European Journal of Pharmaceutical Sciences 19 (2003) 211221 www.elsevier.

com / locate / ejps

Characterisation and deposition studies of engineered lactose crystals with potential for use as a carrier for aerosolised salbutamol sulfate from dry powder inhalers
a a, b a Hassan Larhrib , Gary P. Martin *, David Prime , Christopher Marriott
a

Department of Pharmacy, King s College London, 150 Franklin-Wilkins Building, Stamford Street, London SE1 9 NN, UK b GlaxoSmithKline, Park Road, Ware, Hertfordshire SG12 0 DP, UK Received 30 July 2002; received in revised form 21 March 2003; accepted 9 April 2003

Abstract Lactose particles with different elongation ratio, roundness, polymorphic form and crystallinity were prepared by a one-step crystallisation process using varying ratios of acetone / water. The crystals were characterised using image analysis optical microscopy, scanning electron microscopy, differential scanning calorimetry and X-ray powder diffraction. The elongation ratio was found to increase with increasing acetone ratio which therefore, appears to accelerate the growth in length rather than width and / or thickness. The crystallinity and polymorphic forms were also acetone-concentration dependent. For example, the crystals formed using 6580% v / v acetone were almost all of the a-form whereas at 85% v / v a small amount of b-form was precipitated, as detected by a peak at the reection angle 2u 510.4 in the X-ray diffractogram. When 90% v / v acetone was incorporated a mixture of a- and b-forms were produced in almost equal quantity, whereas, with 95% v / v acetone the b-form predominated. At high acetone concentration (90 and 95% v / v), the crystallisation proceeded rapidly leading to the creation of some amorphous content. The 6390-mm sieve cut of either commercial grade lactose (CL) or crystallised lactose was mixed with salbutamol sulfate and dispersibility was determined using the twin stage liquid impinger. All the formulations containing carrier particles generated by crystallization from solvent showed higher dispersibility and ne particle fraction (FPF) of the drug compared to the formulation made containing CL. The carrier that showed the highest elongation ratio (produced from an 85% acetone 15% water solution), when mixed with salbutamol sulfate produced the highest dispersibility (38.5%) and highest FPF (29.24%). These parameters were six times higher than the values obtained with the formulation containing CL. 2003 Elsevier B.V. All rights reserved.
Keywords: Dry powder inhalers; Lactose; Polymorphic forms; Crystallinity; Elongation ratio; Salbutamol sulfate; Dispersibility; Fine particle fraction

1. Introduction Dry powder inhalation aerosols (DPIs) are generally formulated by mixing cohesive, micronised drug particles (size 15 mm), which are often dosed only in microgram quantities, with large carrier particles (size 50200 mm). The carrier particles should be physiologically inert, not impair the drug bioavailability, be chemically compatible with the drug and be able to aid the ow and dispersion of the highly cohesive drug particles. The majority of previous studies have focused on lactose-based formulations because lactose, as a carrier, fulls most of the require* Corresponding author. Tel.: 1 44-20-7848-4791; fax: 1 44-20-78484800. E-mail address: gary.martin@kcl.ac.uk (G.P. Martin). 0928-0987 / 03 / $ see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016 / S0928-0987(03)00105-2

ments discussed above, and additionally its hygroscopicity is lower compared to other sugars, such as sorbitol, dextrose and maltose (Macritchie, 1998). In addition, lactose is available in many grades and thus it remains the carrier of choice for the formulation of dry powder aerosol formulations. However, the fraction of deeply respirable drug particles from most carrier-based DPI formulations at present is relatively low since sometimes as little as 10% of the total administered from a DPI dose is delivered to the lower airways, the site of action for most aerosolised drugs. Various attempts to improve the delivery efciency of drug particles to the lung have been made including: the use of different grades of lactose (Larhrib et al., 1999a), combining different grades of lactose (Karhu et al., 2000), adding ternary materials to the blends (Ganderton, 1992; Staniforth, 1996; Zeng et al., 1998; Tee et al., 2000) and

