The Struggle is Real: Drug Molecules are Getting Bigger. Needles Aren’t.

High-viscosity drugs can be problematic at both ends of the syringe: Not only do they potentially cause more pain at the injection site, but they require more force to inject, which could compromise the delivery device. 

To address the pain points associated with injecting high-viscosity drug formulations, Becton, Dickinson and Company (BD) has developed an in-silico simulation model that predicts the injection force through a prefilled syringe (PFS) or injection time through a prefilled syringe/autoinjector (PFS/AI) system. The findings can optimize the development of pharmaceutical injection systems and improve understanding of device/drug compatibility.

To learn more about BD’s research and its implications for drug delivery and packaging design, Packaging Digest reached out to Julien Singer, PhD, R&D Senior Engineer. Singer worked on the development of BD’s simulation model and will present his findings at the American Association of Pharmaceutical Scientists’ (AAPS) annual meeting on October 22, 2024.

Why did BD develop a model to predict the injection force through prefilled syringes or injection time through PFS/AI systems?

Singer: The drivers were the same as those for any intent to model physical systems: understand, predict, improve. We use modeling to get a step ahead in the comprehension of our systems, to make predictions during product development, and finally, to improve their behavior within their use cases.

Related:4 Observations from Pharmapack Europe 2024

Is it difficult for healthcare workers to know if they’ve delivered a full dose during an injection? If so, why?

Singer: Regarding the full dose, there is always a way to look at the plunger stopper position during and after injection, making it easy to know whether the dose was entirely delivered or not.

Is BD trying to lower the force needed to inject a drug in a PFS? Is injection force a problem, causing injections to be less than a full dose?  

Singer: The trend to inject larger doses of high-viscosity encapsulated drug formulations (such as lipid nanoparticles) causes the required injection force and/or time to increase. If injection devices are not optimized, injections can become longer than recommended or require the healthcare worker to apply too much force. If this required force cannot be held by the user during the whole injection process, the injection could be interrupted, indeed. BD products are optimized to ensure full-dose delivery in acceptable conditions (time, force) based on their expected applications.

If drugs are getting more viscous, why not just make the needle/needle hole larger?

Singer: The needle external diameter is related to patient perception and pain, although this highly varies between individuals. Making the needle hole larger while keeping the same external dimension is an option BD is currently working on, and our models help during these development phases.

What will be the result of having a robust and predictive numerical model?

Singer: Qualitative models are still useful if they are used with perspicacity. But if you want to use a model to predict the behavior of your device on a wide range of conditions and use these results to choose the best suited device for a given drug, you need to have a model that allows you to quantitatively predict the injection force and the injection time — and a model robust enough so that it remains predictive over a wide range of conditions (drug and device characteristics).

What is the ideal application of such modeling along the drug/device development journey and why?

Singer: Once we get this high level of confidence in the model, we use it to predict the performance of different device configurations when associated to different drugs. This allows us to get accurate results for many configurations much faster, easier, and more cost-effectively than if all data were built through experiments.

The global, basic structure of BD’s predictive model uses simple laws, with more complexity to come. (Image: Singer/BD)

What are some of the limitations of modeling tools?

Singer: Some limitations include the simulation of complex rheologies: injection of shear-thinning fluids can be estimated with BD’s model, though it still needs to be assessed versus experimental data; while shear-thickening fluids are still out of the scope of the model. Stopper-barrel friction is an input of the model that needs to be characterized or estimated by other means (either experimentally or based on other types of simulations).
More globally, one could say “garbage in, garbage out:” Inputs should be carefully chosen to get the expected result and insightfully used to draw the right interpretation of the outputs.

Where is the industry now in using simulation to optimize device/packaging design?

Singer: It is not easy to generalize, but from what I can see here at BD and around, we are moving on the next step. Physical modeling and numerical simulation are used more and more, from project development through industrialization and post-market surveillance.

Can you share one insight/result of your work that you’ll be presenting in October?

Singer: We will share why taking into account the right physics as well as the right device data (inputs) lead to near-to-perfect correlation. Two aspects in particular will be highlighted: In some conditions, laminar flow cannot be neglected, and turbulence should be included in the model; the second point will be about the barrel-stopper friction, a major input of the model that cannot be considered as a single, constant value.