Patent Application
20150297458
This invention relates to rapid reconstitution packages designed using computational fluid dynamics techniques to optimize the fluidic structure for mixing and delivery of lyophilized drugs.
Emergency medications in ambulatory settings are formulated and administered often in time-critical situations, leading to an increase in the probability of errors for therapeutic treatment [1-4]. Reconstituting lyophilized drugs is an operational challenge in demanding environments, e.g. disaster zones, battlefield support, and rural areas. Heat insulation is critical for biological drugs, e.g. antibodies, for which the efficacy of the active pharmaceutical ingredient (API) heavily relies on biomolecular conformality. Small molecule drugs consisting of pure chemical formulation are often deployed in liquid form in hermetically packaged glass ampules. In order to avoid hydrolysis, biologics are stored in lyophilized form in hermetically sealed glass vials that require subsequent reconstitution upon delivery. Lyophilization process involves freeze-drying the active pharmaceutical ingredient (API) combined with excipients, rendering a powder form with specific physical properties that include mass density, solubility, crystal size and packing factor.
The steps for drug reconstitution in vials require: injecting specific diluents into the vials, manual shaking and visually confirming homogeneity [5]. Various pre-filled dual-chamber solutions exist for on-demand reconstitution. These solutions typically rely on chemical solubility between a diluent and lyophilized drug and still require manual shaking. The design of these devices typically consists of two chambers: one for the drug in lyophilized form (solute) and another for the diluent (solvent). Dual-chamber devices have been utilized for small API payloads with a limited number of pharmacological therapies, such as human growth hormone (HGH) commercially known as Genotropin (Pfizer, Inc.) developed by Vetter. See, DE 3736343A1, CA 2397829C, and DE 102004056617A1 [6].
In a first aspect, the invention is a rapid reconstitution package including a substantially cylindrical structure including a diluent chamber and a drug chamber communicating through a valve, the cylindrical structure configured to retract under axial force. The retraction causes the valve to open to allow the diluent to flow through the valve into the drug chamber for mixing and reconstitution of the drug, wherein the drug chamber shape is tuned to enhance mixing. An embodiment of this aspect of the invention includes a plunger operatively connected to the valve to open the valve under the axial force. It may be preferred that the cylindrical structure be sized to fit within a standard syringe body for activation. A separate mixing chamber may be provided to communicate with the drug chamber to provide for enhanced mixing. In another embodiment of this aspect of the invention, the drug chamber shape deflects a jet of diluent downwards for enhanced mixing. The jetting effect can be tuned both in intensity and spatial direction depending on the drug that is required to be reconstituted. In another embodiment the RRP disclosed herein can be used not only as a cartridge but as an injector or even an auto-injector when it is coupled to a pre-loaded spring. See, US 2013/0178823, US 2012/0101475, U.S. Pat. No. 4,394,863 and WO 1991016094.
A second aspect of the invention is a method for designing a rapid reconstitution package including prototyping and testing physical properties of a package structure. Computer computational fluid dynamics is used to model and numerically process results of the testing. Experimental data is used to modify fluidic structures to achieve enhanced mixing. In a preferred embodiment, the computational fluid dynamics provides quantitative techniques to optimize mixing parameters for drug mixing, advection and diffusion phenomena. In yet another preferred embodiment, the modified fluidic structure increases inlet velocity into the drug chamber in the package. The CFD analysis allows a change in the drug chamber silhouette so that it acts as both a storage and mixing chamber while indices of vorticity and concentration of the output as a function of time are monitored. Furthermore, depending on a targeted release profile it is possible to adjust the silhouette combined with activation time to adjust for an overall release profile per pharmacological therapy.
a is a cross-sectional, isomeric view of an embodiment of a rapid reconstitution package disclosed herein shown within a standard syringe.
b, c, and d are cross-sectional views of an embodiment of the rapid reconstitution package disclosed herein.
a, b, and c show velocity magnitude contours showing the velocity of jet stream flow through several embodiments of the rapid reconstitution package disclosed herein.
a shows vorticity simulations for an embodiment of the rapid reconstitution package disclosed herein.
b is a graph of average vorticity versus time over a two second inject for different embodiments of rapid reconstitution packages as disclosed herein.
a shows simulation results from numerical models.
b is a graph of concentration versus time showing simulated outflow concentration profiles at three different positions over four activation times.
c is a graph of drug exhausted versus activation time showing the percentage of tPA released over simulated activation times.