212

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221

reducing the carrier particle size (French et al., 1996; Srichana et al., 1998). However, work has not only focused on simple processing of the powder in an effort to improve its efciency of delivery but also upon the approach of engineering or modifying the crystal habit or particulate structure of therapeutic agents and carriers. For example, most commercially available carriers used in DPI formulations present irregularities or crevices on their surfaces (Larhrib et al., 1999a; Tee et al., 2000). These crevices are sometimes larger than the drug particles, and any drug entrapped is unlikely to be dislodged by energy imparted by the patients inhalation, resulting in that drug being unavailable for deposition in the respiratory tract. Considerable efforts have been made to engineer carrier lactose particles with improved surface texture and crystallinity compared to that routinely available commercially. For example, reducing the rugosity of lactose by means of crystallization from Carbopol gels, resulted in a more reproducible delivery of salbutamol sulfate as compared with control lactose (Zeng et al., 2001). The overall carrier morphology has also been shown to affect the deposition proles of salbutamol sulfate from DPIs, for example, increasing the elongation ratio of lactose, improved the ne particle fraction (FPF) of the drug (Zeng et al., 2000a). The lactose used in an earlier study was prepared using different crystallisations processes (Zeng et al., 2000b). The use of ethanol as a non-solvent to engineer needle shaped lactose crystals as a carrier has been shown to increase the deposition proles of salbutamol sulphate in vitro by more than ve times compared to the commercial grade lactose (Larhrib et al., 1999b). The principal aims of the present study were to prepare and characterise lactose particles by a crystallization process using a different non-solvent, namely acetone. It was then proposed to employ crystals obtained from the various different lactose batches to produce a series of formulations containing salbutamol sulfate as a model drug with a view to determining the effects of the physical characteristics of the carrier particle on drug dispersion and deposition in vitro.

Acetone technical grade was obtained from BDH, Poole, UK.

2.1. Crystallisation of lactose

Lactose (30 g) was dissolved in distilled water (300 ml) at 55 8C. After cooling to room temperature, the lactose solution was added to acetone without stirring such that the nal proportions of acetone, in combination with the 10% w / v lactose solutions, were 65, 75, 85, 90, 95% v / v of the 100-ml total volume. Immediate precipitation was observed as the concentration of acetone increased from 80 to 95% v / v, whereas the remaining solutions, i.e. 6575% acetone remained initially clear. The beakers containing different concentrations of acetone / lactose solution were covered tightly with paralm and left unstirred for 12 h. After that time interval, the resultant crystals were ltered and allowed to dry in an oven at 55 8C overnight. The dried crystals were poured into a 90-mm sieve, which had been placed on a 63-mm sieve. The particles were then sieved manually and slowly for 1 h so as not to rupture any crystals. The crystals (6390 mm) were then transferred to a sealed vial and placed in a desiccator over silica gel until required for further investigation. Commercial grade lactose (CL) was sieved in the same manner as a control and the various 6390-mm size fractions were then characterised and employed to produce the formulations.

2.2. Characterisation of particle size and shape of lactose crystals by image analysis optical microscopy and scanning electron microscopy
A small amount of powder was suspended in mineral oil (Sigma Chemical Co, St. Louis, USA) and the suspension was spread onto a microscope slide. A cover slip was applied, allowing the suspension to settle homogeneously between the two glass surfaces. Particle size and shape were assessed in tandem using image analysis software (designed in-house at Kings College London) installed on an Archimedes computer, which was attached to an optical microscope (Nikon Labophot, Tokyo, Japan) via a miniature video camera. Four hundred particles were measured in each sample and the surface volume mean diameter, roundness and elongation ratio were recorded, the latter two factors being dened as follows: Perimeter 2 Roundness 5 ]]]] 4 3 p 3 Area Maximum Feret Diameter Elongation ratio 5 ]]]]]]] Minimum Feret Diameter (1) (2)

2. Materials and methods Lactose [a-lactose monohydrate (Lactochem lactose, Batch no. S648090)] was obtained from Borculo Whey Ltd., Chester, UK whilst b-lactose (Batch 47H0490) was obtained from Sigma Chemical Co, Poole, UK. Micronised salbutamol sulfate (Batch no. 540330, BP 1993-USP XXII) was purchased from Allchem International, Maidenhead, Berkshire, UK and was further micronised using an air jet microniser (JM-80, M & M Fryma Ltd., Herts, UK) with the nozzle pressure set to 6 bar. Ventolin Rotahaler and gelatin capsules (size 3) were supplied by GlaxoSmithKline, Research and Development, Ware, UK.

where the Minimum and Maximum Feret Diameters were calculated from 16 calliper measurements at 68 intervals around the particle. These two measurements were not necessarily at right angles to each other. For the 10 3

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221

213

objective, employed in these studies, the pixel resolution of the digitised image used for measurement was 1.13 mm per pixel in the x axis and 2.26 mm per pixel in the y axis. Double-sided adhesive tape was placed on an aluminium stub and after stripping off the protective covering, a small number of particles were scattered on the stub and dispersed by tapping lightly on the edge of the stub with a spatula to break up any agglomerates. The particles were then coated with approximately 1520 nm gold using a sputter coater (Polaron E5100, Polaron Equipment Ltd., Watford, UK) at an electrical potential of 2.0 kV and a current of 20 mA. Several photomicrographs were produced by scanning elds, selected randomly at different magnications under a Philips SEM501B scanning electron microscope (Einhoven, Holland).

