Effect of polymers in moisture sorption and physical stability of polymorphic olanzapine

Accepted Manuscript Effect of polymers in moisture sorption and physical stability of polymorphic olanzapine

Maria C. Paisana, Martin A. Wahl, F. João PII: DOI: Reference:

S0928-0987(16)30516-4 doi: 10.1016/j.ejps.2016.11.023 PHASCI 3809

To appear in:

European Journal of Pharmaceutical Sciences

Received date: Revised date: Accepted date:

26 July 2016 31 October 2016 25 November 2016

Please cite this article as: Maria C. Paisana, Martin A. Wahl, F. João , Effect of polymers in moisture sorption and physical stability of polymorphic olanzapine. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Phasci(2016), doi: 10.1016/j.ejps.2016.11.023

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Effect of polymers in moisture sorption and physical stability of polymorphic olanzapine

iMed.ULisboa – Research Institute for Medicines and Pharmaceutical Sciences,

RI

1

PT

Maria C. Paisana1, Martin A. Wahl2, João F. Pinto1

SC

Faculdade de Farmácia, Universidade de Lisboa, Av. Prof. Gama Pinto, P-1649-003

2

NU

Lisboa, Portugal.

Pharmazeutisches Institut, Eberhard Karls Universität Tübingen, Auf der

CE

PT E

D

MA

Morgenstelle 8, D-72076 Tübingen, Germany.

AC

* Corresponding author

Departamento de Farmácia Galénica e Tecnologia Farmacêutica Faculdade de Farmácia da Universidade de Lisboa Av. Prof. Gama Pinto, P-1649-003 Lisboa, Portugal E-mail address: [email protected] Phone / Fax: +351-217946434 1

ACCEPTED MANUSCRIPT Abstract

The study aims to elucidate the transformations of anhydrous olanzapine Form I (OLZ FI) into the hydrate forms, when stored at a high relative humidity or suspended in an aqueous media,

PT

in the presence of polymers. OLZ FI and physical mixtures (3:1 and 1:1, as powders or compacts) of olanzapine with polyethylene glycol (PEG-6000), polyvinylpyrrolidone (PVP K25)

RI

and hydroxypropylcellulose (HPC-LF) were stored (75%RH/25⁰C, 75%RH/40⁰C and 93%RH/25⁰C)

SC

for 28 days. OLZ FI and the physical mixtures were also suspended in water under stirring (200rpm/60min). Samples were collected at different time points and vacuum filtered. OLZ FI

NU

showed to hydrate at 75%RH/25⁰C when stored in the presence of HPC and PEG. At 93%RH all

MA

polymers affected the kinetics of hydration of OLZ FI with PVP as the only polymer with the ability to minimize the formation of the hydrate. When olanzapine was suspended in water

D

with HPC and PVP the formation of the hydrate was inhibited. Compaction of the powders

PT E

before storage led to an increase of the hydrate conversion rate of olanzapine on the first week of storage, due to a partial amorphisation of olanzapine present at the tablet surface. When

CE

stored at high humidity environments OLZ FI converted into dihydrate D and, when exposed to aqueous environments in the presence of different polymers converted into dihydrates B and E.

AC

From an industrial point of view, this study highlighted the importance of the excipient’s choice for OLZ formulations, so that a final OLZ medicine can have a consistent quality and performance throughout the entire medicine’s shelf life.

Keywords: hydroxypropylcellulose; olanzapine; polyethyleneglycol; polymorphism; polyvinylpyrrolidone; powder; tablet

2

ACCEPTED MANUSCRIPT 1. Introduction A pharmaceutical solid dosage form is a combination of an active pharmaceutical ingredient (API) and excipients which are added to the API to promote the manufacturability of the formulation into a dosage form which

PT

should be easy to handle by patients and maximize the bioavailability of the API.

RI

Furthermore assurance of both physical and chemical, together with

SC

microbiological, stabilities of the final product is paramount to the usefulness of

NU

the dosage form (Campisi et al., 1998). During manufacture and subsequent storage of the solid dosage form both API(s) and excipients can take up moisture

MA

which is capable of altering the solid dosage form’s stability and likely the physical

D

and chemical stabilities of the API(s) (Heidemann and Jarosz, 1991). For example,

PT E

the physical instability of an API in a tablet might be reflected by a polymorphic or pseudopolymorphic transition resulting in a change of the APIs properties (e.g.

CE

flowability, compressibility or rate of dissolution in water), which impact directly

AC

or indirectly on the bioavailability of the drug in the final dosage form (Zhang et al., 2004, Malaj et al., 2010). Pseudopolymorphic transitions are triggered by different causes, being environmental moisture an important cause for such transition. Therefore, it is important to know the different moisture sorption properties of both API and excipients in a certain formulation for an adequate selection of the latter to be considered in the formulation and for the 3

ACCEPTED MANUSCRIPT manufacturing process to be followed. Only optimized formulation and process allow the assurance of the dosage form’s stability and an accurate prediction of the medicine’s shelf-life (Marsac et al., 2008). The interaction of a crystalline material with water may mainly occurs by four different mechanisms: adsorption

PT

of water on the surface of the particles, incorporation into microporous regions of

RI

the crystalline lattice by capillary condensation, formation of a crystal hydrate and

SC

deliquescence (Marsac et al., 2008, Ahlneck and Zografi, 1990, Airaksinen, et al.,

NU

2005). Overall, the crystalline structure, solubility in water, porous structure and the ability to form crystal hydrates are going to determine the mechanism of

MA

water sorption by the solid material (Airaksinen, et al., 2005). On the other hand,

D

the water vapor, is absorbed by amorphous materials in their amorphous regions

PT E

and not simply adsorbed on their surfaces: in these amorphous solids the amount of water uptake is not directly related to the specific surface area of the solid

CE

(Zografi, 1988). The amount of water absorbed by amorphous solids is commonly

AC

much larger than the one absorbed by non-hydrated crystalline materials below their critical relative humidity (RH0) (Rumondor et al., 2009). Furthermore, water acts as a plasticizer, increases the molecular mobility of the solid lattice due to the breakage of hydrogen bonds between molecules thus, promoting the absorption of more water (Zhang and Zografi, 2001, Kontny and Zografi, 1995).

4

ACCEPTED MANUSCRIPT Olanzapine (OLZ) was the model drug selected for this study because of both clinical relevance and ability to convert into different polymorphic and pseudopolymorphic forms (Reutzel-Edens et al., 2003). Olanzapine is an antipsychotic drug required for long term therapies in patients with poor

PT

compliance. It follows that small variations on the solubility of olanzapine due to

RI

its polymorphism may delay the fraction of the drug in a tablet dissolved in the

SC

gastro-intestinal tract, thus delaying its absorption with a negative impact on the

NU

patients’ behavior and compliance. Recent case studies (Galarneau, 2013, Samuel et al., 2013) point out problems with some generic formulations, such as

MA

undesired side effects or lack of efficacy, where authors claim the cause to be the

D

ability of OLZ to undergo polymorphic transformations. These observations

PT E

support the need for a proper selection of excipients once their role on promoting the stability of olanzapine in a dosage form has been determined. For this study,

CE

different amorphous and partially amorphous excipients were selected due to

AC

their ability to sorb large amounts of water than purely crystalline materials and due to their established use in formulations. Two groups of water soluble polymers were selected for the study, namely, a partially amorphous polyethyleneglycol (PEG) and amorphous hydroxypropylcellulose (HPC) and polyvinylpyrrolidone (PVP).