In order to effectively reconstitute lyophilized or powder-based drugs with large payloads (e.g. antibiotics and biologic drugs in the order of tens of milligrams) and low solubility APIs, numerical models have been considered taking into account the physical properties of drugs and diluents. The challenge lies in the design of devices to perform optimal reconstitution incorporating parameterization models of APIs and diluents, and engineering fluidic structures in order to optimize mixing homogeneity and response time as a function of input shear flows [7]. The next generation of reconstitution devices for large payloads (10 mg-250 mg) and low solubility APIs will therefore require a more comprehensive design consolidation of physical parameters to optimize mixing.
Rapid Reconstitution Packages (RRPs) were designed with numerical models integrating microfluidic structures to dramatically enhance mixing of large payloads to optimize mixing in the 1-7 sec range. RRPs were designed using Computational Fluid Dynamics (CFD) models that incorporate API and diluent parameterization of physical properties. CFD modeling provided insights into the flow distribution variables of drug dissolution, such as turbulence, solubility and shear stress.
Device performance was evaluated by selecting Tissue Plasminogen Activator (tPA) as the initial model drug. tPA is a commonly used thrombolytic for treating ischemic stroke and myocardial infarction (MI) [8, 9]. tPA is a large peptide-based molecule, reportedly unstable in liquid form (up to 8 hours in solution) [10]. The drug was selected for investigation as it is normally administered in emergency situations, requiring refrigeration and not able to be reliably stored in liquid form. The tPA used for experiments was Cathflo Activase recombinant (Genentech, Inc.). The present invention may be used with drugs or powders (foods, cosmetics, mixers for pediatric drugs).
RRPs disclosed herein were designed from a clinical perspective to improve logistics for emergency and ambulatory applications in a single-step process. RRPs can potentially provide an impact on the delivery of APIs in wide variety of applications for emergency and ambulatory settings.
a is an isometric view of an embodiment 10 of the rapid reconstitution package disclosed herein. In
As shown in
The RRP design disclosed herein was initially conceived as the cylindrical cartridge 10 that can be activated by inserting into a standard syringe 12 and compressed with the syringe plunger. This mode of operation was selected to help preserve the standard operating procedure of drug delivery using syringes. RRPs are designed to fit in 10, 20 and 50 mL syringes, depending on the target pharmacological therapy that corresponds to payload volume. The RRP was designed for activation in three steps. First, the syringe plunger is removed from the syringe barrel. Second, the RRP is placed inside of the syringe barrel as a cartridge. Finally, the syringe plunger is reinserted into the syringe barrel and put into firm contact with the RRP. The system is then ready for operation. The device relies on the telescopic operation of an internal plunger which opens a valve between the diluent and drug chambers, forcing the diluent into the drug chamber. The drug-diluent product is then further mixed in the mixing chamber and then released.
In terms of materials, the device disclosed herein can be made out of polymers or glass (when possible), or a combination of both, for example by deposition of a conformal layer of SiO2 along the surface area of polymers. In addition, Paralyne could be used as an example of a conformal film deposited to prevent leeching or material interference with the drug (API) or diluent.
In terms of manufacturing, as a method, it could be implemented with 3D printing or Stereo Lithography Apparatus (SLA), but ultimately injection molding or other mass production will be needed. The inner back-loop exhaust cylindrical channel (tube) can be defined quite small such that in the case the RRP is implemented as a syringe cartridge, when the end tip of the inner plunger can be inserted without effort as close as possible inside of the syringe inner channel that accesses the needle, preventing any overhead volume and reduces bubbles.
Computational Fluid Dynamics (CFD) techniques were employed to quantitatively optimize RRP reconstitution performance.
Considering device symmetry, simulations were run on a 1/12th cylindrical domain, at the expense of neglecting angular flow to limit computational load since the speed of simulations for rapid iteration cycles was desirable. The mesh adopted for simulations is shown in
Preliminary analyses focused only on diluent flow across structures. The primary concerns for these simulations were increasing inlet velocity into the drug chamber, minimizing stagnation-prone regions, and streamlining the fluid flow for increasing advection. Inlet velocity information was utilized to increase turbulence and vorticity in the drug chamber for improving mixing. Stagnation-prone regions were to be avoided to maximize drug exhaustion. Streamlining was crucial for the overall design optimization.