4000 Multiple Solvent Delivery System, LDC Analytical Inc., FL, USA), a multiple wavelength UV detector (SpectroMonitor 3100, LDC Analytical Inc., FL, USA) and a 30 cm 3 4.6 mm i.d. column packed with 5 mm Novapack C18 (Waters, Milford, MA, USA), which was maintained at 60 8C. The retention times for salbutamol sulfate and the internal standard were 6 and 10.6 min, respectively.

2.6. Measurement of dose uniformity

The homogeneity of the blends was examined by analysing the quantity of salbutamol sulfate in aliquots (3360.8 mg) of sampled powder, the amount of powder in each capsule. Each aliquot of the blend was placed in a 100-ml volumetric ask and made up to volume with the HPLC mobile phase containing the internal standard. Ten aliquots were taken randomly from each blend and each solution was assayed in duplicate using the HPLC method described above. The mean actual drug content (determined by assay) expressed as a percentage of the expected drug content was determined (% uniformity) and the coefcient of variation (% CV) of these values was used to assess the homogeneity of the blends.

2.3. Characterisation of lactose particles by differential scanning calorimetry

Differential scanning calorimetry (DSC) was employed to determine the crystal form of lactose. The calorimeter used was a Mettler TA 4000 (Mettler Instruments AG, CH-8608 Greifensee, Switzerland) thermal analysis system, with a DSC20 furnace. Thermograms were analysed using Mettler GraphWare TA72PS.1 software. An empty aluminium pan (40 ml) was used as the reference for all measurements. The instrument was calibrated using tin, indium and gallium as standard materials. A sample (26 mg) was weighed into an aluminium pan, which was hermetically sealed and placed in the pre-equilibrated DSC furnace (25 8C). Each sample was allowed to equilibrate for 5 min at 25 8C before being heated to 300 8C at a heating rate of 10 8C / min. Each sample was analysed in triplicate.

2.7. Deposition test of salbutamol sulfate

Deposition of salbutamol sulfate from each blend was determined using a twin stage liquid impinger after aerosolisation of three capsules at 60 l / min via a Rotahaler. Seven millilitres and 30 ml, respectively, of the mobile phase containing the internal standard was introduced into the upper and lower stages of a twin stage liquid impinger (British Pharmacopaeia, 2000). The capsule to be tested was placed in the inhaler device (Rotahaler , GlaxoWellcome, Ware, UK), which had been tted into a moulded rubber mouthpiece attached to the throat piece of the impinger. Once the assembly had been checked and found to be airtight and vertical, the pump was switched on, the dose released and the pump allowed to run for 7 s at 60 l / min before switching off. The capsule shell was then removed from the inhaler device and the deposition test was repeated so that two more capsules were actuated in the same manner. The capsule shells were washed ve times with the mobile phase containing internal standard and the solution was made up to a xed volume (50 ml). The inhaler device was washed with the same solvent and the solution made up to volume (50 ml). The upper and lower stages of the twin stage impinger were washed individually and the solution was made up to volume (100 ml). All the samples obtained were analysed for salbutamol sulfate using the HPLC method outlined above. Deposition of salbutamol sulfate from each formulation was determined at least ve times and a number of parameters were employed to characterise the deposition

2.4. Preparation of powder formulations

Salbutamol sulfate and lactose were mixed in a ratio of 1:67.5 w / w in accordance with the ratio employed in commercial Ventolin Rotacaps. Stoppered vials (20 ml), containing the separate blends of salbutamol sulfate with lactose (5 g total powder weight), were placed in a Turbula mixer (Glen Greston Ltd., Middlesex, UK) and mixing was carried out for 30 min at 42 rev / min. All blends were then manually lled into hard gelatin capsules (size 3) such that each capsule contained 481.75611.68 mg salbutamol sulfate.

