Non-invasive, continuous oral delivery of solid levodopa-carbidopa for management of Parkinson’s disease Authors: William L Murch, John Spiridigliozzi, Adam Heller, Ephraim Heller Published: 2024 ABSTRACT AbstractWhen short plasma half-life drugs act only briefly, they require frequent or continuous administration. We report the engineering of a non-invasive oral drug delivery system for long-term, continuous administration of these drugs. Their non-invasive, long-term, continuous administration at daily doses exceeding 100 mg has, for many years, been considered an insurmountable challenge. We show that over 1200 mg/day of 4:1 levodopa-carbidopa (LD-CD) can be non-invasively and continuously extruded when formulated as a semisolid paste, loaded with 63%w/w of the solid drugs. The drug delivery system comprises a reusable orthodontic retainer with a co-molded pocket into which the patient inserts after each meal a new 1 mL propellant-driven, prefilled, disposable, drug delivery extruder. The paste is delivered to the lingual side of the teeth where it is mixed with saliva and swallowed. As reported elsewhere, a 15-day, 16 patient open label clinical trial of the drug delivery system continuously extruding LD-CD paste significantly reduces the variability of the plasma LD concentration and alleviates symptoms of advanced Parkinson’s disease (PD) as compared to LD-CD tablets.Subject terms: Drug delivery, Pharmaceutics, Parkinson's disease, Materials for devices, Rheology, Biomedical engineering, Chemical engineering Introduction Non-invasive and minimally invasive continuously drug administering systems delivering short half-life drugs, exemplified by transdermal patches and skin-adhered insulin pumps, can provide clinical outcomes that are superior to intermittent administration, but are limited to daily doses of less than 100 mg. The continuous delivery of larger doses, for example with ambulatory pumps, can be inconvenient, and can have adverse local effects, e.g., on the skin. Here we describe the engineering of a heavily solid drug loaded paste and of its extrusion system, providing non-invasive, safe, and efficacious continuous oral delivery of more than 1200 mg daily of a solid drug1. The need for continuous delivery of a short half-life drug was recognized in the 1930s when it was found that insulin injections cause adverse glycemic fluctuations2,3. To overcome these, Pickup et al. introduced in 1978 a miniature, battery-driven syringe pump continuously subcutaneously infusing insulin4. Two years earlier, Judah Folkman and Robert Langer continuously released other short-lived proteins and biochemically active macromolecules from implanted non-inflammatory polymers for more than 100 days5. Their study defined a major direction of the massively impacting lifetime work of Langer. In the mid-1980s, transdermal patches continuously delivering short-lived small molecules like nicotine, nitroglycerine, and scopolamine were introduced6–8, followed by numerous other short-lived drugs9. Although these and other continuous delivery methods significantly impacted clinical practice, they were confined to the delivery of daily doses below 100 mg. For short-lived drugs that require continuous chronic administration in daily doses greater than 100 mg, exemplified by levodopa (LD), the most widely used drug in the management of Parkinson’s disease (PD), the few available delivery options are invasive. Advanced PD patients require more than 300 mg LD daily, some requiring 1,000 mg or more. Because the plasma half-life of LD is about 60 min, and when extended by a peripheral decarboxylase inhibitor like carbidopa (CD) about 90 min, the administration of LD-CD tablets is associated with adverse LD plasma concentration fluctuations. The patients lapse into the OFF state when the LD concentration is low10–12, and develop dyskinesia, LD-tolerance and multiple quality-of-life affecting complications when pulsatile and periodically high13. Continuous LD-CD administration extends the periods in which the concentration is optimally therapeutic11,14,15. In present practice, plasma LD fluctuation is reduced by intrajejunal infusion of an aqueous 20 mg/mL LD and 5 mg/mL CD gel16–19 and by a less invasive system that recently became available in Europe which subcutaneously infuses for 24 h daily a solution of 240 mg/mL of foslevodopa, an LD prodrug, with 12 mg/mL of foscarbidopa, a CD prodrug20. Three earlier studies indicated that in advanced PD patients, frequent or continuous non-invasive oral delivery of LD-CD also alleviates adverse plasma LD concentration fluctuations. Patients dosed with 5 mg 4:1 LD-CD micro-tablets significantly benefited from their frequent and individualized administration21–23; continuous oral intake of a LD solution with concomitant CD administration reduced the OFF-time relative to intermittently taken tablets, while providing stable LD serum levels approximating the improvement seen with continuous intravenous infusion