., Chaubal, M., Papadopoulos, P., Rabinow, B., James Kipp, J., Ning,

J., 2008. Adv. Drug Delivery Rev. 60, 939–954.

Wu, Y., Loper, A., Landis, E., Hettrick, L., Novak, L., Lynn, K., Chen, C., Karen Thompson, K.,

Higgins, R., Batra, U. 2004. The role of biopharmaceutics in the development of a clinical

nanoparticle formulation of MK-0869: a Beagle dog model predicts improved bioavailability and

diminished food effect on absorption in human. Int. J. Pharm. 285, 135–146.

Zhang, H., Hollis, C.P., Zhang, Q., Li, T., 2011. Preparation and antitumor study of camptothecin

nanocrystals. Int. J. Pharm. 415, 293– 300.

Chapter I

Page 36 of 182

Chapter II

Page 37 of 182

Chapter II

Miconazole nanosuspensions: Influence of formulation variables on particle size reduction and physical stability1

1. Introduction

Many new drug substances are only very slightly soluble or even practically insoluble. A substantial

portion (40%) of these drugs fails full development, because of their poor and highly variable

bioavailability (Gardner et al., 2004; Riley, 2006). Upon peroral administration, very slightly water-

soluble or practically water-insoluble drugs have a limited and variable or erratic oral absorption

(Crison, 2000). Further, the low solubility of such drugs limits their parenteral use. Low solubility of

drug substances may result from hydrophobicity or high lattice energy. Highly hydrophobic drug

substances possess insufficient capacity of molecular interactions with water, whereas molecules with

high lattice energy resist to the weakening of the lattice upon molecular interactions with water (Kipp,

2004). According to the law of Noyes-Whitney, low solubility yields a low concentration gradient

towards the bulk of the solution and, thereby, a low dissolution rate. Therefore, absorption and

bioavailability of perorally administered drugs that possess good permeability, but low solubility, can

be improved by increasing either the solubility or the surface area of the drug substance, both resulting

in increased dissolution rates. A very common way to increase drug substance surface area is by

micronization, which produces particles in the size range of 2 µm to 5 µm. However, when the

solubility of a drug is very low, i.e., below approx. 1 mg/ml, micronization is generally insufficient to

increase adequately the drug dissolution rate and absorption in the gastro-intestinal tract (Muller et al.,

2001).

Nanonisation has become a popular approach to produce particles in the size range of 200 nm to 400

nm, to improve both the dissolution rate and the solubility of the compound (Liversidge et al., 1992).

The latter phenomenon is due to the well known dependency of solubility on particle size as described

by the Ostwald-Freundlich equation. Breakage of micron-sized drug crystals into nanoparticles creates

1

The work in this chapter has been published as Cerdeira, A.M., Mazzotti, M., Gander, B., 2010. Miconazole nanosuspensions: Influence of

formulation variables on particle size reduction and physical stability. Int. J. Pharm. 396, 210–218.

Chapter II

Page 38 of 182

an increased particle surface area, which is thermodynamically unfavourable. Thus, nanosized particles

tend to agglomerate to reduce their surface area. Particle agglomeration in nanosuspensions can be

prevented by steric and electrostatic stabilization using polymeric and/or surfactant excipients

(Rabinow, 2004). Most nanosuspensions are thus composed of an aqueous medium (e.g., purified

water), a nanosuspended drug substance of maximal 400 mg/ml, and adequate excipients for

nanogrinding and particle stabilization (Merisko-Liversidge et al., 2003). Both the type and

concentration of excipient(s) are important for particle size reduction and physical stabilization of the

formulations. Physically stable nanosuspensions are obtained at drug substance-to-excipient ratios of

20:1 to 2:1 (Merisko-Liversidge et al., 2003). Therefore, inadequate types or amounts of excipients may

either cause particle agglomeration due to the high surface energy of the nanoparticles or crystal growth

due to the drug substance solubility increase. Electrostatic and steric mechanisms are mediated by

combining ionic surfactants and polymers (Rabinow, 2004). Most commonly used polymeric excipients

for nanosuspensions include cellulose ethers (e.g., hydroxypropylcellulose,

hydroxypropylmethylcellulose), povidone, and poloxamers (Liversidge et al., 1992; Merisko-Liversidge

et al., 2003; Kesisoglou et al., 2007). The surfactant excipients can be non-ionic, such as polysorbate

(Tween® 80), or anionic, such as sodium dodecyl sulfate or sodium docusate. Cationic surfactants are

less frequently used (Kesisoglou et al., 2007).

Physical stabilization of nanosuspensions is a major challenge. Alterations in particle size distribution,

polymorphic and solvate forms need to be carefully analyzed and monitored during storage (Kipp,

2004). Nanosuspension stability depends on: (i) the solid state properties of the nanoparticles (density,

hardness, number and type of lattice defects), (ii) the interfacial properties (wetting and interfacial

energy between nanoparticles and medium, structure of the solid-liquid interface), and (iii) the

properties of the suspending medium (viscosity, drug solubility, presence of micelles and their

interaction with the dissolved and solid drug). Instability of nanosuspensions may manifest by a shift of

particle size distribution to larger sizes, irreversible agglomeration, or solid phase transformation (Kipp,

2004).

Despite the numerous challenges, nanosuspensions represent a very promising, rather general

formulation approach to increase solubility and dissolution rate of very slightly soluble or practically

water-insoluble solid drug substances. Furthermore, nanosuspensions are suitable for administration by

various routes (parenteral, oral, ophthalmic and nasal), which is an eminent advantage over other

dosage forms. Although numerous studies have explored drug substance nanogrinding and

nanosuspensions, the parameters affecting nanogrinding and particle stabilization during storage are

still not well understood (Augustijns et al., 2008). The aim of this study was, therefore, to evaluate the

importance of the concentration of miconazole drug substance and of the type and concentration of

Chapter II

Page 39 of 182

surfactant and polymeric excipients on the physical characteristics of miconazole nanosuspensions

during milling and storage. Miconazole was selected as it is practically insoluble in water, but possesses

high permeability, which makes it an excellent candidate for nanogrinding. Miconazole is a well-known

imidazole, used as base or nitrate salt, for treatment of superficial candidiasis, dermatophytosis, and

pityriasis versicolor (Sweetman, 2006).

2. Materials and Methods

2.1. Materials

Miconazole (lot # R018134PUC701, Janssen Pharmaceutica, Geel, Belgium) used for nanogrinding had

a volume-based mean diameter of d50 = 27 µm, and 10% and 90% undersize percentiles of d10=14 and

d90 = 49 µm, respectively. Sodium dodecyl sulfate [SDS] (Texapon® K12P, Cognis, Düsseldorf,

Germany), sodium docusate [DS] (Cytec Industries, Belmont West Virginia, USA), benzalkonium

chloride [BK] (Sigma-Aldrich, Schnelldorf, Germany), hydroxypropylcellulose [HPC-LF, HPC-EF]

(Klucel® LF, Klucel® EF, Hercules, Doel, Belgium), povidone [PVP] (Plasdone® K29/32, ISP

Technologies, Texas City, US), poloxamer [poloxamer] (Pluronic® F68, BASF, Ludwigshafen,

Germany), hydroxypropylmethylcellulose [HPMC] (Hypromellose 2910, Methocel® E15 LV,

Colorcon, Dow Chemicals, Dartford, UK), and were all used as received.

2.2. Production of nanosuspensions

For nanogrinding miconazole, solutions of surfactant and polymer stabilizers in purified water were

first prepared. Miconazole (d10 = 14 µm; d50 = 27 µm; d90 = 49 µm) was then dispersed in the stabilizer

solution. Initial experiments were designed to screen most suitable surfactant and polymer stabilizers

(Table 1). Nanogrinding was performed in a high-energy mill (LabStar, Netzsch, Selb, Germany) filled

(to 83%, v/v) with yttrium-stabilized zirconium oxide beads (0.8 mm in diameter). Nanogrinding was

performed in circulation mode using 300 g of suspension, a pump-speed of 41 rpm, and a stirrer-tip-

speed of 3400 rpm (10 m/s); the duration of the process was up to 60 min.

Chapter II

Page 40 of 182

Table 1: Miconazole suspension formulations used for screening polymeric and surfactant stabilizers.

Miconazole concentrationa) (%, w/w)

Surfactant type and concentrationa) (%, w/w)

Polymer type and concentrationa) (%, w/w)

5

SDS, 0.05

HPC-EF, 1.25 HPMC, 1.25 PVP, 1.25 Poloxamer, 1.25 (without SDS)

10

SDS, 0.10

HPC-EF, 2.5 HPC-LF, 2.5 HPMC, 2.5 PVP, 2.5 Poloxamer, 2.5 (without SDS)

10

No surfactant SDS, 0.1 DS, 0.1 BK, 0.1

HPC-LF, 2.5

a) Single line information in each row applies to all variables of this row.

2.3. Design of experiments (DOE)

A 23 DOE (Table 2) was used to study the effects and interactions of miconazole, SDS, and HPC-LF on

particle size distribution. A centre point with replicate was introduced in the design to estimate the

curvature and the pure error. The data were fitted according to the following polynomial equation:

Y= ��+ ��Xi + ��Xj +��Xk + ��XiXj + ��XiXk + ��XjXk + ��XiXjXk (1)

where a0 is the overall mean response (mean particle size), ai, aj, and ak are the main effect coefficients,

aij, aik, ajk, and aijk are the coefficients of the interaction effects (first and second order), and Xi, Xj and Xk

are the factors (miconazole, SDS, HPC-LF). The statistical design and evaluation of the obtained

experimental data was carried out with the software Minitab® 15 (Minitab Inc.). The model was reduced

by removing non-significant coefficients (α = 0.05). The significance and validity of the model was

estimated by analysis of variance (ANOVA). Additional experiments were performed to explore the

corners of the DOE and optimize the formulations.

Table 2: Formulation factors and levels according to a 23 experimental design.

Level Miconazole (%, w/w) SDS (%, w/w) HPC-LF (%, w/w)

(+1) 20 0.2 5

(-1) 5 0.05 1.25

(0) 12.5 0.125 3.125

2.4. Particle size distribution by laser light diffraction

Particle size distribution (volume based) was measured by laser light diffraction (Mastersizer 2000,

Malvern Instruments, Worcestershire, UK), using the small volume dispersion unit (Hydro 2000 Micro

Chapter II

Page 41 of 182

Precision). The Mie theory (dispersant refractive index = 1.33; real particle refractive index = 1.55;

imaginary part of the particle refractive index = 0.001) was used for particle size calculation. The

nanosuspensions were diluted with purified water to obtain an appropriate obscuration. Particle sizes

were expressed by the volume-based 50% (d50) and 90% (d90) diameter percentiles.

2.5. Optical microscopy

Optical microscopic pictures of miconazole suspensions in different stabilizer solutions were taken

before milling (Zeiss Axiophot, Zürich, Switzerland).

2.6. Determination of contact angle

The dynamic contact angle between miconazole powder compacts and water was measured by the

sessile drop method (Krüss DSA100, Hamburg, Germany and software Krüss drop shape analysis

DSA1, Hamburg, Germany). The static contact angle between miconazole and stabilizer solutions was

measured in triplicates using the powder method. The contact angle was calculated using the Washburn

equation (Aulton, 2007).

2.7. Quantification of miconazole solubility in the nanosuspensions

Miconazole nanosuspensions were centrifuged (ultracentrifuge Sorvall, Thermo Fisher Scientific,

Waltham, USA) at 50,000 rpm (minimum of 3 h), and the supernatant assayed for drug content by

HPLC. Centrifugation was preferred over ultrafiltration, because of the difficulty experienced with

filtering some of the relatively viscous (Fig. 5) suspensions, which caused filter clogging. Miconazole

was assayed by a validated method using reversed phase HPLC with UV detection at 230 nm (Waters

Alliance HPLC system, Milford, MA, US) and a C18 column (Zorbax®, 10 cm length, 4.6 mm ID, 3.5

µm particle size). The drug was eluted with 10 mM di-sodium hydrogen phosphate of pH 7.5 (solvent

A) and acetonitrile (solvent B) according to the gradient reported in Table 3. Linearity was confirmed

between 0.5 and 700 µg/ml, and the accuracy was ± 30% for 0.5 to 1 µg/ml, ± 20% for 1 to 10 µg/ml; ±

10% for 10 to 200 µg/ml, and ± 3% for 200 to 700 µg/ml.

Table 3: HPLC solvent gradient sequence for assaying miconazole.

Time (min) 0 30 35 40

Phase:

A (%, v/v) 80 35 80 80

B (%, v/v) 20 65 20 20

Chapter II

Page 42 of 182

2.8. Viscosity measurement of the nanosuspensions

The viscosity of the nanosuspensions was determined using a rheometer (Rheostress® RS600 Haake,

ThermoScientific, Waltham, USA). The measurements were performed at 20 ± 0.1 °C using rotational

mode (constant shear rate of 100 s-1) and cone-plate geometry (plate diameter 60 mm, cone angle 1°).

To avoid shear history effects, samples were kept at rest for 5 min after their application on the sensor

system.

2.9. Zeta-potential measurement of the nanosuspensions

Zeta-potential was measured using a Zetasizer (Zetasizer Nano ZS, Malvern Instruments,

Worcestershire, UK). The samples were adequately diluted with deionised water and placed in an

electrophoretic cell. The mean zeta-potential was calculated from the electrophoretic mobility using the

Smoluchowski equation (Aulton, 2007).

2.10. Particle size stability during storage

Nanosuspensions were stored in glass bottles (type I glass) with polypropylene caps, at 5 °C and 25

°C/60 %RH for a period of up to 6 months. The stability was assessed in terms of particle size

distributions (at 0 and 6 months), and drug solubility and zeta potential (at 0 and 3 months).

3. Results

3.1. Excipient screening for miconazole nanogrinding

To select a suitable polymeric excipient, miconazole (5 and 10%, w/w) was nanoground using SDS

(0.05 and 0.1%, w/w) and different types of polymers (1.25 and 2.5%, w/w) (Table 1; Figs. 1 and 2).

The higher concentrations of miconazole and excipients promoted particle size reduction (Fig. 1 versus

Fig. 2). The polymeric stabilizers HPMC and HPC were found to be highly effective for nanogrinding

miconazole, whereas poloxamer (non-ionic polymeric surfactant used without SDS) and PVP/SDS were

ineffective (Figs. 1 and 2). The reasons for the inefficiency of poloxamer and PVP/SDS remain

unknown. Nonetheless, microscopic observation of unground miconazole dispersions revealed large

aggregates in poloxamer and PVP/SDS solutions, but well dispersed particles with the other excipients,

as illustrated exemplarily for the poloxamer and HPC-LF/SDS formulations in Fig. 2.

Chapter II

Page 43 of 182

Figure 1: Effect of the type of polymeric excipient on the efficiency of miconazole nanogrinding in terms of

volume distributions of particle sizes. The formulations contained 5% (w/w) miconazole, 1.25% (w/w) polymer,

and 0.05% (w/w) SDS. SDS was not present when poloxamer was used as polymeric stabilizer.

To select a suitable surfactant, miconazole (10%, w/w) was milled using HPC-LF (2.5%, w/w) and

different types of surfactants (0.1%, w/w) (Table 1, Fig. 3), because HPC-LF was found in the previous

experiment to be a highly effective polymeric excipient for nanogrinding of miconazole. While all

surfactants facilitated nanogrinding and yielded nanoparticles with similar d50 values (d50 of 155 to 175

nm) (Fig. 3, inset), SDS was the most effective in minimizing the large particle size fraction (2 to 10

µm) of the bimodal size distribution (Fig. 3). Using SDS, the particle size fraction of 2 to 10 µm

represented less than 1% (volume) of the entire particle population. Here again, the extent of particle

size reduction seemed to be predetermined by the degree of dispersion of unground miconazole in the

different stabilizer solutions, as observed microscopically and illustrated exemplarily for the HPC-

LF/DS and HPC-LF/SDS formulations in Fig. 3.

0

2

4

6

8

10

12

10 100 1,000 10,000 100,000

Vo

lum

e (%

)

Particle size (nm)

HPC-EF / SDS

HPMC / SDS

PVP / SDS

Poloxamer

MIC before milling

Chapter II

Figure 2: Effect of the type of polymeric excipient on the efficiency of miconazole

volume distributions of particle sizes. Micrographs (insets) illustrate two extreme cases of

native drug particles in the aqueous medium before

miconazole, 2.5% (w/w) polymer, and 0.1% (w/w) SDS. SDS was not present when poloxamer was used as

polymeric stabilizer.

A key requisite for nanogrinding is adequate particle wetting. While the contact angle between

miconazole and pure water was above 140°, this value was reduced to 86° by addition of 2.5% HPC

to the aqueous medium (Table 4). Wetting was mostly enhanced b

Solutions of DS, BK and SDS all reduced the contact angle to values in the range of 43° to 56° (Table

4). Similar contact angles were observed between miconazole and the solutions containing SDS (0.1%,

w/w) in combination with either HPMC, or HPC

of both PVP/SDS and poloxamer presented a rather high contact angle with miconazole, which was

similar to that of HPC-LF alone. The poor wetting of miconazole by PVP/SDS and po

coincides with the inefficient nanogrinding in the presence of these excipients (Figs. 1 and 2).

0

2

4

6

8

10

12

10 100

Vo

lum

e (%

)

HPC-LF / SDS

HPMC / SDS

Figure 2: Effect of the type of polymeric excipient on the efficiency of miconazole nanogrinding

volume distributions of particle sizes. Micrographs (insets) illustrate two extreme cases of

native drug particles in the aqueous medium before nanogrinding. The formulations contained 10% (w/w)



miconazole and pure water was above 140°, this value was reduced to 86° by addition of 2.5% HPC

to the aqueous medium (Table 4). Wetting was mostly enhanced by the use of surfactants (0.1%, w/w).

, BK and SDS all reduced the contact angle to values in the range of 43° to 56° (Table


ion with either HPMC, or HPC-EF, or HPC-LF (2.5%, w/w). Surprisingly, solutions


LF alone. The poor wetting of miconazole by PVP/SDS and po


1,000 10,000

Particle size (nm)

Poloxamer

MIC before milling

HPC-EF / SDS

PVP / SDS

HPMC / SDS

Page 44 of 182

nanogrinding in terms of

volume distributions of particle sizes. Micrographs (insets) illustrate two extreme cases of dispersibility of the

. The formulations contained 10% (w/w)



miconazole and pure water was above 140°, this value was reduced to 86° by addition of 2.5% HPC-LF

y the use of surfactants (0.1%, w/w).

, BK and SDS all reduced the contact angle to values in the range of 43° to 56° (Table


LF (2.5%, w/w). Surprisingly, solutions


LF alone. The poor wetting of miconazole by PVP/SDS and poloxamer


100,000

MIC before milling

Chapter II

Figure 3: Effect of the type of surfactant on the efficiency of miconazole

distributions of particle sizes. Micrographs (insets)

particles in the aqueous medium before

(w/w) HPC-LF, and 0.1% (w/w) surfactant. One batch was prepared without surfactant.

Surface active excipients not only provide particle wetting, but can also form micelles and, thereby,

solubilise water-insoluble compounds. This

Ostwald ripening during storage (Merisko

excipients on miconazole solubility was assessed (Table 4). In pure water and in

without surfactant, miconazole solubility

the addition of BK and DS. In contrast, SDS increased substantially the miconazole solubility, i.e., to 95

µg/ml. In agreement with the observations of wetting, addition of PVP to the SDS soluti

interaction with the drug, thereby lowering the solubility (61

inefficient for miconazole nanogrinding, remained also inefficient to increase the drug solubility.

Finally, the drug solubility was similar in

Despite the important solubility of miconazole in SDS/HPC

0

1

2

1,000

Vo

lum

e (%

)

HPC

HPC

Figure 3: Effect of the type of surfactant on the efficiency of miconazole nanogrinding

distributions of particle sizes. Micrographs (insets) illustrate two extreme cases of dispersibility of the native drug

particles in the aqueous medium before nanogrinding. The formulations contained 10% (w/w) miconazole, 2.5%

surfactant. One batch was prepared without surfactant.


insoluble compounds. This might be critical in nanosuspension

Ostwald ripening during storage (Merisko-Liversidge et al., 2003). Therefore, the effect of the different

excipients on miconazole solubility was assessed (Table 4). In pure water and in

without surfactant, miconazole solubility was very low (3 µg/ml), and it was increased only slightly by

. In contrast, SDS increased substantially the miconazole solubility, i.e., to 95

µg/ml. In agreement with the observations of wetting, addition of PVP to the SDS soluti

interaction with the drug, thereby lowering the solubility (61 µg/ml). Poloxamer alone, which was


Finally, the drug solubility was similar in all solutions containing SDS and the different cellulose ethers.

Despite the important solubility of miconazole in SDS/HPC-LF, these excipients were selected for

10,000

Particle size (nm)

HPC -LF

HPC -LF / DS

HPC -LF / BK

HPC -LF / SDS

0

2

4

6

8

10

12

10 100

Vo

lum

e (%

)

Particle size (nm)

Page 45 of 182

nanogrinding in terms of volume

illustrate two extreme cases of dispersibility of the native drug

. The formulations contained 10% (w/w) miconazole, 2.5%

surfactant. One batch was prepared without surfactant.


in nanosuspensions, because of potential

Liversidge et al., 2003). Therefore, the effect of the different

excipients on miconazole solubility was assessed (Table 4). In pure water and in HPC-LF solutions

was very low (3 µg/ml), and it was increased only slightly by

. In contrast, SDS increased substantially the miconazole solubility, i.e., to 95

µg/ml. In agreement with the observations of wetting, addition of PVP to the SDS solution lowered the

g/ml). Poloxamer alone, which was


all solutions containing SDS and the different cellulose ethers.

LF, these excipients were selected for

HPC -LF / SDS

HPC -LF / DS

1,000 10,000

Particle size (nm)

Chapter II

Page 46 of 182

optimizing the nanogrinding of miconazole, as nanosuspensions are expected to be stable if the

solubility of the drug substance is less than 1 mg/ml (Merisko-Liversidge et al., 2003).

Table 4: Solubility and contact angle of miconazole in stabilizer solutions before nanogrinding.

Excipient Miconazole solubility

(µg/ml)

Contact angle

mean ± sd (°)

Surfactant: 0.1% (w/w); HPC-LF: 2.5% (w/w)

None 3 86 ± 1

SD 4 46 ± 1

BK 7 56 ± 3

SDS 95 43 ± 3

Polymer: 2.5% (w/w); SDS: 0.1% (w/w)

Poloxamer (without SDS) 0.5 89 ± 0

PVP 61 89 ± 0

HPMC 93 65 ± 1

HPC-LF 95 43 ± 3

HPC-EF 102 53 ± 4

Thus far, the data suggest that good indicators for the suitability of nanogrinding media for miconazole

are: (i) a low contact angle between the process solution and miconazole powder, and (ii) absence of

microscopically visible agglomerates in the drug dispersions prior to nanogrinding, with the latter

indicator being derived from subjective though consistent observations.

3.2. Optimization of miconazole nanogrinding

Miconazole nanogrinding was further optimized by a 23 DOE using SDS in combination with HPC-LF

(Table 2). The three factors and their levels (concentration) were miconazole (5 and 20%, w/w), SDS

(0.05 and 0.2%, w/w), and HPC-LF (1.25 and 5%, w/w). High miconazole and HPC-LF concentrations

promoted particle size reduction, as expressed by the d50 and d90 values (Table 5).

Chapter II

Page 47 of 182

Table 5: Miconazole particle undersize diameters d50 and d90 achieved by nanogrinding using the miconazole and

the excipients HPC-LF and SDS according to a factorial 23 design with a repeated centre-point. For comparison

the particle size parameters before nanogrinding were: d50 = 27,000 nm; d90 = 49,000 nm.

Miconazole

(%, w/w)

SDS

(%, w/w)

HPC-LF

(%, w/w)

Particle size d50

(nm) ± sd

Particle size d90

(nm) ± sd

5 0.05 1.25 164 ± 3 3,300 ± 138

5 0.2 1.25 163 ± 0 3,157 ± 62

5 0.05 5 150 ± 2 445 ± 21

5 0.2 5 169 ± 8 1,044 ± 52

12.5 0.125 3.1251 150 ± 1 536 ± 11

12.5 0.125 3.125 145 ± 1 507 ± 1

20 0.05 1.25 156 ± 1 1,433 ± 28

20 0.2 1.25 153 ± 2 781 ± 57

20 0.05 5 140 ± 1 413 ± 10

20 0.2 5 152 ± 1 595 ± 20

Conversely, the effect of SDS concentration on d50 and d90 differed depending on the polymer

concentration; increasing SDS concentration caused particle size (d90) decrease at low HPC-LF

concentration (1.25%, w/w), but particle size (d50 and d90) increase at high HPC-LF concentration (5%,

w/w). Both d50 and d90 values were fitted to a polynomial equation (Eq. 1) after normalizing the

coefficients, in order to evaluate the main effects and interactions affecting the d50 (Eq. 2; R2 = 0.95)

and d90 values (Eq. 3; R2 = 0.99). The polynomial equations include only statically significant variables,

except the factor -1.7WSDS of Eq. 3, which could not be removed due to two significant interactions with

SDS (-115.7WMICWSDS and 197WHPC-LFWSDS). The obtained polynomial equations were:

��= 156 − 5.6W�� + 3.4W�� – 3.1W�� + 4.4W��W�� (2)

��= 1396 − 590.5W�� − 1.7W�� – 771.8W�� − 115.7W��W�� + 470.2W��W�� +197W��W�� (3)

where W corresponds to:

!= (#�,�,� − %&'()&*+,'(%+'%&'()�(,+') (%+'%&'()�(,+')�'.& 2)⁄⁄ (4)

with Xi,j,k being the corresponding concentrations (in %, w/w) of miconazole, SDS, and HPC-LF,

respectively.

Chapter II

Page 48 of 182

Miconazole and HPC-LF concentrations exerted the main effects on miconazole particle size reduction

with the HPC-LF concentration being more important for reducing the larger particles (d90). SDS

concentration, on the other hand, exerted a detrimental effect on the the d50 value (p < 0.05), but no

significant effect on the larger particle size fraction (d90) (p > 0.05). Besides the main effects, significant

(α = 0.05) positive interactions were observed between HPC-LF and SDS for both the d50 and d90 values

as well as between miconazole and HPC-LF for the d90 value. By contrast, a significant negative

interaction was determined between miconazole and SDS for d90 (p < 0.05).

The DOE was an important tool to explore and consolidate the effects of the formulation variables on

miconazole particle size reduction. For confirming and further exploring the beneficial effects of high

miconazole and HPC-LF concentrations and the interaction effect of HPC-LF and SDS, the milling

experiments were extended using additional concentrations (Table 5, Fig. 4). The additional miconazole

concentrations of 12.5% and 25%, at fixed SDS (0.05%) and HPC-LF (5%) concentrations, yielded

consistent particle size values in comparison with the previous experiments using 5 and 20% of drug

substance (Fig. 4A, B). The variation of SDS concentration from 0 to 0.2%, at fixed concentrations of

miconazole (12.5 or 20%) and HPC-LF (5%), revealed that a minimal amount of SDS (0.0125%) was

necessary for efficient nanogrinding (Fig. 4C, D). However, increasing the SDS concentration from

0.0125 to 0.125% resulted in significantly (α = 0.05) larger particle sizes (for d50 and d90). As shown

before in the DOE experiments, 0.2% SDS was detrimental for nanogrinding. This data is consistent

with the fact that HPC-LF interacts with SDS above a critical aggregation concentration of 1.5 mM

SDS (0.0433%), thereby probably reducing the adsorption of HPC-LF to the miconazole particles

(Berglund et al., 2003).

For increasing the HPC-LF concentration above 5%, the miconazole concentration had to be kept at

maximal 12.5% to avoid excessive suspension viscosity hampering the processing. Increasing the HPC-

LF concentration from 1.25% to 6.25% (at 0.05% SDS) lowered the d50 and d90 to minimal values of

approx. 150 nm and 400 nm, respectively (Fig. 4E, F). The data obtained with the higher HPC-LF and

miconazole concentrations suggests that an adequately high viscosity may be one of the parameters that

promote particle breakage. Conversely, an upper viscosity limit seemed to exist above which particle

breakage was no further promoted.

To analyze the effect of viscosity on miconazole particle size reduction, the d50 and d90 values were

plotted against the viscosity of all the starting suspensions used for nanogrinding (Fig. 5). The viscosity

of the starting suspensions affected mainly the size reduction of the larger particles (d90). The results

indicate that a minimum viscosity of approx. 50 mPa•s to 100 mPa•s was required to obtain appropriate

Chapter II

Page 49 of 182

d90 values. On the other side, suspension viscosities exceeding 1300 mPa•s hampered the processability

of the suspensions in the actual stirred media mill.

Figure 4: Reduction of the miconazole particle sizes, expressed as d50 (panels A, C, E) and d90 (panels B, D, F),

upon nanogrinding as a function of miconazole (A, B), SDS (C, D), and HPC-LF (E, F) concentrations.

130

140

150

160

170

180

190

0 10 20 30

Pa

rtic

le s

ize

d50

(nm

)

Miconazole concentration (%, w/w)

HPC-LF: 5 %

SDS: 0.05 %

A)

300

500

700

900

1100

1300

1500

1700

0 10 20 30

Pa

rtic

le s

ize

d90

(nm

)

Miconazole concentration (%, w/w)

HPC-LF: 5 %

SDS: 0.05 %

B)

130

140

150

160

170

180

190

0.00 0.10 0.20 0.30

Pa

rtic

le s

ize

d50 (

nm

)

SDS concentration (%, w/w)

MIC: 12.5 %

HPC-LF: 5 %

MIC: 20 %

HPC-LF: 5 %

C)

300

500

700

900

1100

1300

1500

1700

0.00 0.10 0.20 0.30

Pa

rtic

le s

ize

d90

(nm

)


MIC: 12.5 %

HPC-LF: 5 % MIC: 20 %

HPC-LF: 5 %

D)

130

140

150

160

170

180

190

0 2 4 6 8

Pa

rtic

le s

ize

d50

(nm

)

HPC-LF concentration (%, w/w)

MIC: 20 %

SDS: 0.05 %MIC: 12.5 %

SDS: 0.05 %

E)

300

500

700

900

1100

1300

1500

1700

0 2 4 6 8

Pa

rtic

le s

ize

d90

(nm

)


MIC: 20 %

SDS: 0.05 %

MIC: 12.5 %

SDS: 0.05 %

F)

Chapter II

Page 50 of 182

Figure 5: Nanoground miconazole particle sizes, expressed as d50 (A) and d90 (B), as a function of the viscosity of

the miconazole suspensions before milling.

3.3. Stability of miconazole nanosuspensions during storage

Storage of the nanosuspensions at 5 °C for 6 months generally caused only minor particle growth, in

contrast to storage at 25 °C (Fig. 6). High miconazole concentration (20%; in presence of 0.05% SDS

and 5% HPC-LF) stabilized much better the particle sizes during storage than the low (5%) drug

substance concentration (Fig. 6A, B). Similarly, higher SDS concentration (0.2%) seemed to be

preferable for storage of the nanosuspensions at 25 °C, as 0.05% SDS could not prevent substantial

particle size growth, especially in the large particle size fraction (d90) (Fig. 6D). Interestingly,

microscopic observation of the nanosuspensions stored at 25 °C revealed that the increase of the large

particle size fraction was probably mainly caused by crystal growth (Ostwald ripening). On the contrary

nanosuspensions formulated without SDS presented an important fraction of particle aggregates, both

before and after storage at 25 °C (data not shown). As for miconazole and SDS, the higher

concentration of HPC-LF (5% versus 1.25%) attenuated the particle size growth during storage at 5 °C

and 25 °C (Fig. 6E, F). When comparing the differences in particle size growth of formulations stored

at 5 °C and 25 °C, it appears that the steric stabilization, as provided by HPC-LF, was slightly more

sensitive to the increased temperature than the electrostatic repulsion, as provided by the SDS

(Rabinow, 2004).

100

120

140

160

180

200

0 200 400 600 800 1000 1200 1400

Pa

rtic

le s

ize

d50

(nm

)

Viscosity before milling (mPas)

A)

0

500

1000

1500

2000

2500

3000

3500

4000

0 200 400 600 800 1000 1200 1400

Pa

rtic

le s

ize

d90

(nm

)

Viscosity before milling (mPas)

B)

Chapter II

Page 51 of 182

Figure 6: Changes of the particle size parameters d50 (panels A, C, E) and d90 (panels B, D, F) of miconazole

nanosuspensions during storage over 6 months at 5 and 25 °C. Effects of miconazole (A, B), SDS (C, D), and

HPC-LF (E, F) concentrations.