2.5. HPLC analysis of salbutamol sulfate

Salbutamol sulfate was analysed by HPLC employing a mixture of methanol and 0.25% w / v 1-heptane sulfonic acid sodium salt (40:60, v / v) as the mobile phase running at a ow rate of 0.9 ml / min, p -hydroxybenzoic acid ethyl ester (1 mg / ml) as an internal standard and UV detection at 238 nm. The HPLC system consisted of a pump (CM

214

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221

proles of the drug. The recovered dose (RD) was the sum of the weights of drug ( mg) recovered from the capsule shells, the inhaler device and the upper and lower stages of the twin stage impinger, whilst the emitted dose (ED) was the dose emitted from the inhaler device and depositing in the upper the upper and lower stages of the twin stage impinger. Fine particle dose (FPD) was the amount of drug recovered from the lower stage of the impinger, which has a diameter less than the cut-off diameter of the upper stage of a twin stage impinger (drug particles , 6.4 mm at an air ow rate of 60 l / min). The ne particle fraction (FPF) was calculated as the ratio of the FPD to RD and the dispersibility as the ratio of FPD to ED (both expressed as a percentage). The percent recovery was calculated as the ratio of RD to the expected dose of salbutamol sulfate in the capsules and the percent emission as the ratio of ED to RD (both also expressed as percentages).

growth of the longest axis of the crystals at the expense of an increase in width and thickness. This hypothesis, developed from image analysis optical microscopy data (Table 1) and was supported by the qualitative assessment of scanning electron micrographs (Fig. 1). The roundness value also generally showed a concomitant increase as acetone concentration was increased (Table 1), although this factor also reached a maximum when 85% v / v acetone was employed.

3.2. Characterisation of polymorphic forms of lactose

Fig. 2 shows the DSC scans of a-lactose monohydrate, b-lactose and crystallised lactose using different combinations of acetone / lactose solutions. Fig. 2be show DSC scans similar to that of a-lactose monohydrate (Fig. 2a). The endothermic transition starting at about 130 8C corresponds to the dehydration of crystallisation water, whereas the endothermic peak at about 225 8C is the melting endotherm of a-lactose monohydrate. However, the samples crystallised using 90 and 95% acetone (Fig. 2f,g) contain some b-lactose. This is inferred from the DSC scan of pure b-lactose (Fig. 2a), which showed a melting peak at about 240 8C. The X-ray diffraction patterns of a-lactose monohydrate, crystallised lactose and pure b-lactose are shown in Fig. 3. The reection intensity at the angle 2u 5 10.4 corresponds to b-lactose (Fig. 3h). It can be seen that the samples prepared from the solutions containing 65, 75 and 80% acetone do not show such a peak, suggesting a lack of b-lactose in these samples. A small amount of b-lactose was apparent in the sample prepared from 85% acetone, which could not be detected using DSC. The samples prepared from 90 and 95% acetone showed more b-lactose (Fig. 3f,g) and this concurred with the DSC data. The samples crystallised from 90 and 95% acetone also showed a broad and small endothermic transition at about 130 8C corresponding to the dehydration of water of crystallisation (Fig. 2f,g). The number of moles of water per mole of anhydrous lactose (n) was calculated for all the samples (Table 2) using Eq. (3) (Khankari et al., 1992): DHd x RMM lactose n 5 ]]]]]]] sDHv 2 DHddx RMM water (3)

3. Results and discussion

3.1. Characterisation of particle size and morphology of lactose

The various batches of lactose crystals, prepared from solutions containing differing acetone concentrations, were of different shape, size and surface texture, as can be seen from the scanning electron micrographs (Fig. 1). For example, the crystals obtained from 65% v / v acetone were either tomahawk or pyramid shaped, whereas when 75% acetone was used, elongated particles with a rough surface were formed; in the latter instance, it is possible that the crystals formed initially served as a nucleus on which other crystals grew leading to the formation of numerous needles on the surface of the primary crystals. The crystals prepared using 80 and 85% acetone were less rough and more regular in shape compared to those obtained using 75% acetone as the non-solvent. When 90 and 95% acetone was employed, individual crystals were formed with smooth surfaces but the crystals were porous and more brittle, as indicated by the presence of a larger number of fractured faces (Fig. 1). The concentration of acetone also had effects on the size, roundness and elongation ratio of the resultant particles (Table 1). The particle size generally decreased with increasing concentration of acetone. For example, the mean particle diameter decreased from 109.77 mm when acetone was present at 65% v / v to 31.09 mm at 95%, when upon addition of acetone a cloudy solution was rapidly formed. The elongation ratio of lactose particles was clearly dependent upon the concentration of acetone in the crystallization medium. Increasing the concentration of acetone increased the elongation ratio (Table 1) of lactose particles, up to 85% v / v acetone. Therefore, generally higher concentrations of acetone would appear to accelerate the

where DHd is the enthalpy of dehydration obtained from the dehydration endotherm (J / g); DHv is the enthalpy of vaporisation of water, 2261 J / g (Stark and Wallace, 1976), and RMM lactose and RMM water are the relative molecular masses of anhydrous lactose (340.3) and water (18.0), respectively. The lactose crystallised using 90 and 95% of acetone appeared to contain less than 1 mol water of crystallisation per mole anhydrous lactose (Table 2). This was due to a decrease in the amount of a-lactose monohydrate and an increase in the b-lactose form as shown by DSC (Fig.