5

ACCEPTED MANUSCRIPT The aim of this work was the study of the phase transitions of olanzapine when exposed to different amounts of water in the presence of different excipients. Furthermore the study investigated how different excipients may change the rates of solid phase transitions of olanzapine when exposed to water

PT

vapor according to different mechanisms of water sorption. The study evolved by

RI

considering the application of pressure to the formulations, and OLZ in particular,

SC

by the production of compacts, which is the basis for tableting and dry

NU

granulation, prior to exposing them to different humidity conditions in order to evaluate how pressure may affect the formation or prevention of olanzapine

MA

hydrates. The study also aimed at providing clarification on the mechanism of

D

interaction between olanzapine and various excipients with water, when aqueous

CE

PT E

suspensions were produced, likewise in dissolution tests.

Experimental section

2.1.

Materials

AC

2.

Olanzapine anhydrous form I (OLZ FI, molecular weight = 312.43 g/mol, density = 1.300 g/cm3) was purchased from Pharmorgana, India. The chemical structure and the three-dimensional structure of anhydrous OLZ FI are represented in Figure 1. The different polymers considered in the study were 6

ACCEPTED MANUSCRIPT Polyethyleneglycol

(PEG,

PEG

6000,

Sigma-Aldrich,

Germany),

Hydroxypropylcellulose (HPC, LF pharm grade, Klucel™, Ashland, Germany), Polyvinylpyrrolidone (PVP, PVP K25, BASF Chemicals, Germany). Other chemicals, such as, lithium chloride, magnesium nitrate, sodium chloride, potassium nitrate

PT

were supplied by Sigma Aldrich (Germany) whereas demineralized water was

SC

RI

produced in house (Milli-Q system, Merck Millipore, USA).

Methods

2.2.1.

Preparation of olanzapine dihydrate D

NU

2.2.

MA

In order to characterize and quantify olanzapine transformations during

D

storage at different conditions of temperature and humidity, the dihydrate D was

PT E

produced. OLZ FI (1 g) was suspended in water (10 ml) at room temperature for 7 days under stirring before filtration under vacuum for collection of the crystals

Preparation of samples for moisture sorption studies: PEG, PVP and HPC

AC

2.2.2.

CE

formed (Reutzel-Edens et al., 2003) (CSD refcode: AQOMAU).

were sieved to eliminate agglomerates with collection of particles below 180 μm diameter. Approximately 1g of each polymer was placed in a 20 ml glass vial and stored at 11% RH (lithium phosphate), 25 oC, for 7 days prior to prosecution of the vapor sorption experiments to decrease and normalize the water content in samples. Thereafter, these polymers were used to prepare physical mixtures with 7

ACCEPTED MANUSCRIPT olanzapine (olanzapine: polymer, 25:75 w/w and 50:50 w/w, Table 1) by mixing olanzapine and each polymer in glass vials rotating in a roller mixer for 10 min. 200 mg mixtures were prepared in triplicate and split in halves. One half (100 mg) was used to obtain compacts with 10 mm in diameter at a compaction force of 40

PT

kN with an upper punch speed of 10 mm/min, using an universal testing

RI

equipment (Lloyd Instruments LR 50K, UK). Overall 36 samples (powder mixtures

SC

and compacts) were prepared for each polymer. 6 samples (3 powder mixtures

NU

and 3 compacts) of each olanzapine:polymer ratio were stored at 3 different conditions: 75%RH/ 25⁰C, 75%RH/40⁰C and 93%RH/25⁰C for 28 days (see 2.2.3 for

MA

details) and analyzed periodically (7, 14, 21 and 28 days). Each polymer and

D

olanzapine (raw materials) were also stored under the same conditions and used

2.2.3.

PT E

as controls.

Gravimetric water vapor sorption: Water vapor sorption was determined

CE

gravimetrically by storing samples at 25 °C for 28 days in desiccators containing

AC

saturated salt solutions to maintain the relative vapor pressure, namely, sodium chloride (75.29 ± 0.12 %, i.e., 75 %RH), and potassium nitrate (93.48 ± 0.55 %, i.e., 93 %RH). The different formulations were also stored in a climatic chamber (Vötsch, VC2033, Germany) with controlled humidity (75%RH) and temperature (40⁰C). During this period the masses (by weight) of each sample were periodically (7, 14, 21, 28 days) determined with a Mettler–Toledo analytical 8

ACCEPTED MANUSCRIPT balance. Simultaneously, spectral (FTIR), calorimetric (DSC) and X-ray (XRPD) analysis were performed. 2.2.4.

Preparation of samples for solvent-mediated conversion studies: Polymers

(25 mg) were dissolved in 12.5 ml demineralized water (n=3). Olanzapine (75 mg)

PT

was suspended in both water and polymer solutions (12.5 ml) for 60 min under

RI

agitation (200 rpm) (Table 1). The sediment was recovered by vacuum filtration to

SC

remove the excess of water and immediately transferred to a difractometer for

2.2.5.

NU

XRPD analysis.

X-ray powder diffraction (XRPD) analysis:

Olanzapine, polymers and

MA

physical mixtures of olanzapine and polymers were analyzed before and after

D

storage at different RH and temperatures. The diffractograms of compacted OLZ

PT E

were obtained from the gentle milled compacts powders. The structural characterization of the different samples was carried out by X-ray powder

CE

diffraction (XRPD) on an Analytical X’Pert PRO apparatus set with a vertical PW

AC

3050/60 goniometer (ɵ-2, ɵ mode) equipped with a X’Celerator detector and with automatic data acquisition (X’Pert Data Collector software, v2.0b), using a monochromatized CuKα

radiation as incident beam, 40 kV–30 mA.

Diffractograms were obtained by continuous scanning in a 2ɵ -range of 7–35° with a step size of 0.017° and scan step times of 20 s.

9

ACCEPTED MANUSCRIPT 2.2.6.

Differential scanning calorimetry (DSC): Thermograms of the polymers and

physical mixtures were obtained before and after storage at different RHs for different time periods. The OLZ particles recovered from the different polymeric solutions were allowed to dry at 53%RH/25°C for 24h prior to the analysis. 2 to 3

PT

mg were placed in pin-holed crucibles. The analysis was performed in a

RI

differential scanning calorimeter (TA Instruments, Q200, USA). Dry N2 was used as

SC

the purge gas (Air Liquide, 50 ml/min). Samples were heated at 10C/min within

2.2.7.

NU

the 0-210C temperature range.

FT-Infrared spectroscopy (FTIR): Polymers and physical mixtures before and

MA

after storage were mixed with KBr at a concentration of 0.5 w/w. The mixtures

D

(about 200 mg) were compressed into compacts of 12 mm diameter each. The

PT E

infrared spectra were obtained at the absorption mode (IR Affinity-1 Shimadzu spectrophotometer, Japan) with 64 scans with a resolution of 2 cm-1. Comparisons

CE

between FTIR spectra of the olanzapine:polymer samples after storage at

AC

different conditions of humidity and temperature were made possible by using partial least square regression (PLS), a chemometric technique often used for this purpose (Tantishaiyakul et al., 2008). PLS was employed to detect and quantify the anhydrous form I and dihydrate D and polymers contents in the physical mixtures of OLZ stored for different periods at different RHs. The spectral region of 875-915 cm-1 was selected for the PLS analysis because in this region both 10

ACCEPTED MANUSCRIPT polymorphic forms can be distinguished withpeaks at 886 cm-1 and 904 cm-1, which are indicators of the existence of the anhydrous Form I or the dihydrate D (the one found after storage), respectively. This region was also chosen for analysis because the polymers’ spectra showed not to interfere in this spectral

PT

region (Chakravarty et al., 2010). Physical mixtures of OLZ F I and dihydrate D of

RI

compositions ranging from 0% to 100% were used as the calibration set. The

SC

spectra were pretreated by standard normal variate transformation (The

NU

Unscrambler X, Camo software, Norway). From this, a set of spectra collected at each concentration were set aside to form the test set. Internal cross validation

Near infrared spectroscopy (NIR): The polymers which were stored for a

D

2.2.8.

