(M1430-01-03) Development of a Coextrusion Process to Prepare Etonogestrel Long-Acting Implant

Purpose: Ethylene vinyl acetate (EVA) based etonogestrel long-acting implants modeled after Nexplanon^® were prepared using a coextrusion process. With the exception for the absence of barium sulfate, our prototype implant possessed the same dimensions and drug loading as Nexplanon®. The implant was composed of an EVA 28 core (28% vinyl acetate content) imbedded with etonogestrel. Surrounding the core was a rate-controlling EVA 15 skin (15% vinyl acetate content). This study aimed to develop a manufacturing process to control the physicochemical properties of etonogestrel implant to achieve the desired quality attributes.

Methods: The D90 of etonogestrel was 12.3 µm. EVA was supplied by Celanese as 3x3 mm pellets. Implants were prepared using a two-step double melt extrusion process. In the first step, core materials were extruded using an 18 mm twin-screw corotating extruder (TSE) (Leistritz) and cut into pellets. In the second step, the core pellets and EVA 15 skin pellets were coextruded using the TSE and a 19 mm single-screw extruder (SSE) (Davis Standard), respectively. In the TSE, the core melt passed through a gear pump before converging with the implant skin at a co-axial die to form a co-axial strand. The molten strand was drawn vertically into a water bath by a wheel puller and subsequently cooled and solidified. The strand diameter was measured in-line using a laser micrometer (ODAC). Finally, the solidified strand was pulled by a downstream belt puller and collected in a winder prior to cutting to 4 cm. The total output rate was 1.50 kg/hour with core and skin output rates of 1.34 kg/hour and 0.16 kg/hour, respectively. The implant structure was studied by analyzing implant cross sections under a light microscope. To evaluate etonogestrel supersaturation in the core, in-vitro drug release from skin was conducted by sealing both ends with etonogestrel-impermeable Loctite^® 4011 glue. The degree of drug supersaturation in core was calculated based on mass balance and compared with that of Nexplanon^® [1].

Results: Content Uniformity: To ensure uniform distribution of etonogestrel in the core, the double extrusion was used to prepare the core mixture of cryogenically milled EVA 28, micronized etonogestrel, and magnesium stearate. The weight ratio of implant skin to core (skin thickness) was determined by the feed rate ratio of SSE to TSE (0.16 kg/hour to 1.34 kg/hour). Because the TSE was starve-fed, a gravimetric feeder was used to control its feed rate. To ensure a consistent feed rate of core materials (± 5% of target, Figure 4A), the core blend (containing half of the total magnesium stearate) was extruded and cut into 2x2 mm pellets. The pellets were then mixed with the remaining magnesium stearate to improve flowability (Table 1). Because the SSE was flood-fed, its feed rate was determined by its screw speed. Based on a linear calibration curve, the screw speed was set at 4.8 rpm to achieve a feed rate of 0.16 kg/hour of skin material. The desired skin thickness (60 µm) was achieved for our implant (56.0 ± 5.0 µm, n=8) (Figure 3). Dimension Control: The swollen strand exiting the die was drawn down by a wheel puller (Figure 1B) to the target diameter (2 mm) with less than 3% fluctuation (Figure 4D). Stability of the inlet melt pressure at the co-axial die was critical for dimension control. Shown in Figure 4B, the fluctuating melt pressure at the TSE outlet (± 23%) was attenuated by the gear pump to obtain a stable core inlet melt pressure at the die (± 6%). The fluctuation of skin inlet melt pressure at die was maintained within ± 7% (Figure 4C). Drug Supersaturation and its Effect on Drug Release: Due to mechanical energy input by the co-rotating screws of TSE (Figure 1A), the implant core was subjected to elevated temperature. At the extrusion temperature (145 °C, Figure 2), etonogestrel solubility in EVA 28 is 5 wt.%, a 20-fold higher than solubility at storage temperature (25°C) [2]. Therefore, etonogestrel was supersaturated in the implant core and skin, leading to an initial burst release. The prototype implant, tested one month after extrusion, exhibited a higher burst release rate compared to that of Nexplanon^® (Figure 5). This was attributed to a higher degree of supersaturation in the prototype implant (6.4) compared to Nexplanon^® (1.4). To minimize supersaturation, implants will be manufactured in a future study with reduced energy input by reducing the number of kneading elements and increasing the specific rate (feed rate/screw speed).

Conclusion: We have developed a co-extrusion process to prepare EVA-based etonogestrel long-acting implant. The extrusion process was designed to ensure content uniformity, accurate and precise dimension of etonogestrel implant. The prototype implant demonstrated a higher degree of drug supersaturation compared to that of Nexplanon^®, resulting in a higher burst release rate. The degree of supersaturation in prototype implant can be reduced to match Nexplanon^® by lowering the energy input.

References: 1. Zhong, Z., et al., Reverse engineering of Nexplanon® contraceptive implant: physicochemical properties and in-vitro drug release. AAPS PharmSci360, 2024.
2. Van Laarhoven, J., et al., Effect of supersaturation and crystallization phenomena on the release properties of a controlled release device based on EVA copolymer. Journal of Controlled release, 2002. 82(2-3): p. 309-317.

Acknowledgements: This work was supported by the Broad Agency Announcement (BAA) Contract # 75F40122C00019 from the U.S. Food and Drug Administration (FDA). The content reflects the views of the authors and should not be construed to present FDA’s views or policies. The authors would like to thank Evestra for providing technical support and access to its coextrusion line. The authors would also like to thank Celanese for providing the EVA.


Table 1. Composition of etonogestrel implant, 55 mg (1st extrusion process shaded).
Figure 1. Schematic illustration of (A) screw profile of twin-screw extruder and (B) coextrusion line for implant preparation


Figure 2. Process parameters of coextrusion line.
Figure 3. Light microscope image of cross section of the etonogestrel implant, 55 mg.


Figure 4. Process control and deviation (red dashed line) during coextrusion: (A) deviation of feed rate of implant core from the target value, (B) twin-screw extruder outlet melt pressure and core inlet melt pressure at the die, (C) skin inlet melt pressure at the die and (D) deviation of implant diameter from the target value.
Figure 5. Skin release rate of prototype implant and Nexplanon® as a function of time (SD, n=3) and amount of drug supersaturated in implant.