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221

215

Fig. 1. Scanning electron micrographs of lactose crystals and recrystallised lactose using different combinations of acetone / lactose solution.

216

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221

Fig. 1. (continued )

2f,g). In addition these fractions of lactose would also appear to have an increased amorphous content, as suggested by the small deection in the base line of the DSC thermogram (Fig. 2f) and the appearance of a small exothermic peak at about 172 8C (Fig. 2g), which can be attributed to the crystallisation of amorphous lactose (Lerk et al., 1994; Darcy and Buckton, 1997). There was a concomitant qualitative change in the physical characteristics of the lactose as higher concentrations of acetone were employed since a more brittle material was produced at 90 and 95% levels of solvent.
Table 1 The surface-volume mean diameter, roundness and elongation ratio for sieved (6390 mm) commercial lactose (CL) and sieved recrystallised lactose crystals measured by optical microscopy image analysis (n 5 400) Carrier (6390 mm) CL 65% Acetone / 35% LS 75% Acetone / 25% LS 80% Acetone / 20% LS 85% Acetone / 15% LS 90% Acetone / 10% LS 95% Acetone / 5% LS Diameter ( mm) 106.12 109.77 64.67 64.77 80.45 47.34 31.09 Roundness 1.4460.02 1.6060.09 1.8460.07 2.1660.11 2.5960.09 2.0660.13 2.2160.17 Elongation ratio 1.6960.05 1.9660.09 2.3860.11 2.7860.75 3.7060.42 2.5660.23 3.3460.85

The crystallinity of lactose crystallised from 95% acetone / 5% lactose solution was calculated from the heat of crystallisation of amorphous lactose, obtained from the DSC thermograms (Saleki-Gerhardt et al., 1994). Given the specic heat of crystallisation of amorphous lactose, DHc (J / g), then the content of amorphous lactose (% amorphous w / w) in the crystalline lactose can be calculated using Eq. (4): Dh c % amorphous lactose 5 ]] 3 100 DHc (4)

where Dh c is the heat of crystallisation of amorphous lactose in J / g lactose crystals and DHc is taken as 32 J / g for amorphous lactose (Sebhatu et al., 1994). The heat of crystallisation of lactose crystallised from 95% acetone / 5% lactose solution was 4.7 J / g, which on the basis of Eq. (4) corresponds to 14% amorphous lactose.

3.3. Content uniformity

Table 3 shows the percentage uniformities and coefcient of variation (% CV) in salbutamol sulfate content obtained for the formulations containing lactose crystals and recrystallised lactose. The salbutamol sulfate content of samples varied from 93.1761.56 for the formulation containing lactose prepared from 65% acetone as a carrier

Recrystallisation was carried out from solutions containing different proportions of acetone and 10% w / v lactose solution (LS).

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221

217

Fig. 2. DSC thermograms of (a) a-lactose monohydrate and b-lactose, and lactose prepared from solutions containing acetone at the following concentration (b) 65% v / v, (c) 75% v / v, (d) 80% v / v, (e) 85% v / v, (f) 90% v / v, and (g) 95% v / v. The original concentration of lactose was 10% w / v.

to 106.1962.44% for the formulation containing lactose prepared from 75% acetone. All the formulations showed a very low coefcient of variation (% CV), which was less than 2.7%. These results suggest that the overall process of mixing, sampling and analysis was accurate and reproducible, and a uniform mixing was achieved using the employed mixing procedure.

3.4. Deposition proles of salbutamol sulfate from formulations containing lactose crystals and recrystallised lactose
The deposition data presented in Table 4 were calculated after the actuation of the contents of three capsules at 60 l / min into a two-stage twin impinger. It can be seen from

218

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221

Fig. 3. X-Ray diffraction patterns of (a) a-lactose monohydrate and lactose prepared from solutions containing acetone at the following concentration (b) 65% v / v, (c) 75% v / v, (d) 80% v / v, (e) 85% v / v, (f) 90% v / v and (g) 95% v / v, as well as (h) b-lactose (graph A). The original concentration of lactose was 10% w / v.