MA

(leave-one-out) was performed to the calibration samples.

PT E

period of 1 month at different RHs were analyzed by an Antaris FT-NIR spectrometer (Thermo Scientific, USA) equipped with an InGaAs detector. Each

CE

NIR spectrum was recorded between 10 000-4000 cm-1 with a resolution and

2.2.9.

AC

number of scans of 2 cm-1 and 64 cm-1, respectively. Scanning electron microscopy (SEM): Physical mixtures of OLZ and

polymers were observed by scanning electron microscopy. The samples were mounted on aluminum stubs, coated with gold by ion sputtering (JEOL JFC-1200 Fine Coater, Japan) and observed by a scanning electron microscope (Jeol, JSM5200 LV, Japan). 11

ACCEPTED MANUSCRIPT 3. Results 3.1.

Characterization of the starting materials The characteristic XRPD diffraction patterns of all 4 excipients were

observed using XRPD in the angular range of 7–35 (2Ɵ). Figure 2 (a) shows the

PT

XRPD diffraction patterns obtained for the crystalline (OLZ), partially amorphous

RI

(PEG) and amorphous (PVP and HPC) compounds. Crystalline OLZ exhibits strong

SC

diffraction peaks, while materials that lack long-range molecular order (PVP and

NU

HPC) scatter X-rays to produce a diffuse or ‘‘halo’’ pattern. The evaluation of the different water sorption behaviors of each excipient

MA

contributes to elucidation on how they may alter the kinetics of hydration of

D

olanzapine in formulations due to the likely competitiveness to water molecules.

PT E

The moisture uptake patterns of the various materials when exposed to 75%RH/25⁰C, 75%RH/40⁰C and 93%RH/25⁰C for 4 weeks is represented in Figure

CE

2 (b). For instance, partial crystalline PEG showed to sorb low amounts of water at

AC

75%RH, regardless of the temperature (2.8±0.8% at 40°C and 2.5±0.9% at 25°C after 28 days). However, the change of PEG powder’s mass recorded at 93%RH was as high as 53% w/w. The amorphous and hydrophilic PVP showed to sorb considerable amounts of water when exposed to both 75% and 93%RH. HPC showed also to be able to take up water vapor at a humidity of 75% RH. However, at 93% RH, this polymer was less hygroscopic than both, PVP and PEG. 12

ACCEPTED MANUSCRIPT The calorimetric analyses carried out in pin-holed crucibles lids showed a broad endotherm below 100⁰C for olanzapine stored at 93 %RH (Figure 2c). This endotherm was related to the water loss, which was confirmed by thermogravimetric analysis (data not shown). After the dehydration of the

PT

sample, one could observe the peak for the melting of the anhydrous form. The

RI

thermograms of the polymers on heating showed the apearance of a broad band,

SC

between 75-120⁰C, due to the presence of water in the sample, except for PEG.

NU

PEG’s thermogram showed the disapearance of the peak of melting which resulted from the ability of the crystalline regions of PEG to dissolve into the

MA

water absorbed by the amorphous regions of the polymer (Marsac et al., 2008).

D

The effect of moisture sorption on all excipients was investigated by NIR

PT E

spectroscopy after a 4 week storage under different RH and temperature conditions. As a result of the water vapor sorption, a common feature to all

CE

spectra was the increased absorbance with an increase in water activity for

AC

wavelengths higher than 1300 nm (Figure 3, a-d). This trend was especially evident in the 1400-1450 nm (first overtone of -OH stretch) and 1900-1950 nm (combination of -OH stretch with -OH deformation) water absorption regions. For PVP and HPC polymers a significant increase of the absorbance observed in the spectra at the region between 1400-1450 nm could be identified for the samples stored at both 75% and 93%RH. With regard to the position of the band due to 13

ACCEPTED MANUSCRIPT water molecules, the 1850-1950 nm region, the maximum absorption is strongly dependent on the hydrogen-bonding reflecting the molecular mobility of water molecules. Therefore, the broadening of these bands and a decrease on their maximums’ wavelengths indicated an increase of free water molecules (higher

PT

molecular mobility) when polymers were stored at high humidity conditions. The

RI

water absorbed by PEG was seen as a distinct absorption maximum at 1933 nm

SC

after storage at 75%RH at both 40 ⁰C and 25 ⁰C. At 93%RH, free water showed an

NU

absorption maximum at around 1900nm. Furthermore, the bands related to the

Phase transformation of olanzapine during storage

D

3.2.

MA

PEG’s structure have shown to decrease their intensity at 93%RH.

PT E

The effect of humidity on the different olanzapine samples is shown in the Figure 4. The OLZ samples maintained at 75%RH/40⁰C for 28 days showed no

CE

traces of OLZ hydrate formation (Figure 4 a). At 75%RH/25⁰C it was possible to

AC

observe changes on the diffraction pattern of OLZ when it was physically mixed with both PEG and HPC (B1 and C1, Figure 4 b), due to the appearance of new peaks in the diffractogram due to crystalline domains (grey bar). Although 3 different dihydrates of OLZ are known (dihydrate B, D and E) (Reutzel-Edens et al., 2003), the characteristic peaks of OLZ dihydrate D (at 2 = 9.27o, 9.49o, 11.62o, 12.01o) appeared along with the peaks of Form I in the diffractograms of these 2 14

ACCEPTED MANUSCRIPT olanzapine:polymer samples. The dihydrate D specific peaks were more evident in the OLZ:PEG XRPD profile, due to the higher content of the dihydrate D form in this sample. The presence of characteristic peaks of both anhydrous and dihydrate forms of OLZ suggested that the dihydrate conversion was not

PT

complete within the period of test.

RI

At 93%RH, the X-ray powder pattern of OLZ, which was stored without any

SC

excipient, revealed a reduction of the intensity of the peaks, characteristic of the

NU

anhydrous form I, and the presence of new peaks, characteristic of the dihydrate D (Figure 4 c). Regarding Error! Reference source not found.OLZ which was

MA

physically mixed with different excipients and stored at 93%RH, its hydration

D

showed to have increased or decreased, depending on the polymer in which the

PT E

drug was mixed. For instance, the XRPD pattern of D1 sample showed only the specific peaks of anhydrous OLZ F I, whereas the XRPD pattern of OLZ:PEG (1:1)

CE

showed only the characteristic diffraction pattern of OLZ dihydrate D.

AC

The characteristic XRPD pattern of PEG, which presented 2 broad bands with maximum peaks at 2 = 19.7 ° and 23.8 ° (Figure 2a), was visible in B1 samples stored at 75%RH, but was no longer visible in B1 samples stored at 93%RH. This may result from the dissolution of the polymer on the absorbed water, which is a phenomenon that can occur with crystalline and highly soluble material (Baird et al., 2010). 15

ACCEPTED MANUSCRIPT The different OLZ forms can be easily distinguished by FTIR analysis, as a result of changes on the frequency of several IR bands scattered throughout the spectrum (Figure 5). For example, the bands describing the deformations of the piperazinyl group coupled to Me (2) (1009 cm−1) or to the azepine and thiophene

PT

moieties (965 cm−1) in OLZ Form I, change to 1004 cm−1 and 971 cm−1 in OLZ

RI

dihydrate D. At lower wavenumbers, the main components of the vibrational

SC

modes are the deformation and torsions of the rings, giving rise to highly coupled

NU

movements (Ayala, et al., 2006). In this region it is possible to distinguish both crystalline forms. When OLZ Form I was converted into the OLZ dihydrate D form,

MA

the infrared bands at 886 cm−1 and 745 cm-1 were shifted to 904 cm-1 and 750 cm1

D

, respectively. The hydration of OLZ was not completed by the end of storage (28

PT E

day at 93%RH) according to the presence of characteristic bands of both OLZ anhydrous and dihydrate forms (Figure 5). At 75%RH, no changes were visible on

CE

the olanzapine spectra at both temperatures (25⁰C and 40⁰C).