The drug chamber shape is tuned per physical properties of the drugs, excipients and diluents, including: solubility, packing factor, viscosity of the diluent, viscosity of the product. The size of the drug chamber depends on payload; the size of the diluent chamber depends on required volume to dilute. Furthermore, in case that simulations indicate that a higher viscosity (effective one) is required, one can add nano-particle or micro-particles either or both in the diluent or drug chamber as excipients that enhance reconstitution by means of convective forces. Suitable nano- or micro-particles are PLGA, silica, iron oxide and other biocompatible materials.
Furthermore, one can potentially encapsulate the API with nano-particles or micro-particles that will enhance the reconstitution process by creating nano or micro domains of hydrophilicity. Hydrophobicity of drugs is one of the key physical properties to consider. The challenge in the simulations is to combine convection and diffusion at once. As the drugs are being ‘hit’ by the diluent (with or without jetting) at high velocity, the drugs start to be dissolved as they are mechanically moved down to the drug chamber (convection). This invention could be used as cartridge, injector, or auto-injector.
A more comprehensive CFD model with diluent-drug interactions was introduced to simulate the drug reconstitution process in the preliminarily simulated designs. Physical parameters of the drug and excipients were incorporated into the model.
Cathflo Activase in its delivered form was composed of a 1:40 API to excipient, with 2 mg tPA and 78 mg excipients (sucrose, L-arginine, polysorbate 80, and phosphoric acid), summing up to 80 mg per payload. The entire payload was modeled as a homogeneous compound of sucrose (the major component) with a diffusivity value in water equal to 5.10−8 m2/s and a mass density of 1.587 mg/mm3. A homogenous mixture of tPA and excipient was assumed to quantify API output concentration as 1/40th of the total output scalar concentration.
The packing factor due to lyophilization was taken into account to incorporate the stacking of the drug into a singular annular volume defined in the drug chamber. The packing factor was estimated and approximated as follows [21]:
Where N is the number of crystals, Vsphere is the crystal volume approximated as sphere of diameter, d, and Vdrug is the volume of the drug, distributed as a ring in the drug chamber. The density of the lyophilized drug taking into account the packing factor was estimated at 1.2 mg/mm3.
Distilled ultrapure water was selected as the model diluent. The modeled water was assumed to be Newtonian and incompressible. Gravity and surface tension effects were neglected for the Newtonian fluid as the Froude number, defined as the ratio of inertia to gravity forces, and Weber number, defined as ratio of inertia to surface tension forces, were both much larger than unity.
The effect of flow on drug dissolution upon initial device activation was introduced as a change in effective viscosity, estimated by approximating the resulting solution as a dilute suspension of small rigid spherical particles, estimated as follows [22]:
Where μeff is the effective viscosity of the diluent, μf is the viscosity of the fluid without solute and φ is the volume fraction, defined as the diluent volume to the mixed volume.
To assess the device behavior under possible operation conditions, numerical simulations were carried out for three different locations of the drug (tPA) within the drug chamber and four distinct activation times of the RRP defined as follows: 1 s, 2 s, 4 s and 7 s. These activation times were chosen to emulate real operation times of syringes. The location of drug was modeled at different positions to account for its random distribution within the drug chamber. The three different positions are shown in
RRPs were initially kept at 20° C., and subsequently subjected to a fixed wall temperature of 65° C. for 3 hours. Samples were kept in 6 different storage conditions for 24 hours prior to assay preparation: (1) Standard Cathflo Activase lyophilized in glass vial at 20° C.; (2) Cathflo Activase reconstituted in Millipore filtered water and kept in glass vial at 20° C.; (3) lyophilized Cathflo Activase loaded into RRPs and kept at 20° C.; (4) Cathflo Activase lyophilized in glass vial at 65° C.; (5) Cathflo Activase reconstituted in Millipore filtered water and stored in glass vial at 65° C., and (6) Cathflo Activase loaded into RRP and kept at 65° C. After 24 hours each sample in lyophilized form, including those stored in the RRP, was reconstituted in 2 mL Millipore water, and apportioned for subsequent use in ELISA, UV-Vis and HPLC assays.
