Impact of drug loading on release from levonorgestrel intrauterine systems

Levonorgestrel intrauterine systems (LNG-IUSs) are polydimethylsiloxane (PDMS) based non-biodegradable complex drug-device combination products providing efficacy for up to several years based on the strength. A large amount of LNG (e.g., 52 mg in Mirena and Liletta) must be loaded in the LNG-IUS products to maintain the long-acting effect even though LNG is a potent hormone. However, the high amount of LNG not only poses the potential risk of dose dumping, but also leads to drug waste due to incomplete drug utilization close to the end of usage. It has been unclear whether the duration of usage of these products should be extended for full drug utilization or products with lower drug loading should be developed. Therefore, it is critical to understand the impact of strength (or drug loading) on drug release from LNG-IUSs. In the current study, drug reservoirs with a broad range of drug loading (from 0.5% w/w to 50% w/w) were prepared and assembled into LNG-IUSs. Different accelerated release conditions were used to perform release testing of LNG-IUSs with different drug loading. 5% to 10% variation in excipient of the LNG-IUSs did not significantly alter the drug release profiles of the LNG-IUSs. The release rate of LNG-IUSs is inversely proportional to their drug loading at high drug loading (10% w/w, 25% w/w and 50% w/w). Drug release was incomplete for LNG-IUS with low drug loading (2.5% w/w and 1% w/w) and no drug release could be detected for the LNG-IUS with 0.5% w/w drug loading. In addition, the burst effect of the LNG-IUSs with different drug loading was investigated. This is the first research report covering ultra-long duration (more than four years) of real-time drug release from LNG-IUSs with different drug loading (0.5%-50% w/w). The amount of excipient (PDMS) used in the reservoir of LNG-IUSs was determined to be not a critical quality parameter in the formulation design since LNG-IUSs (50% w/w drug loading) with up to 10% variation in excipient did not show significant differences in their release profiles. The drug release kinetics/mechanism remained the same for LNG-IUSs with drug loading ranging from 1% to 50%. In addition, the accelerated release testing methods were confirmed to be representative of the real-time release profiles and this can give confidence in extending the duration of usage for these products provided that the device remains physically intact (no tearing or damage in the outer membrane) and the release rate is within the therapeutic window. It is recommended to perform both real-time and accelerated release testing simultaneously for LNG-IUSs to understand the burst effect as well as the complete release characteristics. Lastly, drug/polymer interaction may play a role when designing LNG-IUS formulations with low drug loading (<5% w/w) since drug/polymer interaction is significant when only a small amount of drug present.