AC

The quantification of a hydrate conversion can be challenging. Fortunately, the FTIR spectra between 850 cm-1 and 915 cm-1 enabled the quantification of the OLZ hydrate conversion in OLZ:Polymer mixtures when a multivariate analysis test (PLS, partial least square method) was used. The PLS best model contained 1 PLS component for the analysis in binary mixtures. This component captured 94% of the variance in the X-variables, demonstrating that 16

ACCEPTED MANUSCRIPT the information contained in the spectral data was effectively used in calibration models (R2 = 0.995 and Q2) = 0.994). The root mean square errors of calibration (RMSEC = 1.89%) and cross validation (RMSECV = 2.24 %), together with the value for the root mean square error of prediction (RMSEP = 2.73%) were small

PT

confirming the choice of that region in the spectra for OLZ hydrate quantification

RI

due to the absence of excipient’s interference. No hydrate conversion occurred

SC

for any samples at 75%RH, 40⁰C, whereas, at 75%RH, 25⁰C, OLZ dihydrate D

NU

became evident in OLZ:PEG (1:1 and 3:1) and OLZ:HPC (1:1) samples. The dihydrate D content of these 3 samples after 7, 14, 21 and 28 days are shown in

MA

Table 2. After 28 days, all samples presented mixtures made of OLZ anhydrous

D

Form I and dihydrate D. At 93%RH, PEG has shown to increase the rate of hydrate

PT E

formation of OLZ; within 2 weeks OLZ was completely hydrated, regardless of PEG content in the sample (Table 2). The OLZ hydrate conversion rate in the presence

CE

of lower contents of HPC (OLZ:HPC, C2) showed to increase in comparison to the

AC

OLZ sample which was stored without the addition of any excipient (Table 2). Higher contents of the same polymer (OLZ:HPC, Table 2, C1) showed to reduce the hydrate conversion rate significantly when these mixtures were stored at 93%RH. This suggests that changes on the HPC due to humidity may cause changes on the hydrate conversion rate.

17

ACCEPTED MANUSCRIPT The thermograms of the physical powder mixtures before storage showed that OLZ’s enthalpy was largely reduced when physically mixed with either PEG or PVP (Figure 6, a and c), showing the ability of both polymers to dissolve OLZ when the mixtures were heated, contrarily to HPC (Figure 6b). The samples stored at

PT

different humidity showed a broad band due to water evaporation. These bands

RI

were broader for OLZ mixed with PEG, PVP and HPC stored at 93%RH due to the

SC

increased water sorption ability of these polymers at this high humidity.

NU

Figure 6 (a) shows the elimination of the melting peak associated with PEG when these samples were stored at 93%RH, which is indicative of a polymer

MA

crystallinity loss. The B1 samples (Table 1, 75%RH/25⁰C), presented the

D

appearance of a broad band due to water evaporation after the melting of the

PT E

polymer. By opposition the thermogram (DSC) showed that PEG’s melting temperature increased without visible elimination of water (75%RH/40⁰C). Figure

CE

6 (b) revealed two overlapping broad bands between 70-110 ⁰C only when

AC

OLZ:HPC mixtures were exposed to 75%RH/25⁰C and 93%RH/25⁰C. These two broad bands may have been originated after the disappearance of the water molecules present in both HPC and OLZ dihydrate D structures (Figure 6b). All OLZ:HPC stored samples have shown broad melting peaks for OLZ, with the OLZ Tmonset starting about 2 ⁰C below the physical mixture. The water sorption by PVP

18

ACCEPTED MANUSCRIPT also led to the broadening of OLZ melting peak, with the Tmonset of OLZ registering a decrease of about 10 ⁰ C for all the stored D formulations (Figure 6 c). The observation of the particles' surface by electronic microscopy (SEM) revealed that OLZ raw material presented an irregular shape. The conversion of

PT

OLZ into the hydrate did not lead to any change of the particles’ shape (Figure 7)

RI

which suggests that the phase transition occurred in the solid-state without

SC

passing through an intervening transient liquid phase.

NU

The mixtures of OLZ with the 3 polymers stored at 93%RH showed that the amount of water absorbed by the polymers resulted in coverage of OLZ particles

MA

by the plasticized polymers, which failed to allow the observation of changes on

D

the particles’ surfaces. The optical microscopy with polarized light was useful on

PT E

the observation of the OLZ particles even in the presence of the polymers. It was noticeable that OLZ particles in PEG samples completely changed their

CE

morphology - the irregular particle shape was converted into a lath shaped

AC

morphology, suggesting that hydrate formation of OLZ in this sample occurred by a different mechanism.

3.3.

Solid-phase transformation in compacts Thermograms of powders and freshly prepared compacts of formulations

A, B2, C2 and D2 (Table 1) are presented in Figure 8 (a and b). It was possible to 19

ACCEPTED MANUSCRIPT observe a reduction of melting temperature (Tm) and melting enthalpy (∆H) of OLZ when OLZ or the physical mixtures of OLZ:Polymer were compacted. Figure 8 (a) exhibits a reduction of the intensity of the crystalline peaks of OLZ (Ap) when it was compacted into tablets (At), reflecting a loss of crystallinity of the sample.

PT

Figure 9 (a and b) shows the dihydrate D content in different formulations

RI

(powders, P and compacts, C) after storage for 7 and 28 days at 93 % RH. OLZ,

SC

compressed alone, showed to have an increased dihydrate D content after 7 days

NU

of storage, as compared to the OLZ raw material stored as a powder. However, after 28 days the dihydrate D content present in the OLZ compact showed to be

MA

about 12%, which was much lower than the dihydrate D content in OLZ powder

D

sample (62 %).

PT E

For compacts containing mixtures of drug and polymer, a larger percentage of dihydrate D in the compacts when compared to the uncompressed material

CE

was found after 7 days of storage, reaching 100% in the case of OLZ:PEG

AC

compacts. As found for OLZ alone, the compacts made of physical mixtures have shown a slight increase or complete stagnation of hydrate formation between the 7th and the 28th days.

20

ACCEPTED MANUSCRIPT 3.4.

Conversion kinetics of OLZ in aqueous suspensions Up to this point the water content in the systems was present as absorbed

or adsorbed water but turning the systems into suspensions by increasing the amount of water was deemed relevant to promote the understanding of the

PT

stability of the different forms of OLZ in aqueous environments. Figure 10 shows

RI

the FTIR spectra of OLZ within the 720 cm-1-1020 cm-1 range. The spectra display

SC

OLZ powders filtered after being suspended in water or polymeric solutions for

NU

different periods (15, 45 and 60 min). Experiment E showed a mixture of anhydrous OLZ and dihydrate B after 45 min (Figure 10), whereas in experiment F

MA

only anhydrous OLZ was found after 45 min. This result revealed a slower

D

hydration rate of OLZ when suspended in PEG solutions. The hydrates formed in

PT E

this experiment were different from the ones obtained during storage at 93%RH.

AC

CE

Table 4 shows the viscosity of the 3 solutions used in the experiment.