Miconazole solubility and zeta potential values were also examined before and after storage of the

nanosuspensions at 5 °C and 25 °C, although data are available only for 3 months storage (Table 6). As

expected, the higher miconazole (20%), SDS (0.2%), and HPC-LF (5%) concentrations all increased the

0

50

100

150

200

250

300

350

5 20

Pa

rtic

le s

ize

d50

(nm

)

Miconazole (%, w/w)

HPC-LF: 5 %

SDS: 0.05 %

Applies to all plots

0 m

on

ths

6 m

at

5 °

C

6 m

at

25

°C

A)

0

500

1000

1500

2000

2500

3000

5 20

Pa

rtic

le s

ize

d90

(nm

)

Miconazole (%, w/w)

HPC-LF: 5 %

SDS: 0.05 %

B)

0

50

100

150

200

250

300

350

0 0.05 0.2

Pa

rtic

le s

ize

d50

(nm

)


MIC: 20 %

HPC-LF: 5 %

C)

0

500

1000

1500

2000

2500

3000

0 0.05 0.2

Pa

rtic

le s

ize

d90

(nm

)


MIC: 20 %

HPC-LF: 5 %

D)

0

50

100

150

200

250

300

350

1.25 5

Pa

rtic

le s

ize

d50

(nm

)


MIC: 20 %

SDS: 0.05 %

E)

0

500

1000

1500

2000

2500

3000

1.25 5

Pa

rtic

le s

ize

d90

(nm

)


MIC: 20 %

SDS: 0.05 %

F)

Chapter II

Page 52 of 182

miconazole solubility in the nanosuspensions before storage. The low drug solubility (1 µg/ml) in the

medium containing low concentrations of SDS (0.05%) and HPC-LF (1.25%) must be ascribed partly to

the limited solubilising capacity of both excipients at low concentration and partly to the fraction of

large particles (d90 of approx. 1,400 nm, see Table 5) found in this formulation in comparison to the

other formulations shown in Table 6 (with d90 values in the range of 400 to 600 nm, see Table 5). Upon

storage, miconazole solubility tended to decrease slightly, although this could not be confirmed

statistically. The zeta potential values were mostly influenced by the SDS concentration of the

nanosuspensions and did not change during the three month storage at 5 °C and 25 °C. Finally, neither

the miconazole solubility nor the zeta-potential data could be related to the particle size stability results.

Table 6: Miconazole solubility and zeta-potential in nanosuspensions before and after 3 months of storage.

Component (%, w/w) Miconazole solubility (µg/ml) before

and after 3 months of storage

Zeta-potential,

x1 ± s. d., n=6 (mV)

before and after 3

months of storagea)

Miconazole SDS HPC-LF 0 months 3 months

at 5 °C

3 months

at 25 °C

5 0.05 5 21 38 21 -15 ± 0

20 0.05 5 58 47 15 -12 ± 1

20 0.20 5 86 52 56 -19 ± 1

20 0.05 1.25 1 1 1 -12 ± 1

a) The zeta potential did not alter during storage at 5 and 25 °C over 3 months.

4. Discussion

Nanogrinding is a complex process requiring the selection of adequate formulation and process

parameters to obtain appropriate particle size reduction and stability of nanosuspensions. In this study,

we focussed on the importance of formulation and related physical-chemical parameters. The selection

of appropriate excipients is governed by two main functional criteria: (i) wetting of the drug substance

(Merisko-Liversidge et al., 2003), and (ii) steric and/or electrostatic stabilization of the nanoparticles

(Rabinow, 2004). As the literature does not provide any rational criteria for the selection of excipients

and process conditions, formulation development was done empirically. The initial screening of surface

active and polymeric excipients revealed that HPC (both HPC-LF and HPC-EF) in combination with

SDS was the most adequate, while poloxamer alone or the mixture of PVP and SDS were unsuitable for

miconazole nanogrinding (Figs 1 and 2). Good nanogrinding results were also obtained with the

excipient mixtures of HPMC/SDS and HPC-LF/benzalkonium chloride (Fig. 3). While the possibility of

substituting HPC by HPMC does not surprise, given their similar structure and properties, the suitability

Chapter II

Page 53 of 182

of both the cationic benzalkonium chloride and the anionic SDS suggests that the interaction between

polymer and drug substance was more important than the interaction between surfactant and polymer.

Such conclusion has already been made previously with SDS and benzethonium chloride (cationic

surfactant), which were found to be equally effective for nanogrinding with both improving the

interaction between polymer and 11 different drugs, thus promoting particle size reduction (Lee et al.,

2008). However, the replacement of SDS by sodium docusate (also an anionic surfactant) in our study

did not provide comparatively effective particle size reduction (Fig. 3). With sodium docusate, a

relatively important fraction of coarse particles (diameters of 130 µm) remained in the nanosuspension

after milling.

The results of the excipient screening study are in general agreement with reports in the literature, in

particular regarding the usefulness of combining SDS and cellulose ethers (Rabinow, 2004; Lee et al.,

2008; Van Eerdenbrugh et al., 2009). In a large formulation screening study using different drug

substances, surfactants, and polymers, SDS promoted the nanogrinding in eight cases, but caused

particle growth in five formulations (Lee et al., 2008). In fact, SDS was found efficient in combination

with HPC, PVP and poloxamer 407 for the drug substances prednisolone acetate, nifedipin,

hydrocortisone acetate, and itraconazole. SDS and HPC-LF are known to interact with each other and

form polymer-surfactant aggregates (Evertsson and Nilsson, 1997; Berglund et al., 2003; Lee et al.,

2008). Through such interaction, SDS may have facilitated the adsorption of HPC-LF on miconazole,

thus promoting the formation of an entropic barrier preventing aggregation of nanoground drug

particles (Choi et al., 2005). Moreover, the interaction between SDS and HPC-LF reduces the self-

repulsion of the anionic SDS molecules thereby affording a greater particle surface coverage (Rabinow,

2004). The inferior suitability of PVP for producing nanosuspensions has already been described

earlier, when HPC, HPMC, poloxamer, PEG, and PVP were compared as nanosuspension stabilizers at

concentrations of approx. 17% relative to the concentration of drug substance (Lee et al., 2008). More

recently however, higher PVP concentrations, i.e., 25% to 100% relative to the drug substance,

produced more favourable results (Van Eerdenbrugh et al., 2009). The authors concluded that PVP is a

valuable stabilizer when used at high concentration; use of high PVP concentrations is perfectly feasible

thanks to the modest viscosifying capacity of PVP. With regards to poloxamer, previous studies

described this polymer as highly versatile, as the hydrophobic PPO block adsorbs efficiently on

hydrophobic surfaces of insoluble nanoparticles (Lee et al., 2008). Possibly other poloxamer types

would be more suitable for miconazole nanogrinding than the poloxamer 188 used in the present study.

It is noteworthy that the screening experiments revealed two phenomenological parameters that

predicted quite well the success of miconazole nanogrinding: (i) the contact angle between the drug

Chapter II

Page 54 of 182

substance and the stabilizer solution, which had to be relatively low; (ii) the dispersibility of the starting

drug particles in the stabilizer solution, which had to be high and free of microscopically visible

agglomerates.

Besides the selection of appropriate excipients, the optimization of their and the drug substance

concentration in the suspension is equally important for nanogrinding (Table 5, Fig. 4) and

nanosuspension stability (Fig. 6). First, the concentration of drug substance in the suspension for

nanogrinding needs to be sufficiently high to ascertain an elevated frequency of drug particle capture in

the active grinding zone between beads; thereby, the milling energy of the beads is adequately

transferred onto the drug particles (Stenger et al., 2005). Second, a minute amount of SDS (0.0125%)

was required to provide adequate drug particle wetting for efficient nanogrinding. Higher amounts of

SDS (> 0.05%) were found to be detrimental for nanogrinding when HPC-LF was present at elevated

concentration (Table 5, Fig. 4). This observation might be explained by the competitive displacement of

adsorbed HPC-LF by increasing SDS concentration (Evertsson and Nilsson, 1997; Berglund et al.,

2003; Lee et al., 2008). In a model system using the hydrophobic poly(dimethyl siloxane) as adsorbant,

HPC-LF adsorption was maximal at 1 mM SDS (0.03%, w/v) and decreased at higher SDS

concentrations (up to 6 mM SDS; 0.17%) (Berglund et al., 2003). Displacement of the polymeric

stabilizer from the drug surface likely lowers the steric stabilization provided by HPC-LF. This was

highly detrimental in the case of miconazole as steric stabilization by HPC-LF was found to be crucial

for efficient particle size reduction and nanosuspension stabilization, which is in agreement with other

reports using other drugs (Lee 2003; Ain-Ai and Gupta 2008; Kobierski et al., 2009). The amount of

polymeric excipient (e.g., HPC) not only determines its adsorption onto the drug substance particles,

but also contributes to the viscosity of the suspension medium, thereby increasing the diffusion barrier

for particle-particle interaction (Ploehn and Russel, 1990). For the miconazole nanogrinding, good

results were obtained at HPC-LF concentrations of 3.125% and higher.

The viscosity of the miconazole suspensions for nanogrinding containing SDS and HPC-LF determined

mainly the fraction of larger particles present after nanogrinding (Fig. 5). A minimal viscosity was

required for efficient nanogrinding; under the present conditions, the lower critical viscosity was

approx. 100 mPa•s. Incidentally, the viscosity of the suspension resulted from the amounts of drug

substance and polymeric excipient present in the suspensions. However, above an upper critical

viscosity (not determined in this work), nanogrinding became less effective, which was explained by

hindered movement of the milling beads (Kwade, 1999). Lower and upper critical viscosity values

depend on the actual compound to be ground and the milling equipment and conditions. While in the

present work, the upper critical viscosity appeared to be above 1,000 mPa•s (Fig. 5), such value was in

Chapter II

Page 55 of 182

the range of 170 mPa•s to 470 mPa•s for the inorganic fused corundum (α-Al2O3), used at

concentrations above 20% (Stenger et al., 2005).

Finally, storage stability of the produced nanosuspensions in terms of particle size growth was

acceptable when the nanosuspensions were stored at 5 °C for 6 months (Fig. 6). By contrast, when

stored at 25 °C for 6 months, two types of larger particles became microscopically visible, i.e., particle

agglomerates and larger individual crystals. Particle agglomerates were present in the absence of SDS,

and crystal growth had occurred in the presence of SDS. The former phenomenon can be explained by

the zeta-potential value that was close to zero in the absence of SDS, whereas crystal growth, also

known as Ostwald ripening, was likely caused by increased drug substance solubility at 25 °C and

increased SDS concentration (Verma et al., 2009). Under such supersaturation conditions, some of the

dissolved drug re-precipitated onto the larger particles with a lower surface energy.

5. Conclusions

As rational predictions for optimal excipients for nanogrinding are presently not available, this work

provides simple empiric tools to orient experiments for fast achievement of stable nanosuspensions.

Efficient particle size reduction by nanogrinding requires the use of excipients that provide proper

wetting and physical stabilization (steric and electrostatic) of the practically water-insoluble drug

substances. We found that a low contact angle between drug substance and dispersion medium in

combination with an excellent agglomerate-free dispersibility of the micronized drug particles in the

medium (before milling) can provide an indication of the suitability of excipients. For nanogrinding

miconazole, the combination of 0.025% to 0.05% SDS and 5% HPC-LF was most suitable providing a

synergistic effect for particle size reduction (d50 = 145 nm) and nanosuspension stabilization. SDS

mediated wetting and facilitated the adsorption of HPC-LF onto the miconazole particles; HPC-LF

adsorbed extensively onto the nanoparticles, thereby affording steric protection from agglomeration and

crystal growth. The present findings may facilitate and accelerate the nanogrinding of other drug

substances, as we have shown that prediction of particle size reduction and nanosuspension stability

may be feasible, to some extent, from simple preformulation experiments. The study also emphasizes

the importance of formulation development before process parameters should be optimized. Finally,

appropriate particle size reduction and nanosuspension stability of practically water-insoluble drugs are

important not only for enhancing the dissolution rate and bioavailability of the drug, but also for the

safe use of the medicament by different administration routes such as the oral, nasal, ophthalmic and

parenteral routes.

Chapter II

Page 56 of 182

Acknowledgments

The authors thank Dr. Elke Walter for valuable comments and suggestions, and Sarah Barthold for the

experimental contributions.

Chapter II

Page 57 of 182

References

Ain-Ai, A., Gupta, P.K., 2008. Effect of arginine hydrochloride and hydroxypropyl cellulose as

stabilizers on the physical stability of high drug loading nanosuspensions of a poorly soluble

compound. Int. J. Pharm. 351, 282–288.

Augustijns, P., Van den Mooter, G., Van Eerdenbrugh, B., 2008. Top-down production of drug

nanocrystals: nanosuspension stabilization, miniaturization and transformation into solid products.

Int. J. Pharm. 364, 64–75.

Aulton, M.E., 2007. Aulton’s Pharmaceutics. The design and manufacture of medicines, 3rd Ed.,

Churchill Livingstone, Edinburgh, pp. 65, 78.

Berglund, D.K., Przybycien, T.M., Tilton, R.D., 2003. Coadsorption of sodium dodecyl sulfate with

hydrophobically modified nonionic cellulose polymers. 1. Role of polymer hydrophobic

modification. Langmuir 19, 2705–2713.

Choi, J-Y, Yoo, J.Y., Kwak, H-S., Nam, B.U., Lee, J., 2005. Role of polymeric stabilizers for drug


Crison J.R., 2000. Biopharmaceutical aspects of water-insoluble drugs for oral drug delivery. In: Rong,

L. (Ed.), Water insoluble drug formulation. Interpharm / CRC, Boca Raton, pp. 98–108.

Evertsson, H., Nilsson, S., 1997. Microviscosity in clusters of ethyl hydroxyethyl cellulose and sodium

dodecyl sulfate formed in dilute aqueous solutions as determined with fluorescence probe

techniques. Macromolecules. 30, 2377–2385.

Gardner, C.R., Walsh, C.T., Almarsson, O., 2004. Drugs as materials: valuing physical form in drug

discovery. Nat. Rev. Drug Discov. 3, 926–934.

Kesisoglou, F., Panmai, S., Wu, Y., 2007. Nanosizing – oral formulation development and

biopharmaceutical evaluation. Adv. Drug Del. Rev. 59, 631–644.

Kipp, J.E., 2004. The role of solid nanoparticle technology in the parenteral delivery of poorly water-

soluble drugs. Int. J. Pharm. 284, 109–122.

Kobierski, S., Ofori-Kwakye, K, Müller, R.H., Keck, C.M., 2009. Resveratrol nanosuspensions for

dermal application--production, characterization, and physical stability. Pharmazie. 64, 741-747.

Kwade, A., 1999. Wet comminution in stirred media mills – research and its practical application.

Powder Technol. 105, 14–20.

Lee, J., 2003. Drug nano and microparticles in solid dosage forms. J. Pharm. Sci. 92, 2057-2068.

Lee, J., Choi, J.-Y, Park, C.H., 2008. Characteristics of polymers enabling nano-comminution of water

insoluble drugs. Int. J. Pharm. 355, 328–336.

Liversidge, G.G., Cundy, K.C., Bishop, J., Czekai, D., 1992. Surface modified drug nanoparticles. US

Patent 5,145,684.

Chapter II

Page 58 of 182

Merisko-Liversidge, E., Liversidge, G.G., Cooper, E.R., 2003. Nanosizing: a formulation approach for


Muller, R.H., Jacobs, C., Kayser, O., 2001. Nanosuspensions as particulate drug formulations in therapy

rationale for development and what we can expect for the future. Adv. Drug Delivery Rev. 47, 3–19.

Ploehn, H.J., Russel, W.B., 1990. Interactions between colloidal particles and soluble polymers. Adv.

Chem. Eng. 15, 137–228.


Riley, S., 2006. Innovation in drug delivery. Business Insights.

Stenger, F., Mende, S., Schwedes, J., Peukert, W., 2005. Nanomilling in stirred media mills. Chem.

Eng. Sci. 60, 4557–4565.

Sweetman, S., 2006. Martindale. The complete drug reference, 35th ed., Pharmaceutical Press, London.

Van Eerdenbrugh, B., Vermant, J., Martens, J. A., Froyen, L., Van Humbeeck, J.V., Augustijns, P., Van

den Mooter, G., 2009. A screening study of surface stabilization during the production of drug

nanocrystals. J. Pharm. Sci. 98, 2091–2103.



Chapter III

Page 59 of 182

Chapter III

Role of milling parameters on nanogrinding of drug substances of similar mechanical properties2

1. Introduction

Nanosuspensions, defined as colloidal dispersions of nanosized (200–400 nm) drug particles, represent

a recent approach to enhance the solubility and bioavailability of drug substances of limited solubility

(Lipinski, 2002; Merisko-Liversidge et al., 2003; Müller et al., 2004; Patravale et al., 2004).

Nanosuspensions can be produced either by precipitation of the drug substance dissolved in an organic

solvent (bottom-up approach) or by size reduction of the native drug particles (top-down approach)

(Rabinow, 2004; Verma et al., 2009). Efficient particle size reduction can typically be achieved by wet-

nanogrinding using high-energy mills (Kwade, 1999a). Here, particle size reduction depends on: (i) the

milling parameters; (ii) mechanical properties of the (drug) particles; (iii) appropriate particle

stabilization and, hence, nanosuspension composition (Kwade, 1999a; Merisko-Liversidge et al., 2003).

To achieve stable nanosuspensions, formulation and process parameters require careful optimization for

each individual drug substance. Firstly, excipients have to be determined that provide adequate wetting

of the drug particles for nanogrinding and particle stabilization after milling. The latter issue is

particularly critical, as the nanogrinding increases greatly the specific surface area and surface energy

(Gibbs free energy) of the drug particles (Rabinow, 2004), which may cause particle aggregation and

changes in suspension viscosity (Mende et al., 2003). Upon adsorption of stabilizers onto the drug

nanoparticles, their aggregation can be prevented or at least slowed down by electrostatic and/or steric

effects. Thus, selection of an appropriate type and amount of stabilizer is a prerequisite to promote

particle size reduction and generate physically stable nanosuspensions (Merisko-Liversidge et al.,

2003). In nanogrinding, the particles are subjected to stress, resulting in their breakage. The applied

stress is concentrated on the cracks already present in the material, which causes crack propagation

leading to fracture (Bernotat et al., 2003). Thus, the surface morphology, i.e., presence of cracks or

other inhomogeneities or deficiencies, can be of great importance in the explanation of different

2 The work in this chapter has been published as Cerdeira, AM., Mazzotti, M., Gander, B., 2011. Role of Milling Parameters and Particle

Stabilization on Nanogrinding of Drug Substances of Similar Mechanical Properties. Chem. Eng. Tech. 34, 1427–1438.

Chapter III

Page 60 of 182

breaking behaviors of different drug substances (Kwade, 1999a). According to Hooke’s law, stress is

directly proportional to strain. The proportionality constant E (Young’s modulus or modulus of

elasticity) is a measure of the hardness, stiffness, or rigidity of the solid, i.e., its resistance to

deformation. In nanogrinding, various process parameters define the milling efficiency and the

achievable particle sizes. Wet-nanogrinding using high-shear mills and milling beads is typically

performed with circumferential stirrer speeds of 8 ms–1 to 20 ms–1 in either batch or circulation mode

under controlled temperature (Kwade, 1999a; Date et al., 2004). In the circulation mode, the milling

beads are held back in the milling chamber by either a sieve or a defined gap between rotating and static

parts. Commonly, 70% to 85% of the grinding chamber volume of the mill is filled with beads of sizes

ranging from 0.2 mm to 4.0 mm; it should be noted that the percentage is generally expressed in terms

of apparent bead volume relative to the true grinding chamber volume. A minimum milling time of 30

min to 60 min is generally required to obtain nanosuspensions with a monomodal particle size

distribution and a mean diameter of below 200 nm. With optimized formulation and process

parameters, the batch-to-batch reproducibility of particle size distributions is generally high (Date et al.,

2004). Particle size reduction by nanogrinding is essentially determined by the number of collisions

between drug particles and beads, and by the intensity of such collisions (Kwade, 1999b). Thus, the

nanogrinding process in stirred media mills can be described by the parameters stress number (N),

stress energy of grinding beads (SEB), and specific energy input (SEI). In a batch process, the stress

number N can be determined by Eq. (1) (Kwade et al., 2002):

2

B

B

d

tn

Cε)}(1{1

ε)(1 N

B−−

−∝

ϕ

ϕ

(1)

where n is the number of stirrer revolutions per unit time, t is the milling time, ϕB is the volume fraction

of the grinding beads relative to the volume of the grinding chamber, ε is the porosity of the bulk of the

beads, dB is the diameter of the grinding beads (m), and C is the drug substance concentration in the

suspension (gm–3). Eq. (1) indicates that a very important parameter influencing the stress number is the

size of the grinding beads, i.e., smaller beads yield higher stress numbers.

The stress energy of grinding beads SEB is defined as the energy transferred to a product particle during

one stress event (Stenger et al., 2005); therefore, it is the maximum energy that can be transferred from

two colliding beads to the particle being caught in-between them. The stress energy of the grinding

beads is determined by the kinetic energy, density and size of the beads, and the rotation rate of the

stirrer, according to Eq. (2):

SEB ∝ dB3 ρB ν

2 (2)

Chapter III

Page 61 of 182

where ρB is the density of the beads (gm–3) and ν is the rotational speed of the stirrer tip (ms–1). Here,

the diameter of the beads and the speed of the stirrer affect most strongly the SEB. Finally, the specific

energy input SEI is the total energy input relative to the total mass (or volume) of suspension and it can

be calculated by (Stenger et al., 2005):

( )

B

0t

IEm5.0m

dtNNS

∆+

−∫=

(3)

where m is the mass of drug substance (g), Nt is the power consumption (kW) at time t, N0 is the power

consumption of the unloaded mill (kW), and ∆mB is the wear at time t of the grinding beads (g). The SEI,

which is proportional to N SEB, can be measured by the power consumption over time (Kwade et al.,

2002).

The present work focuses on the importance of process parameters of wet-nanogrinding for the particle

size reduction of the poorly water-soluble antifungal compounds miconazole and itraconazole, and the

antiviral compound etravirine, all having similar mechanical properties in terms of elastic and plastic

deformation and formulated according to a previously optimized composition for miconazole (Cerdeira

et al., 2010).


2.1. Materials

Miconazole (mean particle size d50: 15 µm; lot # R018134PUC701, Janssen Pharmaceutica N.V., Geel,

Belgium), itraconazole (mean particle size d50: 21 µm; lot # ZR051211PUK401, Janssen Pharmaceutica

N.V., Geel, Belgium), etravirine (mean particle size d50: 11 µm; lot # R165335PUA121, Janssen

Pharmaceutica N.V., Geel, Belgium), sodium dodecyl sulfate [SDS] (Texapon® K12P, Cognis,

Düsseldorf, Germany), and hydroxypropylcellulose [HPC-LF] (Klucel® LF, Hercules, Doel, Belgium)

were used as received.

2.2. Surface Morphology of Unprocessed Drug Particles

Unprocessed miconazole, itraconazole, and etravirine crystals were dispersed onto sticky carbon tabs

and coated with 7–10 nm of gold for scanning electronic microscopy (SEM) (electron microscope Leo

1430VP, Carl Zeiss, Germany). The unprocessed drug powders were also immobilized using a thin

Chapter III

Page 62 of 182

layer of epoxy resin on a metal AFM stub for atomic force microscopy (AFM). The images were

recorded in tapping mode using a Digital Instruments Nanoscope IIa equipped with silicon cantilevers

with a resonant frequency of approx. 260 kHz; scan rates were typically at 1–2 Hz.

2.3. Mechanical Properties of the Drug Substances

Young’s modulus values were determined by elastically indenting drug crystal surfaces using an AFM

tip. Force-distance curves were obtained with a Digital Instruments Nanoscope IIIa AFM equipped with

diamond-like carbon-coated probes (TESPD) with a spring constant of ~30Nm–1, which was determined

accurately prior to the sample measurements. The approach/retract rate was kept at 1.0 Hz and the

contact load at 250–400 nN. Size (radius of indenter R) and shape of the AFM tips used for indentation

were monitored both before and after the force-distance measurements, using a tip selfimaging process.

All measurements were obtained under ambient conditions. For each drug substance, elastic

deformation was monitored on at least 6 different particles, with 16 force distance-curves being

registered for each particle at different locations across the surface. Curves that showed evidence of

anomalies or slippage related to surface roughness were not used for calculation of Young’s modulus.

Young’s modulus (E) values were calculated from force-indentation data using inhouse software

(Molecular Profiles, Nottingham) and the Hertz calculation model:

( )213

2

)8(4

13E

R

L s

δ

υ−= (4)

where L is the applied load (N), νs the Poisson’s ratio (estimated to be 0.3; mid-range), δ the indentation

depth (mm), i.e. the maximum penetration depth into the sample under elastic deformation, as

determined from the force-distance curves, and R is the radius of the indenter (mm).

Plastic deformation of the drug particles was determined from force-distance curves using AFM

equipped with a three sided pyramid-shaped diamond indenter mounted at the end of a large steel

cantilever (spring constant = 226 Nm–1). Particles were first located by the optical microscope of the

AFM, and an initial AFM image was taken in tapping mode using a standard tip. The tapping mode tip

was then replaced with the diamond indenter, which was aligned over the region of interest of the

particle. For assessing plastic deformation, three larger particles with a relatively flat exposed surface

were chosen to increase the ease of the alignment of the diamond indenter with the area suitable for

measurements. The load for indentation was varied from 1 to 19 µN, and the loading rate was kept

constant at 2 µm s-1. Nine force-distance curves per particle were recorded along a grid pattern. Such

Chapter III

Page 63 of 182

group of nine indentations was termed set and was repeated three times on each particle surface at

varying locations, so that a total of 27 measurements per particle were performed. After completion of

the sets, the diamond indenter was replaced by a conventional tapping mode tip, and the area was re-

imaged. Force curves were processed in a similar manner as described above for Young’s modulus data

to calculate the maximum loading force at each data point. Indent depths were measured from images

using line profiles taken across the indents with the analysis software provided with the instrument.

2.4. Miconazole, Itraconazole, and Etravirine Formulations for Nanogrinding

In a first set of experiments, the effect of milling parameters on particle size reduction of miconazole

was evaluated using a previously optimized formulation containing 20%, w/w, miconazole, 5%, w/w

HPC, and 0.025%, w/w SDS. In a second set of experiments, adequate stabilizers for itraconazole and

etravirine were determined for extending the process parameter study on these drug substances.

Amongst different stabilizers tested (SDS, poloxamers, and cellulose ethers), the combination of SDS

and HPC was found to be also most suitable for itraconazole and etravirine. While the concentration of

HPC appeared to be less critical for particle stabilization of the three drug substances, the SDS

concentration revealed to be crucial for etravirine. Thus, nanogrinding experiments were performed

with 20%, w/w drug substance, 5%, w/w HPC, and increasing concentrations from 0% to 0.2%, w/w of

SDS.

Thirdly, based on the results obtained in the formulation optimization, fixed compositions containing

20%, w/w drug substance, 5%, w/w HPC, and 0.025%, w/w SDS were used to study the importance of

the milling parameters on the particle size reduction of all three drug substances. The suspensions for

nanogrinding were prepared by dissolving SDS in purified water, adding HPC under mechanical

stirring, and maintaining stirring for 30 min until all HPC was dissolved. The raw drug substance was

then dispersed in the excipient solution by means of mechanical stirring for minimum 60 min.

2.5. Mill and Grinding Media

Nanogrinding was performed in a high-energy mill (LabStar, Netzsch, Germany) in circulation mode.

The mill was refrigerated to keep the product temperature below 37 °C. Yttrium stabilized zirconium

oxide beads with diameters of 0.4 and 0.8 mm were used as grinding media. The milling parameters

investigated were: (i) bead size (0.4 and 0.8 mm); (ii) filling volume fraction of the apparent volume of

grinding beads relative to the volume of the grinding chamber (v/v) (81%, 83% and 85%); (iii) product

flow rate (~ 97, 113, 183 gml–1); (iv) stirrer-tip speed (2400, 2800, 3000, 3400, and 3600 rpm); (v)

milling time (15, 30, 60 min); see Tabs. 2 and 3 in Sect. 3.3.

Chapter III

Page 64 of 182

2.6. Particle Size Measurement

Particle size distribution (volume-based) was measured by laser light diffraction (Malvern 2000,

Malvern Instruments, UK), using the small-volume dispersion unit. The Mie theory served for particle

size calculation, using the following refractive indices: dispersant, 1.33; the real and imaginary particle

refractive indices used were, respectively: 1.55 and 0.001 for miconazole, 1.64 and 0.001 for

itraconazole, and 1.64 and 0.001 for etravirine. Particle sizes were expressed by the 50% and 90%

volume percentiles (d50 and d90). Two samples per batch were analyzed, and the measurements were

repeated three times for each sample. The results were presented as average ± standard deviation of the

six measurements.

2.7. Viscosity of the Nanosuspensions

The viscosity of the nanosuspensions was determined using a Rheometer (Rheostress® RS600 Haake,

ThermoScientific, USA). The measurements were performed at 20 °C ± 0.1 °C, using rotational mode

(constant shear rate of 100 s–1) and a cone-plate measuring system (plate diameter 60 mm, cone angle

1°). To avoid shear history effects, samples were kept at rest for 5 min after their application on the

measuring system and before starting the measurement.

3. Results

3.1. Particle Size and Surface Morphology of the Drug Particles

Untreated miconazole particles showed roundish irregular morphology (Fig. 1A), sized between 10 and

40 µm. Itraconazole and etravirine particles possessed elongated rectangular like shapes with flat and

smooth surfaces (Figs. 1B and C). Itraconazole particles had a maximum length of 100 µm and widths

of 5 µm to 15 µm, while etravirine particles had maximal lengths of 20 µm and widths of 2 µm to 5 µm;

smaller (1 µm to 5 µm) roundish particles of itraconazole and etravirine were also present (Figs. 1B and

C).

Closer inspection by AFM (Fig. 2A) revealed that the miconazole particles were layered or flaky with

the surface of the layers being quite smooth and their edges rounded (Fig. 2A). AFM images of

itraconazole and etravirine particles (Figs. 2C and E) confirmed the structural similarity between the

two drugs, which were characterized by densely packed crystalline terrace-like topographies (Figs. 2C

and E).

Chapter III

Page 65 of 182

A) Miconazole

B) Itraconazole

C) Etravirine

Figure 1: SEM images of nonmilled miconazole, itraconazole, and etravirine. Bar represents 10 µm for all panels.

3.2. Young’s Modulus and Plastic Deformation of Miconazole, Itraconazole, and

Etravirine Drug Particles

The Young’s modulus (E) values of miconazole (E = 3 ± 2 GPa), itraconazole (E = 5 ± 2 GPa), and

etravirine (E = 2 ± 1 GPa) were found to be comparable and typical for organic materials of

intermediate hardness. Unfortunately, the uneven surface at large scale of miconazole (Fig. 1A) limited

the number of successful measurements, whereas the relatively flat and smooth surfaces of itraconazole

and etravirine (Figs. 1B and C) offered a higher number of usable areas for force-distance

measurements.

Microscopic observation of the indented areas revealed a sink-in behavior of miconazole (Fig. 2B),

whilst itraconazole and etravirine showed piling-up (Figs. 2D and F). The plastic deformation behavior

of the three drug substances was also found to be comparable amongst the three drug substances (Figs.

3A–C), although miconazole showed a variable deformation behavior. For one out of the three

miconazole particles surveyed, plastic deformation was similar to that of itraconazole, whereas the other

two particles deformed slightly more readily (Fig. 3A). Taken together, the data suggest that the three

drug substances should require similar energies for particle breakage during nanogrinding.

Chapter III

Page 66 of 182

A) Miconazole B) Miconazole with

AFM tip identations

C) Itraconazole D) Itraconazole with tip

identations

E) Etravirine F) Etravirine with tip

identations

Figure 2: AFM pictures of the particle surfaces of the studied drug substances; (left) before applying mechanical

stress and (right) after applying mechanical stress by AFM tip resulting in indentations

3.3. Effect of Milling Parameters on Miconazole Particle Size Reduction and

Nanosuspension Viscosity

A previously optimized formulation (Cerdeira et al., 2010) containing 20%, w/w miconazole, 5%, w/w

HPC, and 0.025%, w/w SDS was selected to evaluate the effect of the milling parameters on

miconazole particle size reduction. The results reveal that miconazole particle size reduction depended

Chapter III

Page 67 of 182

primarily on the specific energy input, a value that was measured by the energy consumption during the

grinding process (Figs 4–6). This data is consistent with theory, which predicts that the breaking up of

particles is essentially determined by the stress number N and the stress energy of the grinding beads,

SEB, whose product is proportional to the specific energy input SEI (Eqs. 1 and 2). Interestingly, a

plateau of minimal miconazole particle sizes (d50 ~ 130 nm; d90 ~ 270 nm) was achieved at a specific

energy input of approximately 15 MJ/kg (bead size 0.4 mm), irrespective of the volume fraction of the

grinding beads in the milling chamber (Fig. 4), of the product flow rate of the continuous process (Fig.

5), and of the stirrer tip speed (Fig. 6).