Table 4 that powder formulations containing lactose crystals and recrystallised lactose as the carrier produce differences in the deposition of salbutamol sulfate. The recovered dose (RD) of salbutamol sulfate varied from 456 mg for the formulation containing lactose crystals as the carrier to 495 mg for the blend that incorporated lactose

crystallised from a solution containing 95% acetone / 5% lactose solution. The RD from all formulations were within the range of the mean62 S.D. of the theoretical dose (482612 mg), suggesting that the method of washing and analysing was accurate and reproducible. The emission of salbutamol sulfate from the Rotahaler varied from 71%

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221

219

Table 2 The enthalpy of dehydration and moles of water of crystallisation per mole of anhydrous lactose of sieved commercial lactose (CL) and recrystallised lactose obtained from the DSC thermograms Sample lactose (n) CL 65% Acetone / 35% LS 75% Acetone / 25% LS 80% Acetone / 20% LS 85% Acetone / 15% LS 90% Acetone / 10% LS 95% Acetone / 5% LS Enthalpy of dehydration (w / g) 141.2263.81 142.4763.63 140.1561.51 141.1261.58 138.5262.49 105.3064.15 94.4262.98 Moles of water / mole of anhydrous lactose 1.2660.04 1.2760.03 1.2560.01 1.2660.01 1.2360.03 0.9260.04 0.8260.02

Recrystallisation was carried out from solutions containing different proportions of acetone and 10% w / v lactose solution (LS).

Table 3 % Uniformity (mean6S.D.) and coefcient of variation (CV) in salbutamol sulfate content obtained from the formulations containing sieved commercial lactose (CL) and recrystallised lactose (n 5 10) Carrier (6390 mm) CL 65% Acetone / 35% LS 75% Acetone / 25% LS 80% Acetone / 20% LS 85% Acetone / 15% LS 90% Acetone / 10% LS 95% Acetone / 5% LS % Uniformity 98.2061.14 93.1761.56 106.1962.44 95.0162.50 100.9261.78 101.8461.99 105.9561.38 % CV 1.16 1.67 2.30 2.63 1.76 1.95 1.30

Recrystallisation was carried out from solutions containing different proportions of acetone and 10% w / v lactose solution (LS).

for the formulations containing recrystallised lactose, prepared using 90% or 95% acetone to 94% for the blend composed of recrystallised lactose, prepared using 80% acetone. The emission of the drug from the inhaler device was signicantly lower when the formulation containing lactose prepared using 95% acetone was employed. These results indicate that approximately 30% RD of the drug from the formulation containing the lactose recrystallised from 90 and 95% acetone was retained in the capsules and / or in the inhaler device. The lower emission of salbutamol sulfate from the latter formulation could be due in part to a decrease in the ow properties since the lactose particles are elongated. In addition the batches of lactose

precipitated using 90 or 95% acetone had the highest b-lactose content as well as the greatest amorphous contents. These factors are likely to affect the interaction of drug with the carrier, which in turn will affect the resultant deposition prole. The blend containing CL produced a markedly lower dispersibility of salbutamol sulfate compared to all formulations containing recrystallised lactose (Table 4), indicating that most of the emitted dose of salbutamol sulfate was deposited in the upper stage of the twin stage impinger, allowing only a small fraction of the salbutamol sulfate to reach the lower stage. This was mirrored by the concomitantly lower values of FPD (2367 mg) and FPF (5.161.6%) of the drug (Table 4). Despite the lower emission of the drug from the formulation containing 95% acetone, a higher FPF (27.161.6%) was produced due to a higher proportion of the emitted dose depositing in the lower stage of the twin stage impinger compared with that from the formulation containing CL (Table 4). One way analysis of variance showed that the formulations produced from crystallised lactose prepared using 85% acetone produced a higher FPF of drug than formulations composed from the batch of lactose prepared using lower concentrations of acetone. There was no signicant difference (ANOVA, P . 0.05) in FPF from the blends composed of crystallised lactose prepared using 75% and 80% acetone and no signicant difference (ANOVA, P . 0.05) in FPF between the blends containing crystallised lactose

Table 4 Recovered dose (RD), emitted dose (ED), ne particle dose (FPD), ne particle fraction (FPF), dispersibility, recovery and emission of salbutamol sulfate from formulations containing either sieved commercial lactose (CL) or recrystallised lactose (mean6S.D., n 5 5) Lactose (6390 mm) CL 65% Ac / 35% LS 75% Ac / 25% LS 80% Ac / 20% LS 85% Ac / 15% LS 90% Ac / 10% LS 95% Ac / 5% LS RD ( mg) 456610 471614 48563 464629 47467 474620 49567 ED ( mg) 365626 368612 398611 434622 359612 337622 351611 FPD ( mg) 2367 9863 12068 106611 139612 12766 134610 FPF (%) 5.161.6 20.761.1 24.861.7 22.860.9 29.262.2 26.961.8 27.161.6 Dispersibility (%) 6.462.1 26.561.2 30.262.1 24.461.2 38.562.2 37.860.8 38.161.8 Recovery (%) 93.062.1 105.063.1 94.960.7 102.066.3 97.461.5 96.564.0 96.961.4 Emission (%) 80.167.0 78.162.2 81.961.8 93.661.1 75.861.6 71.265.1 71.061.8

Recrystallisation was carried out from solutions containing different proportions of acetone (Ac) and 10% w / v lactose solution (LS).