4. Discussion

In multi-component systems the sorption of solvents (e.g. water in a wet granulation process) may exhibit a complex and non-additive behavior (Baird et al., 2010). Water may induce unexpected phase transitions in both API and excipients acquiring new properties in the respective particulates’ solid-state 21

ACCEPTED MANUSCRIPT lattices impacting on the API’s bioavailability and materials and dosage form stabilities. (Heidemann and Jarosz, 1991, Otsuka and Matsuda, 1994, Airaksinen et al., 2003). It has been observed that hygroscopic excipients (e.g. HPC, PVP and PEG) are prone to retard the kinetics of hydration of APIs, by a competing

PT

mechanism (Salameh and Taylor, 2006). The understanding of the interaction

RI

between those polymers and olanzapine by retarding, or not, its solid state

Polyetyleneglycol (PEG)

MA

4.1.

NU

control the hydration kinetics of olanzapine.

SC

transitions, contributes to the selection and use of a polymer in formulations to

D

Polyethyleneglycol is a semi-crystalline material presenting a large water

PT E

surface adsorption ability, particularly in its amorphous regions (Marsac et al.,

2001).

CE

2008) increasing the disordered fraction of PEG (Baird et al., 2010, Byrn et al.,

AC

PEG displayed small weight changes at 75% humidity (2.5-2.8) by opposition to 93%RH (53% w/w). The water uptake was dependent on the number of favorable polymer–water hydrogen bonds created, forming a first film of adsorbed water molecules on the polymer’s surface (Thijs et al., 2007). This led to the PEG chains to acquire a different conformation further increasing the number of hydrogen bonds with water molecules allowing the uptake of more 22

ACCEPTED MANUSCRIPT water. PEG’s ability to sorb large amounts of moisture results from the presence of ether oxygen atoms (–O–) in the oxyethylene polymer backbone, as well as, hydroxyl (–OH) end groups (Baird et al., 2010). Results from the calorimetric experiments showed a decrease and even

PT

elimination of the melting transition, i.e., loss of crystallinity, when PEG was

RI

exposed at 75-93%RH (25⁰C for 4 weeks). Chan and Kazarian (2004) have shown

SC

that, when oxygen atoms acting as electron donors for H-bonds with water

NU

molecules are available (RH<70%), water molecules can enter into the PEG matrix forming hydrogen bonds with oxygen atoms without formation of condensed

MA

clusters of water. At RH> 70%, all oxygen atoms in the polymer chain are taken

D

thus, additional water molecules will interact with the ones already absorbed to

PT E

PEG forming clusters (Chan and Kazarian, 2004). Furthermore, absorbed water also acts as a plasticizer for the polymer, increasing the mobility of polymer

CE

chains, promoting swelling and increasing the space available for the growing

AC

cluster of water molecules (Chan and Kazarian, 2004). In the limit, RH>90% water can dissolve the PEG (Thijs et al., 2007, Kitano et al., 2001). The visual inspection of samples (during the 28 days at 93% RH) has shown a solid-to-solution phase transformation indicating that PEG started to dissolve in the condensed water. The results from the current study revealed that OLZ could undergo hydrate formation at 75% RH when mixed with PEG, which was never observed 23

ACCEPTED MANUSCRIPT when OLZ was stored alone. At 93% RH, OLZ samples could fully hydrate and, increased PEG contents were followed by faster hydrate conversions. The complete conversion of anhydrous OLZ into the dihydrate D form (weight increase of 10.3% w/w) was complete at day 15, regardless of PEG content. The

PT

hydration of OLZ at lower relative humidity and at faster rates when physically

RI

mixed with PEG may be explained by the formation of condensed water (water

SC

clusters) within the mixture, which were available to interact with OLZ particles.

NU

The complete change of morphology of the crystals of OLZ indicated that the transformation could potentially occur (at least partially) in solution (Marsac et

MA

al., 2008), i.e., anhydrous OLZ was able to dissolve into the bulk water and the

D

nucleation and crystal growth of the more stable dihydrate form has occurred.

PT E

The higher rates of hydrate conversion of OLZ in compacts with PEG may be explained by the partial loss of crystallinity of the sample and the more intimate

CE

contact between the drug and the polymer, resulting in a complete hydrate

4.2.

AC

formation in one week for both OLZ:PEG ratios used.

Polyvinylpirrolidone (PVP) Polyvinylpirrolidone, an amorphous and hygroscopic polymer (Kiekens et

al., 2000), at 25⁰C absorbs approximately 5-/10% moisture at 10%RH and about 55% at 93%RH (based on the dry mass of the sample) reflecting a significant 24

ACCEPTED MANUSCRIPT increases (Fitzpatrick, 2002). Water molecules can influence the free volume due to the breakage of hydrogen bonds between the molecules in the solid lowering the glass transition temperature (Tg) (Kiekens et al., 2000). It was found that PVP rapidly absorbed moisture when stored at the

PT

different humidity environments and reached an apparent equilibrium (i.e.,

RI

becomes saturated with vapor) after one week of storage (data not shown). The

SC

rate of moisture sorption appears to be much smaller than the observed

NU

timescale for hydrate transformation, once PVP was able to inhibit hydrate conversion during the 28 days of storage at different environmental conditions.

MA

This suggests that a desiccant effect of PVP was not the only mechanism

D

responsible for hydrate inhibition of OLZ. Taking into consideration that the

PT E

viscoelasticity and the volume of the polymer changes due to the glassy to rubbery transition (Kiekens et al., 2000), the physical state of the polymer at high

CE

humidity needs to be considered. As one could observe by both visual inspection

AC

and by SEM, PVP tended to form a ‘‘viscous liquid like’’ structure (Salameh and Taylor, 2006) when stored at high RHs due to sorption of water, refleted on the decrease of the tensile strength of compacts which became soft and deformable (Kiekens et al., 2000). This highly plasticized and mobile structure covered OLZ particles once in contact with them, and contrary to PEG, no hydrate conversion was verified. In samples with PEG the dissolution of the polymer in the condensed 25

ACCEPTED MANUSCRIPT water enhanced hydration of OLZ by providing an aqueous medium . In the case of PVP different explanations can de identified for hydration to be inhibited. Firstly, the extent of OLZ dissolution into PVP is reduced due to the high viscosity of the plasticized polymer that showed to present as “gel-like” structure (visual

PT

inspection) with reduction of the contact area with the API and, consequently, the

RI

hydrate formation (Salameh and Taylor, 2006). Secondly, in the aqueous slurries

SC

of OLZ in PVP and PEG solutions (high contents of water available), no hydrate

NU

transformation of OLZ was visible in the presence of PVP, showing that this polymer is a more effective inhibitor than PEG in aqueous environments. The

MA

higher hydrogen bond ability of this polymer relatively to PEG (Tian et al., 2007)

D

may justify the larger inhibitory effect of PVP. The formation of hydrogen bonds

PT E

between polymers and the drug inhibits the latter from establishing hydrogen

Hydroxypropylcellulose (HPC)

AC

4.3.

CE

bonds with water molecules.

In celluloses water is only absorbed in the non-crystalline regions (Mihranyan et al., 2004). The higher content of amorphous domains in HPC than in other other celluloses (e.g. microcrystalline cellulose) (Alvarez-Lorenzo et al., 2000) result in its higher water absorption. In this study, OLZ showed to hydrate faster in the first week when mixed with HPC. This was not antecipated because 26

ACCEPTED MANUSCRIPT moisture is commonly selectively taken up by the most hygroscopic component. Water-vapor sorption-desorption isotherms of HPC showed to fit the YoungNelson model (Young and Nelson, 1967) which assumes that moisture can be absorbed by a solid at 3 locations: monolayer adsorption (bound water),

PT

externally adsorbed moisture (normally condensed) and internally absorbed

RI

moisture. The internally absorbed moisture results from the water diffusion

SC

through pores into the core of the particle from the adsorbed monolayer

NU

(Alvarez-Lorenzo et al., 2000, Joshi and Wilson, 1993). The absence of protection of OLZ hydration by HPC may be justified by the interaction beyween the

MA

condensed water molecules encountered on the surface of the polymer (water

D

molecules above the monolayer moisture, i.e, externally adsorbed moisture) and

PT E

OLZ molecules.

