Particle engineered inhalable dry powders of rifampicin

Abstract

Tuberculosis (TB) is an infectious disease caused by Mycobacterium tuberculosis. Main site of infection are the lungs, but infection of other organs can occur in later stages. First line therapy includes antibiotic therapy using a combination of three out of four drugs (rifampicin, isoniazid, ethambutol, and pyrazinamide), for a period of at least 26 weeks. In the case of rifampicin, fairly high daily doses of 10 mg/KGBW (600 mg max.) are to be administered, which often times is accompanied by unwanted side effects that might not only affect the patient´s overall status, but could also compromise compliance and adherence to the therapeutic regimen. In order to achieve therapeutically effective concentrations at the site of infection (the lungs) whilst minimizing systemic exposure to avoid unwanted side effects, inhalable antibiotic therapy is desirable. Much effort has already been made to develop respirable dry powder dosage forms of rifampicin, and spray drying was shown as suitable method to produces rifampicin particles with favorable aerosol performance. Two solvent free polymorphs as well as several solvates are described in literature. This work systematically investigated aerosol properties of formulations spray dried from solutions and suspension of rifampicin in common solvents, in order to assess if different (pseudo-) polymorphs show suitability for a potential aerosol therapy. In a first study, formulations spray dried from ethanol, methanol, isopropyl alcohol, acetone, dichloromethane, and water were manufactured and assessed for their aerodynamic (NGI), as well as their solid state (XRPD) properties. Suitable candidates were found in samples spray dried from ethanol, methanol, and water, yielding three crystalline and one amorphous formulation. Additionally an amorphous reference formulation showing the characteristic shape of collapsed spheres, often seen when spray drying solutions, was manufactured. Selected formulations were repeated and investigated more closely with special focus on crystallographic properties or mechanism of particle formation in the case of crystalline, or amorphous formulations respectively. All formulations investigated in the second study showed good aerosol properties with fine particle fractions (FPF, being the fraction of the dose with an aerodynamic diameter ≤ five µm, which is indicative for lung deposition) of about 40%, but one formulation spray dried from aqueous solutions of rifampicin in water showed excellent aerosol properties with FPFs as high as 90%. Crystallographic studies showed that suspensions spray dried from water are most likely a member of an isostructural series of solvates that was firstly described as rifampcin pentahydrate. Contrary, samples spray dried from ethanol and methanol were shown to be members of a common isostructural series, but could not clearly be allocated to other (pseudo-) polymorphs already reported in literature. In order to investigate the mechanism of particle formation when spray drying samples from watery or isopropylalcoholic solutions, drying kinetics of individual droplets were investigated using an acoustic levitator. It was found that in watery solution rifampicin displays early crust formation, which detaches from the liquid core that eventually is removed from the system. As a consequence remaining particles consist of the highly collapse shell only, creating a powder of low density with excellent aerosol properties. <br /> For efficient formulation design, it is highly desirable to develop suitable <i>in vitro</i> tools being predictive of <i>in vivo</i> performance of the formulation. It is known from literature that aerodynamic diameters smaller than five µm are indicative for lung deposition, but standard <i>in vitro</i> test methods as for example the NGI, which are designed to provide a suitable test system in an QC environment, fail to provide accuracy in full dose assessment necessary. Amongst others, lack of physiological relevance of the NGIs induction port, has been identified as a reason for this. Much effort has already been made to develop more biorelevant deposition models, but limited information on the impact of modifications of the IP on aerosol data using the NGI is available. Additionally, approaches most intensely investigated represent idealized physiological conditions of an averaged patient collective, which is a reasonable approach in many applications, but a more dynamic model might be desirable to investigate lung deposition in special patient populations (e.g. children or patients with pathophysiological alterations) or to investigate fundamental mechanisms of particle deposition in the relevant regions. In this study we presented a modified induction port to be used with the NGI that was manufactured based on geometries (trachea) derived from a computer tomographic scan of a patient, and 3D printed using FDM technique. In a first study it was investigated which types of aerosol formulations are critical in terms of tracheal deposition, by assessing the aerodynamic properties of preparations of salbutamol, formulated as pressurized metered dose inhaler (pMDI), dry powder inhaler, and solution for nebulization, respectively. Additionally, the impact of using a valved holding chamber was investigated. It was found that using the modified induction port offered no additional information in the case of the dry powder, and nebulizer formulation. Contrary, the pMDI formulation showed increased deposition, which consequently translated into a lower FPF, which was found to correlate better to <i>in vivo</i> data reported in literature. In a second study, additional pMDI formulations, which were selected on availability of <i>in vivo</i> data of lung deposition in literature, were assessed and results of <i>in vitro</i> / <i>in vivo</i> (IVIVC) correlation were found more accurate than ones obtained using the regular induction port, which is monographed in the European and United States pharmacopoeia. Additionally, another modified induction port was manufactured to get a more accurate understanding of the factors responsible for increased tracheal deposition in certain formulations, and it was confirmed that adding physiologically more accurate geometries to the induction port is beneficial for IVIVC. It was concluded that besides a physiologically relevant trachea, the next model should also include more biorelevant mouth/throat region, as this was found relevant in certain formulations. <br /> The novel induction port was also used to estimate <i>in vivo</i> performance of the rifampicin formulations investigated. Though increased deposition in the induction port, which correlated to a decrease in expected lung deposition, was found, differences were found at a magnitude, which probably cannot be resolved in an <i>in vivo</i> setting. Thus good to excellent aerosol performance achieved could translate into good performance <i>in vivo</i>.