RRPs were loaded with Cathflo Activase (2 mg of tPA and 78 mg of excipients). RRPs were placed in a 25 cc syringe (Exel International, Co.). The syringe was then fixed into a syringe pump (Customized 4X PhD Series, Harvard Apparatus, Inc.) that would activate the RRP at a rate of choice. The syringe pump was set to inject a standard 20 cc volume syringe at a flow rate of 260 mL/min for a total inject time of two seconds. As the flow began to emerge from the outlet of the syringe, a customized linear actuator collected the elution. This linear actuator was configured with a collection tray consisting of 14 wells with a fixed width, such that each well represented a time increment of approximately 0.14 s. An illustration of the experimental setup is shown in
Stability experiments were performed to characterize and compare the RRP with standard vials at various storage conditions. ELISA assays were used to test for stability of samples (Human tPA Platinum ELISA Kit, eBioscience, Inc.) that underwent two temperature conditions, T=20, 65° C. HPLC assays were not sufficient in assaying stability for tPA as the low pH buffer required denatured tPA regardless of conformation or partial denaturation already induced by the storage condition.
The samples to be tested were injected into each ELISA well and left to incubate. After a wash step, a second antibody, horseradish peroxidase (HRP)-conjugated anti-tPA, was injected into each sample well. The HRP-conjugated anti-tPA bound onto the opposing side of the tPA molecule already bound onto the plate-fixed antibody. After a second wash step, a color inducing substrate (TMB substrate) was then injected into the wells, which bound onto the HRP to induce a measureable gradient representing the amount of tPA bound. In order to prevent oversaturation of the antibodies, the samples were diluted to 10 ng/mL to fall within the assay sensitivity range of 2 pg/mL-10 ng/mL. Optical signal intensity from assay plates were measured with spectrophotometer at wavelength 450 nm (PowerWave HT Microplate Spectrophotomer, BioTek Instruments, Inc.). Each sample was normalized against a calibration curve derived from a 5-parameter logistic fit of tPA standards ran concurrently with the samples.
HPLC (1100 Model, Agilent Technologies, Inc.) was performed with each sample referenced against a calibration curve for the drug being measured. Assays were used with a reversed-phase column (ZORBAX SB-CN, 4.6 mm×250 mm column with 5 μL pores, Agilent Technologies, Inc.). The mobile phase was a gradient with an initial hold for 5 min in a ratio of 70:30:0.1 water:acetonitrile (ACN):trifluoroacetic acid (TFA), brought to 50:50:0.1 water:ACN:TFA over 80 min, and subsequently to 100:0.1 ACN:TFA over 15 min. An acidic environment (created by the use of TFA, pKa=0.23) was necessary to separate tPA from solution. Flow rate was 1 mL/min and injection volume was 250 μL. tPA presence was detected by UV spectrometry at 280 nm.
Simulation results with flow characterization models allowed for incremental, yet significant design changes to improve flow properties for three subsequent design versions.
Another important aspect is to reduce overhead volume, meaning volume of the diluent, or mixture hung up the in the chambers. There will always be a bit, but also that's another parameter to take into account when optimizing the design. For example, in another proposed embodiment, we can implement a valve that uses sidewall bypasses, maintaining the overall cross section of the device. The way the valve 24 works is that it goes from one small cross section to a larger cross section allowing the o-ring to lose grip from the walls as it enters the drug chamber 18. That is how the jetting effect is included. Because we are adjusting a silhouette and curvature for jetting and mixing, the inner RRP plunger 22 cannot go all the way, and some volume remains. To reduce this, we have changed the valve mechanism. So, instead of going from small cross section to large cross section (barrel), bypasses are spatially defined between barrel in the wall and/or moving plunger. Then, the drug chamber is defined simply straight. A straight (no silhouette) chamber is widely used, e.g. Vetter, etc, but limited to a number of drugs due to solubility and payload. The overhead, volume, however, is much reduced since the inner plunger advances all the way. So, in other words, we can implement a similar concept with telescopic back-loop or without back-loop.
Another significant optimization in the RRP was the elimination of the secondary chamber, which was initially conceived as an additional mixing chamber. This chamber mixing effect was proven insignificant through the simulations. In fact, a large secondary chamber increased the volume of stagnant regions and decreased flow efficiency, as shown in
To perform parametric comparison of the designs we define an “average vorticity”, obtained by making a cell volume weighted average of the absolute value of vorticity of the fluid for every fluid cell. A visual representation of average vorticity across the chambers for three subsequent versions is shown in
A design requirement that became apparent from observing flow streamlines through the RRP was the addition of a compartmentalized mechanism that would prevent drug powder from exiting the drug chamber too soon. Implementation of such a solution was achieved by adding a secondary valve in order to keep the drug in the drug chamber during the initial transient mixing.