When water molecules penetrate the HPC structure they move towards

CE

and into the hydrogen-bonded links between adjacent polymer chains increasing

AC

their mobility and space between them reflected by swelling in a time dependent process (Dios et al., 2015)(Alvarez-Lorenzo et al., 2000). As a result, the chains are extended, resulting in more locations available for hydrogen bonding with OLZ and further molecular entanglements (Ju et al., 1995) justifying the abrupt reduction of the hydrate formation rate, when higher contents of HPC were used in the formulation. This hydrogen bonding with OLZ may hinder the hydrogen 27

ACCEPTED MANUSCRIPT bonding of OLZ with water molecules. When OLZ was suspended in HPC solutions, OLZ remained anhydrous as it had occurred with PVP. The hydrogen bonding between the drug and the excipients has been pointed out as an important factor on hydration inhibition in aqueous solutions (Tian et al., 2007).

PT

Therefore, the presence of both hydrogen bond donor and acceptor groups in the

RI

HPC ring structure may have facilitated the interaction with OLZ molecule.

SC

Furthermore, the plasticization of the polymer seemed to form a thick gel layer

NU

around the OLZ particles, which was observed by SEM and visual inspection. The formation of this gel may also hinder OLZ particles’ dissolution into condensed

MA

water and/or prevent moisture penetration into the OLZ crystal when higher

4.4.

PT E

D

contents of HPC were used.

Pseudopolymorphic transformations in compacts

CE

Tableting is a process which may potentiate the occurrence of phase

AC

transformations in either the drug or the excipients. Most data (e.g. caffeine, sulfabenzamide and maprotiline hydrochloride suggests that the transformations from one polymorphic form into another during tableting encompasses an intermediate amorphous phase (Morris et al., 2001) (Chan and Doelker, 1985). The diffractograms of compacted OLZ neither showed missing nor new diffraction peaks of OLZ, which means that no polymorphic transformation 28

ACCEPTED MANUSCRIPT occurred under compression. However, it was observed that, upon compression, diffraction experiments showed a loss of intensity and broadening of OLZ peaks. This was probably due to the induction of lattice disorder in the surface of anhydrate promoted by compaction. The reduction of the melting temperature

PT

enthalpy of OLZ also revealed loss of crystallinity of the drug induced by

RI

application of pressure. This phenomenon was observed for both drug and

SC

drug:polymer mixtures. OLZ in compacts showed a faster hydration within 7 days

NU

than in the physical powder mixtures (93%RH). In the solid state, nucleation tends to proceed from dislocations in a crystal because of their high free energy. For

MA

this reason, the activation energy barrier for a polymorphic transformation is

D

lower at these sites (Chan and Doelker, 1985) which is emphasized during

PT E

compression in which the plastic flow induces dislocational strains in a crystal facilitating nucleation. However, the OLZ compacts showed only a minimal

CE

increase on the dihydrate D content between 7 and 28 days. The reduced hydrate

AC

rate conversion may be justified by the low specific surface area of the compacts in relation to the powders and by the low diffusion rate of water molecules through the OLZ crystal lattice particularly within the core of the tablet. This fact resulted on a reduction of the number of molecules of OLZ in direct contact with the water vapor.

29

ACCEPTED MANUSCRIPT 4.5.

Conversion of OLZ in aqueous suspensions Data has shown that no hydration occurred with OLZ suspended in PVP and

HPC solutions, whereas hydrate conversion was observed in PEG solutions. PVP and PEG solutions showed to have similar viscosities, which led us to conclude

PT

that the protector effect of PVP is not related to the viscosity. Previous studies

RI

have shown that the effectiveness of a polymer as a crystallization inhibitor is

SC

highly dependent on its structure, due to the formation of specific intermolecular

NU

interactions between the polymers and the drugs (Otsuka et al., 2015, Ilevbare et al., 2013). These interactions have been found to contribute to the adsorption of

MA

the polymer molecules onto solids and thus, controlling a number of interfacial

D

processes such as crystal growth. Specific intermolecular interactions (such as

PT E

hydrogen-bonding) and non-specific hydrophobic drug–polymer interactions appeared to be important in determining the impact of a given polymer on crystal

CE

growth effectiveness (Otsuka et al., 2015). The adsorption of the polymer

AC

molecules to the crystal surface interface promotes steric hindrance to the access of the solute molecules to the crystal terrace. All of these polymers have the ability to form hydrogen bonds with OLZ. PEG has ether oxygen groups in its chain and two hydroxyl groups at the two ends of the polymer chain; PVP presents one acceptor hydrogen bonding group in its ring structure. The oxygen in the carbonyl group is able to form stronger 30

ACCEPTED MANUSCRIPT hydrogen bonds than the ether oxygen groups of PEG, explaining its stronger protection against crystal growth (Tian et al., 2007). In a different study with acetaminophen, the stronger inhibitory effect of PVP was also attributed to the lower mobility of the functional groups involved in specific polymer-drug

PT

interactions when compared to PEG. HPC has both hydrogen bond donor and

RI

acceptor groups in its ring structure which may form hydrogen bonds with OLZ.

SC

Furthermore, the hydrophobic nature of HPC can also lead to adsorption of this

NU

polymer to the OLZ hydrophobic surfaces, blocking the crystal growth by preventing the access of water molecules to OLZ crystal terrace (Tian et al.,

D PT E

5. Conclusions

MA

2007, Gift et al., 2008).

This study provided an insight into the role of hygroscopic polymeric

CE

additives on the stabilization of OLZ anhydrous Form I. The study showed that the

AC

stabilization of anhydrous OLZ in humid environments was dependent on the type of polymer and the fraction of polymer present in the formulation. For example, both PEG and HPC showed to induce hydrate transformations of OLZ at a relative humidity of 75%. Since OLZ cannot hydrate at this humidity when it is stored alone, this result highlighted the importance of designing the formulation by the right choice of excipients in order to maintain the drug stable in the medicine. At 31

ACCEPTED MANUSCRIPT 93%RH, PEG showed to absorb high contents of moisture, which allowed OLZ to undergo a solvent-mediated transformation, accelerating the hydrate conversion kinetics of OLZ. PVP also showed to uptake high moisture contents but, contrary to PEG, PVP was able to protect OLZ from undergoing hydrate transformation. In

PT

aqueous environments, when a higher content of free water molecules was

RI

available to bind OLZ, PVP also showed to protect OLZ from hydration. It was

SC

therefore possible to extrapolate that PVP could form specific interactions with

NU

the drug, which inhibited OLZ to form hydrogen bonds with the water molecules. Regarding HPC, the protector effect of this polymer seemed to be related to the

MA

content of this polymer within the sample. Since during the swelling of the

D

polymer there are more hydrogen bond locations available to bind with the OLZ

PT E

molecule, this process was amplified when there were more molecules of HPC in the system available to bind with the OLZ molecule. One cannot exclude the

CE

possibility of different conformations of these molecules depending on the

AC

environmental water content. Compaction of OLZ showed to slightly reduce its crystallinity, inducing a faster hydrate formation in the first days of storage. Compaction also led to a closer contact between OLZ and polymers’ molecules, which resulted in an acceleration of hydrate formation when OLZ was mixed with PEG.

32

ACCEPTED MANUSCRIPT Overall, since the choice of the formulation design is important, in order to produce a product with a consistent quality performance throughout the product’s shelf life, the study confirms the attention that must be payed to the choice of excipients used in the formulation, helping on avoiding the formation of

PT

metastable phases that may affect the quality of the final product. Furthermore,

RI

being OLZ a drug commonly used in the treatment of schizophrenia, an illness

SC

which requires patients to be kept clinically stable and in good compliance with

NU

the treatment, minor changes on the bioavailability could affect the stability of

D

MA

the patients, with a decrease on their compliance.