of LD24; and the LD plasma concentration fluctuation was reduced when advanced PD patients sipped every 5 to 10 min small volumes of LD-CD suspended in water25. Here we describe a system for continuous, long-term, non-invasive, oral administration of solid LD-CD, applicable also to other drugs. The system, which extrudes a concentrated semisolid drug paste into the mouth, extends continuous non-invasive delivery from daily doses of < 100 mg to over 1200 mg. It consists of a conventional orthodontic retainer with a co-molded pocket, the pocket holding a 1 mL extruder, that extrudes at a constant rate the semisolid paste comprising 63% by weight of micronized LD-CD, such that the weight and volume are comparable to those of standard excipient-containing LD-CD tablets. The paste is delivered through a tube to the lingual side of the molars, where it is rapidly mixed with saliva. Every time that the patient swallows, a small dose is delivered into the stomach. Absorption is via the traditional route in the small intestine; it is estimated that less than 1% of the administered drug is absorbed via the buccal mucosa. The extruder is titanium-walled and hydrofluorocarbon propellant-driven like the 100-fold larger refillable, subcutaneously-implanted, Infusaid™ and SynchroMed™ pumps that intravenously26 or intrathecally27 deliver short-lived drug solutions. Results of a clinical safety and efficacy study of the oral drug delivery system of this article for treatment of PD were reported by Olanow et al.1. Sixteen motor-fluctuations-experiencing PD patients receiving daily 400–1200 mg of LD in 4:1 LD-CD tablets participated in a 15-day open-label clinical trial. Each patient received his or her standard LD-CD daily dose, either as tablets, or as continuously orally extruded paste. The patients did not receive LD-CD at night and consequently were in the OFF-state in the morning. On Day 1 the patients were on LD-CD tablets; on Day 2, they were for 12 h on continuously orally extruded paste; on the morning of Days 3 to 15 the patients took a typical morning dose of LD-CD tablets to rapidly raise their blood LD level, then were for 12 h on continuously orally extruded LD-CD paste. On Days 1, 2, 3 and 15 patients were at clinics, while on Days 4–14, they were at home. Their LD plasma concentration was measured at 30–60-min intervals for 12 h on Days 1, 2, and 3 and their motor state was evaluated at 30-min intervals for 12 h on Days 1, 3, and 15. There were no dropouts and no clinically significant adverse events during the two-week continuous oral paste trial. In comparison to LD-CD tablets, the continuous oral extrusion of LD-CD paste significantly reduced the variability in plasma LD concentration and improved activities of daily living. For example, the fluctuation index of the LD plasma concentration (Cmax – Cmin)/Cavg (reported as mean ± standard deviation) was reduced from 2.2 ± 0.6 on Day 1 to 1.0 ± 0.5 on Day 3, and the mean plasma levodopa concentration difference, Cmax – Cmin, was reduced from 3152 ± 980 ng/mL on Day 1 to 1400 ± 694 ng/mL on Day 3. The OFF time, normalized for 16 awake hours, was reduced from 3.2 ± 2.2 h on Day 1 to 1.5 ± 0.4 h on Day 15, and patients’ ON time without severe dyskinesia correspondingly increased by 1.7 ± 0.4 h between Day 1 to Day 15. In the 11-patient subgroup experiencing at least 2 h OFF time on Day 1, the OFF time was dramatically reduced by 72% from 4.3 h on Day 1 to 1.2 h of Day 15. For all 16 patients the Activities of Daily Living score, known as the UPDRS Part II score, improved from Day 1 to Day 15 by 2.1 ± 0.8 points. Overall, the clinical efficacy of the non-invasively continuously paste-delivering oral drug delivery system, on the engineering of which this article focuses, is projected to match that of the highly invasive intrajejunal and moderately invasive subcutaneous continuous delivery systems. Results Design of the drug delivery system and the extruder The drug delivery system, shown in Fig. 1, comprises a reusable, molded, copolyester orthodontic retainer with a co-molded pocket (Fig. 1a) into which the patient inserts a new, single-use, prefilled, disposable extruder (Fig. 1b and c). The titanium-walled extruder contains the below-described semisolid levodopa-carbidopa paste and a small volume of pressure-liquefied propellant in separate compartments. The paste extrudes at a constant rate. The patient replaces each extruder as it approaches depletion with a new extruder, typically after breakfast, lunch, and dinner. The system is removed for meals. The drug delivery system. Photographs of (a) the orthodontic retainer having a co-molded pocket; (b-c) the retainer with the disposable extruder in its pocket; and (d) the retainer residing in the cheek pocket with its delivery tube delivering paste on the lingual side of the upper molars, where it is mixed with frequently swallowed saliva. The retainer and extruder are not visible when worn and have no more impact on speech and