A)

B)

0

20

40

60

80

100

120

140

160

0 5 10 15 20 25 30

Ind

ent

dep

th (

nm

)

Applied force (µN)

Particle 1

Particle 2

Particle 3

Miconazole E = 3 ± 2 GPa

0

20

40

60

80

100

120

140

160

0 5 10 15 20 25 30

Ind

ent

dep

th (

nm

)

Applied force (µN)

Particle 1

Particle 2

Particle 3

Itraconazole E = 5 ± 2 GPa

Chapter III

Page 68 of 182

C)

Figure 3: Plastic deformation of miconazole (A), itraconazole (B), and etravirine (C) by AFM

The latter factors exerted minor effects only at low specific energy input (Figs. 4 and 6). Milling of

miconazole suspensions for 60 min under stirrer tip speeds higher than 2800 rpm (8.3 ms–1) became

ineffective, because no further particle size reduction was achieved (Fig. 6). Thus, under the actual

experimental conditions, the process parameters time t (Eq. 1) and the size of the grinding beads, dB,

(Eqs. 1 and 2) exerted via the stress number N and the stress energy of the grinding beads SE the

dominant effects on particle size reduction. Calculated values of N and SEB for the different process

conditions studied and a milling time of 60 min are summarized in Table 1. As the milling time not only

determines the stress number N (Eq. 1), but very directly also the experimentally measured specific

energy input SEI (Eq. 3), its effect on particle size reduction was examined under constant SEB (3.8 · 10–5

Nm), using a bead size of 0.4 mm and a stirrer tip speed of 3400 rpm (10 ms-1). Both the d50 and d90

values decreased with longer milling time, with the d90 parameter indicating a more pronounced

decrease (Table 2). This series of experiments confirmed the previously observed optimum specific

energy input value of ~ 15 MJ/kg (Figs. 4–6), where the limit of particle size reduction is achieved for

miconazole and the particular mill used. The effect of the size of the grinding beads (0.4 or 0.8 mm)

was examined at stirrer tip speeds of 2800 rpm (8.3 ms-1) and 3400 rpm (10.0 ms-1) (Table 1). At a

given specific energy input, the smaller grinding beads produced slightly smaller drug particles with

narrower particle size distributions (Fig. 7), at both stirrer tip speeds.

0

20

40

60

80

100

120

140

160

0 5 10 15 20 25 30

Ind

ent

dep

th (

nm

)

Applied force (µN)

Particle 1

Particle 2

Particle 3

Etravirine E = 2 ± 1 GPa

Chapter III

Page 69 of 182

Figure 4: Influence of specific energy input and of the fraction of the apparent volume of volume fraction of

grinding beads (ƒb of 81, 83, and 85%) relative to the total volume of the grinding chamber on particle size

reduction (d50 in main panel; d90 in inset) of miconazole nanosuspensions (n = 6 ± sd).

Figure 5: Influence of specific energy input and product flow rate (97, 113, and 183 gml–1) on particle size

reduction (d50 in main panel; d90 in inset) of miconazole nanosuspensions (n = 6 ± sd).

125

130

135

140

145

150

155

0 10 20 30 40

Pa

rtic

le s

ize

d5

0(n

m)

Specific energy input (MJ/kg)

81

83

85

Miconazole

125

130

135

140

145

150

155

0 10 20 30 40

Part

icle

siz

e d

50

(nm

)


97

113

183

Miconazole

0

250

500

750

1000

1250

1500

1750

2000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


818385

0

250

500

750

1000

1250

1500

1750

2000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


97113183

Chapter III

Page 70 of 182

Figure 6: Influence of specific energy input and of stirrer tip speed (2400 to 3600 rpm) on particle size reduction

(d50 in main panel; d90 in inset) of miconazole nanosuspensions (n = 6 ± sd).

Figure 7: Influence of grinding bead size (0.4 and 0.8 mm) on particle size distribution of miconazole

nanosuspensions after 60 min of nanogrinding at two stirrer tip speeds (n = 2800 rpm; n = 3400 rpm); n = 6 ± sd.

125

130

135

140

145

150

155

0 10 20 30 40

Pa

rtic

le s

ize

d5

0(n

m)


2400

2800

3000

3400

3600

Miconazole

0

2

4

6

8

10

12

10 100 1,000 10,000

Volu

me

(%)

Particle size (nm)

0.8 mm0.4 mm

Miconazolen = 2800 rpm

Bead size

0

2

4

6

8

10

12

10 100 1,000 10,000

Volu

me

(%)

Particle size (nm)

0.4 mm

0.8 mm

Miconazolen = 3400 rpm

Bead size

0

250

500

750

1000

1250

1500

1750

2000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


24002800300034003600

Chapter III

Page 71 of 182

Table 1: Milling parameters for nanogrinding miconazole in a stirred media mill and particle size distribution (d50

and d90). The stress number N was calculated for 60 min milling time.

Table 2: Effect of milling time on miconazole particle size reduction by nanogrinding.

Milling time

(min)

Stress energy of grinding beads, SEB

(MJ/kg)

Stress number, N

(x 10-5) (mm-2)

d50 ± sd

(nm)

d90 ± sd

(nm)

0 0 0 16,745 ± 290 3,1124 ± 350

15 7.8 19 140 ± 2 873 ± 15

30 12.6 37 135 ± 0 365 ± 2

60 24.0 74 133 ± 0 291 ± 1

Viscosity of miconazole nanosuspensions was measured because of its importance for milling and

nanosuspension stability. The viscosity decreased significantly (from mPa•s to 182 mPa•s, p < 0.05)

over 60 min of milling (Fig. 8). As a control, the milling experiment was also performed without drug

particles to evaluate whether the observed viscosity drop was due to changes in the structure of the

excipients HPC and SDS. The results (Fig. 8) demonstrated that milling did not alter significantly the

viscosity of the excipients solution.

Bead size

(mm)

Product

flow rate

(gml-1)

Stirrer-

tip-speed

(rpm)

Fraction of

apparent

bead volume

(%)

Stress number,

N (x 10-5)

(mm-2)

Stress energy of

grinding

beads, SEB

(x 105)

(J)

d50 ± sd

(nm)

d90 ± sd

(nm)

Variable: Volume fraction of grinding beads relative to mill chamber volume

0.4 83 3400 81 67 3.8 132 ± 0.6 294 ± 2.1

0.4 149 3400 83 71 3.8 133 ± 0.6 290 ± 1.7

0.4 113 3400 85 74 3.8 131 ± 0.6 282 ± 2.6

Variable: Product flow rate

0.4 97 3400 85 74 3.8 132 ± 0.6 286 ± 1.5

0.4 113 3400 85 74 3.8 132 ± 0.6 287 ± 1.5

0.4 183 3400 85 74 3.8 132 ± 0.0 337 ± 3.0

Variable: Stirrer tip speed

0.4 134 2400 85 53 1.9 132 ± 0.0 337 ± 3.0

0.4 134 2800 85 61 2.7 129 ± 0.0 269 ± 1.0

0.4 132 3000 85 66 3.0 132 ± 0.6 287 ± 1.5

0.4 113 3400 85 74 3.8 131 ± 0.6 282 ± 2.6

0.4 143 3600 85 79 4.3 132 ± 0.6 287 ± 2.5

Variable: Bead size

0.4 134 2800 85 61 2.7 129 ± 0.0 269 ± 1.0

0.8 138 2800 85 15 21 148 ± 0.0 625 ± 2.6

0.4 113 3400 85 74 3.8 131 ± 0.6 282 ± 2.6

0.8 105 3400 85 18 31 148 ± 0.6 474 ± 3.2

Chapter III

Page 72 of 182

Figure 8: Viscosity change of a 20% miconazole (MIC) nanosuspension and the corresponding suspension

medium during the milling process (HPC, hydroxypropylcellulose; SDS, sodium dodecyl sulfate)

3.4. Importance of the SDS Concentration in the Formulations for Particle Size

Reduction and Nanosuspension Viscosity

For extending the process parameter study on itraconazole and etravirine, appropriately stabilized

formulations of these two drug substances had to be determined first. Amongst different stabilizers

tested (SDS, poloxamers, and cellulose ethers), the combination of SDS and HPC was found to be also

most suitable for itraconazole and etravirine (data not shown). While the concentration of HPC (1.25%,

3.75%, 5.0%) appeared to be less critical for particle stabilization of the drug substances (data not

shown), the SDS concentration revealed to be crucial for etravirine. Thus, nanogrinding experiments

were performed with 20%, w/w drug substance, 5%, w/w HPC, and increasing concentrations of 0% to

0.2%, w/w of SDS. Here, all the experiments were performed under fixed process conditions (stirrer tip

speed of 3400 rpm; grinding bead size of 0.8 mm; Fig. 9). In the absence of SDS, only itraconazole

could be milled to the desired particle size range, provided that the specific energy input was high (Fig.

9A). Nanomilling of miconazole in the absence of SDS left behind a fraction of relatively large particles

(d90 = 1200 nm), in agreement with data of a previous study, where a minimal quantity of SDS

(0.0125%) was found to be necessary for efficient nanogrinding of this drug substance (Kwade et al.,

2002). Etravirine was not millable at all at 0% or 0.025% SDS in the formulation, as the mill had to be

stopped after 5 to 10 min or 30 min, respectively, due to improper viscosity increase. Efficient

etravirine nanogrinding required minimum 0.05% SDS in the drug suspensions.

0

100

200

300

400

500

600

0 20 40 60

Vis

cosi

ty (

mP

as)

Milling time (min)

MIC / HPC / SDS

HPC / SDS

Chapter III

Page 73 of 182

0

100

200

300

400

0 10 20 30 40

Pa

rtic

le s

ize

d5

0(n

m)

Specific energy input (MJ/Kg)

MIC ITR ETR

A) SDS: 0%

0

100

200

300

400

0 10 20 30 40

Part

icle

siz

e d

50

(nm

)


MIC ITR ETR

B) SDS: 0.025%

0

100

200

300

400

0 10 20 30 40

Pa

rtic

le s

ize

d5

0(n

m)


MIC ITR ETR

C) SDS: 0.05%

0

1000

2000

3000

4000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


MIC

ITR

ETR

0

1000

2000

3000

4000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


MIC

ITR

ETR

0

1000

2000

3000

4000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


MIC

ITR

ETR

Chapter III

Page 74 of 182

Figure 9: Influence of specific energy input and SDS concentration (0–0.2%) on particle size reduction (d50 and

d90) of miconazole (MIC), itraconazole (ITR), and etravirine (ETR) during nanogrinding with stirrer tip speed of

3400 rpm (n = 6 ± sd). In top left panel, no data is shown for etravirine, because milling had to be stopped at a

specific energy input of 7MJ/kg (after ~15 min of milling) due to excessive viscosity increase; particle sizes at this

point were: d50 = 4412 ± 236 nm; d90 = 9998 ± 823 nm.

The crucial importance of SDS for etravirine nanogrinding could also be observed in the viscosity

change of the suspension during milling (Fig. 10). While the viscosity of the miconazole and

itraconazole suspensions decreased substantially and significantly upon milling at all SDS

concentrations (0% to 0.2%; Figs. 10A and B), the viscosity of the etravirine suspensions increased

during nanogrinding when the SDS concentration in the formulations ranged between 0% and 0.05%

(Fig. 10C). Concomitantly with this viscosity increase, particle agglomeration was observed, leading to

the stopping of the mill either a few minutes after the start of the process (at 0% SDS) or 30 min later

0

100

200

300

400

0 10 20 30 40

Pa

rtic

le s

ize

d5

0(n

m)


MIC ITR ETR

D) SDS: 0.125%

0

100

200

300

400

0 10 20 30 40

Part

icle

siz

e d

50

(nm

)


MIC ITR ETR

E) SDS: 0.200%

0

1000

2000

3000

4000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


MIC

ITR

ETR

0

1000

2000

3000

4000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)

Specific energy (MJ /kg)

MIC

ITR

ETR

Chapter III

Page 75 of 182

(when 0.025% SDS was applied). At an SDS concentration of 0.05%, etravirine nanogrinding was

possible for 60 min, although the viscosity of the etravirine nanosuspension increased and visible

agglomerates were formed. Efficient nanogrinding of etravirine required SDS concentrations of at least

0.125%, under which conditions the viscosity of the suspension decreased during milling and no visible

agglomerates were formed.

0

200

400

600

800

1000

1200

1400

0 0.025 0.05 0.125 0.2

Vis

cosi

ty (

mP

a•s

)

SDS (%, w/w)

Before milling

After milling

A) Miconazole

0

200

400

600

800

1000

1200

1400

0 0.025 0.05 0.125 0.2

Vis

cosi

ty (

mP

a•s

)

SDS (%, w/w)

Before milling

After milling

B) Itraconazole

Chapter III

Page 76 of 182

Figure 10: Nanosuspension viscosity before and after nanogrinding of the three drug substances miconazole,

itraconazole, and etravirine as a function of SDS concentration.

3.5. Effect of Milling Parameters on Particle Size Reduction of Miconazole,

Itraconazole, and Etravirine

As etravirine required at least 0.05% SDS in the formulation for being millable under the above set

process conditions (stirrer tip speed 3400 rpm; grinding bead size 0.8 mm), it was investigated whether

variation of process conditions could improve the millability of this drug substance in a critical

formulation. Thus, the effects of stirrer tip speed (2800 and 3400 rpm corresponding to 8.3 and 10.0 ms–

1, respectively) and milling bead size (0.4 and 0.8 mm in diameter) on the nanogrinding of etravirine in

comparison to miconazole and itraconazole, in formulations of 20% drug substance, 5% HPC, and

0.025% SDS (Fig. 9) were determined. Nanogrinding efficiency was similar for miconazole (d50 = 15

µm and d90 = 28 µm before milling) and itraconazole (d50 = 21 µm and d90 = 52 µm before milling) (Fig.

11), although miconazole was ground slightly more efficiently than itraconazole at specific energy input

levels of below 15 to 20 MJ/kg, with the difference being more important for a stirrer tip speed of 2800

rpm and a bead size of 0.8 mm (Fig. 11A versus Figs. 11B–D). By contrast, etravirine particles (d50 = 11

µm and d90 = 24 µm before milling) and nanosuspensions behaved very differently in the nanogrinding

of this critical formulation with respect to millability, viscosity change, and particle size reduction. Two

phenomena hampered the nanogrinding of etravirine: (i) formation of aggregates and (ii) viscosity

increase during milling under all tested milling conditions. Aggregates became clearly visible under the

microscope with the aggregates causing high d90 values (Figs. 11A–C), panels on the right hand side),

although mean particle sizes similar to those of miconazole and itraconazole were achieved under

optimal milling conditions (d50 of ~140 nm; Fig. 11B). The viscosity increase of the etravirine

0

200

400

600

800

1000

1200

1400

0 0.025 0.05 0.125 0.2

Vis

cosi

ty (

mP

a•s

)

SDS (%, w/w)

Before milling

After milling

C) Etravirine

Chapter III

Page 77 of 182

suspensions indicated an inverse behavior to that of the other two drug substances. Under milling at

2800 rpm, the viscosity of the suspension increased from the initial 470 mPa•s to 711 mPa•s (0.8 mm

beads) or even to 1100 mPa•s (0.4 mm beads). When milling at 2800 rpm with 0.4 mm beads (Fig.

11B) or at 3400 rpm with 0.8 mm beads (Fig. 11C), the viscosity increase forced the stopping of the

process at a specific energy input of ~15 MJ/kg. Milling of etravirine at 3400 rpm with 0.4 mm beads

(Fig. 11D) was not possible at all due to the immediate and important increase of viscosity. Altogether,

the substantial viscosity increase of the etravirine nanosuspension during nanogrinding impaired or

inhibited efficient particle size reduction. Only under relative mild nanogrinding conditions (2800 rpm,

Figs. 11A and B), the d50 and d90 sizes were reduced to values of below 200 and 2000 nm, respectively.

So, in terms of remaining large-sized particles, etravirine was much less efficiently milled than

miconazole and itraconazole, for which d90 values of ~260 nm were achieved.

100

200

300

400

500

0 10 20 30 40

Part

icle

siz

e d

50

(nm

)


MIC

ITR

ETR

A) n: 2800 rpmBead size: 0.8 mm

100

200

300

400

500

0 10 20 30 40

Pa

rtic

le s

ize

d5

0(n

m)


MIC

ITR

ETR

B) n: 2800 rpmBead size: 0.4 mm

0

1000

2000

3000

4000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


MIC

ITR

ETR

0

1000

2000

3000

4000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


MIC

ITR

ETR

Chapter III

Page 78 of 182

Figure 11: Influence of specific energy input and bead size (0.8 and 0.4 mm) on particle size reduction (d50 and

d90) of miconazole (MIC), itraconazole (ITR), and etravirine (ETR) during nanogrinding. Stirrer tip speed: n =

2800 and 3400 rpm (n = 6 ± sd). In lower right panel, no data is shown for etravirine, because milling had to be

stopped at a specific energy input of 7MJ/kg (after ~15 min of milling) due to excessive viscosity increase;

particle sizes at this point were: d50 = 1652 ± 13 nm and d90 = 5934 ± 40 nm.

4. Discussion

The manufacturing of nanosuspensions by nanogrinding is influenced by particle breakage and

nanoparticle interactions, with the latter becoming increasingly important since particle size decreases

below 1 µm due to the high particle surface energy and Brownian motion (Knieke et al., 2009).

Therefore, formulation and process parameters as well as material properties are of crucial importance

for achieving nanosized particles and stable nanosuspensions. In the present work, standard

100

200

300

400

500

0 10 20 30 40

Pa

rtic

le s

ize

d5

0(n

m)


MIC

ITR

ETR

C) n: 3400 rpmBead size: 0.8 mm

100

200

300

400

500

0 10 20 30 40

Part

icle

siz

e d

50

(nm

)


MIC

ITR

D) n: 3400 rpmBead size: 0.4 mm

0

1000

2000

3000

4000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


MIC

ITR

ETR

0

1000

2000

3000

4000

0 10 20 30 40

Part

icle

siz

e d

90

(nm

)


MIC

ITR

Chapter III

Page 79 of 182

formulations were used to evaluate the influence of milling parameters on the particle size reduction of

miconazole, itraconazole, and etravirine, three drug substances of similar mechanical properties. For

etravirine, the concentration of SDS in the formulation was found to be critical so that formulations

with increasing amounts of SDS were used for comparing the millability of the three drug substances as

a function of process parameters. With proper and stable formulations, such as the miconazole

formulation used first, particle size reduction depended mostly on the overall parameter SEI, which was

determined by measuring the electric energy consumption of the process. As SEI is proportional to the

product NSEB, with the latter depending directly on the process parameters, the contribution of

individual process parameters on the efficiency of miconazole nanogrinding was evaluated. It appeared

that above a moderate energy input of 10 MJ/kg to 15 MJ/kg only the milling time (affecting N) and the

bead size (affecting N and SEB) influenced significantly the grinding efficiency. Interestingly,

irrespective of all independently varied process parameters, minimal miconazole particle sizes of d50

130 nm and d90 ~270 nm were achieved at a specific energy input of ~15 MJ/kg when using beads of

0.4 mm in diameter. Higher energy input, as achieved mainly by higher stirrer tip speed and longer

milling times, remained ineffective in further reducing the miconazole particle size. Additional energy

input was lost by heat development, which required additional energy for cooling of the product.

Further, higher stirrer tip speeds or longer milling times are likely to result in a higher wear of the

grinding beads causing product contamination.

Besides the milling time, bead size was the second process parameter that influenced the particle size

reduction of miconazole distinguishably from the specific energy input (Fig. 9).

The bead size influences N and SEB in an opposite way and to a different extent (Eqs. 1 and 2). Smaller

beads provide a higher number of stress events, but their lower stress energy is less effective for

breaking up the particles, so that multiple stressing is required. Conversely, larger grinding beads exert

higher stress energy on the drug particles, though at lower frequency. As already demonstrated with

other materials, the specific energy consumption can be minimized if optimal bead sizes are used, e.g.,

beads of 0.4 to 0.8 mm in diameter for limestone (Kwade et al., 1996). This agrees with our results

demonstrating that lower energy input was necessary with the 0.4 mm beads than with the 0.8-mm

beads to obtain minimally sized drug particles. Thus, beads of 0.4 mm in diameter transferred sufficient

energy to break up the miconazole particles of both and itraconazole. Smaller beads (e.g., 0.2 mm)

could not be tested because of equipment limitations. Yet, it would be interesting to test whether 0.2

mm beads would still transfer sufficient stress energy to break up drug particles to achieve a faster

(possibly with less specific energy input) particle size reduction due to the higher number of stress

events. The particle size achievable by wet nanogrinding is determined by two opposing phenomena: (i)

particle breakage and (ii) particle-particle interactions potentially leading to particle agglomeration. The

two phenomena define the true and apparent grinding limits (Knieke et al., 2009). The true grinding

Chapter III

Page 80 of 182

limit is determined by the mechanical properties of the drug particles, the process parameters, and the

equipment configuration, while the apparent grinding limit is determined by the stabilization of the

particles during milling. During milling, lattice defects are generated within the particles. After reaching

a critical size, it is assumed that the defect domains are so small that the lattices can no longer store

defects. At this point, the stress energy provided by the grinding beads is insufficient to break the

particles. In the present case of miconazole ground under the selected conditions and equipment

configuration, it is assumed that the true grinding limit was probably achieved. This assumption is

supported by the absence of a viscosity increase, which would indicate increased particle-particle

interactions and insufficient stabilization (Greenwood et al., 1997). Although etravirine showed similar

mechanical properties to both miconazole and itraconazole, its milling behavior differed substantially

from that of miconazole and itraconazole, when the formulation contained only 0% to 0.05% SDS. For

example, the viscosity of etravirine nanosuspensions increased during milling, whereas it decreased

with miconazole and itraconazole nanosuspensions. The exact reason for this opposed behavior remains

elusive as the viscosity of suspensions depends on several parameters such as the particle size, shape

and concentration, the viscosity of the suspension medium itself, and the solids concentration

(Bernhardt et al., 1999). One possible explanation for the viscosity decrease might be a partial loss of

molecular weight of HPC during milling. However, the milling of a drug particle-free control

formulation under identical milling conditions as used for the drug substance nanosuspensions did not

result in a viscosity decrease. Another explanation may be based on the adsorption of the viscosifying

and stabilizing agent HPC on the milled drug nanoparticles, whose surface increases substantially

during nanogrinding. Adsorption of HPC on the ground nanoparticles not only lowered the polymer

concentration in the continuous phase (Greenwood et al., 1997; Bernhardt et al., 1999), but may have

also lessened the friction between the particles upon shearing, with both of these phenomena resulting

in a lowering of the nanosuspension viscosity. Further, particles rearrange under deformation; thereby,

they increase the fraction of maximum packing clusters so that more free volume becomes available,

which reduces again the interparticle frictions (Greenwood et al., 1997; Olhero et al., 2004).

Viscosity increase under nanogrinding, as observed for etravirine, has already been described by other

authors (Greenwood et al., 1997; Bernhardt et al., 1999; Olhero et al., 2004) and been explained by

particle agglomeration. In agreement with such former observations, here indeed such particle

agglomeration during etravirine milling was observed. The combined effect of particle agglomeration

and viscosity increase hampered efficient nanogrinding, so that a relatively large fraction of etravirine

particles in the micrometer size range remained in the nanosuspensions when SDS was present at low

concentrations (0% to 0.05%). This stresses the importance of nanosuspension stabilization.

Stabilization with 0.025% to 0.05% SDS and 5% HPC seemed to be adequate for miconazole and

itraconazole, but not for etravirine. Increasing the SDS concentration to 0.125% or 0.2% prevented the

Chapter III

Page 81 of 182

viscosity increase and particle agglomeration in the etravirine nanosuspension upon milling.

Incidentally, SDS is known to stabilize nanoparticles via electrostatic repulsion and HPC via steric

hindrance. Interestingly, in stabilized nanosuspensions, i.e., in formulations containing 0.05% to 0.2%

SDS (Figs. 9C–E), etravirine and itraconazole were both less efficiently milled than miconazole at low

energy input as both drug substances presented an important fraction of relatively large particles (as

expressed by d90). Although the reason for this difference is not known, one may speculate that surface

roughness and mechanical behavior under stress may have an influence, as discussed below. At the

generally used SDS concentration of 0.025%, the comminution behavior of miconazole and

itraconazole was quite similar, although miconazole required a slightly lower specific energy input

(corresponding to lower number of stress events) than itraconazole for achieving the same particle size,

independently of the size of the milling beads used. The larger particle size of itraconazole before

milling could be one of the reasons for the higher specific energy required by itraconazole since larger

particles need a higher number of stress events or higher stress energy to achieve similar sizes of

nanoparticles. However, for etravirine, which presented the smallest initial particle size of all three drug

substances, a higher specific input energy was required than for miconazole for achieving the same

particle size. Another factor that defines the required energy input for nanogrinding is the mechanical

properties of the particles. The three drug substances tested not only showed similar elastic properties

(Young’s modulus values of 3 ± 2 GPa for miconazole, 5 ± 2 GPa for itraconazole, and 2 ± 1 GPa for

etravirine), but also similar plastic deformation behavior, although miconazole particles exhibited

variable plastic deformation (Fig. 3).

One out of the three miconazole particles tested for plastic deformation by AFM deformed similarly to

itraconazole and etravirine particles (indentation depth of 5 to 20 nm under an indentation force of 10

N), whereas the other two miconazole particles were weaker (indentation depth of 40 to 80 nm under

same force). The more efficient millability of miconazole in comparison to itraconazole and etravirine

might be related to two observed phenomenological properties: surface roughness and sink-in type

plastic deformation. Miconazole exhibited the highest roughness of all three drug substances, as

evidenced by AFM images (Fig. 2) and determined by the AFM roughness measurement system (Meli,

2002) (data not shown). Roughness is in many cases related to friction (Meli, 2002), which promotes

particle breakage. Moreover, miconazole indicated a tendency towards a sink-in type of plastic

deformation, while piling-up was observed with itraconazole and etravirine. Sinkin or pile-up behavior

is related to the so-called work-hardening of the materials under plastic deformation. Therefore,

miconazole appeared to undergo work-hardening under plastic deformation, while itraconazole and

etravirine did not. Work-hardening is defined as an increase in mechanical strength due to plastic

deformation. Materials that readily work-harden under an indenting force will show sink-in, whereas

materials that had been work-hardened previously or that do not readily exhibit work-hardening will

Chapter III

Page 82 of 182

show pile-up due to plastic flow. To which extent work-hardening affects particle breakage under

nanogrinding conditions is unknown. More generally, there is only scarce information available in the

literature on the relationship between mechanical properties of materials and their breakage under

nanogrinding conditions, with the few studies in this area focusing primarily on inorganic materials

(Kwade et al., 2002; Knieke et al., 2009). Upon adequate stress, brittle materials can break up directly

in very fine fragments, which can hardly be further milled, whereas other materials may break up in

only a few fragments that can be further ground. Therefore, the true grinding limit may be as low as 5–

100 nm for brittle materials, but be in the range of a few micrometers for plastic materials (Bernotat et

al., 2003; Knieke et al., 2009; Gers et al., 2010). For example, under the milling conditions used by

Knieke et al., the ductile calcium carbonate reached a minimum size of 55 nm, while the very brittle

zirconium oxide could be milled to 5 nm. The energy transferred to brittle materials is used for particle

fracture, whereas ductile materials also undergo plastic deformation (Knieke et al., 2009). When

correlating Young’s modulus (E) with the minimal achievable particle size of CaCO3, SiO2, SnO2,

ZrO2, and Al2O3, Knieke et al. demonstrated that higher E-values (above 50 GPa) tend to promote

particle size reduction (Knieke et al., 2009). As the three drug substances studied here have very low E-

values (2 GPa to 5 GPa), a substantial amount of energy was likely lost for elastic deformation, so that

the mean particle sizes could not be lowered to values below 100 nm, as feasible with inorganic

materials.

5. Conclusions

The present work focused on the influence of milling parameters on particle size reduction of

miconazole, itraconazole, and etravirine, three drug substances of similar mechanical properties.

To facilitate comparison between these substances, in a first set of experiments, the composition of the

nanosuspensions was kept constant, namely 20% drug substance, 0.025% SDS, and 5% HPC. The use

of optimally sized grinding beads offered the possibility to minimize the specific energy input required

to achieve minimal particle sizes. A minimal energy input should be beneficial for drug substance and

formulation stability (less heat exposure), material wear (less abrasion), and product contamination (less

abrasion). Miconazole and itraconazole not only possessed similar mechanical properties, but also

similar comminution behavior, with minimal mean particle sizes of 120 nm to 130 nm being achievable

at an energy input of about 15 MJ/kg. Etravirine nanosuspensions exhibited a viscosity increase and

particle agglomeration during milling when the SDS concentration was below 0.125%, which hindered

efficient nanogrinding, yielding an important fraction of micrometer-sized particles. This study

demonstrated that the knowledge of particles’ mechanical properties and use of optimized milling

Chapter III

Page 83 of 182

parameters, while being important, may alone not warrant efficient particle size reduction, but an

appropriate formulation is needed for adequate nanosuspension stabilization, as only properly stabilized

particles can be ground efficiently.

Symbols used

DmB [kg] wear of the grinding beads at the given time of milling

C [kgm–3] concentration of drug substance in suspension

d50 [m] mean particle diameter

d90 [m] 90% undersize particle diameter

dB [m] diameter of grinding beads

E [Pa] Young’s modulus

L [N] applied load in atomic force microscopy

m [kg] mass of drug substance

n [ms–1] stirrer speed

N [mm–2] stress number of grinding beads

N0 [W] power consumption of the unloaded mill

Nt [W] power consumption at time t of milling

R [m] radius of atomic force microscope indenter

SEB [Nm] stress energy of grinding beads

SEI [J kg–1] total energy input relative to the total mass

t [s] milling time

Greek symbols

ν [ms–1] rotational speed of the stirrer tip

ε [–] porosity of the bulk of beads

ρB [kgm–3] density of grinding beads

ϕB [–] volume fraction of the grinding beads relative to the volume of the grinding chamber

δ [m] indentation depth in atomic force microscopy measurements under elastic deformation

νs [–] Poisson’s ratio (for atomic force microscopy measurements)

Chapter III

Page 84 of 182

References

Bernhardt, C., Reinsch, E., Husemann, K., 1999. The influence of suspension properties on ultra-fine

grinding in stirred ball mills. Powder Technol. 105, 357–361.

Cerdeira, A.M., Mazzotti, M., Gander, B., 2010. Miconazole nanosuspensions: Influence of formulation

variables on particle size reduction and physical stability. Int. J. Pharm. 396, 210–218.

Date, A.A., Patravale, V.B., 2004. Current strategies for engineering drug nanoparticles. Curr. Opin.

Coll. Interf. Sci. 9, 222–235.

Gers, R., Climent, E., Legendre, D., Anne-Archard, D., Frances, C., 2010. Numerical modelling of

grinding in a stirred media mill: hydrodynamics and collision characteristics. Chem. Eng. Sci. 65,

2052–2064.

Greenwood, R., Luckham, P.F., Gregory, T., 1997. The effect of diameter ratio and volume ratio on the

viscosity of bimodal suspensions of polymer lattices. J. Coll. Interf. Sci. 191, 11–21.

Knieke, C., Sommer, M., Peukert, W., 2009. Identifying the apparent and true grinding limit. Powder

Technol. 195, 25–30.

Kwade, A., Blecher, L., Schwedes, J., 1996. Motion and stress intensity of grinding beads in a stirred

media mill. Part 2: Stress intensity and its effect on comminution. Powder Technol. 86, 69–76.

Kwade, A., 1999a. Wet comminution in stirred media mills – research and its practical application.

Powder Technol. 105, 14–20.

Kwade, A., 1999b. Determination of the most important grinding mechanism in stirred media mills by

calculating stress intensity and stress number. Powder Technol. 105, 382–388.

Kwade, A., Schwedes, J., 2002. Breaking characteristics of different materials and their effect on stress

intensity and stress number in stirred media mills. Powder Technol. 122, 109–121.

Lipinski, C., 2002. Poor aqueous solubility – an industry wide problem in drug discovery. Am. Pharm.

Rev. 5, 82–85.

Meli, F., 2002, in Proc. of the 3rd euspen Conf., European Society for Precision Engineering and

Nanotechnology, Cranfield Vol. 3, p. 533.

Mende, S., Stenger, F., Peukert, W., Schwedes, J., 2003. Mechanical production and stabilization of

submicron particles in stirred media mills. Powder Technol. 132, 64–73.



Muller, R.H., Keck, C.M., 2004. Challenges and solutions for the delivery of biotech drugs – a review

of drug nanocrystal technology and lipid nanoparticles. J. Biotech. 113,151–170.

Olhero, S.M., Ferreira, J.M.F., 2004. Influence of particle size distribution on rheology and particle

packing of silica-based suspensions. Powder Technol. 139, 69–75.

Chapter III

Page 85 of 182

Patravale, V.B., Abhijit, A.D., Kulkarni, R.M., 2004. Nanosuspensions: a promising drug delivery

strategy. J. Pharm. Pharmacol. 56, 827-840.


Stenger, F., Mende S., Schwedes J., Peukert W,. 2005. Nanomilling in stirred media mills. Chem. Eng.

Sci. 60, 4557–4565.