PT E

Acknowledgments

Authors acknowledge “Fundação para a Ciência e a Tecnologia”, Lisbon,

CE

Portugal for providing the financial support to this work (PTDC/CTM/098688/2008

AC

and SFRH/BD/90118/2012).

33

ACCEPTED MANUSCRIPT References

Ahlneck C., Zografi G., 1990. The molecular basis of moisture effects on the physical and chemical stability of drugs in the solid state. Int. J. Pharm. 62, 87-

PT

95.

RI

Airaksinen S., Karjalainen M., Shevchenko A., Westermarck S., Leppanen E.,

SC

Rantanen J., Yliruusi J., 2005. Role of water in the physical stability of solid

NU

dosage formulations. J. Pharm. Sci. 94, 2147-2165.

MA

Airaksinen S., Luukkonen P., Jorgensen A., Karjalainen M., Rantanen J., Yliruusi J., 2003. Effects of excipients on hydrate formation in wet masses containing

PT E

D

theophylline. J. Phar. Sci. 92, 516-528. Alvarez-Lorenzo C., Gomez-Amoza J.L, Martinez-Pacheco R., Souto C., Concheiro

CE

A. 2000. Interactions between hydroxypropylcelluloses and vapour/liquid

AC

water. Eur. J. Pharm. Biopharm. 50, 307-318. Ayala A.P., Siesler H.W., Boese R., Hoffmann G.G., Polla G..I, Vega D.R., 2006. Solid state

characterization

of

olanzapine

polymorphs

using

vibrational

spectroscopy. Int. J. Pharm. 326, 69-79.

34

ACCEPTED MANUSCRIPT Baird J.A., Olayo-Valles R., Rinaldi C., Taylor L.S., 2010. Effect of molecular weight, temperature, and additives on the moisture sorption properties of polyethylene glycol. J. Pharm. Sci. 99, 154-168.

PT

Byrn S.R., Xu W., Newman A.W., 2001. Chemical reactivity in solid-state

RI

pharmaceuticals: formulation implications. Adv. Drug Del. Rev. 48, 115-136.

SC

Campisi B., Chicco D., Vojnovic D., Phan-Tan-Luu R., 1998. Experimental design for a pharmaceutical formulation: optimisation and robustness. J. Pharm. Biomed.

NU

An. 18, 57-65.

MA

Chakravarty P., Govindarajan R., Suryanarayanan R., 2010. Investigation of

D

solution and vapor phase mediated phase transformation in thiamine

PT E

hydrochloride. J. Pharm. Sci. 99, 3941-3952. Chan H.K., Doelker E., 1985. Polymorphic transformation of some drugs under

CE

compression. Drug Dev. Ind. Pharm. 11, 315-332.

AC

Chan K.L.A., Kazarian S.G., 2004. Visualisation of the heterogeneous water sorption in a pharmaceutical formulation under controlled humidity via FT-IR imaging. Vibrational Spect. 35, 45-49. Dios P., Pernecker T., Nagy S., Pal S., Devay A., 2015. Influence of different types of low substituted hydroxypropyl cellulose on tableting, disintegration, and 35

ACCEPTED MANUSCRIPT floating behaviour of floating drug delivery systems. Saudi Pharm. J. 23, 658666. Fitzpatrick S., McCabe J.F., Petts C.R., Booth S.W., 2002. Effect of moisture on

PT

polyvinylpyrrolidone in accelerated stability testing. Int. J. Pharm. 246, 143-151.

SC

generic olanzapine. The Ochsner J. 13, 550-552.

RI

Galarneau D., 2013. A case of teeth discoloration upon transition from zyprexa to

NU

Gift A.D., Luner P.E., Luedeman L., Taylor L.S., 2008. Influence of polymeric excipients on crystal hydrate formation kinetics in aqueous slurries. J. Pharm.

MA

Sci. 97, 5198-5211.

D

Heidemann D.R., Jarosz P.J., 1991. Pre-formulation studies involving moisture

PT E

uptake in solid dosage forms. Pharm. Res. 8, 292-297.

CE

Ilevbare G.A., Liu H., Edgar K.J., Taylor L.S., 2013. Impact of polymers on crystal growth rate of structurally diverse compounds from aqueous solution. Mol.

AC

Pharm. 10, 2381-2393. Joshi H.N., Wilson T.D., 1993. Calorimetric studies of dissolution of hydroxypropyl methylcellulose E5 (HPMC E5) in water. J. Pharm. Sci. 82, 1033-1038.

36

ACCEPTED MANUSCRIPT Ju R.T., Nixon P.R., Patel M.V., Tong D.M., 1995. Drug release from hydrophilic matrices. 2. A mathematical model based on the polymer disentanglement concentration and the diffusion layer. J. Pharm. Sci. 84, 1464-1477.

PT

Kiekens F., Zelko R., Remon J.P., 2000. Effect of the storage conditions on the tensile strength of tablets in relation to the enthalpy relaxation of the binder.

SC

RI

Pharm. Res. 17, 490-493.

Kitano H., Ichikawa K., Ide M., Fukuda M., Mizuno W., 2001. Fourier Transform

NU

Infrared Study on the State of Water Sorbed to Poly(ethylene glycol) Films.

MA

Langmuir 17, 1889-1895.

D

Kontny M., Zografi G., 1995. Sorption of water by solids, in: Brittain H. (Ed.),

York, pp. 387-418.

PT E

Physical characterization of pharmaceutical solids. Marcel Dekker Inc., New

CE

Malaj L., Censi R., Gashi Z., Di Martino P., 2010. Compression behaviour of

AC

anhydrous and hydrate forms of sodium naproxen. Int. J. Pharm. 390, 142-149. Marsac P.J., Romary D.P., Shamblin S.L., Baird J.A., Taylor L.S., 2008. Spontaneous crystallinity loss of drugs in the disordered regions of poly(ethylene oxide) in the presence of water. J. Pharm. Sci. 97, 3182-3194.

37

ACCEPTED MANUSCRIPT Mihranyan A., Llagostera A.P., Karmhag R., Stromme M., Ek R., 2004. Moisture sorption by cellulose powders of varying crystallinity. Int. J. Pharm. 269, 433442.

PT

Morris K.R., Griesser U.J., Eckhardt C.J., Stowell J.G., 2001. Theoretical approaches to physical transformations of active pharmaceutical ingredients during

SC

RI

manufacturing processes. Adv. Drug Del. Rev. 48, 91-114.

Otsuka M., Matsuda Y., 1994. The Effect of humidity on hydration kinetics of

NU

mixtures of nitrofurantoin anhydride and diluents. Chem. & Pharm. Bull. 42,

MA

156-159.

D

Otsuka N., Ueda K., Ohyagi N., Shimizu K., Katakawa K., Kumamoto T., Higashi K.,

PT E

Yamamoto K., Moribe K., 2015. An insight into different stabilization mechanisms of phenytoin derivatives supersaturation by HPMC and PVP. J.

CE

Pharm. Sci. 104, 2574-2582.

AC

Reutzel-Edens S.M., Bush J.K., Magee P.A., Stephenson G.A., Byrn S.R., 2003. Anhydrates

and

hydrates

of

olanzapine: 

crystallization,

solid-state

characterization, and structural relationships. Crystal Growth & Des. 3, 897907.

38

ACCEPTED MANUSCRIPT Rumondor A.C., Marsac P.J., Stanford L.A., Taylor L.S., 2009. Phase behaviour of poly(vinylpyrrolidone) containing amorphous solid dispersions in the presence of moisture. Mol. Pharm. 6, 1492-1505.

