(T0930-03-14) Role of Excipients on Drug Release from Long-Acting Intrauterine Systems

Purpose: The need for efficacious long-term reversible contraception, improved patient-compliance, and safety has motivated the development of levonorgestrel intrauterine systems (LNG-IUSs) to improve women’s health. IUSs are a combination of a monolithic (drug-polymer matrix) and a reservoir system (covered with an outer polymer membrane) that provide drug release for up to 3-7 years. Incomplete understanding of the drug release behavior and the impact of excipients are some of the long-standing challenges in the development of IUSs, which have likely contributed to the limited availability of these drug products on the market. Accordingly, the objectives of this work were to investigate the impact of excipients on formulation attributes and in vitro drug release from LNG-IUSs, elucidate drug release mechanisms, and thereby obtain improved product understanding.

Methods: IUSs with a drug loading of 50% w/w LNG were prepared by mixing the drug and polydimethylsiloxane (PDMS) prepolymers using a planetary mixer and subsequent processing via a custom-designed extrusion device containing polycarbonate molds. The formulations were cured at 150°C to ensure complete crosslinking and were covered with an outer PDMS membrane. Four formulations each containing a different excipient viz: resin, silica, silicone oil, and diatomaceous earth (DE) were prepared. In addition, formulations with different silica content and silicone oil viscosities were prepared. Drug loading was determined via UPLC. The microstructure and molecular properties of the formulations were characterized using DSC, TGA, SEM, and ¹³C solid-state NMR. The glass transition temperature (Tg) and moduli properties were determined using dynamic mechanical analysis (DMA). Investigation of real-time in vitro release kinetics was conducted for up to 4 months.

Results: The apparent drug loading of the formulations containing DE was slightly lower than the formulations containing resin, silica, and silicone oil (Fig. 1A). Interestingly, the addition of different excipients significantly altered the crystallization kinetics of the crosslinked polymer. The lowest polymer crystallization temperature (Tc) was recorded for the formulation containing DE, indicating high crystallinity (Fig. 1B). On the other hand, the absence of a crystallization exotherm in the formulations containing resins indicated that the polymer was largely in an amorphous state. The lowest weight loss and degradation were recorded for the formulations containing resins. This was attributed to their highly branched hydrophobic structure. All the excipients also affected the IUS microstructures as evident from the SEM images (Fig. 1C). Formulations containing resins had the highest porosity followed by those containing silica. Distinct drug particles were seen on the surface of the formulations containing DE with minimal or no porosity. Incorporating silicone oil (lubricant) formed smooth polymer matrices coating the drug particles (which were observed as bumps on the surface). The in vitro release profiles revealed the following rank order of drug release, according to the excipients included: resins > silicone oil > silica > DE (Fig. 2A). The presence of resins drastically increases the hydrophobicity of the polymer matrix which may help to dissolve the lipophilic drug (LNG) and increase the drug permeability through the matrix. In addition, the amorphous polymer chains and relatively high porosity may also contribute to higher drug solubilization in the polymer matrix. The lowest drug release rates observed with DE may be attributed to its high hydrophilicity (polar -OH groups on the surface), high crystallinity, and low porosity. Silica particles will interact with the polymer via the formation of hydrogen bonds, thereby strengthening the matrix and increasing the viscosity. This was validated using DMA which showed a drastic Tg shift in these formulations. Accordingly, it was observed that a high amount of silica led to slower drug release rates when compared with formulations with low silica content (Fig. 2B). Inclusion of silicone oil provides a low-viscosity matrix which reduces the diffusion barrier to the drug causing faster drug release than from formulations without silicone oil. However, silicone oil with 350cST and 1000cST viscosities did not show a significant difference in drug release for up to 4 months (Fig. 2C). ¹³C ssNMR revealed additional peaks at approximately 120 ppm, 155 ppm, and 170 ppm suggestive of possible drug-excipient interaction (Fig. 3). The peak intensities were relatively low in the formulation containing resins which showed fast drug release rates. On the contrary, formulations containing silica and DE which exhibited slow drug release rates had higher peak intensities. These results underpin the influence of drug-excipient interactions on drug release from LNG-IUSs.

Conclusion: The critical physicochemical attributes contributing to drug release from LNG-IUSs are: hydrophobicity, viscosity, crystallinity, and microstructure. Specifically, the release properties may be largely dictated by excipient hydrophobicity. Drug-excipient intermolecular interactions may also be a key determinant in the dissolution performance of IUSs. Overall, this research highlights the significance and role of excipients in tailoring drug release from long-acting IUSs which can help guide formulation development. Furthermore, the present study may also have implications for the development and optimization of other silicone-based controlled release products.

References: 1. Fanse S., et al. Impact of polymer crosslinking on release mechanisms from long-acting levonorgestrel intrauterine systems. Int. J. Pharm., 2022. 612, 121383.
2. Bao Q., et al. Drug release testing of long-acting intrauterine systems. J. Controlled Release, 2019. 316, 349-358.
3. Bao Q., et al. Manufacturing and characterization of long-acting levonorgestrel intrauterine systems. Int. J. Pharm., 2018. 550, 447-454.

Acknowledgements: Funding for this project was made possible by a U.S. Food and Drug Administration grant (1U01FD005443-01). This abstract reflects the views of the authors and should not be construed to represent FDA’s views or policies.


Fig. 1. A) Drug loading in formulations with different excipients (determined using tetrahydrofuran extraction and tested via UPLC) (mean±SD, n=3); B) representative DSC thermograms indicating crystallization exotherms for formulations containing different excipients; C) microstructural changes in formulations containing different excipients (determined via SEM, 1000X magnification).


Fig. 2. A) In vitro drug release profiles of formulations containing different excipients (mean±SD, n=3); B) in vitro drug release profiles of formulations containing varying amounts of silica (mean±SD, n=3); C) in vitro drug release profiles of formulations with and without silicone oil (of different viscosities) (mean±SD, n=3).


Fig. 3. 13C CP/MAS (crosspolarization/ magic angle spinning) solid state NMR spectra of pure drug (LNG), blank formulation (containing polymer and additives but no drug), and individual formulations containing different excipients. (Note: Separate blank formulations were prepared and tested for the individual formulations. However, a representative spectrum for the blank is depicted for data visualization and easy comparison).