Chapter III

Page 86 of 182

Chapter IV

Page 87 of 182

Chapter IV

Simultaneous quantification of polymeric and surface active stabilizers of nanosuspensions by using near-infrared

spectroscopy3

1. Introduction

Nanosuspension technology is gaining wide interest for the formulation of very slightly water-soluble

(0.1–1.0 mg/ml) or practically water-insoluble (< 0.1 mg/ml) drug compounds. Drug particles in

nanosuspensions possess an increased specific surface area (SSA), which should enhance the

dissolution rate; increased dissolution rate may finally enhance the bioavailability, especially of drug

substances with high membrane permeability (class 2 drugs of the biopharmaceutical classification

system) (Kesisoglou et al., 2007; Van Eerdenbrugh et al., 2009). One of the major challenges of

formulating nanosuspensions resides in the selection of suitable pharmaceutical excipients for

efficiently reducing particle size and affording adequate formulation stability against crystal growth and

nanoparticle aggregation. Nanosuspension stability is usually mediated by steric and electrostatic

effects as provided by adsorbed polymers and surfactants (Merisko-Liversidge et al., 2003; Rabinow,

2004; Cerdeira et al., 2010; Bhakay et al., 2011). For a polymer to provide effective steric stabilization,

full coverage of the drug substance particles and stable physical adsorption are required (Ploehn et al.,

1990). Thus, successful development of nanosuspensions requires careful evaluation of stabilizer type

and concentration (Van Eerdenbrugh et al., 2008a). However, due to lack of understanding of the

mechanisms governing nanoparticle stabilization, selection of stabilizers remains largely empirical

(Merisko-Liversidge et al., 2003; Van Eerdenbrugh et al., 2008b). For future more rational selection of

stabilizing excipients, we will have to improve our knowledge on nanoparticle stabilization

mechanisms, particularly on quantitative stabilizer–stabilizer and stabilizer–particle interactions (Lee,

2003). For example, knowledge of the amounts of stabilizers adsorbed on drug nanoparticles of

different physical-chemical properties should furnish important information on the interactions between

the drug particles and stabilizing excipients (Sepassi et al., 2007; Van Eerdenbrugh et al., 2009;

3 The work in this chapter has been published as Cerdeira, AM., Werner, I.A., Mazzotti, M., Gander, B., 2012. Simultaneous quantification of

polymeric and surface active stabilizers of nanosuspensions by using near-infrared spectroscopy. Drug Dev. Ind. Pharm. DOI:10.3109/03639045.2011.650864.

Chapter IV

Page 88 of 182

Zimmermann et al., 2009). Thus, there is interest to develop easy methods for quantifying

simultaneously the concentrations of several excipients in nanosuspensions. NIR provides a simple

alternative to the current analytical methods used for assaying mixtures of excipients.

Ideal analytical methods for quantifying nanosuspension stabilizers should offer several features: (1)

simultaneous quantification of different types of materials such as hydrophilic polymers and

surfactants; (2) quantification over a broad range of excipients concentrations, so that nanosuspensions

need no or only little dilution; (3) quantification of excipients in the presence of dissolved drug, i.e.

absence of interference by dissolved drug. In the present work, we investigated if near-infrared (NIR)

spectroscopy would fulfill such requirements for assaying SDS and HPC in a miconazole

nanosuspension. NIR spectroscopy is widely used in the pharmaceutical industry, as the NIR signals

provide information on both the chemical composition and physical properties of the samples (Blanco

et al., 1999; Otsuka et al., 2010). NIR in combination with fiber optical probes is very commonly used

for the identification of raw materials. In addition, different authors have also proposed NIR

spectroscopy for quantitative analysis of drug substance and excipients in pharmaceutical formulations

by using chemometric multivariate calibration and partial least-squares regression (Moffat et al., 2000;

López-Arellano et al., 2009). Most of the quantitative NIR analyses described in the literature were

applied to solid dosage forms (Moffat et al., 2000; López-Arellano et al., 2009), whereas less attention

has been paid to liquid formulations (Moffat et al., 2000). For validation of assay methods, official

guidelines are available, although most of them concern the validation of chromatographic methods and

univariate calibration. The official guidance documents recommend demonstration of the specificity,

linearity, accuracy, precision, and range of the analytical method (Mark et al., 2002; EMEA, 2003;

PASG, 2001).

In the present study, we developed and evaluated the applicability of a NIR method for the

simultaneous quantification of sodium dodecyl sulfate (SDS) and hydroxypropylcellulose (HPC) in

miconazole nanosuspensions.


Miconazole (diameter D [4,3] ~ 20 µm; lot # R018134PUC701, Janssen Pharmaceutica N.V., Geel,

Belgium) was used as received. SDS (Texapon® K12P, Cognis, Düsseldorf, Germany), HPC (Klucel®

LF Pharm, Hercules, Doel, Belgium), all from single batches, were used as received.

Chapter IV

Page 89 of 182

2.1. Preparation of miconazole nanosuspensions

The nanosuspensions were prepared according to previously optimized composition and protocols

(Cerdeira et al., 2010). Briefly, SDS (0%, 0.025%, 0.05%, 0.125% or 0.2%, w/w) and HPC (5% w/w)

were first dissolved in purified water; miconazole (12.5% or 20%, w/w) was subsequently dispersed in

the stabilizer solution (Table 1). The microsuspensions were processed in a high-energy mill (LabStar,

Netzsch, Selb, Germany) filled to 83% (v/v) with yttrium-stabilized zirconium oxide beads (0.8 mm in

diameter). Nanogrinding was performed in circulation mode using 300 g of suspension, a pump speed

of 41 rpm, and a stirrer tip speed of 3400 rpm (10 m/s); the duration of the process was 60 min.

Table 1: Particle size (D[4,3]), specific surface area (SSA), and zeta-potential (ζ) of miconazole nanoparticles, and

miconazole solubility in nanosuspension formulations containing increasing concentrations of SDS, 5% HPC and

12.5% or 20% miconazole (MIC) (all % as w/w).

SDS conc. (%)

D[4,3] of MIC in nanosuspensions (nm) ± sd

SSA of MIC in nanosuspensions (m2/g)

Solubility of MIC in nanosuspensions (µg/ml)

ζ of MIC in nanosuspensions (mV) ± sd

M i c o n a z o l e c o n c e n t r a t i o n (%)

12.5 20 12.5 20 12.5 20 12.5 20

0 525 ± 3 524 ± 5 40 40 23 16 0 0

0.025 283 ± 3 207 ± 3 48 50 84 54 -7 ± 0 -7 ± 0

0.05 281 ± 5 213 ± 1 47 49 71 58 -8 ± 0 -12 ± 1

0.125 218 ± 4 217 ± 1 50 48 47 47 -17 ± 0 -15 ± 1

0.2 249 ± 8 248 ± 1 45 46 74 64 -20 ± 0 -19 ± 0

2.2. Nanosuspensions dispersed phase characterization: particle size and surface area,

particle morphology, and ζ-potential

Particle size distribution (volume-weighted) was measured by laser light diffraction (Malvern 2000,

Malvern Instruments, Worcestershire, UK) using the small volume dispersion unit. The Mie theory

served for particle size calculation using the following refractive indices: dispersant: 1.33; real and

imaginary particle refractive indices of miconazole: 1.55 and 0.001. Particle sizes were expressed by the

volume moment mean diameter, D[4,3]. In view of determining the amount of polymeric and surface

active stabilizers adsorbed on the nanoparticles, the SSA, of the drug nanoparticles was calculated from

the determined D[3,2] value according to the commonly used Eq. 1

667 = 9∑;<=<>∑?< = 9>�[A,B] (1)

Chapter IV

Page 90 of 182

where Vi is the relative volume in class i with a mean class diameter of di, ρ is the density of the

material, and D[3,2] is the surface area weighted mean diameter. This calculation is based on the

assumption of spherical particles with smooth surfaces.

Particle morphology was examined by SEM (Zeiss Gemini Supra 40; Carl Zeiss SMT, Oberkochen,

Germany) at 2 kV. Samples were prepared by placing one droplet of fourfold diluted (with distilled

water) nanosuspension on a gold-coated membrane filter with a pore size of 500 nm (RapID, Berlin,

Germany).

ζ-potential was measured using a Zetasizer (Zetasizer Nano ZS, Malvern Instruments, Worcestershire,

UK). The samples were adequately diluted with deionized water and placed in an electrophoretic cell.

The mean ζ-potential was calculated from the electrophoretic mobility using the Smoluchowski

equation.

2.3. Nanosuspensions continuous phase characterization: solubility of drug substance

and surface tension

The solubility of miconazole was determined in miconazole nanosuspensions stabilized with 5% (w/w)

HPC and increasing amounts of SDS (Table 1). Miconazole nanosuspensions were centrifuged

(ultracentrifuge Sorvall, Thermo Fisher Scientific, Waltham, USA) at 50,000 rpm (300,000 x g) for 3.5

h, and the supernatant assayed for drug content by HPLC. Centrifugation was preferred over

ultrafiltration, because of the difficulty experienced with filtering some of the relatively viscous high

concentration suspensions, which caused filter clogging. Miconazole was assayed by a validated

method using reversed phase HPLC with UV detection at 230 nm (Waters Alliance HPLC system,

Milford, MA, US) and a C18 column (Zorbax®, 10 cm length, 4.6 mm ID, 3.5 µm particle size). The

drug was eluted according to the gradient reported in Table 2. Linearity was confirmed between 0.5 and

700 µg/ml, and the accuracy was ±30% for 0.5 µg/ml to 1 µg/ml; ±20% for 1 µg/ml to 10 µg/ml; ±10%

for 10 µg/ml to 200 µg/ml; and ±3% for 200 µg/ml to 700 µg/ml.

Table 2: HPLC eluent gradient sequence for assaying miconazole.

Time (min)

Solvent

0 30 35 40

10 mM Na2HPO4, pH 7.5 (%, v/v) 80 35 80 80

Acetonitrile (%, v/v) 20 65 20 20

Surface tension was measured using the rod method (Kruss Tensiometer K100) for small volume

samples (2 ml) at a temperature of 23°C according to the instructions provided in the tensiometer

manual. The nanosuspensions were ultracentrifuged following the conditions described above (for

Chapter IV

Page 91 of 182

solubility measurements) and the surface tension measured in the supernatant. The surface tension was

calculated using the following equation:

D = ��.EFGH (2)

where σ is the surface or interfacial tension; F is the force acting on the balance; L is the wetted length

and θ is the contact angle.

Evaluation of NIR spectrometry to quantify simultaneously SDS and HPC in aqueous solution and

miconazole nanosuspension The spectra were recorded on a multi-purpose Fouriertransform NIR (FT-

NIR) analyzer (MPA FT-NIR, Bruker Optics, Ettlingen, Germany) in transmittance mode using a flat-

bottom glass vessel with a volume of 1 ml. NIR-spectra were collected in the wave number range of

4,000 cm−1 to 12,500 cm−1 (800 nm to 2,500 nm). To account for the effects caused by the presence of

water in the analyzed solutions and drug suspensions, we applied spectral pretreatment (second

derivatives using the Quant module in Bruker OPUS). The second derivative of the NIR absorption was

used for multivariate analysis using partial least-squares algorithm and cross-validation. The optimal

model for NIR-signal transformation was selected according to satisfactory correlation coefficient, the

root mean square error of cross-validation (RMSECV), the prediction bias (average of difference

between actual value and predicted value) and the number of latent variables. Each sample was

measured at least in triplicate, as specified under Results. The NIR method for quantifying

concomitantly SDS and HPC in aqueous solutions was first examined for linearity and accuracy using

SDS and HPC standard solutions. In view of using the developed NIR method for quantifying SDS and

HPC in miconazole nanosuspensions, the method repeatability and precision between analysts as well

as the method specificity were then assessed using miconazole nanosuspensions stabilized with the two

excipients. Finally, the reporting threshold was deduced from the precision, accuracy, and linearity

study for the lower levels of the validated range for both SDS and HPC. Aqueous solutions of 0.03% to

0.3% (w/w) SDS and 0.75% to 7.5% (w/w) HPC were prepared for calibration. The concentrations were

selected to cover the actual SDS (0.03% to 0.2%) and HPC (1.25% and 5%) concentrations that were

previously found useful for preparing miconazole nanosuspensions by wet media milling (Cerdeira et

al., 2010). Randomly selected combinations of SDS and HPC concentrations (0.03% to 0.3% SDS;

0.75% to 7.5% HPC) were used to generate the calibration samples. An external validation was also

performed using a set of samples constructed in the same way as the calibration samples (Table 3). The

accuracy of the measured SDS and HPC concentrations was determined from aqueous solutions

Chapter IV

Page 92 of 182

containing known amounts of SDS and HPC as specified in Table 4. For this, six individual samples of

each concentration were prepared (Table 4).

The repeatability and precision between analysts was determined using two nanosuspension

formulations that had previously been developed (Cerdeira et al., 2010). For assaying SDS and HPC,

the miconazole nanosuspensions (Tables 5 and 6) were centrifuged at 50,000 rpm (300,000 x g) for 3.5

h (ultracentrifuge Sorvall, Thermo Fisher Scientific, Waltham, US), and the clear supernatants collected

for analysis by NIR. With each nanosuspension, six independent measurements were made on two

different days and by two different analysts using the same equipment.

Table 3: Partial least square model parameters used for constructing the calculation model for SDS and HPC

concentrations from 2nd derivative of NIR-spectra.

Parameters SDS HPC

Number of samples used to build the model

69 69

Data treatment Second derivative Second derivative

Spectral regions used in the model (wavenumbers in cm-1)

6150 to 5740 8760 to 8355

6033 to 5724

Number of latent variables 2 1

Correlation coefficient (r) 0.991 0.999

Root-mean-square error of cross-validation (RMSECV)

0.0138 0.0944

Bias 0.00076 -0.0036

Residual predictive deviation (RPD)

7.49 22.2

Table 4: Accuracy of concomitant SDS and HPC quantification at concentrations of interest.

SDS concentration HPC concentration

Theoretical % (w/w)

Mean recovery (%)

RSDa (%)

Theoretical % (w/w)

Mean recovery (%)

RSDa (%)

0.025 98 1 5 101 6

0.04 104 13 2.5 100 0.3

0.05 105 3.5 0.75 110 1.1

0.1 103 4.1 6.25 98 0.2

0.2 102 1.6 1.25 104 0.5

0.25 105 0.7 5 98 0.3

aRSD: Relative standard deviation of mean recovery.

Chapter IV

Page 93 of 182

The miconazole nanosuspension of interest contained, besides the suspended drug nanoparticles, water,

SDS, HPC, and a minimal amount of dissolved drug substance.besides the suspended drug

nanoparticles, water, SDS, HPC, and a minimal amount of dissolved drug substance. Thus, we wanted

to examine whether the miconazole present in the supernatant of the ultracentrifuged nanosuspensions

interfered with the quantification of SDS and HPC. As miconazole solubility increases with SDS and

HPC concentrations, solutions of 0.2% SDS, 1.25 or 5% HPC and different miconazole concentrations

were prepared to test the method specificity (Table 7). The current analytical method validation

guidelines do not require establishing the reporting threshold in assay methods for the quantification of

major components (PASG, 2001; EMEA, 2003; Roggo et al., 2007). However, as SDS is present at low

concentrations (0.03% to 0.3%, w/w), the reporting threshold was established based on the precision,

accuracy, and linearity data that was verified with an independent set of samples of both SDS and HPC

using the constructed model (PASG, 2001; EMEA, 2003; Roggo et al., 2007).

Table 5: Repeatability (samples 1 to 6) and intermediate precision between analysts (analyst 1 and analyst 2) for

concomitant quantification of SDS (nominal 0.025%) and HPC (nominal 5.0%) in miconazole nanosuspensions

(all concentrations as %, w/w).

athe measured SDS and HPC concentrations were consistently higher than the nominal values; this can be ascribed to that fact that the concentrations were expressed as % w/w; as the supernatants used for analysis were free of solid miconazole, the relative weight fractions of the SDS and HPC in this phase were consequently increased.

bRSD: relative standard deviation (%), relative to average values.

cCI: confidence interval.

Sample Measured SDS (%)a Measured HPC (%)a

Analyst 1 Analyst 2 Analyst 1 Analyst 2

1 0.030 0.026 5.29 5.17

2 0.029 0.029 5.24 5.23

3 0.025 0.022 5.30 5.24

4 0.034 0.032 5.25 5.24

5 0.029 0.025 5.31 5.23

6 0.025 0.025 5.26 5.17

Repeatability between samples

Average 0.029 0.026 5.28 5.21

RSDb (n=6) 12 14 1 1

Precision between analysts

RSDb (n=12) 14 1

p-value (95%, CIc for difference of means between analysts) 0.259 0.009

95% CIc for difference of means between analysts -0.00223 – 0.00731 0.0198 – 0.1024

Chapter IV

Page 94 of 182

Table 6: Repeatability (samples 1 to 6) and intermediate precision (analyst 1 and analyst 2) for concomitant

quantification of SDS (nominal 0.2%) and HPC (nominal 1.25%) in miconazole nanosuspensions (all

concentrations as %, w/w).

Sample Measured SDS (%) Measured HPC (%)

Analyst 1 Analyst 2 Analyst 1 Analyst 2

1 0.119 0.120 1.0 0.9

2 0.109 0.118 0.9 1.0

3 0.121 0.106 0.9 0.9

4 0.121 0.111 0.9 0.9

5 0.119 0.106 0.9 0.9

6 0.122 0.119 0.9 1.0

Repeatability between samples

Average 0.119 0.114 0.9 0.9

RSDa (n=6) 6 4 4 5

Precision between analysts

RSDa (n=12) 4 4

p-value (95%, CIb for difference of means between analysts) 0.153 0.887

95% CIb for difference of means between analysts -0.002 – 0.013 -0.056 – 0.049

aRSD: relative standard deviation (%), relative to average values.

bCI: confidence interval.

2.4. Quantification of HPC and SDS adsorbed to nanoparticles in nanosuspensions

Miconazole nanosuspensions containing 12.5 or 20% miconazole, 0.025, 0.05, 0.125 or 0.2% SDS, and

5% HPC in purified water were used to determine simultaneously SDS and HPC by NIR in the

dispersion medium. The amount of SDS and HPC adsorbed onto miconazole nanoparticles upon

nanogrinding was derived from the differences in the measured concentrations of the medium before

adding the drug substance and after its nanogrinding; in the latter case, the nanosuspensions were

centrifuged at 50,000 rpm (300,000 x g) for 3.5 h (ultracentrifuge Sorvall) to assay the excipients in the

particle-free supernatant. The results were presented as quantity of SDS or HPC adsorbed per unit

surface of the miconazole nanoparticles. The SSA was calculated from the particle size distribution as

determined by laser light diffraction (Mastersizer 2000, Malvern Instruments, Worcestershire, UK).

To verify if the ultracentrifugation could have impact on the HPC and SDS quantification, one sample

was both filtrated and ultracentrifuged and the results compared.

Chapter IV

Page 95 of 182

3. Results

For developing the NIR method for simultaneous quantification of HPC and SDS in miconazole

nanosuspensions and determining the amounts of SDS and HPC adsorbed on the drug nanoparticles,

pertinent nanosuspension characteristics had to be determined first. The studied miconazole

nanosuspensions were prepared according to previously optimized formulation and process parameters,

where minimal amount of SDS was found to be necessary for efficient nanogrinding (Cerdeira et al.,

2010; Cerdeira et al., 2011).

3.1. Miconazole nanoparticle size, ζ-potential, and solubility

Standard miconazole nanosuspensions that contained 12.5 or 20% drug substance, 5% HPC and

variable amounts of SDS (0.0% to 0.2%) exhibited particles of mostly spherical shape as revealed by

scanning electron microscopy (data not shown). The mean particle size (D[4,3]) was in the range of 200

nm to 300 nm, corresponding to a SSA of 45 m2/g to 50 m2/g (Table 1). The absolute ζ-potential value

of the miconazole nanoparticles increased with increasing SDS concentration, i.e. from approx. 0 to −20

mV. The miconazole solubility increased when SDS was present in the nanosuspensions (from approx.

20 µg/ml at 0% SDS to 60–80 µg/ml at 0.2% SDS), although the solubility increase was unsteady. Both

ζ-potential and solubility were analyzed in nanosuspensions containing 5% HPC (Table 1).

3.2. Critical micellar concentration of SDS and SDS–HPC aggregation concentration

For appropriate interpretation of the quantitative NIR data, it seemed important to determine the critical

micellar concentration (CMC) of SDS as well as the critical aggregation concentration (CAC) of SDS in

the presence of 5% (w/w) HPC. The values were determined by surface tension measurements using

both aqueous solutions of SDS and HPC and supernatants of nanosuspensions. In pure aqueous SDS

solutions, the CMC value was in the range of 0.2–0.25% (7–9 mM) (Fig. 1), which agrees with data in

the literature reporting a CMC value of 8.2 mM (Mukerjee et al., 1971). The presence of 5% HPC in the

SDS/HPC solution caused a strong drop of surface tension at low SDS concentrations and moderated

the decrease in surface tension with increasing SDS concentration. At SDS concentrations of above

0.05% (w/w) (1.7 mM), the surface tension of the SDS/HPC-solutions remained nearly constant within

the SDS concentration range studied (up to 0.25% w/w). This discontinuity of surface tension drop of

SDS/HPC solutions likely represents the SDS–HPC aggregation concentration, at which the surfactant

starts to bind to the cellulose ether, which was reported to be approx. 1–1.5 mM for SDS–HPC

solutions (Stefansson, 1998; Berglund et al., 2003). Supernatants of 20% miconazole nanosuspensions

exhibited a slightly higher surface tension than pure SDS/HPC solutions of comparable concentrations

Chapter IV

Page 96 of 182

up to approx. 0.125%, whereas the values merged at higher SDS concentrations (0.2%) (Fig. 1).

Further, the surface tension values of the 12.5% miconazole nanosuspension supernatants were only

very slightly, though consistently lower than the values of the of 20% miconazole nanosuspensions at

same SDS/HPC concentrations and within the SDS concentration range tested (0.025% to 0.125%

SDS).

Figure 1: Surface tension of pure aqueous solutions of SDS alone, SDS and 5% (w/w) HPC, and supernatants of

nanosuspensions, all as a function of SDS concentration; surface tension of water = 72.8 mN/m; surface tension of

HPC 5% (w/w) = 40.4 mN/m, all at 23 °C. CMCSDS: critical micellar concentration of SDS; CACSDS-HPC: critical

aggregation concentration between SDS and HPC.

3.3. NIR spectrometry for quantification of SDS and HPC in miconazole

nanosuspensions

Calibration and linearity with SDS and HPC standard solutions

The present NIR method was developed as a quantitative tool to predict SDS and HPC concentrations

in aqueous nanosuspension formulations from NIR-spectra. The model was built on calibration sets and

tested using a second external data set. Verification of an NIR method and applied calculation is

commonly done by comparing NIR data against results obtained from official methods.

20

25

30

35

40

45

50

0.001 0.010 0.100 1.000

Su

rface

ten

sio

n (

mN

/m)

Conc of SDS (%)

MIC 12.5%/ HPC 5%: nanosuspension supernatant

MIC 20%/HPC 5%: nanosuspension supernatant

HPC 5%/SDSin water

SDS in water

CMCSDS

~7mM (0.2%)

CACSDS-HPC

~1.7 mM (0.05%)

Chapter IV

Page 97 of 182

Figure 2: NIR-spectra of HPC/SDS solutions used for calibration. (A) NIR-spectra over entire wavenumber range.

(B) Second derivative of absorption signals in the region of 8760-8355 cm-1 as used for HPC calibration. (C)

Second derivative of absorption signals in the region of 6150-5724 cm-1 as used for calibration of HPC and SDS.

In the present study, however, only the calibration and validation by an external set of samples were

performed as the objective of the present work was not to fully validate the analytical method, but

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4,0006,0008,00010,00012,000

Ab

sorb

an

ce

Wavenumber (cm-1)

A)

-1.5E-05

-1.0E-05

-5.0E-06

-6.0E-19

5.0E-06

1.0E-05

8,3008,4008,5008,6008,7008,8002n

dd

eriv

ati

ve

of

Ab

sorb

an

ce

(Arb

itra

ry u

nit

s)

Wavenumber (cm-1)

B)

-6.0E-05

-4.0E-05

-2.0E-05

-2.5E-18

2.0E-05

4.0E-05

6.0E-05

8.0E-05

1.0E-04

5,7005,8005,9006,0006,1002n

dd

eriv

ati

ve

of

Ab

sorb

an

ce

(Arb

itra

ry u

nit

s)

Wavenumber (cm-1)

C)

Chapter IV

Page 98 of 182

simply to assess its suitability to measure SDS and HPC in nanosuspensions. The NIR-spectra (Fig. 2A)

showed, as expected overlapping bands that were unspecific and poorly resolved. Nonetheless,

characteristic bands for HPC, SDS, and water (Walling et al., 1986; Langkilde et al., 1995) were also

present, such as those related to −CH3, =CH2, and ≡CH groups (at approx. 8500 and 5700 cm−1),

−ROH groups (at 6700, 6250, and 4800 cm−1), and water (at approx. 6900 and 5200 cm−1) (Fig. 2B and

2C). After resolving from the water bands, SDS shows characteristic bands between approx. 5900 and

5600 cm−1 which has been described already a long time ago (Walling et al., 1986). Simultaneous

quantification of SDS and HPC in aqueous solution was achieved by FT-NIR in combination with

multivariate analysis of the light absorption in the range of wave numbers as indicated in Table 3. The

mathematical pre-processing of the NIR spectral data was an important step in the chemometric

analysis to enhance spectral features, remove or reduce unwanted bands, and standardize these effects

of variation prior to the development of the calibration model (Otsuka et al., 2010; PASG, 2001). Thus,

the raw data were transformed into second derivatives (Figs. 2B and C) to improve the resolution

between overlapping bands and minimize baseline offset for fitting the values according to partial least-

squares algorithm with cross validation. Both SDS and HPC (Table 3) calibrations yielded highly

satisfactory correlation coefficients (r > 0.99). The low values obtained for the RMSECV (Eq. 3) and

for bias support the adequacy of the calculation model. The low number of latent variables (Table 3)

(correspondent to the minimum RMSECV) also confirms the method validity.

IJ6KLM = NO (PEQR�PR)SEBTU

�VW (3)

Here, Ycvi represents the cross-estimate of sample i, Yi the actual concentration of the same sample, and

Nc the number of calibration samples. The method validity was confirmed by plotting an external set of

independent samples (known concentrations) against the predicted values calculated by the developed

model (Figs. 3A and B).

Chapter IV

Page 99 of 182

Accuracy with SDS and HPC solutions

For determining accuracy, known amounts of SDS and HPC were co-dissolved in water at randomly

selected concentrations covering the whole concentration range of interest. For SDS, the mean recovery

rates (six independent samples of each concentration) lay between 98 and 105%, and the maximum

relative standard deviation was 13% (Table 4). The mean HPC recovery rate of six independent samples

varied between 98 and 110%, and the relative standard deviation was 6%.

Table 7: Influence of dissolved miconazole on SDS (nominal 0.2%) and HPC (nominal 1.25% and 5%)

quantification.

Miconazolea (µg/ml) SDS, 0.20% (w/w) HPC, 1.25% (w/w)

0 0.19 1.1

42 0.20 1.1

69 0.20 1.1

Average 0.20 1.1

Pooled standard deviation 0.0132 0.0242

p-value (95%, CIb) 0.428 0.961

Miconazolea (µg/ml) SDS, 0.20% (w/w) HPC, 5.0% (w/w)

0 0.20 4.8

89 0.20 4.8

152 0.19 4.8

Average 0.20 4.8

Pooled standard deviation 0.011 0.032

p-value (95%, CIb) 0.195 0.954

aMiconazole concentration was determined by HPLC bCI: confidence interval

Chapter IV

Page 100 of 182

Figure 3: Correlation between observed and predicted concentrations of SDS (A) and HPC (B) using an external

set of samples.

Preparation of nanosuspension supernatant for assaying SDS and HPC by NIR

For the preparation of miconazole nanosuspension supernatant for quantification of SDS and HPC, the

methods of ultracentrifugation and filtration (0.22 µm) were considered. The two methods were then

evaluated for their suitability to quantify reliably SDS and HPC in the supernatants. The results

indicated that the two particle peparation methods did not influence on the determined SDS and HPC

concentrations of a test sample: HPCcentrifuged = 5.6 ± 0.02%, HPCfiltered = 5.6 ± 0.03%; SDScentrifuged = 0.2

± 0.01%, SDSfiltered = 0.2 ± 0.01%. Hence, the ultracentrifugation method was selected for the further

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0 0.05 0.1 0.15 0.2 0.25 0.3

SD

S c

on

cen

tra

tio

n p

red

icte

d (

%)

SDS concentration measured by NIR (%)

A)

R2 = 0.99489

0

1

2

3

4

5

6

7

0 1 2 3 4 5 6 7

HP

C c

on

cen

trati

on

pre

dic

ted

(%

)

HPC concentration measured (%)

B)

R2 = 0.99996

Chapter IV

Page 101 of 182

experiments, because the moderately elevated viscosity (approx. 5–350 mP s) of the nanosuspensions

and the small particle sizes rendered the filtration process quite difficult.

Repeatibility and precision between analysts of SDS and HPC quantification in miconazole

nanosuspensions

The repeatability and precision between analysts of simultaneous quantification of SDS and HPC by

FT-NIR was determined with 20% miconazole nanosuspensions, which contained the two excipients as

stabilizers at concentrations of 0.03 or 0.2% SDS and 1.25 or 5% HPC (Tables 5 and 6). The relative

standard deviations of the determined SDS and HPC concentrations were between 1 and 14%, both in

the repeatability and the precision between-analysts tests. Thus, no significant difference was generally

found between the measurements when performed by different analysts on different days (two sample t-

test, p > 0.05), except for HPC at a concentration of 5% (two-sample t-test, p < 0.05).

Specificity of SDS and HPC quantification in miconazole nanosuspensions

The miconazole nanosuspension of interest contained water, SDS, HPC, and a minimal amount of

dissolved drug substance. Thus, we wanted to examine whether the miconazole present in the

supernatant of the ultracentrifuged nanosuspensions interfered with the quantification of SDS and HPC.

As miconazole solubility increases with SDS and HPC concentrations (Cerdeira et al., 2010), solutions

of 0.2% SDS, 1.25 or 5% HPC, and different miconazole concentrations were prepared to test the

method specificity (Table 7). No significant differences (ANOVA, p > 0.05) in the amounts of SDS or

HPC were detected between the different solutions containing different amounts of excipients and

dissolved drug (Table 7). Therefore, according to the current ICH guidelines, the applied NIR method is

specific with respect to the drug substance.

Reporting thresholds

The limit of quantitation is the minimum concentration in the sample that can be quantified with

precision and accuracy. Based on the accuracy (Table 4), precision (Tables 5 and 6), and linearity data

(Figs. 3A and 3B), the reporting thresholds of the method were 0.025% SDS and 1.25% HPC.

3.4. Quantification of SDS and HPC adsorbed onto miconazole nanoparticles in

nanosuspension

The amounts of SDS and HPC adsorbed onto miconazole nanoparticles were quantified in miconazole

nanosuspensions of different compositions (all as %, w/w): 12.5 or 20.0% miconazole, 0.025, 0.05,

Chapter IV

Page 102 of 182

0.125, or 0.2% SDS, and 5% HPC. The amounts of adsorbed excipients were calculated from the

concentration difference in the medium before adding the drug substance and after its nanomilling.

Consistent with common isothermal adsorption behavior, SDS adsorption onto the miconazole

nanoparticles increased with increasing SDS equilibrium concentration from 11 to 75 µg/m2 (0.38 ×

10−7–2.6 × 10−7 mol/m2) and from 11 to 122 µg/m2 (0.38 × 10−7–4.2 × 10−7) in the 20% and 12.5%

miconazole nanosuspensions, respectively, all containing also 5% HPC (Fig. 4A). On the other side,

increasing the SDS concentration from initial 0.025% to 0.2% lowered the adsorption of HPC onto the

miconazole nanoparticles from 898 to 713 µg/m2 (23 × 10−7–18 × 10−7 mol repeating unit/m2) and from

1,071 to 839 µg/m2 (27 × 10−7–21 × 10−7) in the 20% and 12.5% miconazole nanosuspensions,

respectively (Figure 4B); for calculating the molar quantities of adsorbed HPC, a molecular weight of

391.5 was assigned to the 2-propoxysubstituted glycosyl unit, assuming a degree of molecular

substitution of 3.8 as specified by the manufacturer.

0

20

40

60

80

100

120

140

0 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0.2

Am

ou

nt

of

SD

S a

dso

rbed

on

to

nan

op

art

icle

s (µ

g/m

2)

SDS equilibrium concentrations (%, w/w)

Supernatant ofMIC 12.5%/HPC 5%

Supernatant of MIC 20%/HPC 5%

A)

SDS

Chapter IV

Page 103 of 182

Figure 4: Amounts of SDS (A) and HPC (B) adsorbed onto miconazole nanoparticles in 12.5% and 20%

miconazole nanosuspensions as a function of SDS equilibrium concentration in the suspension medium. All

nanosuspensions contained 5% HPC.