Simulation results initially showed that a jet of solvent entering the drug chamber collided with the drug crystals as the RRP was activated. This step proved to be a complex process to model due to the combination of drag forces, dissolution and forced convection. Convection was considered the dominant mass transport mechanism for modeling dissolution. The resulting mixed solution was modeled using effective viscosity from Eq. (2) and constant density calculated from Eq. (1). The drug mixing models showed that most of the reconstitution occurred in the drug chamber. The inlet jet effect caused by the opening of the valve between the drug chamber and the diluent reservoir increased reconstitution performance. The fraction of drug that does not exit the chamber dragged by the initial flow goes into localized stagnation regions and slowly diffuses through advection towards the exhaust.
Snapshots of the flow patterns for Position 2 is shown in
c shows the total amount of drug that was eluted from the RRP after full activation. The total output amount from the RRP was not significantly changed by payload size, initial drug location, or rate of injection, indicating that flow rate may indeed have an effect in concentration profiles but not in the total mass of drug delivered. Simulations for various amounts of tPA were carried out with 1, 4, 10, 30, 60, and 100 mg payloads, which resulted in similar total normalized outputs, as shown in
In order to characterize the concentration profile from the RRPs, a 2 s injection time was chosen to emulate a typical injection time as delivered by a clinician. Results from the experimental studies compared to the modeled results are shown in
The RRP design was optimized by numerical models using CFD models ultimately to maximize reconstitution. The numerical models allow simulations of reconstitution performance for specific pharmacological formulation with different injection times and payloads. Experimental results showed the reliability of simulated results, within the same order of magnitude, in concentration outputs as a function of response time. Differences can be attributed to errors in the discrete sample collection method, primarily related to variability in volume amounts per sample. Another source of variability can be attributed to the slightly different locations of lyophilized APIs between different generations of RRPs, as the drug powder cannot be localized precisely within the drug chamber. A continuous measurement method will be implemented in future work to characterize release kinetics with higher precision.
Each pharmacological formulation requires a set of customized parameters that take into account the physical properties of the API, excipients, and diluents. Experimentally determined concentration profiles are critical for performance and safety. For example, the value of the concentration peak can serve as a limiting factor in the design and optimization of the RRP so it does not reach a specific concentration of toxicity. Conversely, the models can be parameterized to the desired delivery profile in order to inform the design of the fluidic structures. For example, the relative position of the payload within the drug chamber can be considered a design parameter to adjust according to the targeted release kinetics. Simulation results have shown that the Position 3 under activation time of 1 s can be regarded as a bolus application for a drug, whereas Position 1 and activation time of 4 s can be useful for a more continuous release of the same drug.
The payload released has a critical impact in controlling the total dosage delivered in a given pharmacological application. It must be emphasized that the proposed reconstitution model assumes that drug dissolves instantaneously, hence not taking into account drag effects that are experienced by a small portion of solid particles. A more comprehensive model would be required to describe particles that are reconstituted while they are transported.
The presented model allows for simulations results to be incorporated back into the design process, which in turn are quickly implemented using a 3D printing technology. The integrated use of 3D printing technology for the construction of simulated structures provided an advanced approach for accelerating prototyping of microstructures. The computational-experimental iterations provide a method for validating and optimizing reconstitution performance based on structural dimensions and physical parameters. Overall, simulations provide a predictive analytical method, rendering engineering optimization of the RRPs.
Experimental methods including ELISA, HPLC, and flow characterization provided quantitative results required to inform design decisions and ensure design robustness. The integration of CFD models for design of RRPs could be extended to other lyophilized drugs that are required for emergency applications.
The numbers in square brackets throughout this document refer to the references listed herein. The contents of all of these references are incorporated herein by reference in their entirety.
It is recognized that modifications and variations of the present invention will be apparent to those of ordinary skill in the art and it is intended that all such modifications and variations be included within the scope of the appended claims.
This application claims priority to provisional application Ser. No. 61/719,012 filed on Oct. 26, 2012, the contents of which are incorporated herein by reference.
This invention was made with government support under Contract No. W911NF-13-D-0001 awarded by the Army Research Office. The government has certain rights in the invention.
| Number | Date | Country | |
|---|---|---|---|
| 61719012 | Oct 2012 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | PCT/US2013/066758 | Oct 2013 | US |
| Child | 14696788 | US |