PT

Salameh A.K., Taylor L.S., 2006. Physical stability of crystal hydrates and their

RI

anhydrates in the presence of excipients. J. Pharm. Sci. 95, 446-461.

SC

Samuel R., Attard A., Kyriakopoulos M., 2013. Mental state deterioration after switching from brand-name to generic olanzapine in an adolescent with bipolar

NU

affective disorder, autism and intellectual disability: a case study. BMC Psych.

MA

13, 244.

D

Tantishaiyakul V., Permkam P., Suknuntha K., 2008. Use of Drifts and PLS for the

PT E

determination of polymorphs of piroxicam alone and in combination with pharmaceutical excipients: A technical note. AAPS PharmSciTech 9, 95-99.

CE

Thijs H.M.L., Becer C.R., Guerrero-Sanchez C., Fournier D., Hoogenboom R.,

AC

Schubert U.S., 2007. Water uptake of hydrophilic polymers determined by a thermal gravimetric analyser with a controlled humidity chamber. J. Mat. Chem. 17, 4864. Tian F., Saville D.J., Gordon K.C., Strachan C.J., Zeitler J.A., Sandler N., Rades T., 2007. The influence of various excipients on the conversion kinetics of 39

ACCEPTED MANUSCRIPT carbamazepine polymorphs in aqueous suspension. J. Pharm. Pharmacol. 59, 193-201. Young J.H., Nelson G.H., 1967. Theory of hysteresis between sorption and

PT

desorption isotherms in biological materials. Trans. ASAE 10, 260-263.

RI

Zhang G.G., Law D., Schmitt E.A., Qiu Y., 2004. Phase transformation

SC

considerations during process development and manufacture of solid oral

NU

dosage forms. Adv. Drug Del. Rev. 56, 371-390.

Zhang J., Zografi G., 2001. Water vapour absorption into amorphous sucrose-

D

Pharm. Sci. 90, 1375-1385.

MA

poly(vinyl pyrrolidone) and trehalose-poly(vinyl pyrrolidone) mixtures. J.

PT E

Zografi G., 1988. States of water associated with solids. Drug Dev. Ind. Pharm. 14,

AC

CE

1905-1926.

40

ACCEPTED MANUSCRIPT Figure Captions

Figure 1- Chemical structures of olanzapine and polymers used in the study and

RI

PT

packing of crystalline olanzapine.

SC

Figure 2 – XRPD patterns of the excipients prior to processing (a), variation in mass after exposure to high levels of humidity (b) and excipients’ thermal behaviors after

MA

NU

28 days of storage at different environmental conditions (c).

Figure 3 – Near infrared (NIR) spectra of olanzapine (a); polyethyleneglycol (b),

PT E

environmental conditions.

D

hydroxypropylcellulose (c) and polyvinylpyrrolidone (d) stored for 28 days at different

[RH (%) / Temperature (⁰C): 11 /25 (__); 75 / 40 (---); 75 / 25 (….); 93 / 25 (__)]

AC

CE

The main changes in the spectra are identified with a grey bar.

Figure 4 - XPRD of physical mixtures (A, B1, C1 and D1, see Table 1) stored for 28 days. [RH (%) / Temperature (⁰C): 75 / 40 (a); 75 / 25 (b) and 93 / 25 (c)].

Figure 5 - Characterization of olanzapine by FTIR after 28 days of storage at different environmental conditions. 41

ACCEPTED MANUSCRIPT

Figure 6 – Thermal analysis of the drug: excipient samples after storage for 28 days at different moisture and temperature conditions.

PT

Figure 7 – Micrographs from electronic (SEM) microscopy and optical microscopy with

RI

polarized light of different materials.

Formulation A before storage (a); formulation A 93%RH/ 28 day (b); formulation B2

NU

SC

93%RH/ 28 day (c); Formulation C2 93%RH/ 28 day (d); formulation 93%RH/ 28 day (e)

MA

Figure 8 – DSC Thermograms and X-ray diffractograms of OLZ (a)In different formulations before (powders, p) and after (compacts, c) compaction; Shows the X-ray diffractograms of OLZ before and after compaction.

PT E

D

(b)

Figure 9 – Content of olanzapine dihydrate D for 7 and 28 days after storage at 93%RH.

CE

(a)Formulations A, B1, C1 and D1; and (b) formulations B2, C2 and D2 (93%RH).

AC

The figure includes the formulations as powders or compacts

Figure 10 – FT-Infrared spectra of powders recovered after 15, 45 and 60 min exposure from different suspensions. The main shifts are represented with a grey bar.

42

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Fig. 1

43

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Fig. 2

44

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Fig. 3

45

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Fig. 4

46

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

Fig. 5

47

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Fig. 6

48

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Fig. 7

49

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Fig. 8

50

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

Fig. 9

51

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Fig. 10

52

ACCEPTED MANUSCRIPT

Samples

Drug/Excipient composition (mg, w/w)

Moisture Sorption Studies

A B1 B2 C1 C2 D1 D2

OLZ OLZ:PEG OLZ:PEG OLZ:HPC OLZ:HPC OLZ:PVP OLZ:PVP

100:0 50:50 75:25 50:50 75:25 50:50 75:25

E F G H

OLZ OLZ:PEG OLZ:HPC OLZ:PVP

PT

Code

solventmediated conversio n studies

Table 1 – Formulations considered in the study.

AC

CE

PT E

D

MA

NU

SC

RI

75:00 75:25 75:25 75:25

53

ACCEPTED MANUSCRIPT Table 2 – Olanzapine dihydrate content in stored samples (75%RH/ 25⁰C). Dihydrate conversion (%) Time (day)

Formulations (API: Polymer proportions) B2 (3:1) B1 (1:1) C1 (1:1) 0.2 ± 0.0

0.5 ± 0.2

0.3 ± 0.1

14

0.5 ± 0.1

9.5 ± 1.6

0.9 ± 0.2

21

7.8 ± 0.7

35.2 ± 1.5

10.6 ± 0.4

28

20.2 ± 0.9

55.2 ± 2.6

14.0 ± 0.7

AC

CE

PT E

D

MA

NU

SC

RI

PT

7

54

ACCEPTED MANUSCRIPT Table 3 – Olanzapine dihydrate content present in stored samples (93%RH, 25⁰C).

Dihydrate conversion (%) Time (day)

Formulation (API: Polymer proportions) B1 (1:1)

B2 (3:1)

C1 (1:1)

C2 (3:1)

D1 (1:1)

D2 (3:1)

7

8.4 ± 0.2

93.1 ± 0.4

76.1 ± 2.7

9.3 ± 0.8

8.7 ± 1.3

2.8 ± 0.5

2.0 ± 0.6

14

29.1 ± 0.2

99.4 ± 1.1

99.8 ± 1.2

25.2 ± 0.5

34.5 ± 2.1

2.6 ± 1.1

1.7 ± 0.4

21

48.2 ± 0.3

100.1 ± 0.6

100.3 ± 0.6

29.6 ± 0.8

56.2 ± 1.2

2.4 ± 1.3

1.6 ± 1.2

28

62.1 ± 0.4

99.8 ± 1.1

100.2 ± 0.4

35.0 ± 0.7

65.6 ± 1.1

2.8 ± 0.8

1.8 ± 0.6

AC

CE

PT E

D

MA

NU

SC

RI

PT

A

55

ACCEPTED MANUSCRIPT Table 4 – Viscosity of solutions of the 3 polymers considered in the study. Viscosity of the polymer solutions (mPa.s)

PEG

1.006

HPC

1.144

PVP

1.009

AC

CE

PT E

D

MA

NU

SC

RI

PT

Polymer (0.2%)

56

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Graphical abstract

57