4. Discussion

Different methods have been described in the literature for quantifying the adsorption of stabilizers onto

nanoparticles. However, most of the methods are complex, which restricts their use in routine

development, and generally unsuitable to quantify combinations of excipients, as frequently used to

achieve nanosuspension stabilization. Indeed, most authors have described the quantification of single

excipients, as for example by thermogravimetry for assaying HPC (Van Eerdenbrugh et al., 2008b),

optical rotatory dispersion assay for hydroxypropylmethylcellulose, or UV-spectroscopy for PVP (Lee,

2003). Adsorption of d-α-tocopherol polyethylene glycol 1000 succinate (TPGS) onto nanoparticles

was analysed by HPLC–UV after hydrolysis of TPGS to d-α-tocopherol (Ploehn et al., 1990). More

recently, size exclusion chromatography combined with evaporative light scattering detection was used

to determine individually the adsorption of various stabilizers (SDS, PEG, poloxamer, PVP, cellulose

ethers) onto siramesine hydrochloride microcrystals (Sepassi et al., 2007). In the present work, a FT-

NIR method was developed to quantify simultaneously SDS and HPC in aqueous miconazole

nanosuspensions and thereby determine indirectly the amount of excipients adsorbed onto the

nanoparticles. SDS and HPC are two very common stabilizers for nanosuspensions; they are frequently

used in combination and known to interact with each other (Evertsson et al., 1997; Berglund et al.,

2003; Lee et al., 2008). The NIR method was found to be linear, accurate, and precise within the SDS

and HPC concentration ranges of interest; the method was also specific with respect to dissolved drug.

300

400

500

600

700

800

900

1,000

1,100

1,200

1,300

0 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0.2

Am

ou

nt

of

HP

C a

dso

rbed

on

to

na

no

pa

rtic

les

(µg

/m2)

SDS equilibrium concentrations (%, w/w)

B)

HPCSupernatant ofMIC 12.5%/HPC 5%

Supernatant ofMIC 20%/HPC 5%

Chapter IV

Page 104 of 182

The higher precision and accuracy for HPC as compared to SDS must be ascribed to the substantially

higher HPC concentration that was of interest in the aqueous solution (0.75–7.5% HPC versus 0.03% to

0.3% SDS). In fact, the quantification of low amounts of excipients (< 1%, w/w) is one of the

limitations of NIR, because NIR absorption bands are typically broad and overlapping, and 10–100

times weaker than in the mid-IR range, which lowers the method sensitivity (Reich, 2005). Another

difficulty of NIR for quantifying substances in aqueous solutions is occasioned by the presence of water

which causes interference from broad vibrational bands (Chen et al., 2004). To eliminate spectral

interferences, it is a common practice and also necessary to pre-process the NIR spectral data of the

aqueous medium devoid of the analytes (Chen et al., 2004). Chemometric data processing via second

derivative prior to multivariate modeling was therefore used to reduce and standardize the effect of

interfering spectral parameters and improve the resolution of the overlapping bands and reduce baseline

offsets (Chen et al., 2004; Reich, 2005; Roggo et al., 2007). The feasibility of using NIR to determine

simultaneously glucose, lactic acid, and ammonia in aqueous fibroblast cultures has already been shown

earlier (Rhiel et al., 2002).

As in our study, the authors had used only parts of the absorption bands of the compounds of interest,

and also applied partial least square regression to construct accurate calculation models.

We could also show that miconazole does not interfere with the quantification of SDS and HPC which

is important as the miconazole solubility increases in the presence of SDS. In general, as the solubility

of candidate drug substances for nanosuspensions should be low (<1 mg/ml), we expect that the

developed NIR method should also be applicable to other drug nanosuspensions.

The FT-IR method was primarily developed for determining indirectly the amounts of SDS and HPC

that were adsorbed onto the micronazole nanoparticles, thereby contributing to the physical stability of

the nanosuspensions.

In a previous work (Cerdeira et al., 2010), we manufactured miconazole nanosuspensions with

increasing amounts of SDS (0.0125% to 0.2%, w/w) and observed that higher amounts of SDS (>

0.05%) were detrimental for nanogrinding when HPC was present at a concentration of above 3%

(w/w).

We hypothesized that increasing amounts of SDS might displace competitively some of the adsorbed

HPC (Evertsson et al., 1997; Berglund et al., 2003; Lee et al., 2008), which may have resulted in partial

loss of steric nanoparticle stabilization by HPC, which in turn was found to be crucial for efficient

miconazole particle size reduction and nanosuspension stabilization (Cerdeira et al., 2010). Competitive

displacement of HPC by SDS was indeed confirmed in the present study.

SDS adsorption on the surface of miconazole nanoparticles increased with SDS equilibrium

concentration, even in the presence of 5% (w/w) HPC (Figure 4A); this agrees with the increasing ζ-

potential values of the miconazole nanoparticles with increasing SDS concentrations in the

Chapter IV

Page 105 of 182

nanosuspensions (Table 1). At SDS equilibrium concentration of 0.05% (1.7 mM), the amount of

adsorbed SDS was 11 µg/m2 (0.38 × 10−7 mol/m2), which increased to 75 µg/m2 (2.6 × 10−7 mol/m2) and

122 µg/m2 (4.2 × 10−7 mol/m2) at 0.17% SDS in the 20% and 12.5% miconazole nanosuspensions,

respectively. Interestingly, the amount of SDS adsorbed onto the miconazole nanoparticles at

equilibrium concentration of 0.17% was of similar magnitude as that recently reported for n-hexadecane

oil droplets in water (3.9 × 10−7 mol/m2) (de Aguiar et al., 2010).

In the case of the miconazole nanosuspensions, SDS adsorption was certainly complicated by the

presence and concomitant adsorption of HPC and/or the known SDS–HPC interaction. Earlier studies

indeed described such SDS–HPC interaction which was reported to occur above a so-called CAC of 1.5

mM (0.043%) SDS (Stefansson, 1998; Berglund et al., 2003). A similar critical SDS concentration of

approx. 1.7 mM (0.05%) for SDS–HPC aggregation was found in the present work by surface tension

measurements using solutions of 5% HPC and increasing amounts of SDS (Fig. 1). On the other side,

SDS and HPC competed for adsorption onto miconazole nanoparticles. Indeed, upon increase of SDS

concentration from 0.025 to 0.2%, some of the HPC was displaced from the nanoparticles with the

adsorbed amount of HPC being lowered from approx. 900 µg/m2 to 700 µg/m2 (23 × 10−7–18 × 10−7

mol/m2) in case of the 20% miconazole nanosuspensions, and from 1000 µg/m2 to 800 µg/m2 (27 ×

10−7–21 × 10−7 mol/m2) with the 12.5% miconazole nanosuspensions (Figure 4B). Thus, although the

apparent amounts of HPC repeating units adsorbed were substantially higher (18–27 × 10−7 mol/m2)

than that of SDS (2–4 × 10−7 mol/m2), increasing amounts of SDS appeared to displace some of the

HPC, possibly because of higher affinity of the low molecular weight surfactant for the miconazole

nanoparticle surface. The adsorption of SDS and HPC onto the miconazole nanoparticles was also

confirmed qualitatively by surface tension measurements, where the surface tension values of

nanosuspensions supernatants was higher than those measured for pure SDS/HPC solutions of

comparable nominal concentrations. Furthermore, the surface tension of the 12.5% miconazole

nanosuspension was very slightly lower, which agrees with the NIR measurements that showed slightly

higher HPC adsorption onto the nanoparticles of the 12.5% miconazole nanosuspension.

Finally, the multiple molecular interactions operating in the studied nanosuspensions (excipients–solid

nanoparticles; excipients–dissolved drug substance; between excipients, i.e. SDS–HPC) not only

affected the adsorption of the excipients on the nanoparticles, but also the solubilizing capacity of SDS

for miconazole. While the addition of SDS to a miconazole suspension (containing 5% HPC) increased

the solubility of the drug substance (Table 1), the increase was not steady over the SDS concentration

range studied (0% to 0.2%), but showed a minimum at 0.125% SDS. We hypothesize that the earlier

described SDS–HPC interaction (aggregation) may cause this solubility minimum. In detail, the data

suggest that miconazole is solubilized by early and possibly mixed micelles upon addition of minimal

amounts (0.025%) of SDS. At 0.05% SDS, which corresponds to the CAC, SDS starts to interact

Chapter IV

Page 106 of 182

strongly with HPC and will no longer be available for solubilizing additional miconazole. Upon further

SDS addition, SDS may be bound even more strongly by HPC, and definitely adsorbs to a very

important extent onto the miconazole particles (Figure 4B). Thus, at a SDS concentration of 0.125%, an

important amount of SDS is bound to HPC and the suspended nanoparticles and less available for

solubilizing miconazole. Only upon further SDS addition (0.2%), the solubilizing capacity increases

again and reaches values similar to those observed at very low SDS concentration (Table 1).

5. Conclusions

A NIR spectroscopic method was developed to quantify simultaneously SDS and HPC in aqueous

solutions for quantifying indirectly the adsorption of the two excipients onto miconazole nanoparticles

in nanosuspensions.

Although NIR presents limitations for quantifying low amounts of excipients in aqueous media, the

developed methodology and data analysis for SDS and HPC in the supernatant of nanosuspensions have

proven to be accurate, precise, specific with respect to the compounds of interest, sufficiently sensitive,

fast, and straightforward.

We expect that the method can be applied to quantify other stabilizing excipients in nanosuspensions of

other drug substances. The determined amounts of excipients adsorbed provided indirect information on

their affinity to the miconazole nanoparticle surface and their quantitative importance for stabilizing the

nanoparticles. Thus, the work provides an elegant analytical tool which may eventually help understand

better the quantitative contributions of individual stabilizers when used in combinations.

Chapter IV

Page 107 of 182

References



modification. Langmuir, 19:2705–2713.

Bhakay, A., Merwade, M., Bilgili, E., Dave, R.N., 2011. Novel aspects of wet milling for the

production of microsuspensions and nanosuspensions of poorly water-soluble drugs. Drug. Dev. Ind.

Pharm., 37, 963–976.

Blanco, M., Coello, J., Eustaquio, A., Iturriaga, H., Maspoch, S., 1999. Development and validation of a

method for the analysis of a pharmaceutical preparation by near-infrared diffuse reflectance

spectroscopy. J. Pharm. Sci. 88, 551–556.

Cerdeira, A.M., Mazzotti, M., Gander, B., 2010. Miconazole nanosuspensions: influence of formulation

variables on particle size reduction and physical stability. Int. J. Pharm., 396, 210–218.

Cerdeira, A.M., Mazzotti, M., Gander, B., 2011. Role of milling parameters and particle stabilization on

nanogrinding of drug substances of similar mechanical properties. Chem Eng Technol, 34:1427–

1438.

Chen, D., Hu, B., Shao, X, Su, Q., 2004. Removal of major interference sources in aqueous near-

infrared spectroscopy techniques. Anal Bioanal Chem, 379:143–148.

de Aguiar, H.B., de Beer, A.G., Strader, M.L., Roke, S., 2010. The interfacial tension of nanoscopic oil

droplets in water is hardly affected by SDS surfactant. J. Am. Chem. Soc., 132:2122–2123.

EMEA (The European Medicines Agency for the Evaluation of Medicinal Products). (2003). Notes for

guidance on the use of near infrared by the pharmaceutical industry and the data requirements for

new submissions and variations, London, 2003.

http://www.tga.gov.au/docs/pdf/euguide/qwp/330901en.pdf. Accessed on 10 January 2012.

Evertsson, H., Nilsson, S., 1997. Microviscosity in clusters of ethyl hydroxyethyl cellulose and sodium

dodecyl sulfate formed in dilute aqueous solutions as determined with fluorescence probe

techniques. Macromolecules, 30:2377–2385.

Kesisoglou, F., Panmai, S., Wu, Y., 2007. Nanosizing–oral formulation development and

biopharmaceutical evaluation. Adv. Drug Deliver Rev. 59, 631–644.

Langkilde, F.W., Svantesson, A., 1995. Identification of celluloses with Fourier-transform (FT) mid-

infrared, FT-Raman and near-infrared spectrometry. J Pharm Biomed Anal, 13: 409–414.

Lee, J., 2003. Drug nano- and microparticles processed into solid dosage forms: physical properties. J.

Pharm. Sci. 92, 2057–2068.

Lee, J., Choi, J.Y., Park, C.H., 2008. Characteristics of polymers enabling nano-comminution of water-

insoluble drugs. Int. J. Pharm., 355:328–336.

Chapter IV

Page 108 of 182

López-Arellano, R., Santander-García, E.A., Andrade-Garda, J.M., Alvarez-Avila, G., Garduño-Rosas,

J.A., Morales-Hipólito, E.A., 2009. Quantification of lysine clonixinate in intravenous injections by

NIR spectroscopy. Vibrational Spectrosc, 51, 255–262.

Mark, H., Ritchie, G.E., Roller, R.W., Ciurczak, E.W., Tso, C., MacDonald S.A., 2002. Validation of a

near-infrared transmission spectroscopic procedure, part A: validation protocols. J. Pharm. Biomed.

Anal. 28, 251–260.



Moffat, A.C., Trafford, A.D., Jee, R.D., Graham, P., 2000. Meeting the International Conference on

Harmonisation’s Guidelines on Validation of Analytical Procedures: quantification as exemplified

by a near-infrared reflectance assay of paracetamol in intact tablets. Analyst 125, 1341–1351.

Otsuka, E., Abe, H., Aburada, M., Otsuka, M., 2010. Nondestructive prediction of the drug content of

an aspirin suppository by nearinfrared spectroscopy. Drug Dev. Ind. Pharm. 36, 839–844.

PASG (Pharmaceutical Analytical Sciences Group). (2001) Guidelines for the development and

validation of near-infrared (NIR) spectroscopic methods, May 2001:1–39.

<http://www.pasg.org.uk/NIRmay01.pdf>. Accessed on 10 January 2012.

Ploehn, H.J., Russel, W.B., 1990. Interactions between colloidal particles and soluble

polymers. Adv. Chem. Eng. 15, 137–228.

Rabinow, B.E., 2004. Nanosuspensions in drug delivery. Nat. Rev. Drug Discov., 3, 785–796.

Reich, G., 2005. Near-infrared spectroscopy and imaging: basic principles and pharmaceutical

applications. Adv. Drug Deliver Rev., 57:1109–1143.

Rhiel, M., Cohen, M.B., Murhammer, D.W., Arnold, M.A., 2002. Nondestructive near-infrared

spectroscopic measurement of multiple analytes in undiluted samples of serum-based cell culture

media. Biotechnol Bioeng, 77:73–82.

Roggo, Y., Chalus, P., Maurer, L., Lema-Martinez, C., Edmond, A., Jent N., 2007. A review of near

infrared spectroscopy and chemometrics in pharmaceutical technologies. J Pharm Biomed Anal,

44:683–700.

Sepassi, S., Goodwin, D.J., Drake, A.F., Holland, S., Leonard, G., Martini, L., 2007. Effect of polymer

molecular weight on the production of drug nanoparticles. J. Pharm. Sci. 96, 2655–2666.

Stefansson, M., 1998. Characterization of cellulose derivatives and their migration behaviour in

capillary electrophoresis. Carbohydr. Res., 312:45–52.

Van Eerdenbrugh, B., Vermant, J., Martens, J.A., Froyen, L., Van Humbeeck, J., Augustijns, P., 2009.

A screening study of surface stabilization during the production of drug nanocrystals. J. Pharm. Sci.

98, 2091–2103.

Chapter IV

Page 109 of 182

Van Eerdenbrugh, B., Vercruysse, S., Martens, J.A., Verman,t J., Froyen L., Van Humbeeck, J., 2008.

Microcrystalline cellulose, a useful alternative for sucrose as a matrix former during freeze-drying of

drug nanosuspensions – a case study with itraconazole. Eur. J. Pharm. Biopharm. 70, 590–596.

Van Eerdenbrugh, B., Van den Mooter, G., Augustijns, P., 2008. Top-down production of drug

nanocrystals: nanosuspension stabilization, miniaturization and transformation into solid products.

Int. J. Pharm. 364, 64–75.

Walling, P.L., Dabney, J.M., 1986. Application of near infrared reflectance spectroscopy to the quality

assurance of surfactants. J. Soc. Cosmet. Chem., 37:445–454.

Zimmermann, A., Millqvist-Fureby, A., Elema, M.R., Hansen, T., Müllertz, A., Hovgaard, L., 2009.

Adsorption of pharmaceutical excipients onto microcrystals of siramesine hydrochloride: effects on

physicochemical properties. Eur. J. Pharm. Biopharm. 71, 109–116.

Chapter IV

Page 110 of 182

Chapter V

Page 111 of 182

Chapter V

Formulation and drying of miconazole and itraconazole nanosuspensions4

1. Introduction

Nanoparticle technology has gained wide interest in the medicinal and pharmaceutical sciences as it has

numerous applications; nanoparticles can be designed to target and/or sustain the delivery of drugs, to

improve oral bioavailability, or for pulmonary, ocular, or parenteral delivery (Patravale et al., 2004;

Mishra et al., 2009; Müller et al., 2011; Merisko-Liversidge et al., 2011; Shegokar et al., 2011; Gao et

al., 2012). For very slightly water-soluble or practically water-insoluble drug substances,

nanosuspensions are of great interest, as they can be formulated with up to 40% drug content in either

aqueous or mixed aqueous-organic solvents, require only small amounts of non-toxic excipients, and

may preserve drug stability better than other formulations (Patravale et al., 2004). As a consequence,

several nanosuspension products have become available, with most of them being manufactured by

media milling and intended for oral administration.

The current engineering processes to obtain nanosuspensions are divided into bottom-up processes such

as nanoprecipitation or nanocrystallization (Chan et al., 2011; D'Addio et al., 2011; Dandagi et al.,

2011) and top-down processes such as high pressure homogenization (Keck et al., 2006) and media

milling (Merisko-Liversidge et al., 2003; Merisko-Liversidge et al., 2011). In media milling

(nanogrinding), for example, desired particle size range and particle stability may be achieved by

optimizing the formulation (e.g., type and concentration of excipients and concentration of drug

substance) (Wu et al., 2011) as well as the process parameters (e.g., size of grinding beads and specific

energy input) (Cerdeira et al., 2011; Merisko-Liversidge et al., 2011; Juhnke et al., 2012).

Whereas process parameters can be optimized relatively easily (Cerdeira et al., 2011; Hennart et al.,

2012), the selection of adequate excipients remains an important challenge that has to be addressed

4 The work in this chapter has been submitted, after revision, for publication to the International Journal of Pharmaceutics.

Chapter V

Page 112 of 182

mostly empirically (Merisko-Liversidge et al., 2011; Wu et al., 2011); the latter is due to lack of basic

knowledge of exact mechanisms of nanosuspension stabilization upon nanogrinding (Lee et al., 2005;

Lee et al., 2008; Merisko-Liversidge et al., 2011). Nanosuspension formulation generally requires

addition of appropriate stabilizers to lower the free surface energy of the nanoparticles and prevent

particle aggregation and/or particle growth. The high surface free energy of nanoparticles is readily

lowered by lowering the solid-liquid interfacial tension upon addition of surfactants (Rabinow, 2004).

Particle aggregation or growth may be efficiently prevented or at least slowed down through adsorption

of stabilizers that form electrostatically repulsive or steric barriers (Merisko-Liversidge et al., 2011; Wu

et al., 2011). So far, only relatively few compounds have proven to be suitable for nanosuspension

stabilization, as for example, sodium dodecylsulfate (SDS), polysorbates, povidones, poloxamers, and

cellulose derivatives (Van Eerdenbrugh et al., 2009). Surfactant / polymer mixtures have often shown

synergistic effects (Lee et al., 2008; Cerdeira et al., 2010).

To enhance long-term stability, nanosuspensions can be converted into dry form, typically by freeze-

drying or spray-drying (Liu et al., 2010; Chaubal et al., 2008; Van Eerdenbrugh et al., 2008a; Choi et

al., 2005). However, drying of nanosuspensions may negatively affect nanoparticle size and

dispersibility (Van Eerdenbrugh et al., 2008a; Chaubal et al., 2008). The generation of aggregates likely

causes alteration of disintegration and dissolution, which may subsequently cause changes in

bioavailability (Chaubal et al., 2008). Therefore, particle size distribution of dried nanosuspensions is a

critical quality attribute, which is primarily affected by the formulation.

Particle aggregation upon drying of nanosuspensions can be minimized by adding so-called matrix

formers such as sugars or sugar alcohols (e.g., sucrose, lactose, mannitol) or insoluble excipients (e.g.,

microcrystalline cellulose, colloidal anhydrous silica) (Van Eerdenbrugh et al., 2008c); matrix formers

fill the gaps between nanoparticles upon water removal and thereby prevent undue close contacts

between the particles (Kim et al., 2010). To select appropriate excipients both for nanogrinding and

subsequent drying it is important to consider drug substance properties and their potential role in the

manufacturing process and final product quality attributes. For example, in a study on nine drug

compounds, which were media-milled in presence of the stabilizer d-tocopheryl polyethylene glycol

1000 succinate (TPGS) and subsequently freeze-dried or spray-dried, Van Eerdenbrugh et al., 2008a

demonstrated that the drug substance hydrophobicity plays a major role in these processes; more

hydrophobic drug substances, including itraconazole, were found to be more difficult to stabilize

against irreversible particle aggregation during drying.

Chapter V

Page 113 of 182

Drying of itraconazole nanosuspensions, obtained by bottom-up or top-down processes, has been

studied by different authors (Chaubal et al., 2008; Lee et al., 2008; Van Eerdenbrugh et al., 2008b; Liu

et al., 2010; Mou et al., 2011). For example, Van Eerdenbrugh et al., 2008b reported that

microcrystalline cellulose was superior to sucrose as matrix former to avoid agglomeration upon freeze-

drying of itraconazole nanosuspensions stabilized with 10%TPGS. Contrarily to itraconazole,

miconazole nanosuspensions have, to our knowledge, only scarcely been considered (Cerdeira et al.,

2010; Cerdeira et al., 2011; Cerdeira et al., 2012). Itraconazole and miconazole are imidazole

derivatives of low solubility (BCS II) and used as antifungals (Piel et al., 1998; Tsutsumi et al., 2011;

Van Eerdenbrugh et al., 2009). The two drug substances differ, however, in their molecular weight

(miconazole: 416.1 g/mol; itraconazole: 705.6 g/mol), melting temperature (miconazole: 83-87 °C for

polymorph I; itraconazole: 168 °C), and water-solubility (miconazole: ~ 1 mg/L; itraconazole: ~ 0.1

mg/L, both in unbuffered water and pH 7 buffer). Therefore, comparison between the two drug

substances in nanogrinding and susbsequent drying appeared to be of interest, as compounds with of

higher molecular weight and melting point, and lower aqueous solubility (itraconazole) were reported to

be easier to formulate as nanosuspensions (Lee et al., 2008).

In the present study, we first aimed at relating the stability of miconazole and itraconazole

nanosuspensions, obtained by media milling, with the adsorption of various stabilizers. Further, we

assessed various matrix formers in spray-drying and freeze-drying of the obtained nanosuspensions in

terms of nanoparticle stability and dissolution.


2.1. Materials

Miconazole (diameter D[4, 3] ~ 20 µm; lot # R018134PUC701, Janssen Pharmaceutica N.V., Geel,

Belgium) and itraconazole (diameter D[4, 3] ~ 20 µm; lot # ZR051211PUK401, Janssen Pharmaceutica

N.V., Geel, Belgium) were used as received. All other materials were also used as received: sodium

dodecyl sulfate [SDS] (Texapon® K12P, Cognis, Düsseldorf, Germany; hydroxypropylcellulose [HPC]

(type LF, Hercules, Doel, Belgium); hydroxypropylmethylcellulose [HPMC] (Hypromellose 2910, E15

LV, Colorcon, Dow Chemicals, Dartford, UK); poloxamers [poloxamer 188 and 407] (Pluronic® F68

and F127; BASF, Ludwigshafen, Germany); mannitol (Pearlitol 160C®, Roquette, Lestrem, France);

microcrystalline cellulose (Avicel PH-105®, FMC BioPolymer, Brussels, Belgium); croscarmellose

sodium (Ac-Di-Sol®, FMC, Philadelphia, PA, US); gelatine capsules (size 0, Capsugel®, Bornem,

Belgium).

Chapter V

Page 114 of 182

2.2. Nanogrinding of the drug substances

Formulation and process parameters for this work were selected according to previous experiments

(Cerdeira et al., 2010, Cerdeira et al., 2011). A first series of experiments aimed at screening polymeric

excipients for their suitability for nanogrinding and stabilizing the two drug substances. For this, HPC,

HPMC, or the poloxamers 188 or 407 were used at a concentration of 5% (w/w) and each in

combination with 0.05% (w/w) SDS. The concentration of drug substance in the suspensions was kept

at 20% (w/w). HPC without SDS was also tested as a stabilizer for itraconazole to compare with

previous experiments with miconazole (Cerdeira et al., 2010). A second series of experiments examined

the effect of the quantitative composition of the formulations on the particle size reduction using 5,

12.5, or 20% miconazole or itraconazole, along with 0.05, 0.125 or 0.2% SDS, and 1.25, 3.125 or 5%

HPC.

For preparing the nanosuspensions, SDS was dissolved in purified water, and the polymer added under

mechanical agitation. The drug substance was then dispersed in the stabilizer solution and kept under

mechanical stirring for 60 min. The suspensions were left overnight to reduce the air incorporated

before starting the nanogrinding process. Nanogrinding was performed in a high-energy mill (LabStar

LS1 MiniCer, Netzsch, Selb, Germany) filled to 83% (v/v; apparent volume of grinding beads relative

to the volume of the grinding chamber) with yttrium-stabilized zirconium oxide beads (0.8 mm or 0.4

mm in diameter). The suspension was first circulated through the milling chamber to adjust the flow to

113 g/min, before turning on the stirrer. Nanogrinding was performed in circulation mode using 300 g

of suspension, a pump-speed of 41 rpm (113 g/min), and a stirrer-tip speed of 3400 rpm (10 m/s). The

stirrer speed was gradually increased during 5 min from 1000 rpm to 3400 rpm; the duration of the

process lasted 60 min. The nanosuspensions that were further dried (see below) were all milled with

beads of 0.4 mm in diameter.

2.3. Spray-drying and freeze-drying of coarse drug suspensions and nanosuspensions

Amounts of suspension (drug substance 20% / HPC 5% / SDS 0.05%, w/w) equivalent to 8 g of drug

substance were diluted with water to obtain drug substance concentration of 10% (w/w). For spray-

drying, coarse suspensions and nanosuspensions were processed either without any matrix former or

supplemented with 8 g of mannitol or MCC (ratio of drug substance : matrix former of 1:1, w/w). For

freeze-drying the suspensions, additional experiments with higher amounts of matrix formers (ratio of

drug substance : matrix former of 1:2) were performed. Because of the limited solubility of mannitol

(22 g/100 ml at 25 °C), the suspensions had to be further diluted with water for dissolving 16 g of

mannitol.

Chapter V

Page 115 of 182

Spray-drying (Mini Spray-Dryer B191, Büchi, Flawil, Switzerland) was performed under the following

conditions: inlet temperature of 92°C; drying air-flow of 60 m3/h (setting: 100%); product feed rate of

2.5 ml/min. During spray-drying, the suspension was continuously stirred with a magnetic stirrer, and

the nozzle cooled with tap water.

Freeze-drying (Minilyo Usifroid, Maurepas, France) of the drug suspensions was performed with and

without an annealing step; the annealing step was necessary to avoid breakage of vials during the

primary drying when mannitol was present. For freeze-drying without annealing, 10 g of suspension

were filled into 20 ml vials and frozen at a shelf-temperature of -50 °C for 2.75 h. Primary drying took

place at a shelf-temperature of -25 °C and a pressure of 0.1 mbar during 74 h. The secondary drying

was done at a shelf-temperature of 35 °C for 13 h and a pressure of 0.08 mbar. For freeze-drying with

an annealing step, 5 g of suspension were filled into 20 ml vials and frozen at a shelf-temperature of -50

°C for 2.75 h followed by annealing at -20 °C for 2.5 h. Primary drying time took place under afore

described temperature-pressure conditions, although it was shortened to 52 h. The selected annealing

temperature was above the glass transition temperature, Tg, of mannitol (Tg1: -32°C; Tg2: -25 °C

(Cavatur et al., 2002) and below the eutectic temperature, Teut, of mannitol (-1.5 °C) (Kim, et al., 1998).

The use of the annealing step and lower volume of suspension in the vials shielded from vial breakage.

2.4. Particle size characterization

The nanosuspensions as well as the spray-dried and freeze-dried products were examined immediately

after production for homogeneity and presence of agglomerates by optical microscopy (Zeiss Axiophot,

Zürich, Switzerland) and for particle size distribution by laser light diffraction (Mastersizer 2000,

Malvern Instruments, Worcestershire, UK). In addition, nanosuspensions were also stored in liquid

form, and their particles analysed after 3 months of storage at 5 ± 3 °C and 25 ± 3 °C. Samples

(nanosuspensions and dried powders) were appropriately diluted with water for measurement; no

ultrasound was used to promote particle desaggregation. Volume weighted particle size distribution was

calculated by means of the Mie theory, using the following refractive indices: dispersant: 1.33; real and

imaginary particle refractive indices of miconazole: 1.55 and 0.001; of itraconazole: 1.64 and 0.001; the

refractive indices were determined by extrapolation method (Saveyn et al., 2002) and Becke Line

microscopy test using Cargille™ Refractive Index Liquids. Particle sizes were primarily expressed by

the volume weighted mean, D[4,3], as this mean is quite sensitive to the presence of large particles and,

therefore, considered to be most suitable for comparing different nanosuspension formulations. Particle

undersize diameters of d50 and d90 are provided in Supplementary materials. The specific surface area

was calculated according to:

Chapter V

Page 116 of 182

SSA = 9∑Z[\[]∑^[ = 9] _[A,B] (1)

where Vi is the particle volume in particle size fraction i with a mean diameter of di, ρ is the density of

the material, and D[3,2] is the surface area weighted mean diameter. Calculation of D[3,2] is based on

the assumption that the particles are spherical and possess a smooth surface, which was largely the case

for both drug substances, as evidenced by scanning electron microscopy (data not included).

2.5. Zeta-potential of the nanosuspensions

Zeta-potential was measured using a Zetasizer (Zetasizer Nano ZS, Malvern Instruments,

Worcestershire, UK). The samples were adequately diluted with Milli-Q water to a final drug particle

concentration of 1 mg/ml and introduced into the electrophoretic cell.

2.6. HPC and SDS adsorption onto miconazole and itraconazole nanoparticles

HPC and SDS adsorption onto the miconazole and itraconazole nanoparticles was quantified indirectly

by measuring their depletion from solution upon nanogrinding. Nanosuspensions were centrifuged at

50,000 rpm (300,000 x g) and constant temperature (8 °C) for 3.5 hours (ultracentrifuge Sorvall,

Thermo Fisher Scientific, Waltham, US), and the liquid fractions collected. HPC and SDS were

measured in the supernatant by NIR (MPA FT-NIR, Bruker Optics, Ettlingen, Germany) in

transmittance mode. The method has previously been validated using miconazole nanosuspensions

(Cerdeira et al., 2012). The fraction of polymer bound to the nanoparticles was calculated through mass

balance. The results were presented as quantity of SDS and HPC adsorbed per unit surface area of the

miconazole or itraconazole nanoparticles. The specific surface area was calculated from the particle size

distribution as determined by laser light diffraction (see above).

2.7. Determination of drug substance content of the dried products

Miconazole content of the dried products was determined by dissolving the equivalent of 20 mg of

miconazole in 50 ml of ethanol (400 µg/ml). The solutions were sonicated for 5 min, and aliquots

filtered using 0.2 µm PVDF membrane filters. Miconazole was quantified by UV-spectrophotometry at

272 nm (Varian Cary 50 scan – cell 281 QS 1000, Walnut Creek, CA, US) based on a calibration curve

in the concentration range of 0 - 100 µg/ml. Three samples of each nanosuspension formulation were

used to determine the content.

Chapter V

Page 117 of 182

Itraconazole content of the dried products was determined by dissolving the equivalent of 5 mg of

itraconazole in 20 ml of dimethylformamide (250 µg/ml). Itraconazole was quantified by reversed

phase HPLC using a C18 column (Hypersil BDS-C18®, 100 x 4.6 mm ID, 3 µm particle size) and UV

detection at 230 nm (Agilent HPLC system, Santa Clara, CA, US); a calibration curve in the

concentration range of 0 - 500 µg/ml was used. In the RP-HPLC, the drug was eluted with 0.02 M

tetrabutylammoniumhydrogensulfate in water (mobile phase A), acetonitrile (mobile phase B), and

water (mobile phase C), according to the gradient reported in Table 1. Three samples of each

formulation were used to determine the content.

Table 1: Composition of HPLC eluent and gradient sequence for assaying itraconazole.