220

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221

obtained using the highest acetone concentrations (85%, 90% and 95% acetone). The FPD and FPF of salbutamol sulfate generally increased with increasing acetone ratio and reached an optimum at the formulation composed of crystallised lactose prepared using 85% acetone (FPD 5 139612 mg, FPF 5 29.262.2%). The difference in the deposition prole from the formulations containing commercial lactose crystals and recrystallised lactose could be attributed partially to the difference in the morphological properties, with all the recrystallised lactose having a higher elongation ratio than lactose crystals (Table 1). Zeng et al. (2000a,b) have shown that the FPF of salbutamol sulfate increases as a function of elongation ratio. This can be explained by the fact that elongated or needlelike particles have aerodynamic diameters almost independent of their length and the diameter is approximately equal to the short dimension of the particle in question (Hinds, 1982). Such particles thus exhibit smaller aerodynamic diameters than spherical particles of similar mass or volume as shown by Hickey et al. (1992). More elongated lactose particles may also be expected to travel a longer distance before impaction occurs in comparison to less elongated carrier particles of similar mass, as a result of the lower relative aerodynamic diameters of the former. More drug particles can adhere to elongated carrier particles in comparison to those that are more spherical and these particles will be subjected to the drag forces of the air stream for a longer period of time. This would be expected to result in a higher proportion of drug being detached from the carrier particle, leading to a higher FPF of the drug (Zeng et al., 2000a,b).

4. Conclusion Lactose particles were prepared using a one-step crystallisation process from acetone / water cosolvent mixtures using acetone as a precipitating agent and water as a solvent. Lactose particles with different elongation ratio, roundness, polymorphic form, particle size and crystallinity resulted as a consequence of varying the proportion of acetone / water (v / v). The crystal growth of lactose using 6575% v / v acetone was slow, with the particles obtained from 65% acetone being either tomahawk- or pyramidshaped. Above 65% acetone needle-shaped lactose was obtained and the elongation ratio and roundness of the particles increased with increasing concentration of acetone. Therefore, an increased acetone concentration appeared to accelerate the growth of length rather than the width and thickness of the crystal. The lactose in the needle-shaped crystals (prepared using concentrations of acetone up to 80%) existed only in the a form as revealed by DSC and X-ray diffraction studies. However, when crystallization took place from 85% acetone, evidence of a small amount of b form was apparent from the X-ray diffractogram but this could not be detected by DSC.

Using 90% acetone a mixture of a and b forms were produced in almost equal proportions (as determined by DSC) whereas, at 95% acetone the b-form was predominant. At 95% acetone, the crystallisation was forced rapidly leading to the formation of some amorphous content (14%) as detected by DSC and the crystals appeared to be brittle in character. After blending with salbutamol sulfate, all the formulations produced showed a very low coefcient of variation (% CV) suggesting that the overall process of mixing, sampling and analysis was accurate and reproducible. Both the elongation ratio and FPF of salbutamol sulfate were found to increase as a function of the amount of acetone used to prepare the lactose crystals with an optimum concentration being approximately 85% v / v. The inuence of elongation ration of the carrier particles on FPF appeared to be more marked than any effects attributable to surface smoothness (as assessed visually). All the formulations containing the engineered lactose produced higher dispersibility and FPF than the formulation incorporating commercial lactose. The carrier that showed the highest elongation ratio (85% acetone / 15% lactose solution) when mixed with salbutamol sulfate produced the highest dispersibility (38.5%) and highest FPF (29.2%). Both these parameters were found to be six times higher than the corresponding values for dispersibility (6.4%) and FPF (5.1%) obtained from the formulation containing commercial lactose. It can be concluded that the use of different concentrations of non-solvent on the controlled crystallization of lactose can lead to particles that might vary not only in particle size and shape, but also in crystal habit and degree of crystallinity. Such manipulation of carrier particles might be a viable means of improving the drug delivery to the lower airways from dry powder inhalers.