Time (min) 0 17 19 25

Phase (%, v/v):

A: 0.02 M Tetrabutylammonium hydrogensulfate 35 35 35 35

B: Acetonitrile 30 58 30 30

C: Water 35 7 35 35

2.8. Determination of discriminative dissolution media

To determine appropriate dissolution media, the highly pH-dependent solubility of non-milled

miconazole and itraconazole was determined in simulated gastric fluid (SGF; pH 1.2; composition: 35

ml of 37% HCl, 10 g NaCl, 0.2% SDS in 500 ml water) and, in case of miconazole, also in 0.05 M

phosphate buffer of pH 4.5 (USP 26). Excess amounts of non-milled miconazole (200 mg) or

itraconazole (200 mg) were added to 20.00 ml SGF or phosphate buffer, and the mixtures were shaken

at 37 °C. After 2 h and 24 h, aliquots were filtered through 0.2 µm PVDF membrane filters. Miconazole

was assayed by UV-spectrophotometry at 272 nm (Varian Cary 50 scan – cell 281 QS 1000, Walnut

Creek, CA, US) using a calibration in the concentration range of 0 µg/ml to 800 µg/ml. Itraconazole

was assayed by UV-spectrophotometry at 225 nm using a calibration in the concentration range of 0

µg/ml to 60 µg/ml. Solubility of both drug substances was determined in triplicate. Non-milled drug

substances were used for solubility determination, because of their easier handling and proximity of

solubility values (15% difference) between non-milled and nanoground substances, as shown recently

by Van Eerdenbrugh et al. (2010).

Chapter V

Page 118 of 182

2.9. Dissolution testing of the dried powders

Dissolution experiments were performed in 500 ml medium at 37 °C using the rotating paddle method

(Pharm. Eur.) with 100 rpm. The dissolution media were identical to those used for solubility

determination, as described above. SGF containing 0.2% (w/v) SDS was used for both drug substances,

whereas the pH 4.5 phosphate buffer was used only for miconazole.

For dissolution testing, amounts of drug product powder equivalent to 25 mg drug substance were

mixed with 5% (w/w) croscarmellose sodium, used as disintegrant, and filled into gelatine capsules.

The capsules were inserted into sinkers to place them at the bottom of the dissolution test vessel.

Samples of 5 ml were taken after 10, 20, 30, 45, and 60 min and filtered through a 0.1 µm PVDF

syringe filter (Millex VV Durapore®, Millipore, Cork, Ireland). The filter pore size was selected

considering the mean nanosuspensions particle size and data from the literature (Juenemann et al.,

2011). The amounts of dissolved drug substances were quantified by reversed phase HPLC, as

described previously for miconazole (Cerdeira et al., 2010) and above for itraconazole. The calculated

amounts of drugs dissolved were corrected for the amounts withdrawn at each sampling point. Each

formulation was tested in triplicate.

2.10. X-ray powder diffraction

The polymorphic structure of untreated (as received) as well as nanoground and subsequently spray-

dried and freeze-dried miconazole and itraconazole powders were assessed by X-ray powder diffraction

measurements. The effect of the presence of mannitol or microcrystalline cellulose in the dried products

was also studied. The powder pattern was recorded in a wide-angle X-ray powder diffractometer (APD

2000, G.N.R., Agrate Conturbia, Italy). For the measurements, samples were filled into aluminium

holders and then exposed to Cu K alpha radiation (40 kV, 30 mA). Counts were monitored for 5 s at

each incremental step of 0.01° 2θ within an angular range from 3°2θ to 40°2θ.

2.11. Statistics

The significance between the mean differences for the different analytical tests was performed using a t-

test (two-sided) for independent samples (p < 0.05, 95% C.I.)

Chapter V

Page 119 of 182

3. Results

3.1. Excipient screening for efficient nanogrinding and stabilization of miconazole and

itraconazole nanoparticles

Miconazole and itraconazole were both milled similarly and most efficiently in presence of HPC and

SDS or HPMC and SDS, yielding D[4,3] values of 200 nm to 300 nm (Table 2). Use of poloxamer

block copolymers instead of the cellulose ethers hampered nanogrinding efficiency, with poloxamer

407 performing better than poloxamer 188 (Table 2). While efficient milling of miconazole required the

presence of both the HPC and SDS, HPC alone was sufficient for itraconazole nanogrinding.

Table 2: Particle size D[4.3] and zeta-potential of miconazole (MIC) and itraconazole (ITR) nanosuspensions

subjected to nanogrinding for 60 min in presence of polymeric excipients and surfactants. The suspensions

contained 20% (w/w) of drug substance, 5% (w/w) of polymer, and 0.05% (w/w) of SDS, unless indicated

otherwise.

Stabilizer type Particle size D[4, 3] nm ± s.d.

ζ ± s.d. n=6 (mV) pH

MIC (20%) ITR (20%) MIC (20%) ITR (20%) MIC (20%) ITR (20%)

HPC 213 ± 1 208 ± 2 -12 ± 1 -22 ± 1 7.7 6.9

HPC1 524 ± 5 214 ± 1 ≈ 0 -20 ± 2 6.0 6.7

HPMC 293 ± 6 297 ± 3 -10 ± 0 -16 ± 0 7.8 7.5

Poloxamer 188 821 ± 130 519 ± 117 n.d n.d 7.7 7.4

Poloxamer 407 439 ± 13 455 ± 5 -23 ± 0 -34 ± 0 7.7 7.6

Hydroxypropylcellulose (HPC); 1Formulations processed without SDS; Hydroxypropylmethylcellulose (HPMC); n.d.: presence of large particles and agglomerates prevent a correct zeta-potential measurement

Based on these results, HPC/SDS and poloxamer 407/SDS were selected for further evaluation. Particle

size distribution of nanogrinded miconazole and itraconazole revealed that the combination HPC/SDS

was more effective than poloxamer 407/SDS in reducing the large-size fraction of drug particles (Fig.

1). This result was confirmed qualitatively by optical microscopy (data not shown).

The zeta-potential of itraconazole was usually higher than the zeta-potential of miconazole when the

same stabilizers and milling conditions were used (Table 2). The biggest difference appeared when the

steric stabilizer HPC was used in absence of SDS; in this situation, miconazole had a surface charge of

approximately 0 mV and itraconazole of -20 mV. With regards to the steric / electrostatic stabilizers

tested, HPC/SDS mediated lower absolute zeta potential values than poloxamer 407/SDS. Due the

formation of agglomerates, the zeta-potential of the poloxamer 188 stabilized nanosuspensions was not

measured.

Chapter V

Page 120 of 182

Figure 1: Particle size distributions of miconazole (A) and itraconazole (B) after 60 min nanogrinding using

HPC/SDS or poloxamer 407/SDS as stabilizers.

0

2

4

6

8

10

12

14

10 100 1,000 10,000

Vo

lum

e (%

)

Particle size (nm)

HPC / SDS

A) Miconazole

Poloxamer 407 / SDS

0

2

4

6

8

10

12

14

10 100 1,000 10,000

Volu

me

(%)

Particle size (nm)

HPC / SDS

B) Itraconazole

Poloxamer 407 / SDS

Chapter V

Page 121 of 182

The pH values (Table 2) of both the miconazole and itraconazole nanosuspensions were similar. With

the miconazole nanosuspension, however, addition of 0.05% SDS increased slightly the pH-value of the

suspension from 6.0 to 7.7, possibly resulting from some solubilized basic miconazole.

In nanogrinding technology, excipients must not only mediate efficient milling, but also stabilize the

produced nanouspension during storage. Both the HPC/SDS and poloxamer 407/SDS provided

adequate particle size stabilization for the miconazole and itraconazole nanosuspensions stored at 5 °C

for three months (Fig. 2). Upon storage at 25 °C, however, poloxamer 407/SDS was less effective for

miconazole and HPC/SDS for itraconazole nanosuspension stabilization (Fig. 2). Yet, taken altogether,

HPC/SDS proved to be the most effective excipient combination for efficient nanogrinding and

stabilizing miconazole and itraconazole nanosuspensions. Hence, further experiments were performed

with HPC/SDS.

Figure 2: Physical stability (change of D[4,3]) of miconazole and itraconazole nanosuspensions stabilized with

HPC/SDS or poloxamer 407/SDS upon storage at 5 ± 3 °C or 25 ± 3 °C for 3 months.

0

500

1000

1500

2000

2500

3000

Part

icle

siz

e D

[4.3

] n

m

Miconazole Itraconazole

Pol 407/SDSHPC/SDS HPC/SDS Pol 407/SDS

Storaget (m) 0 3 3 0 3 3 0 3 3 0 3 3T (°C) - 5 25 - 5 25 - 5 25 - 5 25

Chapter V

Page 122 of 182

3.2. Effect of the amounts of drug substance, HPC, and SDS on miconazole and

itraconazole particle size

The influence of formulation composition on drug particle nanogrinding was studied here for

itraconazole and compared with previously obtained results for miconazole (Cerdeira et al., 2010).

Increasing the drug substance concentration from 5 to 20%, at constant 5% HPC and 0.05% SDS,

enhanced nanogrinding efficiency by yielding smaller particle sizes of both miconazole and

itraconazole (Fig. 3A). Similarly, increasing the HPC concentration (from 1.25 to 5%) facilitated, to a

comparable extent, the particle size reduction of both drug substances (Fig. 3B). By contrast, varying

the SDS concentration from 0.05 to 0.2% exerted only a minor effect on the achievable particle sizes,

although smallest particles (d[4,3] = 207 ± 1 nm for miconazole; d[4,3] = 208±2 nm for itraconazole)

were obtained at 0.05% SDS and significantly (p < 0.05) larger particles at 0.2% SDS (d[4,3] = 248 ± 5

nm for miconazole; d[4,3] = 225 ± 3 nm for itraconazole) (Fig. 3C). The increased miconazole particle

sizes in presence of 0.2% SDS was also reflected by the 50% (d50) and 90% (d90) (Supplementary

information in Appendix AI) undersize diameters (d50. SDS 0.05% = 140 nm versus d50, SDS 0.2% = 152 nm;

d90 SDS, 0.05% = 413 nm versus d90, SDS 0.2% = 595 nm), while the larger itraconazole particle sizes were only

reflected in the d90-value (d50, SDS 0.05% = 136 nm versus d50, SDS 0.2% = 137 nm; d90, SDS 0.05% = 355 nm

versus d90, SDS 0.2% = 382 nm).

100

300

500

700

900

0 5 10 15 20

Part

icle

siz

e D

[4.3

] (n

m)

Drug substance concentration (%, w/w)

Itraconazole Miconazole

HPC: 5% / SDS: 0.05%

A)

Legend common to all plots

Chapter V

Page 123 of 182

Figure 3: Mean particle size, D[4,3], of miconazole and itraconazole upon nanogrinding as a function of drug

substance concentration (A), HPC concentration (B), and SDS concentration (C). Data for miconazole are taken

from a previous study (Cerdeira et al., 2010) and included here for comparison

3.3. HPC and SDS adsorption onto the drug nanoparticles

The amount of HPC (Fig. 4A) and SDS (Fig. 4B) adsorbed onto the drug nanoparticles was determined

as a function of the SDS equilibrium concentration (SDS remaining in solution not adsorbed to the

nanoparticles) in the nanosuspensions, keeping constant the concentration of drug substance (20%) and

HPC (5%). HPC adsorption onto the miconazole and itraconazole nanoparticles did not vary

substantially as a function of SDS concentration, although a very slight tendency towards decreasing

100

300

500

700

900

0 1 2 3 4 5

Pa

rtic

le s

ize

D[4

.3]

(nm

)

HPC concentration (%, w/w)

DS: 20% / SDS: 0.05%

B)

100

300

500

700

900

0 0.05 0.1 0.15 0.2

Part

icle

siz

e D

[4.3

] (n

m)


DS: 20% / HPC: 5%

C)

Chapter V

Page 124 of 182

HPC adsorption at high SDS concentration (0.2%) was noted for both drug substances. Itraconazole

nanoparticles exhibited approximately three times higher adsorption capacity for HPC (2,200 µg/m2)

than miconazole particles (750 µg/m2). SDS adsorption onto the drug nanoparticles increased with

increasing SDS concentration in the nanosuspensions, which mirrors classical physisorption (Fig. 4B).

Oppositely to HPC adsorption, SDS adsorption was similar for both drug substances.

Figure 4: Adsorption of HPC (A) and SDS (B) onto miconazole and itraconazole nanoparticles as a function of

SDS equilibrium concentration (drug substance: 20%, w/w; HPC: 5%, w/w).

0

500

1000

1500

2000

2500

0 0.05 0.1 0.15 0.2

HP

C q

uan

tity

ad

sorb

ed o

nto

n

an

op

art

icle

s (µ

g/m

2)

SDS equilibrium concentration (%, w/w)

Miconazole ItraconazoleA)

0

20

40

60

80

100

120

0 0.05 0.1 0.15 0.2

SD

S q

uan

tity

ad

sorb

ed o

nto

n

an

op

art

icle

s (µ

g/m

2)

SDS equilibrium concentration (%, w/w)

Miconazole ItraconazoleB)

Chapter V

Page 125 of 182

3.4. Spray-dried and freeze-dried drug suspension formulations

3.4.1. Drug substance content

All freeze-dried products contained close to 100% of the expected theoretical drug content, whereas the

spray-dried products showed variable results (Table 3). Spray-drying of the drug nanosuspensions

without and with mannitol as matrix former resulted in the expected drug substance contents. By

contrast, spray-drying of the coarse drug suspensions (not nanoground, no added matrix formers)

yielded significantly lower drug contents than expected, whereas the nanosuspensions spray-dried in

presence of microcrystalline cellulose (MCC) contained significantly more drug substance than

expected. We ascribe the latter findings to particle segregation in the drying column. Due to high

kinetic energy, we observed that larger particles (coarse drug particles or MCC particles) deposited

preferentially in the drying column rather than being carried along with the other formulation

components (nanosized drug substance, HPC, SDS, part of MCC) to the cyclone and collection vessel

of the spray-dryer.

3.4.2. Particle size

Spray-drying and freeze-drying of the different drug formulations generally changed the particle sizes

(Table 4). Spray-drying and freeze-drying of the coarse drug suspensions produced finer drug powders.

With spray-drying, this observation may again be ascribed to the aforementioned segregation occurring

in the drying column of the spray-dryer, where larger particles are more likely deposited due to their

higher kinetic energy; with freeze-drying, larger particles of drug substance must have sedimented and

attached strongly to the bottom of the glass vial during the freezing step (-50 °C, > 2h) resulting in an

inhomogeneous size distribution in the powder cake, which was difficult to remove and homogenise for

particle size analysis.

Drying of the nanosuspensions with and without mannitol as matrix former increased significantly the

particle sizes (Table 4). The nanosuspensions co-spray-dried with MCC presented a large fraction of

MCC particles (size range of 10–30 µm), which was confirmed by microscopy (micrographs not

shown). Interestingly, the aforementioned segregation was noticed once again with the MCC particles,

as larger D[4,3] values were measured before spray-drying (22-30 µm) than after spray-drying (12-15

µm). When freeze-drying in presence of MCC, the measured particle sizes were primarily defined by

the presence of MCC particles with the process having no significant influence.

Chapter V

Page 126 of 182

Table 3: Drug substance content of dried products.

Formulation

Theoretical content

(%, w/w)

Spray-drying Freeze-drying

Actual

miconazole

content

(%, w/w) ± s.d.

(n = 3)

Actual

itraconazole

content

(%, w/w) ± s.d.

(n = 3)

Actual

miconazole

content

(%, w/w) ± s.d.

(n = 3)

Actual

itraconazole

content

(%, w/w) ± s.d.

(n = 3)

Coarse suspension dried

80 62 ± 1 72 ± 3 77 ± 4 88 ± 7

Nanosuspension dried

80 80 ± 2 77 ± 3 81 ± 1 81 ± 2

Nanosuspension dried with 8 g mannitol

44 45 ± 1 44 ± 1 47 ± 0 42 ± 3

Nanosuspension dried with 8 g MCC

44 62 ± 2 56 ± 2 46 ± 2 46 ± 5

Nanosuspension dried with 16 g mannitol

31

Not performed

30 ± 2 30 ± 1

Nanosuspension dried with 16 g MCC

31 32 ± 2 31 ± 1

Table 4: Particle size D[4, 3] of miconazole and intraconazole suspensions before and after spray-drying (SD) and

before and after freeze-drying (FD)

Formulation

Miconazole

particle size

± s.d. (nm) (n=6)

Itraconazole

particle size

± s.d. (nm) (n=6)

Miconazole

particle size

± s.d. (nm) (n=6)

Itraconazole

particle size

± s.d. (nm) (n=6)

Before SD

After SD

Before SD

After SD

Before FD

After FD

Before FD

After FD

Coarse suspension

23,635 ± 422

15,391 ± 192

25,522 ± 240

20,546 ± 407

23,635 ± 422

20,070 ± 354

25,522 ± 240

16,122 ± 217

Nanosuspension 157 ± 1 1,483 ±

71 144 ± 1

4,787 ± 78

182 ± 1 2,039 ±

168 183 ± 2

49,834 ± 6,133

Nanosuspension, with 8 g mannitol

157 ± 1 1,081 ±

30 144 ± 1

1,498 ± 110

182 ± 1 1,988 ±

211 198 ±

10 27,745 ± 1,054

Nanosuspension, with 8 g MCC1

22,280 ± 62

11,499 ± 465

30,512 ± 317

14,905 ± 1,190

22,280 ± 62

22,698 ± 524

30,512 ± 317

33,875 ± 4,372

Nanosuspension, with 16 g mannitol Not performed

182 ± 1 286 ± 5 198 ±

10 11,444 ± 861

Nanosuspension, with 16 g MCC1

23,587 ± 1,910

24,931 ± 808

30,512 ± 317

44,551 ± 2,025

1 particle size results primarily from the presence of MCC-particles

Chapter V

The change in particle size of the nanosuspensions dried in presence or absence of mannitol is more

clearly shown by the particle size distribution plots

range of 1 µm to 10 µm emerged from the resuspended spray

mannitol for spray-drying the nanosuspensions enhanced substantially powder re

was particularly pronounced with itraconazole. Freeze

new particle size population in the range of

itraconazole. Duplicating the amount of mannitol in freeze

particles, which were agglomerates (redispersible) or aggregates (not redispersible).

0

2

4

6

8

10

12

14

10

Volu

me

(%)

Beforespray-drying

A) Miconazole nanosuspension

0

2

4

6

8

10

12

14

10

Vo

lum

e (%

)

Beforespray-drying

B) Itraconazole nanosuspension


clearly shown by the particle size distribution plots (Fig. 5A-D). A new particle size population in the

m emerged from the resuspended spray-dried powders. Incidentally, the use of

drying the nanosuspensions enhanced substantially powder re

articularly pronounced with itraconazole. Freeze-drying of the nanosuspensions also produced a

new particle size population in the range of 1 µm to 10 µm, which extended even up to 100

itraconazole. Duplicating the amount of mannitol in freeze-drying decreased this portion of larger


100 1,000

Particle size (nm)

drying

A) Miconazole nanosuspension

without additive

After spray-drying

with mannitol

100 1,000

Particle size (nm)

B) Itraconazole nanosuspension

with mannitol without additive

After spray-drying

Page 127 of 182


D). A new particle size population in the

dried powders. Incidentally, the use of

drying the nanosuspensions enhanced substantially powder re-dispersibility, which

drying of the nanosuspensions also produced a

, which extended even up to 100 µm with

ng decreased this portion of larger


10,000

without additive

10,000

without additive

Chapter V

Figure 5: Particle size distributions of re

medium was pure water. Miconazole nanosuspensions spray

nanosuspensions spray-dried without or wi

mannitol or with 16 g mannitol (1:1 or 1:2 weight ratio of miconazole : mannitol) (C); itraconazole

nanosuspensions freeze-dried with 8 g mannitol or with 16 g mannitol (1:1 or 1:2 weight ra

mannitol) (D).

3.4.3. X-ray diffraction

The X-ray diffractograms of the spray

(all shifts remained below 0.2°2θ); nonetheless, a slight peak broadening as compared with the original

0

2

4

6

8

10

12

14

10 100

Vo

lum

e (%

)

C ) Miconazole nanosuspension

Beforefreeze-drying

0

2

4

6

8

10

12

14

10 100

Volu

me

(%)

D ) Itraconazole nanosuspension

Beforefreeze-drying

Figure 5: Particle size distributions of re-suspended spray-dried or freeze-dried nanosuspensions; suspension

medium was pure water. Miconazole nanosuspensions spray-dried without or with mannitol (A); itraconazole

dried without or with mannitol (B); miconazole nanosuspensions freeze


dried with 8 g mannitol or with 16 g mannitol (1:1 or 1:2 weight ra

ray diffractograms of the spray-dried and freeze-dried powders did not reveal notable peak shifts

θ); nonetheless, a slight peak broadening as compared with the original

1,000 10,000

Particle size (nm)

C ) Miconazole nanosuspension

After freeze-drying with

16 g of mannitol 8 g of mannitol

1,000 10,000

Particle size (nm)

D ) Itraconazole nanosuspension

After freeze-drying with


Page 128 of 182

dried nanosuspensions; suspension

dried without or with mannitol (A); itraconazole

th mannitol (B); miconazole nanosuspensions freeze-dried with 8 g


dried with 8 g mannitol or with 16 g mannitol (1:1 or 1:2 weight ratio of itraconazole :

dried powders did not reveal notable peak shifts

); nonetheless, a slight peak broadening as compared with the original

100,000


100,000

drying with


Chapter V

Page 129 of 182

drug substance was observed (Supplementary information in Appendix AI). For itraconazole, several

peaks at 4.6°2θ and between 31°2θ and 40°2θ visible with the original substance have mostly

disappeared with the dried products. No differences were observed between the spray-dried and freeze-

dried powders.

3.5. Dissolution behaviour of the spray-dried and freeze-dried products

3.5.1. Spray-dried suspensions

Miconazole dissolution was tested in simulated gastric fluid (SGF) of pH 1.2 supplemented with 0.2%

SDS (miconazole solubility = 650 ± 14 µg/ml) and phosphate buffer of pH 4.5 (miconazole solubility =

40 ± 1 µg/ml) to evaluate the two dissolution media for their discriminative power. In the SGF, which

provided sink condition (maximal amount to be dissolved: 50 µg/ml), the dissolution rates of the

physical mixture (untreated miconazole, SDS and HPC) and spray-dried nanosuspension products were

similar; surprisingly, sole spray-drying of the original miconazole lowered significantly the dissolution

rate (Supplementary information in Appendix AI). Thus, SGF was not a discriminative dissolution

medium for the miconazole formulations. In pH 4.5 phosphate buffer, which did not provide sink

conditions (Fig.6A), a clear difference in the dissolution rate was observed between the different spray-

dried formulations, i.e., the physical mixture (original miconazole, SDS and HPC), the spray-dried

coarse miconazole suspension, and the nanosuspensions co-spray-dried without or with mannitol or

microcrystalline cellulose (Fig. 6A). The spray-dried nanosuspensions with mannitol or MCC exhibited

significantly faster (approximately twofold) dissolution rates as compared with the other formulations.

Dissolution of the spray-dried miconazole nanosuspensions was limited to approximately 70% (35

µg/ml) of the theoretical miconazole concentration (50 µg/ml), which was due to the non-sink condition

in phosphate buffer pH 4.5 (solubility of 40 ± 1 µg/ml).

Chapter V

Page 130 of 182

0

25

50

75

100

0 10 20 30 40 50 60

Mic

on

azo

le d

isso

luti

on

(%

)

Time (min)

Physical mixtureCoarse suspension, spray driedNanosuspension, spray driedNanosuspension, spray-dried with mannitolNanosuspension, spray-dried with MCC

A) Miconazole in pH 4.5 buffer: spray-drying

0

25

50

75

100

0 10 20 30 40 50 60

Itra

con

azo

le d

isso

luti

on

(%

)

Time (min)

Physical mixtureCoarse suspension, spray-driedNanosuspension, spray-driedNanosuspension, spray-dried with mannitolNanosuspension, spray-dried with MCC

B) Itraconazole in pH 1.2 medium with SDS: spray-drying

Chapter V

Page 131 of 182

Figure 6: Dissolution of spray-dried or freeze-dried miconazole nanosuspensions in phosphate buffer of pH 4.5

(A, C) and of spray-dried or freeze-dried itraconazole nanosuspensions in simulated gastric fluid containing 0.2%

SDS (B, D). Physical mixtures of drug substance and matrix formers as well as spray- or freeze-dried coarse drug

suspensions are shown as controls.

0

25

50

75

100

0 10 20 30 40 50 60

Mic

on

azo

le d

isso

luti

on

(%

)

Time (min)

Physical mixtureCoarse suspension, freeze-driedNanosuspension, freeze-driedNanosuspension, freeze-dried with mannitolNanosuspension, freeze-dried with MCC

C) Miconazole in pH 4.5 buffer: freeze-drying

0

25

50

75

100

0 10 20 30 40 50 60

Itra

con

azo

le d

isso

luti

on

(%

)

Time (min)

Physical mixture

Coarse suspension, freeze-dried

Nanosuspension, freeze-dried

Nanosuspension, freeze-dried with 8 g mannitol

Nanosuspension, freeze-dried with 8 g MCC

Nanosuspension, freeze-dried with 16 g MCC

Nanosuspension, freeze-dried with 16 g mannitol

D) Itraconazole in pH 1.2 medium with SDS: freeze-drying

Chapter V

Page 132 of 182

Similar to miconazole dissolution in pH 4.5 phosphate buffer, itraconazole dissolution in simulated

gastric fluid with 0.2% SDS (itraconazole solubility = 163 ± 9 µg/ml) was fastest for the spray-dried

nanosuspensions with mannitol (80% of dose dissolved in 10 min) or microcrystalline cellulose (75% of

dose dissolved in 10 min) (Fig. 6B). The spray-dried nanosuspension without matrix former showed a

similar dissolution rate as did the original itraconazole drug substance and the spray-dried coarse

itraconazole suspension (15-35% of dose dissolved in 10 min; 75% dissolved in 45 min).

3.5.2. Freeze-dried suspensions

The freeze-dried miconazole suspensions showed distinct dissolution profiles in pH 4.5 phosphate

buffer (Fig. 6C). The freeze-dried coarse suspension dissolved slowest (28% after 20 min, 46% after 60

min), the nanosuspension without matrix former a little faster (35% after 20 min, 59% after 60 min),

and the freeze-dried nanosuspensions with the matrix formers mannitol (miconazole/mannitol, 1:1) or

MCC (miconazole/MCC, 1:1) dissolved fastest (52% or 57% after 20 min, 65% or 71% after 60 min,

respectively).

The freeze-dried itraconazole powders with MCC (itraconazole/MCC, 1:1 or 1:2) dissolved, at early

time points (10 and 20 min), slightly faster than the freeze-dried coarse suspension (Fig. 6D); this small

difference disappeared, however, at later time points (beyond 30 min).

4. Discussion

Efficient nanogrinding of very slightly water-soluble or practically water-insoluble drug substances by

stirred media milling requires adequate stabilization of the newly formed drug nanoparticles against

agglomeration and crystal growth (Merisko-Liversidge et al., 2011; Wu et al., 2011). Stabilization by

electrostatic and/or steric effects can be achieved upon adsorption of ionic surfactants, such as SDS, and

non-ionic polymers, such as cellulose ethers or poloxamers, onto the nanoparticles (Rabinow, 2004).

Adsorbed ionic surfactants may primarily increase the particles’ zeta-potential, besides mediating

particle wetting. Thus, adequate particle stabilization by electrostatic and steric effects is a first

requirement for efficient nanogrinding by wet media milling.

In the present study, we showed that the combined use of the anionic surfactant SDS and non-ionic

polymers (HPC or poloxamer 407) mediated efficient nanogrinding and particle stabilization of

miconazole and itraconazole. Although the two very slightly water-soluble and hydrophobic drug

substances (BCS class II) behaved similarly in nanogrinding, during nanosuspensions’ storage, and in

Chapter V

Page 133 of 182

their dissolution from dried nanosuspension formulations, the small differences observed in this study

may be partly related to the physico-chemical properties of the two imidazole derivatives. Relevant

physico-chemical properties include water solubility, pKa, hydrophobicity, and drug crystal surface

roughness, which may all influence crystal growth, molecule’s charge, excipients adsorption, and

impacting on release and stability characteristics (Lee et al., 2008).

Firstly, water solubility of miconazole (~ 1 mg/L) is higher than that of itraconazole (~ 0.1 mg/L)

(Kovács et al., 2009). Higher solubility increases the potential for particle growth during milling and

storage by Ostwald-ripening (Verma et al., 2011). Secondly, the pKa of miconazole is 6.7 (Peeters,

1978), whereas that of itraconazole is 3.7 (Tarsa et al., 2010), which influences the molecules’ surface

charge and, consequently, the zeta-potential, depending on the suspension pH. Thirdly, as described in a

previous work (Cerdeira et al., 2011), scanning electronic microscopy (SEM) micrographs of untreated

miconazole particles showed roundish irregular morphology, while itraconazole particles possessed

elongated rectangular shapes with flat and smooth surfaces. Differences in surface characteristics can

influence the drug substance / polymer interaction (Wang et al., 2012). Last, the surface hydrophobicity

(Van Eerdenbrugh et al., 2008a) and contact angle (Cerdeira et al., 2010) with the dispersion medium

can also influence the milling process.

The differences in zeta-potential and HPC adsorption between miconazole (lower) and itraconazole

(higher) may explain why itraconazole, but not miconazole, could be nanoground efficiently with HPC

in absence of SDS. In absence of SDS, the zeta-potential value of the miconazole nanosuspension was

close to 0 mV, but -19.7 mV with the itraconazole suspension. This difference, in turn, may be

explained by the pKa-values and the polar surface area. The pKa-values of miconazole (=6.7) and

itraconazole (=3.7) indicate that at least some of the miconazole particle surface molecules may be

protonated at pH of 6 to 7, but practically no protonated itraconazole surface molecule will exist at such

pH. In addition, miconazole has a smaller polar surface area (27 Ǻ2) than itraconazole (101 Ǻ2)

(calculated using Advanced Chemistry Development (ACD/Labs) Software V11.02 (©1994-2012

ACD/Labs). Therefore, the low polar surface combined with the presence of at least a small quantity of

positively charged groups on the miconazole particles appears to result in absence of measurable

surface charge, weheras the important polar surface area of itraconazole in combination with absence of

positively charged groups may cause the strongly negative zeta-potential observed with this drug

substance. We would like to stress, however, that the importance of the zeta-potential for

nanosuspension production and storage must be considered with caution. Firstly, the measured zeta

potential may deviate from the true value, as the nanosuspensions were diluted in water for

measurements. Such dilution may alter slightly the surface charge through desorbing some of the

Chapter V

Page 134 of 182

adsorbed stabilizers. Secondly, as shown in a previous study (Cerdeira et al., 2010), satisfactory

miconazole nanosuspensions can be obtained with as little as 0.0125% (w/w) SDS (besides 5% HPC),

which provides a zeta-potential of only +2 mV (pH 6.8). We concluded that additionally to the surface

charge, the reduced contact angle between dispersion medium and drug particles must have facilitated

efficient contact between miconazole and polymer, with the latter forming an adequate layer for steric

hindrance.

Interestingly, more efficient nanogrinding of the drug substances was achieved with HPC/SDS than

with poloxamers/SDS. This finding is in agreement with earlier reports on ibuprofen and siramesin

(Verma et al., 2009; Zimmermann et al., 2009). These authors and others (Rasenack et al., 2003) noted

that the hydrophobicity of excipients and their interaction with hydrophobic particle surfaces promotes

efficient nanogrinding and nanosuspension stability. The authors found that poloxamer adsorbed to a

lesser extent than HPC or HPMC on the studied nanocrystals. With poloxamers, adsorption is mediated

by the hydrophobic PPO segment, while the hydrophilic PEO chains provide steric hindrance against

aggregation (Liu et al., 2010). The lower interaction of poloxamers with particle surfaces in comparison

with HPC or HPMC was explained by the modest hydrophobicity and hydrophobic interaction capacity

of the PPO segment (Rouchotas et al., 2000). This may also explain why poloxamer 407/SDS

performed superior in our study than poloxamer 188/SDS as the first contains a longer

polyoxypropylene block. Other results of our study also confirm the importance of materials

hydrophobicity for particle-excipient interaction. From the zeta-potential values, HPC/SDS (lower zeta-

potential in absolute terms) appeared to be more extensively adsorbed on miconazole and itraconazole

nanoparticles than poloxamer 407/SDS (higher zeta-potential) (Tab. 2). Further, a larger quantity of

HPC adsorbed on the more hydrophobic itraconazole than on the miconazole particles (Fig. 4A). In the

latter case, the zeta-potential values are inconsistent with the HPC-adsorption data, as higher surface

charges were measured for itraconazole than for miconazole (Tab 2). This discrepancy between zeta-

potential and HPC-adsorption suggests that various other parameters affected both the measured zeta-

potential (e.g., dilution medium for measurement; amount and exposed layer of adsorbed excipients)

and amount of adsorbed stabilizers (e.g., polymer-surfactant interaction (Berglund et al., 2003; Clasen

and Kulicke, 2001); particle surface roughness and morphology).

Storage of the nanosuspensions for 3 months at 5 °C did not cause any changes in particle size (Fig. 2).

Upon storage at 25 °C, the miconazole nanosuspension was well stabilized by HPC/SDS, but not by

poloxamer/SDS, whereas the inverse was true for itraconazole (Fig. 2). Although the existing evidence

may not fully explain this data, we may speculate on the following mechanisms. In poloxamer/SDS-

stabilized nanosuspensions, the zeta-potential of itraconazole is higher (-34 mV) than that of

Chapter V

Page 135 of 182

miconazole (-23 mV). In the event that poloxamer/SDS partially desorbs from the particle surface upon

storage at 25 °C, which should be less the case at 5 °C, the higher zeta-potential of itraconazole (-34

mV) may still suffice to protect the particles from agglomeration, which may not be the case for

miconazole with the lower surface charge (Wu et al., 2011). In HPC/SDS-stabilized nanosuspension,

the zeta-potential values are substantially lower (-12 mV with miconazole; -22 mV with itraconazole)

than in poloxamer/SDS stabilized formulations. So, the zeta-potential may not be the cause for the

observed storage stability differences of the two drug substances. More important might be the

difference in the amount of adsorbed HPC, which was much higher for itraconazole. The known

formation of supermolecular structures between HPC and SDS (Berglund et al., 2003; Clasen and

Kulicke, 2001) might have caused changes in the protective steric layer upon storage. As the HPC-layer

on the itraconazole was more important than on the miconazole nanoparticles, formation of such

supermolecular HPC-SDS-structures would have affected primarily the itraconazole particles.

Nanosuspension stability may be further enhanced by transforming liquid formulations into solids, as

applied here by spray-drying and freeze-drying, both without and with matrix formers (mannitol,

microcrystalline cellulose [MCC]). In presence of mannitol, both drying processes yielded satisfactory

products in terms of drug content. Conversely, addition of MCC for drying the nanosuspensions caused

technical issues, as discussed earlier (MCC particle segregation in spray-drying). The observed particle

segregation would, however, become manageable in larger plants of more appropriate geometries and

more flexible atomization equipment.

Nanogrinding and spray-drying or freeze-drying of nanosuspensions may cause changes in crystallinity

and polymorphic forms (Lee, 2003; Kayaert et al., 2012). Drug substances presenting low melting

intervals and different polymorphs, such as miconazole, are particularly prone to such changes. In this

work, the diffractograms of the dried nanosuspensions showed some peak broadening, which can occur

due to smaller particle sizes or stress and strain caused by the milling process (Van Eerdenbrugh et al.,

2008b). In addition to peak broadening, slight loss of cristallinity was also observed, mostly for

itraconazole, where some characteristic peaks disappeared upon drying of the nanosuspensions.

The main purpose of formulating drug nanosuspensions is to enhance the solubility and dissolution rate

of very slightly water-soluble or practically water-insoluble drug substances. Both enhancements result

from the increased particle surface area, improved particle wetting, and partial solubilisation of the drug

substance, with the latter two mechanisms being mediated by surface active and micelle forming

stabilizers (Rasenack et al., 2003; Sinswat et al., 2005). The increased particle surface area may,

however, be partly lost upon storage or drying of nanosuspensions, where agglomeration

Chapter V

Page 136 of 182

(redispersible), aggregation (not redispersible), or crystal growth of primary particles may occur (Wang

et al., 2005; Van Eerdenbrugh et al., 2008c). In the present work, freeze-dried and spray-dried

miconazole nanosuspension products showed higher dissolution rates than the corresponding

itraconazole products, which may be ascribed to the lesser hydrophobicity and observed lower extent of

particle agglomeration of miconazole.

Besides minimizing particle agglomeration, aggregation, and crystal growth, it is also important to

ascertain rapid dispersion of dried nanoparticles upon contact with dissolution media or physiological

fluids. This can be achieved by adding matrix formers to the nanosuspensions prior to drying, which

facilitate powder wetting and hinder intimate contact between nanoparticles upon removal of water

(Kim et al., 2010). In the present work, added mannitol prevented very efficiently the formation of

agglomerates and aggregates, especially in the case of itraconazole nanosuspension. The suitability of

mannitol as matrix former had already been reported by others (Chaubal et al., 2008; Mou et al., 2011).

Interestingly, less agglomerates and aggregates formed when doubling the amount of mannitol for

freeze-drying miconazole and itraconazole nanosuspensions (from drug substance : mannitol ratio of

1:1 to 1:2). The effect of MCC as matrix former on particle size and agglomeration could not be

evaluated, as the particle size of the drug substance was masked by the presence of the larger MCC

particles.

To assess the influence of formulation parameters on drug dissolution rate, it is important to select

discriminative dissolution media. For miconazole, SGF supplemented with 0.2% SDS provided sink-

conditions, but was not discriminative. Thus, dissolution testing of miconazole formulations was done

in pH 4.5 phosphate buffer, which did not provide sink conditions, but was discriminative. Therefore,

due to solubility limit, achieved total amounts dissolved did not reach 100%, but only maximal 70% of

the actual dose. For itraconazole, SGF with 0.2% SDS not only provided near-sink conditions, but was

also discriminative.

Miconazole and itraconazole nanosuspensions co-spray-dried with mannitol or MCC showed enhanced

dissolution as compared to the unmilled drug substances. The total of itraconazole was released from

the co-spray-dried powders within 20 min, in agreement with previous findings (Van Erdenbrugh et al.,

2008b). The slightly slower release of miconazole from the co-spray-dried powders is ascribed to the

very low concentration gradient (limited solubility) in the medium.

Freeze-dried miconazole formulations released at similar rates as did the spray-dried products. On the

contrary, the freeze-dried itraconazole nanosuspension containing the matrix formers mannitol or MCC

Chapter V

Page 137 of 182

dissolved more slowly than the corresponding spray-dried products. The reasons for the observed

differences presently remain elusive. We may speculate that the important agglomeration tendency of

nanoparticulate itraconazole was more favoured by freeze-drying than by spray-drying; further,

laboratory spray-drying inherently produces very fine particles exposing a high contact area, whereas in

freeze-drying can yield collapsed structures are more difficult to disperse.

5. Conclusions

Miconazole and itraconazole particles were efficiently nanoground and stabilized with combinations of

HPC/SDS or poloxamer 407/SDS, which provided steric and electrostatic effects. Nanosuspensions

stabilized with poloxamer/SDS presented, however, a small fraction of large-sized particles (>1000

nm). While SDS adsorbed to similar extent on both drug substances, HPC adsorption was

approximately threefold higher with itraconazole than with the less hydrophobic miconazole. To

preserve nanoparticle stability during storage, the nanosuspensions were transformed into powders by

spray-drying or freeze-drying. For drying, the known matrix formers mannitol or MCC were added to

the nanosuspensions, which prevented nanoparticle agglomeration or aggregation and facilitated

redispersibility and dissolution of the dry product. Although the two drug substances used in this study

had similar properties, the small differences in their aqueous solubility, pKa, zeta potential,

hydrophobicity, and interaction with stabilizers affected the nanogrinding, stability and dissolution of

both compounds. Therefore, our findings provide new and deeper insight into drug substance

characteristics that are critical for selecting appropriate stabilizers both for nanogrinding and subsequent

drying of nanosuspensions. Future work should focus on developing high throughput technologies that

enables to work with very small amounts of drug substances as nanosuspensions are highly desirable

formulations in early drug product development. High throughput technologies should then provide fast

information on optimal qualitative and quantitative compositions to achieve desired nanoparticle sizes

and stability.

Chapter V

Page 138 of 182

References



modification. Langmuir 19, 2705–2713.

Cavatur, R.K., Vemuri, N.M., Pyne, A., Chrzan, Z., Toledo-Velasquez, D., Suryanarayanan, R., 2002.

Crystallization Behavior of Mannitol in Frozen Aqueous Solutions. Pharm. Res. 19, 894–900.

Cerdeira, A.M., Mazzotti, M., Gander, B., 2010. Miconazole nanosuspensions: Influence of formulation

variables on particle size reduction and physical stability. Int. J. Pharm. 396, 210–218.

Cerdeira, AM., Mazzotti, M., Gander, B., 2011. Role of Milling Parameters and Particle Stabilization

on Nanogrinding of Drug Substances of Similar Mechanical Properties. Chem. Eng. Tech. 34, 1427–

1438.

Cerdeira, AM., Werner, I.A., Mazzotti, M., Gander, B., 2012. Simultaneous quantification of polymeric

and surface active stabilizers of nanosuspensions by using near-infrared spectroscopy. Drug Dev.

Ind. Pharm. DOI:10.3109/03639045.2011.650864.

Chan, H-K., Kwok, P.C.L., 2011. Production methods for nanodrug particles using the bottom-up

approach. Adv. Drug Delivery Rev. 63, 406–416.

Chaubal, M.V., Popescu, C., 2008. Conversion of Nanosuspensions into Dry Powders by Spray Drying:

A Case Study. Pharm. Res. 25, 2302–2308.

Choi, J-Y., Yoo, J.Y., Kwak, H-S., Nam, B.U., Lee, J., 2005. Role of polymeric stabilizers for drug


Clasen, C., Kujlicke, W.-M., 2001. Determination of viscoelastic and rheo-optical material functions of

water-soluble cellulose derivatives. Prog. Polym. Sci. 26, 1839–1919.

D'Addio, S.M., Prud'homme, R.K., 2011. Controlling drug nanoparticle formation by rapid

precipitation. Adv. Drug Delivery Rev. 63, 417–426.

Dandagi, P.M, Kaushik, S., Telsang, S., 2011. Enhancement of solubility and dissolution property of

griseofulvin by Nanocrystallization. Int J. Drug Dev. & Res. 3, 180–191.


Controlled Release 160, 418–430.

Hennart, S.L.A., van Hee, P., Drouet, V., Domingues, M.C.,.Wildeboer, W.J, Meesters, G.M.H., 2012.

Characterization and modelling of a sub-micronmilling process limited by agglomeration

phenomena. Chem. Eng. Sci. 71, 484–495.

Kayaert, P., Van den Mooter, G., 2012. Is the amorphous fraction of a dried nanosuspension caused by

milling or by drying? A case study with Naproxen and Cinnarizine. Eur. J. Pharm. Biopharm. 81,

650–656.

Chapter V

Page 139 of 182

Keck, C.M., Müller R.H., 2006. Drug nanocrystals of poorly soluble drugs produced by high pressure

homogenisation. Eur. J. Pharm. Biopharm. 62, 3–16.

Kim, A.I, Akers, M.J., Nail, S.L., 1998. The Physical State of Mannitol after Freeze-Drying: Effects of

Mannitol Concentration, Freezing Rate, and a Noncrystallizing Cosolute. J. Pharm. Sci. 87, 931–

935.

Kim, S., Lee, J., 2010. Effective polymeric dispersants for vacuum, convection and freeze drying of

drug nanosuspensions. Int. J. Pharm. 397, 218–224.

Kovács, K., Stampf, G., Klebovich, I., Antal, I., Ludányi, K., 2009. Aqueous solvent systems for the

solubilization of azole compounds. Eur. J. Pharm. Biopharm. 36, 352–358.

Juenemann D., Jantratid E., Wagner C., Reppas C., Vertzoni M., Dressman J.B., 2011. Biorelevant in

vitro dissolution testing of products containing micronized or nanosized fenofibrate with a view to

predicting plasma profiles. Eur. J. Pharm. Biopharm. 257–264.

Juhnke, M., Märtin, D., Edgar John, E., 2012. Generation of wear during the production of drug

nanosuspensions by wet media milling . Eur. J. Pharm. Biopharm. 81, 214–222.

Lee, J., 2003. Drug nano and microparticles in solid dosage forms. J. Pharm. Sci. 92, 2057–2068.

Lee, J., Lee, S-J, Choi, J.-Y, Yoo, J.Y., Ahn, C-H., 2005. Amphiphilic amino acid copolymers as

stabilizers for the preparation of nanocrystal dispersion. Eur. J. Pharm. Sci. 24, 441–449.

Lee, J., Choi, J.-Y, Park, C.H., 2008. Characteristics of polymers enabling nano-comminution of water

insoluble drugs. Int. J. Pharm. 355, 328–336.

Liu, Y., Suna, C., Haob,Y., Jianga, T., Zhenga, L., Wanga, S., 2010. Mechanism of Dissolution

Enhancement and Bioavailability of Poorly Water Soluble Celecoxib by Preparing Stable

Amorphous Nanoparticles. J. Pharm. Pharm. Sci. 13, 589–606.



Merisko-Liversidge E., Liversidge, G.G., 2011. Nanosizing for oral and parenteral drug delivery: A


Adv. Drug Delivery Rev. 63, 427–440.

Mishra, P.R., Loaye Al Shaal, L.A., Rainer H. Müller, R.H., Keck, C.M., 2009. Production and

characterization of Hesperetin nanosuspensions for dermal delivery. Int. J. Pharm. 371, 182–189.

Mou, D., Chen, H., Wan, J., Xu, H., Yang, X., 2011. Potent dried drug nanosuspensions for oral

bioavailability enhancement of poorly soluble drugs with pH-dependent solubility. Int. J. Pharm.

413, 237– 244.

Müller, R.H., Gohla, S., Keck, C.M., 2011. State of the art of nanocrystals – Special features,

production, nanotoxicology aspects and intracellular delivery. Eur. J. Pharm. Biopharm. 78, 1–9.

Chapter V

Page 140 of 182


strategy. J. Pharm. Pharmacol. 56, 827–840.

Peeters, J., 1978, Determination of ionization constants in mixed aqueous solvents of varying

composition by a single titration. J. Pharm. Sci. 67, 127–129.

Piel, G., Evrard, B., Fillet, M., Llabres, G., Delattre, L., 1998. Development of a non-surfactant

parenteral formulation of miconazole by the use of cyclodextrins. Int. J. Pharm. 169, 15–22.


Rasenack, N., Hartenhauer, H., Müller, B.W., 2003. Microcrystals for dissolution rate enhancement of

poorly water soluble drugs. Int. J. Pharm. 254, 137–145.

Rouchotas C., Cassidy, O.E., G Rowley, G., 2000. Comparison of surface modification and solid

dispersion techniques for drug dissolution. Int. J. Pharm. 195, 1–6.

Saveyn, H., Mermuys, D., Thas, O., Van der Meeren, P., 2002. Determination of the refractive index of

water dispersible granules for use in laser diffraction experiments. Part. Part. Syst. Char. 19, 426–

432.

Shegokar, R., Singh, K.K., 2011. Surface modified nevirapine nanosuspensions for viral reservoir

targeting: In vitro and in vivo evaluation. Int. J. Pharm. 421, 341–352.

Sinswat, P., Gao, X., Yacaman, M.J., Williams, R.O. 3rd, Johnston, K.P., 2005. Stabilizer choice for

rapid dissolving high potency itraconazole particles formed by evaporative precipitation into

aqueous solution. Int. J. Pharm. 302, 113–124.

Tarsa, P.B., Towler, C.S., Woollamb, G., Berghausen, J., 2010. The influence of aqueous content in

small scale salt screening—Improving hit rate for weakly basic, low solubility drugs Eur. J. Pharm.

Sci. 41, 23–30.

Van Eerdenbrugh, B., Froyen, L., Van Humbeeck, J.V., Martens, J. A., Augustijns, P., Van den Mooter,

G., 2008a. Drying of crystalline drug nanosuspensions – The importance of surface hydrophobicity

on dissolution behavior upon redispersion. Eur. J. Pharm. Sci. 35, 127–135.

Van Eerdenbrugh, B. Froyen, L., Van Humbeeck, J.V., Martens, J. A., Augustijns, P., Van den Mooter,

G., 2008c. Alternative matrix formers for nanosuspension solidification: dissolution performance

and X-ray microanalysis as an evaluation tool for powder dispersion. Eur. J. Pharm. Sci. 35, 344–

353.

Van Eerdenbrugh, B., Vercrruysse, S., Martens, J. A., Vermant, Jan., Froyen, L., Van Humbeeck, J.V.,

Van den Mooter, G., Augustijns, P., 2008b. Microcrystalline cellulose, a useful alternative for

sucrose as a matrix former during freeze-drying of drug nanosuspensions – A case study with

itraconazole. Eur. J. Pharm. Biopharm. 70, 590–596.

Chapter V

Page 141 of 182

Van Eerdenbrugh, B., Vermant, J., Martens, J. A., Froyen, L., Van Humbeeck, J.V., Augustijns, P., Van

den Mooter, G., 2009. A screening study of surface stabilization during the production of drug

nanocrystals. J. Pharm. Sci. 98, 2091–2103.

Van Eerdenbrugh, B., Vermants, J., Martens, J.A., Froyen, L., Van Humbeeck, J., Van den Mooter, G.,

Augustijns P., 2010. Solubility increases associated with crystalline drug nanoparticles:

Methodologies and significance. Mol. Pharm. 7, 1858–1870.



Verma, S., Kumara S., Gokhale, R., Burgess, D. J., 2011. Physical stability of nanosuspensions:

Investigation of the role of stabilizers on Ostwald ripening. Int. J. Pharm. 406, 145–152.

Wang, B.H., Zhang, W.B., Zhang, W., Mujumdar, A.S., Huang, L.X., 2005. Progress in drying

technology for nanomaterials. Drying Tech. 23, 7–32.

Wang, G.D., Mallet, F.P, Ricard, F., and Heng, J.Y.Y., 2012. Pharmaceutical nanocrystals. Curr. Opin.

Chem. Eng. 1, 102–107.

Wu, L., Zhang, J., Watanabe, W., 2011. Physical and chemical stability of drug nanoparticles Adv.

Drug Delivery Rev. 63, 456–469.



physicochemical properties. Eur. J. Pharm. Biopharm. 71, 109–116.

Chapter V

Page 142 of 182

Chapter VI

Page 143 of 182

Chapter VI

Conclusions and Outlook

1. Conclusions

Many of the new investigational and marketed drug substances are practically water-insoluble, which

impacts negatively their bioavailability and limits or even prevents their therapeutic use (Patravale et

al., 2004; Mishra et al., 2009; Müller et al., 2011; Merisko-Liversidge et al., 2011; Shegokar et al.,

2011; Gao et al., 2012). Therefore, several approaches have been undertaken to enhance the solubility

and dissolution rate of such drug substances (salt forms, micronization, complexation with

cyclodextrins, and solubilisation by organic solvents, micelles, liposomes or microemulsions). Most of

these formulation types afford, however, low drug loads and raise toxicity concerns related to highly

dosed excipients (Patravale et al., 2004; Liversidge et al., 2011). Nanosuspensions are being used as an

attractive alternative formulation type containing nanosized drug particles with increased dissolution

rate (according to Noyes–Whitney model) and, in some cases, solubility (according to Ostwald–

Freundlich and the Kelvin equations). The increase of drug particle surface area is the main advantage

of nanosuspensions, but also causes major instability problems such as particle agglomeration,

aggregation, and crystal growth, all driven by the increased surface energy of the nanosized particles.

For this reason, it is very important to select appropriately excipients that can stabilize physically the

primary particles during nanogrinding and storage (Rabinow, 2004; Liversidge et al., 2011; Wu et al.,

2011).

Although nanosuspension technology is widely used throughout the pharmaceutical industry,

nanosuspension formulation continues to be based on empirical principles (Liversidge et al., 2003;

Liversidge et al., 2011). The increasing importance of drug nanosuspensions throughout product

lifecycle (from pre-clinical to commercial) calls for a better understanding of the mechanisms of

particle size engineering and stability. For enhancing the efficiency of nanosuspension development it is

crucial to define the most influential factors affecting the critical quality attributes (CQAs) of such

formulations.

In this thesis, we studied the importance of physical-chemical drug substance characteristics as well as

formulation and manufacturing process parameters that affect nanosuspensions CQAs, focusing mainly on

particle size and dissolution. With this objective, we selected three hydrophobic and practically water-

Chapter VI

Page 144 of 182

insoluble drug substances: miconazole (polymorph I), itraconazole and etravirine. The compounds

belong to BCS (Biopharmaceutical Classification System) class II (miconazole and itraconazole both

imidazoles; low solubility and high membrane permeability) and class IV (etravirine anti-viral; low

solubility and low membrane permeability) and are expected, therefore, to benefit from nanosuspension

formulation (Liversidge et al., 2011).

The method of choice for preparing the nanosuspensions was wet media milling (nanogrinding) using a

high energy mill (Labstar® MiniCer) with yttrium stabilized zirconium oxide beads. Throughout the

entire study (Chapter II to V), the mill was operated in circulation mode with a fixed batch size of

300 g, and the product temperature was kept below 34 °C. While such equipment, batch size, and

operational mode are appropriate for pilot-scale and clinical supply, smaller equipment (low energy mill

or high energy mill with smaller volume chamber) and batch sizes would be more appropriate for early

development with ongoing excipient screening. At early development phase, only limited amount of

drug substance is generally available, and numerous experiments are required to define the formulation

design space, which can be optimized at later phase. In the current thesis, a batch size of 300 g was used

for all experiments, because the selected drug substances were available in sufficient quantity.

The used quantity of materials facilitated the control of process parameters and the study of its

influence on the final nanosuspensions quality attributes. Such study is described in Chapter III where

we studied the nanogrinding process parameters that can define the process in stirred media mills (stress

number; stress energy of the grinding beads; specific energy input, with the latter being proportional to

the product between the stress number and stress energy of the grinding beads) using miconazole,

itraconazole, and etravirine. The use of optimally sized grinding beads offered the possibility to

minimize the specific energy input required to achieve minimal particle sizes. A lower energy input was

necessary with the 0.4 mm beads than with the 0.8 mm beads (0.8 mm beads - higher stress energy but

lower number of stress events (stress number) when comparing to the 0.4 mm beads) to obtain

minimally sized drug particles. Thus, beads of 0.4 mm in diameter transferred sufficient energy to break

up the drug substance particles. Smaller beads (e.g., 0.2 mm) could not be tested because of equipment

limitations. Yet, it would be interesting to test whether 0.2 mm beads would still transfer sufficient

stress energy to break up the drug particles to achieve a faster (possibly with less specific energy input)

particle size reduction due to the higher number of stress events (stress number). A minimal energy

input should be beneficial for drug substance and formulation stability (less heat exposure), material

wear (less abrasion), and product contamination (less abrasion). Miconazole and itraconazole not only

possessed similar mechanical properties, but also similar comminution behavior, with minimal mean

particle sizes of 120–130 nm being achievable at an energy input of about 15 MJ kg–1.

Drug substance solubility is a most important characteristic when formulating suspensions, as it affects

not only biopharmaceutical properties, but also processability and stability. For suspensions,

Chapter VI

Page 145 of 182

particularly nanosuspensions, lowest possible solubility is desirable to hinder crystal growth during

nanogrinding and storage. Of importance in this context is the solubility in the actual suspension

medium containing all necessary excipients for stabilizing the nanosuspension rather than in pure water.

For example, miconazole solubility in HPC solution (2.5%) increased from 3 to 95 µg/ml upon addition

of 0.1% of SDS (Chapter II). Evidently, for weakly basic and weakly acidic drug substances, their

solubility also depends largely on their pKa and the pH of the suspension medium; the latter should be

carefully controlled and adjusted to achieve the lowest possible solubility for nanogrinding.

For practically water-insoluble drug substances, the surface hydrophobicity of the dispersed particles in

aqueous media is another important characteristic for nanosuspension formulation. The hydrophobicity

of solid substance can be readily quantified by contact angle measurements using the media of interest.

The hydrophobicity of the three drug substances studied increased in the order: miconazole <

itraconazole < etravirine (Chapter IV and V). A prerequisite for dispersion of hydrophobic particles in

aqueous media is proper wetting of the solid material, which can be achieved by addition of surface

active excipients, e.g., 0.1% SDS. The hydrophobicity of dispersed particles further influences the

adsorption of excipients (and air) and also the processability of the suspension.

Polymorphism and crystallinity are further drug substance characteristics to consider in nanogrinding,

during nanosuspension storage, and further processing such as drying of the suspensions. Miconazole

has three different polymorphs (I: Melting onset / peak = 82.6 °C / 85.2°C; II: Melting onset / peak =

78.9°C / 81.1°C; III: Melting onset / peak = 76.3°C / 78.1°), while itraconazole (165 - 169 °C) and

etravirine (~ 259 °C, decomposition temperature) do not show polymorphs. Despite the differences in

melting points and polymorphism, no major crystallinity changes were observed upon nanogrinding and

nanosuspension drying (Supplementary information in Appendix I and Appendix II). One way to

minimize or avoid such changes is by controlling the nanosuspension temperature during nanogrinding

(Liversidge et al., 2003), as it was done in our work. Although no major crystallinity changes have been

observed shortly after complete processing, this important quality attribute should be monitored over a

longer time scale, because alterations may occur during long-term storage.

Nanosizing increases the surface energy of the particles and the total free energy of the system. To

lower the high energy of the system, the nanoparticles tend to agglomerate and/or aggregate according

to particle–particle interaction theory established by Derjaguin, Landau, Verwey, and Overbeek (DLVO

theory). According to this theory, the fundamental instability of colloidal dispersions is driven by strong

but short-ranged van der Waals attractions between particles, which are countered by the stabilizing

influence of electrostatic repulsions. Such electrostatic repulsion between particles can be modified by

addition of non-ionic and ionic excipients. On the other side, short-ranged van der Waals attraction can

be shielded by steric barriers provided by surface-adsorbed macromolecules (Rabinow, 2004;

Chapter VI

Page 146 of 182

Liversidge et al., 2011). The selection of excipients to stabilize drug nanosuspensions must consider

both technological and safety aspects (safety for specific route of administration, e.g., oral, parenteral).

Among the different types of stabilizers tested (non-ionic polymers: poloxamers [PEO-PPO-PEO],

cellulose ethers, PVP; cationic and anionic surfactants), it was the mixture of hydroxypropylcellulose

(HPC) and sodium dodecyl sulfate (SDS) that stabilized best the three drug substances’

nanosuspensions (Chapter II and Chapter V), followed by the mixture of poloxamer 338 or

poloxamer 407 with SDS (Supplementary information Appendix II). Hydroxypropylcellulose is

more hydrophobic than the poloxamers (Zimmermann et al., 2009 and Rasenack et al., 2003), which

was most likely one of the reasons why this polymer was found to be the most appropriate for the three

hydrophobic drug substances. Amongst the poloxamers, types 338 and 407 (Supplementary

information Appendix II) stabilized better the three drug substances than did poloxamer 188 (Chapter

V). Once again, the higher hydrophobicity of poloxamer 407 (57 PPO units) and poloxamer 338 (44

PPO units) over that of poloxamer 188 (28 PPO units) must have favoured the adsorption onto the

hydrophobic drug particles and their steric stabilization. The importance of the hydrophobicity of

excipients for interaction with hydrophobic drug particles during nanogrinding has already been

previously highlighted (Zimmermann et al., 2009 and Rasenack et al., 2003).

No less important than the type of stabilizing excipients is the concentration of the various ingredients

(stabilizers and drug substance), which have to be optimized to achieve adequate particle size and

stability (Chapter II and V). Higher amounts of drug substance (up to 20%) and HPC (up to 5%)

provided, in the presence of 0.05% (w/w) SDS, more efficient nanogrinding of miconazole and

itraconazole. The maximum concentration of drug substance and polymer was limited by the viscosity

of the suspension. Above an upper critical viscosity, which appeared to be approximately 1000 mPa•s

for miconazole-HPC-SDS, nanogrinding became less effective, which was explained by hindered

movement of the milling beads (Kwade, 1999). For etravirine, the optimal drug concentration tested in

the presence of 5% HPC / 0.05% SDS (when milling time was 60 min) was 12.5% as higher

concentrations produced more agglomerates (Supplementary information Appendix II). An

additional experiment with 5% etravirine (5% etravirine / 5% HPC / 0.05% SDS) where the suspension

was milled for a larger period of time (90 min versus 60 min) confirmed the suitability of

concentrations between 5 and 12.5% of etravirine, as smaller particle size was achieved and no

agglomerates were found (Supplementary information in Appendix II). The particular behaviour of

etravirine during nanogrinding (release of air; high adsorption of HPC; important increase of suspension

viscosity; strong tendency for particle agglomeration) would warrant further studies to elucidate the

physical-chemical causes.

Amongst the three drug substances, only itraconazole did not require the addition of SDS for

nanogrinding. As this drug substance, i.e., the dissolved fraction and surface molecules of particles, is

Chapter VI

Page 147 of 182

largely ionized at the pH of the nanosuspension (pH 7; pKa 3.7), the drug particles showed a relatively

high absolute zeta-potential (~ -20 mV) without added SDS. Conversely, miconazole (pKa 6.7, pH 6)

required the addition of SDS for wetting and provision of surface charge (zeta-potential increased from

0 to -19 mV upon addition of 0.2% SDS) to enable nanogrinding. However, with increasing SDS

concentration (up to 0.2%, w/w), miconazole solubility also increased (from 3 to 86 µg/ml). As a result,

SDS concentrations of > 0.05% caused significant particle growth (Ostwald ripening) of miconazole.

Therefore, a compromise between high zeta-potential and low solubility was required. As noted with

itraconazole, dissolved and particle surface exposed etravirine molecules are ionized in the suspensions

studied (pKa 3.5; pH 7), which is also reflected by the negative zeta-potential (~-14 mV) of the

etravirine particles. Despite its negative zeta potential, etravirine (etravirine concentration 20%)

required addition of SDS for efficient nanogrinding; addition of 0.125% SDS increased the absolute

zeta-potential value from -14 to -26 mV. The minimal required SDS concentration of 0.125% for

etravirine not only afforded efficient nanogrinding, but also prevented a viscosity increase of the

nanosuspension during milling, a phenomenon that had already been observed with miconazole and

itraconazole. Thus, we learned from our observations that the monitoring of nanosuspension viscosity

during nanogrinding is a useful tool to assess proper nanosuspension stabilization, as the viscosity of

inadequately stabilized nanosuspensions invariably increased during the milling.

To obtain more insight into the stabilizing mechanisms of the most efficient excipients SDS and HPC,

we aimed at obtaining quantitative data on their adsorption on the drug nanoparticles. For this, a near-

infrared method was developed to quantify simultaneously and without need for sample dilution the

amount of SDS and HPC adsorbed onto miconazole, itraconazole and etravirine (Chapter IV). While

SDS adsorbed to similar extent to the three drug substances (11, 73 and 75 µg/m2 miconazole; 14, 64

and 101 µg/m2 itraconazole; 14, 80 and 137 µg/m2 etravirine, at 0.05, 0.125 and 0.2% SDS,

respectively), HPC adsorption was lowest with miconazole (700-800 µg/m2), followed by itraconazole

(2100-2300 µg/m2), and etravirine (3100-3500 µg/m2) (etravirine information in Supplementary

information Appendix II). The HPC adsorption differences appear to reflect the differences of

hydrophobicity between the three drug substances. In addition to the excipients adsorption onto the drug

particles, the interaction between HPC and SDS was also found to be important, as we noticed

competitive displacement of adsorbed HPC by increasing SDS concentration above the critical

aggregation concentration (CACSDS-HPC 0.05%) (Chapter IV). Indeed, when increasing the SDS

concentration from 0.05% or 0.125% to 0.2%, the amount of adsorbed HPC decreased as follows: from

769 to 709 µg/m2 miconazole; from 2300 to 2122 µg/m2 itraconazole; from 3506 to 3104 µg/m2

etravirine. Whether or not such HPC desorption can explain the increase of particle size (D[4,3])

observed (miconazole: 213 to 248 nm; itraconazole: 208 to 225 nm; etravirine: 378 to 407 nm), would

require further experiments with other drug substances.

Chapter VI

Page 148 of 182

Storage of optimized miconazole and itraconazole nanosuspensions at 5 °C for up to 6 months caused

only minor changes in particle sizes, whereas storage at 25 °C resulted in particle agglomeration and

crystal growth. Transforming nanosuspensions into powders is one way to increase their stability during

storage; besides, the availability of a solid rather than liquid dosage form for oral administration may

also be attractive to a certain group of patients. On the downside, drying of nanosuspensions can affect

negatively nanoparticle size and redispersibility (Chapter V). Hence, we studied the feasibility of

miconazole and itraconazole nanosuspension spray-drying and freeze-drying, two of the most common

drying methods (spray-drying for oral forms; freeze-drying for parenteral forms); because of safety

concerns, etravirine nanosuspensions were only freeze-dried (Supplementary information Appendix

II). In the absence of any additional excipients for drying, itraconazole particles agglomerated more

importantly than did miconazole, both upon spray-drying and freeze-drying; upon freeze-drying

etravirine and itraconazole agglomerated to similar extent (Supplementary information Appendix II).

Addition of the matrix formers mannitol for drying the nanosuspensions decreased efficiently

agglomerate formation in all cases for miconazole and itraconazole, but not for etravirine where the

dried powders containing mannitol, which inhibited efficiently aggregation in itraconazole, could not be

re-dispersed due to the presence of strong aggregates (Supplementary information Appendix II). To

dry efficiently etravirine without formation of aggregates, further experiments are required using

different ratios etravirine : matrix formers and / or using other matrix formers.