Acknowledgements This work was sponsored by GlaxoSmithKline, Ware, UK. The authors would also like to thank Dr Bob Lancaster (GlaxoSmithKline) for his help in X-ray powder diffraction and Dr Tony Brain from the EM unit (Kings College, London) for scanning electron microscopy.

References
British Pharmacopaeia, 2000. Aerodynamic Assessment of Fine ParticlesFine Particle Dose and Particle Size Distribution. Her Majestys Stationery Ofce; Appendix XII F: 3. Darcy, P., Buckton, G., 1997. The inuence of heating / drying on the crystallization of amorphous lactose after structural collapse. Int. J. Pharm. 158, 157164. French, L., Edwards, D.A., Niven, R.W., 1996. The inuence of formulation on emission, deaggregation and deposition of dry powders for inhalation. J. Aerosol Sci. 27, 769783.

H. Larhrib et al. / European Journal of Pharmaceutical Sciences 19 (2003) 211221 Ganderton, D., 1992. The generation of respirable cloud from coarse powder aggregates. J. Biopharm. Sci. 3, 101105. Hickey, A.J., Fults, K.A., Pillai, R.S., 1992. Use of particle morphology to inuence the delivery of drugs from dry powder aerosols. J. Biopharm. Sci. 3, 107113. Hinds, W.C., 1982. Aerosol Technology: Properties, Behaviour and Measurement of Airborne Particles. Wiley Interscience, New York. Karhu, M., Kuikka, J., Kauppinen, T., Bergstrom, K., Vidgren, M., 2000. Pulmonary deposition of lactose carriers used in inhalation powders. Int. J. Pharm. 196, 95103. Khankari, R.K., Law, D., Grant, D.J.W., 1992. Determination of water content in pharmaceutical hydrates by differential scanning calorimetry. Int. J. Pharm. 89, 117127. Larhrib, H., Zeng, X.M., Martin, G.P., Pritchard, J., 1999a. The use of different grades of lactose as a carrier for aerosolised salbutamol sulphate. Int. J. Pharm. 91, 114. Larhrib, H., Martin, G.P., Marriott, C., Prime, D., 1999b. The use of engineered lactose crystals as a carrier for aerosolized salbutamol sulphate from dry powder inhalers. Proceedings of Drug Delivery to the Lungs X (London), December, pp. 155159. Lerk, C.F., Andreal, A.C., De Boer, A.H., 1994. Alteration of lactose during differential scanning calorimetry. J. Pharm. Sci. 73, 856857. Macritchie, H.B, 1998. The Development of a Laser Diffraction Technique for Use in the Physical Characterization of Dry Powders for Inhalation, PhD thesis, University of London. Saleki-Gerhardt, A., Ahlneck, C., Zogra, G., 1994. Assessment of disorder in crystalline solids. Int. J. Pharm. 101, 237247.

221

Sebhatu, T., Angberg, M., Ahlneck, C., 1994. Assessment of the degree of disorder in crystalline solids by isothermal microcalorimetry. Int. J. Pharm. 104, 135144. Srichana, T., Martin, G.P., Marriott, C., 1998. On the relationship between drug carrier deposition from dry powder inhalers in vitro. Int. J. Pharm. 167, 1323. Stark, J.G., Wallace, H.G., 1976. In: Chemistry Data Book, SI Edn. John Murry, London, p. 122. Staniforth, J.N., 1996. Improvements in dry powder inhaler performance: surface passivation effects. Proceedings of Drugs Delivery to the Lungs VII, pp. 8689. Tee, S.K., Marriott, C., Zeng, X.M., Martin, G.P., 2000. The use of different sugars as ne and coarse carriers for aerosolised salbutamol sulphate. Int. J. Pharm. 208, 111123. Zeng, X.M., Martin, G.P., Tee, S.K., Marriott, C., 1998. The role of ne particle lactose on the dispersion and deaggregation of salbutamol sulphate in an air stream in vitro. Int. J. Pharm. 176, 99110. Zeng, X.M., Martin, G.P., Tee, S.K., Marriott, C., Pritchard, J., 2000a. The inuence of carrier morphology on drug delivery by dry powder inhalers. Int. J. Pharm. 200, 93106. Zeng, X.M., Martin, G.P., Marriott, C., Pritchard, J., 2000b. The effects of carrier size of the dispersion of salbutamol sulphate after aerosolization at different ow rates. J. Pharm. Pharmacol. 52, 12111221. Zeng, X.M., Martin, G.P., Marriott, C., Pritchard, J., 2001. The use of lactose recrystallised from Carbopol gels as a carrier for aerosolised salbutamol sulphate. Eur. J. Pharm. Biopharm. 51, 5562.