Formation of agglomerates or aggregates upon nanosuspension drying may not only hamper the

applicability of the formulation, e.g., for injection, but also the dissolution kinetics. Therefore, it is of

utmost importance to assess the dissolution of such dry reconstitutable nanosuspensions. To obtain

meaningful dissolution data, we had first to determine discriminative dissolution media for the drug

substances. While simulated gastric fluid (pH of 1.2) provided discriminative power and sink conditions

for itraconazole, a pH 4.5 phosphate buffer had to be used for miconazole to warrant discriminative

power, although the latter medium did not afford sink conditions for miconazole.

The dissolution rate of dried miconazole and itraconazole nanosuspension was enhanced when mannitol

or MCC were added as matrix formers, as compared to the dissolution rate of dried coarse drug

suspensions) (Chapter V). Etravirine dissolution was not performed due to the aggregates observed

after reconstitution and the additional formulation work required for drying this drug substance.

As a final conclusion, the particle size achievable by wet nanogrinding is determined by two opposing

phenomena: (i) particle breakage and (ii) particle-particle interactions potentially leading to particle

agglomeration. The two phenomena define the true and apparent grinding limits (Knieke et al., 2009).

(i) The true grinding limit is determined by the mechanical properties of the drug particles, the

process parameters, and the equipment configuration as studied in Chapter III. This study

demonstrated that the knowledge of particles’ mechanical properties and use of optimized

Chapter VI

Page 149 of 182

milling parameters, while being important, may alone not warrant efficient particle size

reduction, but an appropriate formulation is needed for adequate nanosuspension

stabilization, as only properly stabilized particles can be ground efficiently. However, even

if the process parameters as specific energy input as well as specific energy of the grinding

beads and stress number of the grinding beads are only a part of the nanosuspension

formulation, these need to be optimized and known; if these parameters are known, a

reproducible process can be achieved even during up-scaling to commercial if the same

type of mill configuration is kept.

(ii) The apparent grinding limit is determined by the stabilization of the particles during

milling, which is currently, in our view, the field where further investigation is required due

to the still unknown properties that dictate the selection of stabilizers type and

concentration as well as the concentration of the drug substance itself. In the current work

we found that even small differences between drug substances in solubility, ionization and

hydrophobicity, will influence nanogrinding and further processing as drying and the CQAs

of the nanosuspensions as particle size, stability upon storage (crystal growth or

agglomeration), re-dispersibility and dissolution. A final consideration goes to the

hydrophobicity of the drug substance, where the studies have shown evidence of its

primordial influence on nanogrinding and subsequent drying of the nanosuspensions; it is

currently not clear why the most hydrophobic drug substances which adsorb a higher

amount of stabilizer are the most difficult to stabilize as this is somehow contradictory to

the steric stabilization theory. In steric stabilization, the non-ionic polymeric surfactant

covers the surface of the particles with its hydrophobic chain, and permits a hydrophilic tail

to project into the water. Compression of the polymeric coating, as by the approach of a

similarly coated particle, causes loss of entropy and is therefore unfavorable, which

provides the necessary repulsive barrier between two particles (Rabinow, 2004). Additional

studies are needed to determine how to formulate / stabilize such drug substances

overcoming the fact of being more hydrophobic over the less hydrophobic drug substances.

This would need to be a multivariate study considering both drug substance properties and

formulation qualitative and quantitative composition.

In our work we studied in a systematic way the formulation of nanosuspensions and process parameters

and highlighted the most important factors for their formulation. The current thesis provides a useful

tool, to select and optimize the formulation composition and manufacturing process parameters to

obtain the desired particle size during the different stages of the product life-cycle as this is scaled-up

from pre-clinical and clinical to commercial scale.

Chapter VI

Page 150 of 182

2. Outlook

Pharmaceutical nanosuspensions can be used from early development to clinical phases and, if

successful, as commercial drug products. Each of these phases requires a different formulation and

manufacturing process approach (Fig. 1A and B). In early development, the challenges are the fast

turnover and short timelines, low availability and limited characterization of the drug substance. Drug

product development for clinical and commercial phases has to focus on a variety of issues: up-scaling;

evaluation of the critical process parameters and their design space; definition of the material attributes

that impact strongly the nanosuspensions critical quality attributes (CQAs); and establishment of a

quality control strategy (for process parameters and material attributes) to maintain persistent product

quality.

In early development, there is the need of developing high-throughput technologies enabling

investigations with very small amounts of drug substance (e.g., < 100 mg) and providing information on

the best qualitative and quantitative composition to achieve desired particle sizes and nanosuspension

stability. High-throughput technology should further be combined with a best rational strategy of

formulation development that implements knowledge of critical materials properties. High-throughput

nanogrinding may be put into practice using, e.g., shaking microplates or rotating glass vials both armed

with microbeads. Currently, no commercial high-throughput nanomilling system is available such as

those used for protein formulations (Kamerzell et al., 2011). When developing a high-throughput

screening (HTS) system, a platform approach would be most useful, as used in protein formulation

development (Warne, 2011). Such platform approach would require several components: (i) high-

throughput nanogrinding equipment with temperature control; (ii) online measurement systems for

particle size, zeta potential, and pH; (iii) data acquisition and evaluation system. High-throughput

nanomilling systems would not only spare time and materials in early development phases, but also

provide a broad knowledge base to allow for the development of first-principles models of nanosuspension

formulations and related processes and thereby improve our fundamental understanding of the key

mechanisms governing nanosuspensions. Incidentally, if high-throughput technology is not a prime

requirement, but drug substance availability is limited, a small-scale high energy mill (NanoMill® 0.01;

Elan company), which is apparently suitable to process as little as 100 mg drug substance (Liversidge et

al., 2011) may be a useful

In conclusion, I would like to recommend that future work should strive for developing reliable and

accurate first-principles models of nanosuspensions to promote our fundamental understanding of the

key mechanisms governing nanosuspensions. Such models would be extremely useful in research and

industrial practice. I believe that their development is one of the important next research steps in this

Chapter VI

Page 151 of 182

field, which should be optimally made in collaborative efforts between academia due to the multi-

disciplinarity of this technology.

Chapter VI

Page 152 of 182

A)

Chapter VI

Page 153 of 182

Figure 1: Simplified schematic representation of industrial nanosuspension development process: A)

Overview from pre-clinical to commercial; B) Overview of nanosuspension qualitative and quantitative

compositionselection

DS: drug substance; DP: drug product; GRAS: generally regarded as safe; CQAs: critical quality

attributes; CMAs: critical materials attribute; CPPs: critical process parameters

B)

Chapter VI

Page 154 of 182

References


Controlled Release 160, 418-430.

Juenemann D., Jantratid E., Wagner C., Reppas C., Vertzoni M., Dressman J.B., 2011. Biorelevant in

vitro dissolution testing of products containing micronized or nanosized fenofibrate with a view to

predicting plasma profiles. Eur. J. Pharm. Biopharm. 257-264.

Kamerzell, T. J., Esfandiary, R., Joshi, S. B., Middaugh, C. R., Volkin, D. B., 2011. Protein–excipient

interactions: Mechanisms and biophysical characterization applied to protein formulation

development. Adv. Drug Delivery Rev. 63 1118–1159

Knieke, C., Sommer, M., Peukert, W., 2009. Identifying the apparent and true grinding limit. Powder

Technol. 195, 25-30.

Kwade, A., 1999. Wet comminution in stirred media mills – research and its practical application.

Powder Technol. 105, 14-20.


poorly-water-soluble compounds. Eur. J. Pharm. Sci. 18, 113-120.

Merisko-Liversidge E., Liversidge, G.G., 2011. Nanosizing for oral and parenteral drug delivery: A


Adv. Drug Delivery Rev. 63, 427-440.

Mishra, P. R., Loaye Al Shaal, L.A., Rainer H. Müller, R.H., Cornelia M. Keck, C.M., 2009. Production

and characterization of Hesperetin nanosuspensions for dermal delivery. Int. J. Pharm. 371, 182-189.

Müller, R.H., Gohla, S., Keck, C.M., 2011. State of the art of nanocrystals – Special features,

production, nanotoxicology aspects and intracellular delivery. Eur. J. Pharm. Biopharm. 78, 1-9.


strategy. J. Pharm. Pharmacol. 56, 827-840.

Rabinow, B.E., 2004. Nanosuspensions in drug delivery. Nat. Rev. Drug Discov. 3, 785-796.

Rasenack, N., Hartenhauer, H., Müller, B.W., 2003. Microcrystals for dissolution rate enhancement of

poorly water soluble drugs. Int. J. Pharm. 254, 137-145.

Shegokar, R., Singh, K.K., 2011. Surface modified nevirapine nanosuspensions for viral reservoir

targeting: In vitro and in vivo evaluation. Int. J. Pharm. 421, 341-352.

Van Eerdenbrugh, Froyen, L., Van Humbeeck, J.V., Martens, J. A., Augustijns, P., Van den Mooter, G.,

2008. Drying of crystalline drug nanosuspensions—The importance of surface hydrophobicity on

dissolution behavior upon redispersion. Eur. J. Pharm. Sci. 35, 127-135.

Warne, N.W., 2011. Development of high concentration protein biopharmaceuticals: The use of

platform approaches in formulation development. Eur. J. Pharm. Biopharm. 78, 208-212.

Chapter VI

Page 155 of 182

Wu, L., Zhang, J., Watanabe, W., 2011. Physical and chemical stability of drug nanoparticles Adv.

Drug Delivery Rev. 63, 456-469.



physicochemical properties. Eur. J. Pharm. Biopharm. 71, 109-116.

Chapter VI

Page 156 of 182

Appendix I

Page 157 of 182

Appendix I

Supplementary information for chapter V

Figure A.I-1: Median diameter, d50, and undersize diameter, d90, of miconazole and itraconazole upon

nanogrinding as a function of drug substance concentration (A), HPC concentration (B), and SDS

concentration (C). Data for miconazole are taken from a previous study (Cerdeira et al., 2010) and

included here for comparison. In panel (B), the d90 data point for itraconazole processed in presence of

1.25% HPC is not shown, as its value (12,385 nm) lies far outside of the range of the other data points.

A)

HPC: 5%; SDS 0.05%

Appendix I

Page 158 of 182

B)

drug substance: 20%; SDS 0.05%

Appendix I

Page 159 of 182

C)

drug substance: 20%; HPC 5%

Appendix I

Page 160 of 182

Figure A.I-2: X-ray diffractograms of spray-dried miconazole (A) and itraconazole (B)

nanosuspensions, and of freeze-dried miconazole (C) and itraconazole (D) nanosuspensions, all with

and without matrix formers. Diffractograms of untreated drug powders (as received) and of coarse drug

suspensions are shown for reference

Appendix I

Page 161 of 182

0

500

1,000

1,500

2,000

2,500

3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39

Inte

nsi

ty /

cp

s

2 theta / deg

Spray-driedcoarse suspension

Spray-dried nanosuspension

Untreated drug substance

Spray-dried nanosuspension with 8 g mannitol

Spray-dried nanosuspension with 8 g MCC

A) Miconazole spray-dried

Appendix I

Page 162 of 182

0

500

1,000

1,500

2,000

2,500

3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39

Inte

nsi

ty / c

ps

2 theta / deg

Spray-dried coarse suspension

Spray-dried nanosuspension


Spray-dried nanosuspension with 8 g mannitol

Spray-dried nanosuspension with 8 g MCC

B) Itraconazole spray-dried

Appendix I

Page 163 of 182

0

500

1,000

1,500

2,000

2,500

3,000

3,500

3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39

Inte

nsi

ty /

cp

s

2 theta / deg

Untreated

Freeze-dried coarse suspension

Freeze-dried nanosuspension

Freeze-dried nanosuspension with 8 g mannitol

Freeze-dried nanosuspensionwith 16 g mannitol

Freeze-dried nanosuspension with 8 g MCC


C) Miconazole freeze-dried

Appendix I

Page 164 of 182

0

500

1,000

1,500

2,000

2,500

3,000

3,500

3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39

Inte

nsi

ty / c

ps

2 theta / deg








D) Itraconazole freeze-dried

Appendix I

Page 165 of 182

Figure A.I-3: Dissolution of spray-dried miconazole nanosuspensions in SGF pH 1.2 medium with SDS.

Physical mixtures of drug substance as well as spray-dried coarse drug suspensions are shown as controls

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60

Mic

on

azo

le d

isso

luti

on

(%

)

Time (min)

Physical mixture

Coarse suspension, spray-dried

Nanosuspension, spray-dried

Miconazole in pH 1.2 medium with SDS

Appendix I

Page 166 of 182

Appendix II

Page 167 of 182

Appendix II

Supplementary information for chapter VI

Figure A.II-1: Particle size reduction of miconazole, itraconazole and etravirine after 60 min nanogrinding

with 0.8 mm beads, as a function of drug substance concentration (A), HPC concentration (B), and SDS

concentration (C) (Data for miconazole and itraconazole are taken from chapter V and included here for

comparison – etravirine milled under same conditions).

Comments to A.II-1 A): Etravirine mean particle size d[4,3] not shown for the condition etravirine 25%,

w/w due to the high viscosity of the suspension that prevented nanogrinding.

In the case of etravirine, the lowest mean particle size achieved after 60 min of nanogrinding was obtained

with an etravirine concentration of 12.5%, w/w. Two independent experiments were performed for the

formulation etravirine 12.5%/ HPC 5% /SDS 0.05% and etravirine 20%/ HPC 5% /SDS 0.05% to evaluate

the variability in particle size and viscosity and confirm the differences when comparing to miconazole and

itraconazole. The milling time was 60 min:

- Etravirine 12.5%/ HPC 5% /SDS 0.05%

D[4,3] exp. 1 = 192 ± 2 nm and D[4,3] exp. 2 = 176 ± 3 nm.

100

300

500

700

900

0 5 10 15 20 25Pa

rtic

le s

ize

d [

4.3

] (n

m)

Drug substance concentration (%, w/w)

Miconazole

Itraconazole

Etravirine

Legend applies to all plots

HPC 5% / SDS 0.05%

A)

Appendix II

Page 168 of 182

Viscosity exp.1 t0 min = 307 mPa•s; t60 min = 411 mPas and Viscosity exp. 2 t0 min = 308 mPa•s; t60 min =

445 mPa•s.

- Etravirine 20%/ HPC 5% /SDS 0.05%

D[4,3] exp. 1 = 578 ± 53 nm and D[4,3] exp. 2 = 661 ± 120 nm.

Viscosity exp.1 t0 min = 512 mPa•s; t60 min = 744 mPas and Viscosity exp. 2 t0 min = 533 mPa•s; t60 min =

627 mPa•s.

As the increase in viscosity during nanogrinding had previously been found an indication of agglomerate

formation which was confirmed by microscopic observation, the results seemed to indicate that the increase

in etravirine concentration was causing a higher degree of agglomeration. To confirm this observation, an

additional nanogrinding experiment was performed with the formulation containing 5% etravirine to

evaluate wheater increasing the nanogrinding time would result in a decrease in particle size.

The following are the results of particle size and viscosity after 90 min of nanogrinding:

- Etravirine 5% / HPC 5% /SDS 0.05%

PSD = 440 ± 5 nm.

- Viscosity t0 min = 188 mPa•s; t60 min = 154 mPa•s.

The results indicate that for etravirine there is an optimal concentration between 5 and 12.5% of drug

substance where nanogrinding is more efficient and no agglomeration is observed. This optimal drug

substance concentration would need to be determined by further experiments.

Comments to A.II-1 B): Etravirine mean particle size d[4,3] not shown for the condition HPC 1.25% and

3.125%, w/w because after 60 min of nanogrinding the particle size was higher than 1,000 nm.

100

300

500

700

900

0 1 2 3 4 5Pa

rtic

le s

ize

d [

4.3

] (n

m)

HPC concentration (%, w/w)

DS 20% / SDS 0.05%

B)

Appendix II

Page 169 of 182

Comments to A.II-1 C): Etravirine mean particle size d[4,3] () was the lowest when the SDS concentration

was 0.125%, w/w.

Figure A.II-2: Particle size distributions of miconazole (A), itraconazole (B) and etravirine (C) after 60 min

nanogrinding with 0.8 mm, using 5% HPC/0.05% SDS, 5% poloxamer 407/0.05% SDS or 5% poloxamer

338/0.05% SDS as stabilizers (data of miconazole and itraconazole stabilized with HPC/SDS and

poloxamer 407/SDS are include in Chapter V and are presented here for comparison purposes with

etravirine).

100

300

500

700

900

0.00 0.05 0.10 0.15 0.20Pa

rtic

le s

ize

d [

4, 3

] (n

m)


HPC 5% / DS 20%

C)

Appendix II

Page 170 of 182

0

2

4

6

8

10

12

10 100 1,000 10,000

Vo

lum

e (%

)

Particle size (nm)

Poloxamer 338 / SDS

HPC / SDS

Poloxamer 407 / SDS

A) Miconazole nanosuspensions

0

2

4

6

8

10

12

10 100 1,000 10,000

Vo

lum

e (%

)

Particle size (nm)

Poloxamer 338 / SDS

HPC / SDS

Poloxamer 407 / SDS

B) Itraconazole nanosuspensions

Appendix II

Page 171 of 182

Comments to A.II-2: The combination between HPC and SDS resulted in the lowest mean particle size

D[4,3] for the three drug substances (miconazole = 213 ± 1 nm; itraconazole = 208 ± 2 nm; etravirine = 464

± 8 nm). The combination poloxamer 407/SDS and poloxamer 338/SDS resulted in similar particle size

distributions for miconazole (439 ± 13 nm and 441 ± 18 nm, respectively) and itraconazole (455 ± 5 nm

and 471 ± 10 nm, respectively), but not for etravirine (2661 ± 138 nm and 626 ± 69 nm, respectively). The

results for etravirine mean particle size when nanogrinded with poloxamer 407/SDS and poloxamer

338/SDS need to be carefully evaluated, due to the presence of agglomerates and variation between

measurements and are presented just for indication.

0

2

4

6

8

10

12

10 100 1,000 10,000

Vo

lum

e (%

)

Particle size (nm)

Poloxamer 338 / SDS

HPC / SDS

Poloxamer 407 / SDS

C) Etravirine nanosuspensions

Appendix II

Page 172 of 182

Figure A.II-3: Adsorption of HPC (A) and SDS (B) onto miconazole, itraconazole and etravirine

nanoparticles as a function of SDS concentration in the suspension (drug substance: 20%, w/w; HPC: 5%,

w/w) (data of miconazole and itraconazole are include in Chapter V and are presented here for comparison

purposes with etravirine) .

0

500

1000

1500

2000

2500

3000

3500

4000

0 0.05 0.1 0.15 0.2

HP

C q

ua

nti

ty a

dso

rbed

o

nto

na

no

pa

rtic

les

(µg

/m2)

SDS concentration in suspensions (%, w/w)

Miconazole

Itraconazole

Etravirine

A)

0

20

40

60

80

100

120

140

160

0 0.05 0.1 0.15 0.2SD

S q

ua

nti

ty a

dso

rbed

on

to

na

no

pa

rtic

les

(µg

/m2)

SDS concentration in suspensions (%, w/w)

Miconazole

Itraconazole

Etravirine

B)

Appendix II

Page 173 of 182

Figure A.II-4: X-ray diffractograms of freeze-dried etravirine nanosuspensions, with and without matrix

formers. Diffractograms of untreated drug powder (as received) and of coarse drug suspension are shown

for reference

Comments to A.II-4: Similarly to the X-ray diffractograms of the freeze-dried miconazole and itraconazole

powders, etravirine revealed a slight peak broadening as compared with the original drug substance.

Appendix II

Page 174 of 182

0

500

1,000

1,500

2,000

2,500

3,000

3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39

Inte

nsi

ty /

cp

s

2 theta / deg

Untreateddrug substance




Freeze-dried nanosuspensionwith 16 g mannitol

Freeze-driednanosuspensionwith 8 g MCC

Etravirine freeze-dried

Appendix II

Page 175 of 182

Figure A.II-5: Particle size distributions of re-suspended etravirine freeze-dried nanosuspensions;

suspension medium was pure water. Etravirine nanosuspensions freeze-dried with 8 g mannitol or with

16 g mannitol (1:1 or 1:2 weight ratio of etravirine : mannitol).

Comments to A.II-5: Freeze-drying etravirine resulted in irreversible aggregation of the nanoparticles

even with the addition of mannitol in the ratio etravirine : mannitol = 1 : 2, w/w, which had prevented

major agglomeration in the case of itraconazole. Etravirine was the most difficult substance to dry.

Further studies with different ratios etravirine : mannitol or other matrix formers would need to be

experimented to evaluate the possibility of drying etravirine. No dissolution results were performed due

to aggregation.

0

2

4

6

8

10

12

14

10 100 1,000 10,000 100,000

Vo

lum

e (%

)

Particle size (nm)

Beforefreeze-drying

After freeze-drying with16 g mannitol 8 g mannitol

Appendix II

Page 176 of 182

Curriculum Vitae

PERSONAL RECORD

Name: Ana Maria Mendes Cerdeira

Place / date of birth: Lisbon, 9th October 1971

Citizenship: Portuguese

Civil status: Married

Address: Emil Rütti-Weg 4, 8050 Zurich, Switzerland

PROFILE

Proven project management experience (experienced leader of local, international and virtual teams and certified black belt).

Early and late phase development and manufacturing experience, with special expertise in

nanosuspensions, solid dosage forms, enabling technologies (pellets, nanoparticles, solid dispersions,

cyclodextrines) and biologics.

EDUCATION

Higher Education

In Progress Doctorate Degree Pharmaceutical Sciences (PhD), Eidgenössische Technische Hochschule Zürich (ETH Zürich), Switzerland – Production and Stabilization of

nanocrystals® of poorly soluble drug substances (expected October 2012)

1994 Degree in Pharmaceutical Sciences (Licenciatura), University of Lisbon, Faculty of Pharmacy, Portugal

Qualifications

2009 Certified Black Belt in Design Excellence, Johnson & Johnson

1995 Qualified Pharmacist, by the Portuguese Pharmaceutical Society (Ordem dos Farmacêuticos), Portugal

Post-graduation Seminars/Courses

2011 SOFIA Talent Program, Johnson & Johnson, Switzerland

2010 Bioprocess Training, GE and Johnson & Johnson, United States

2009 Flawless Project Execution, Project Management, Johnson & Johnson, Belgium

2008 Quality Management and logistics, Master Industrial Pharmaceutical Sciences, Eidgenössische Technische Hochschule Zürich (ETH Zürich), Switzerland

2004 Mastering Change through Project Management, Lindsay McKenna Limited, Reading, United Kingdom

PRESENT POSITION

2010 Cilag (Johnson & Johnson), Schaffhausen, Switzerland

Drug Product Technical Integrator (Associate Director) in the Pharmaceutical Development and Manufacturing Sciences Group

• Project leader of international multi-disciplinary teams, accountable for the development, technical transfer, process validation and launch support of combination products for biologics

- Management of stakeholders and internal and external team interfaces

- Development of balanced, aligned and motivated teams to achieve common project goals, including investigation teams in crisis management

- Provide team members support to achieve their project objectives

- Provide team members visibility and opportunities for development

- Conflict management between team members and interface groups

- Implementation of new processes for better team performance (development of project planning and execution tools)

- Research and implementation of new technologies enabling project objectives

- Participation in Alliance Projects and partnership for the joint development of medicinal products

- Accountable for the timely supply of the Drug Product to be used in clinical studies and launch

- Accountable for the writing the Drug Product sections of the IND / IMPD and BLA

• Engagement leader participating in the integration of local and international teams, support people’s or groups in problem solving and escalation of issues that need management’s support by participating in pulse check team

• Subject matter expert and management of regulatory questions for marketed products

PROFESSIONAL EXPERIENCE RECORD

2008/10 Cilag (Johnson & Johnson), Schaffhausen, Switzerland

Manager of the Formulation Science & Technology Centre of the Global Technical Services group

• Management of the formulation group, planning and budget

• Organization of the execution of the pilot-plant decommissioning

• Development of line extensions for the Emerging Markets

- Project coordination and alignment between Marketing, New Business Development groups and Development, Regulatory and Clinical groups for business case development and project resource assignment

- Drive line extension projects in Latin-America and Asia-Pacific and facilitation of knowledge transfer between the different groups regions / countries and manufacturing sites

2004/08 Cilag (Johnson & Johnson), Schaffhausen, Switzerland

Group Leader, Principal Scientist in the Marketed Product Support, Formulation Science & Technology Centre

• Planning of the formulation group

• Group leader responsible for the coaching and development, as well as resource management of assigned team members

• Project Manager of international teams working in root-cause analysis of technical manufacturing problems and implementation of optimised processes in Production environment using Process Excellence methodology and tools

2003/04 Janssen Cilag (Johnson & Johnson), Queluz de Baixo, Portugal Project Manager, in the European Brand Support, Project Management Group

• Management of international project teams

- Product transfers between Janssen European Plants

- Regulatory Affairs projects

- Analytical method validation for Cleaning Validation and Drug Product analysis

- Formulation re-development

- Outsourcing projects

1997/03 Labor Qualitas (Tecnimede, S.A. Group), Portugal

Group Leader, Study Director according to the GLP Principles of OECD, in the Pharmaceutical Development Department

• Management of a development team, including coordination with regulatory and clinical as well as third party groups and manufacturing sites

- Development of line extensions / technical processes and analytical method development and validation, as well as production of batches for clinical trials and scale-up for new product introduction (solids for immediate and sustained release, semi-solids, liquids and sterile products including lyophilisation)

• Management, coaching and people development

1995/97 Faculty of Pharmacy, University of Lisbon, Portugal

Post-graduate research student, in a R&D project funded by the European Union (PEDIP II), Portugal

• Development of controlled release formulations and analytical methods

- Spray drying and solvent evaporation for microparticles, extrusion / spheronization, bottom spray bead coating, and layering for pellets, droplet extrusion/precipitation of polymers for granules and cyclodextrin inclusion complexes for solubility improvement

1992/94 Laboratório Militar de Produtos Químicos e Farmacêuticos (National Military

Laboratory), Lisbon, Portugal

Research Student

• Part-time collaboration in the project: Syntheses and Spectrophotometric Characterisation of Bile Acids Fluorescent Derivatives

LANGUAGE KNOWLEDGE

Portuguese Mother Language

English Fluent

French Good knowledge

GermanGood knowledge

Spanish Knowledge

LIST OF PUBLICATIONS

Posters and oral presentations

1. Cerdeira, A.M., Goucha, P. and Almeida, A.J. (1996). Parameters affecting the formulation of NSAIDs-containing gastroresistant beads. 3

rd European Congress of Pharmaceutical Sciences

(Edinburgh, UK). Eur. J. Pharm. Sci. 4 (Suppl.): S157 (poster).

2. Cerdeira, A.M., Goucha, P. and Almeida, A.J. (1996). Mini-grânulos gastro-resistentes contendo AINEs produzidos por precipitação: I. Parâmetros de produção. Proceedings of Congresso

Nacional de Farmacêuticos ‘96/3º Congresso Mundial de Farmacêuticos de Língua Portuguesa (Lisboa, Portugal), pp. 110 (poster).

3. Cerdeira, A.M., Goucha, P. and Almeida, A.J. (1996). Mini-grânulos gastro-resistentes contendo AINEs produzidos por precipitação: II. Caracterização. Proceedings of Congresso Nacional de

Farmacêuticos ‘96/3º Congresso Mundial de Farmacêuticos de Língua Portuguesa (Lisboa, Portugal), pp. 113 (poster).

4. Cerdeira, A.M., Goucha, P. and Almeida, A.J. (1996). Parâmetros que afectam a formulação de grânulos gastro-resistentes contendo AINEs. 1st

Scientific Meeting of the Faculty of Pharmacy (oral presentation).

5. Cerdeira, A.M., Goucha, P. and Almeida, A.J. (1997). In vitro release studies on enteric granular formulations based on HP50. Proceed. II Congreso Hispano-Luso de Liberación Controlada de

Medicamento (Tenerife, Spain), pp. 119-120 (poster).

6. Cerdeira, A.M., Gouveia, L.F., Goucha, P. and Almeida, A.J. (1997). Optimisation of the formulation of enteric beads prepared by the HP50 precipitation method. Proceed. II Congreso

Hispano-Luso de Liberación Controlada de Medicamentos (Tenerife, Spain), pp. 117-118 (poster).

7. Gouveia, L.F., Cerdeira, A.M., Almeida, A.J. and Morais, J.A.G. (1997). Determination of dissolution profiles of etodolac in enteric granules by flow injection analysis. Proceed. II Congreso

Hispano-Luso de Liberación Controlada de Medicamentos (Tenerife, Spain), pp. 41-42 (poster).

8. Cerdeira, A.M., Brandão, A., Goucha, P., Cabral Marques, H.M. and Almeida, A.J. (1998).

Preparation and characterisation of beads containing a nabumetone:dimethyl-β-cyclodextrin complex. Proceed. 17

th Pharmaceutical Technology Conference (Dublin, Ireland), Vol. I pp. 42-51

(oral presentation).

9. Cerdeira, A.M., Goucha, P. and Almeida, A.J. (1999). Crystal size and habit influence on drug release from enteric beads produced by droplet extrusion/precipitation. Proceed. Intern. Symp.

Control. Rel. Bioact. Mater. 26: 1022-1023 (Boston, USA) (poster).

10. Almeida, A.J, Cerdeira A.M. and Goucha, P. (2000) Development of a controlled release particulate formulation for oral delivery of NSAIDs. Proceed. IV Congreso Hispano-Luso de Liberación

Controlada de Medicamentos (Vitoria, Espanha), pp.45-46 (poster).

11. Cerdeira, A.M., Amaral, P. and Girão, R. (2004) Introduction and transfer of new products: Methodology and risk management. Pharmaceutical Society of Industrial Pharmacy, Risk Management (Ofir, Portugal) (oral presentation)

12. Cerdeira, A.M., Gander, B. and Mazzotti, M. (2009) Nanosuspensions of miconazole: Influence of Composition on Particle Size Reduction and Stability Proceed. Controlled release Society. 926 (Copenhagen, Dennark) (poster).

13. Cerdeira, A.M., Gander, B. and Mazzotti, M. (2009) Particle Size Reduction by Nanomilling Correlates with Polymer and Surfactant Adsorption on Nanoparticles. Proceed. Controlled release

Society. 938 (Copenhagen, Dennark) (poster).

14. Cerdeira, A.M., Mammo, T., Henry, K., Labrenz, S. (2012) Identification of Critical Quality Attributes Through Design of Experiments (DOE) 3rd

Annual Protein Formulation Development &

Drug Delivery Forum (Barcelona, Spain) (Invited speaker)

Papers

1. Cerdeira, A.M., Goucha, P. and Almeida, A.J. (1998). Hydroxypropyl methylcellulose phthalate beads containing a model non-steroid anti-inflammatory drug. Int. J. Pharm. 164, 147-154.

2. Cerdeira, A.M., Gouveia, L.F., Goucha, P. and Almeida, A.J. (2000). Drug particle size influence on enteric beads produced by a droplet extrusion/precipitation method. J. Microencapsulation. 17, 733-741.

3. Cerdeira, A.M., Mazzotti, M., Gander, B. (2010). Miconazole nanosuspensions: Influence of formulation variables on particle size reduction and physical stability. Int. J. Pharm. 396, 210-218.

4. Cerdeira, A.M., Mazzotti, M., Gander, B. (2011). Role of Milling Parameters and Particle Stabilization on Nanogrinding of Drug Substances of Similar Mechanical Properties. Chem. Eng. Tech. 34, 1427-1438

5. Cerdeira, A.M., Werner, I.A., Mazzotti, M., Gander, B. (2012). Simultaneous quantification of polymeric and surface active stabilizers of nanosuspensions by using near-infrared spectroscopy. Drug Dev. Ind. Pharm. doi:10.3109/03639045.2011.650864

6. Cerdeira, A.M., Mazzotti, M., Gander, B. Formulation and drying of miconazole and itraconazole nanosuspensions. Int. J. Pharm. (in publication)

Patents

1. Goucha, P., Cerdeira, A.M. and Almeida, A.J. (1999). Process for formulating granular matrices by precipitation of polymers in acidic medium EP1018335

2. Goucha, P., Cerdeira, A.M. and Almeida, A.J. (1999). Apparatus for the preparation of spherical polymer particles EP1018364

3. Cerdeira, A.M. and Goucha, P. (1999). Inclusion aminoacid compounds of benzimidazole derivatives with cyclodextrins, their preparation and pharmaceutical formulations containing them EP1018340

4. Cerdeira, A.M. and Goucha, P. (2001). Prolonged release matricial formulations of pirlindol hydrochloride EP1300140

Zurich, 2012

Ana Cerdeira

Rights / License: Research Collection In Copyright - Non ...6563/eth... · nanosuspension),...

Documents

Transcript of Rights / License: Research Collection In Copyright - Non ...6563/eth... · nanosuspension),...