swallowing than a standard orthodontic retainer. To avoid interference with speech and to minimize change in facial appearance, the co-molded copolyester pocket holding the extruder is in a buccal (cheek) pocket near an upper molar. The paste is, however, not extruded into the buccal cheek pocket, where it would stagnate for tens of minutes, but through a curved plastic delivery tube, to the lingual side of the molars (Fig. 1d). The paste has no taste. Mixed with saliva, the extruded and tasteless paste is swallowed frequently. Photographs of the extruder are shown in Fig. 2. Materials selected for use in the construction of the extruder are restricted to materials with long histories of biocompatibility with the oral cavity. The case is made of 0.38 mm thick commercially pure titanium; being passivated by a film of titanium dioxide, the case does not corrode. The major drug paste containing chamber and the small propellant containing chamber are separated by a 30 µm thick deformable and impermeable silver diaphragm. The external surface of the plastic delivery tube is made of a medical-grade thermoplastic polyurethane (TPU). The drug delivery extruder. Photographs of the (a) buccal side and (b) and lingual side of the extruder. Cross sectional views of the interior of the extruder are shown in the schematic drawings of Fig. 3. During manufacture, the drug paste chamber is filled with 1.04 g (0.84 mL of 1.24 g/mL density) paste through the fill port then plugged with an elastomeric fill plug. The propellant chamber comprises an elastomeric septum, through which 0.22 g (0.18 mL, 1.2 g/ml density) of pressure-liquefied 1,1,1,2-tetrafluoroethane (R-134a) is injected. The constant propellant vapor pressure causes deflection of the diaphragm, forcing constant extrusion of paste through two parallel flow controlling tubes and then into the delivery tube for delivery to the lingual side of the teeth. The plastic components, including the flow controlling tubes and internal surface of the delivery tube, are made of polyethylene terephthalate (PET). The delivery tube has an elongated elastomeric plug (not shown), which prevents delivery during storage and is removed by the patient to initiate flow. Schematic of the drug delivery extruder. (a) A lengthwise cross-section and (b) a crosswise cross-section of the drug delivery extruder. The extrudable semisolid drug paste The extrudable semisolid drug paste is engineered for safety, homogeneity, storage stability, and constancy of the extrusion rate. For safety, all paste components are restricted to common oral excipients and the established drugs LD and CD. The composition of the paste, formulated for years of storage without physical or chemical change, is shown in Table 1. The paste comprises 63% by weight of solid LD and solid CD, in a 4:1 ratio. Its five excipients, all listed on the FDA’s Inactive Ingredient Database and used in other FDA-approved orally-dosed drug products, are the edible oil Miglyol 812 N, a medium-chain triglyceride of 8-carbon caprylic and 10-carbon capric acids; water, the continuous phase of the suspension; Poloxamer 188, an emulsifying triblock polyethylene oxide-polypropylene oxide-polyethylene oxide copolymer, used in pharmaceuticals and cosmetics; antimicrobial benzoic acid; and the chelant and antioxidant edetate disodium (EDTA). Table 1Drug paste formulation.IngredientFunctionComposition (%w/w)Levodopa, micronizedActive49.5%Carbidopa monohydrate, micronizedActive13.4% (12.4% active)Medium chain triglycerides (Miglyol 812 N)Suspending liquid25.2%WaterSuspending liquid6.6%Poloxamer 188Emulsifier4.9%Benzoic acidAntimicrobial0.3%Edetate disodium, dihydrateChelant and antioxidant0.1% Drug paste formulation. Development of unacceptable dose-rate changing physical heterogeneity upon storage (for example, by sedimentation of solid drug particles, by separation of the aqueous and oil phases, or by particle growth by Ostwald ripening) were prevented by the transport-slowing massive apparent viscosity (Fig. 4). Mixing of the components yielded an oil-in-water emulsion matrix, as demonstrated in Supplementary Fig. S1, providing for rapid dispersion of the drug paste in saliva. The emulsion is stabilized against coalescence and phase separation both by the surfactant and by the dispersal of micronized solid particles, as in a Pickering emulsion28. Apparent viscosity of the drug paste. (a) Apparent shear rate dependence at 37 °C (red circles), 22 °C (green triangles), and 5 °C (blue squares), and the fit to a Herschel-Bulkley (‘H-B’) model (for the fitting parameters see Supplementary Table S1). (b) Temperature dependence at a shear rate of 5 s−1. The paste is semisolid, exhibiting solid-like behavior below a defined apparent yield stress, and liquid-like behavior above that yield stress29. The dependence of the apparent viscosity on the shear rate is shown in Fig. 4a. In rheological terms, the paste is a Bingham pseudoplastic, meaning that above a certain yield stress, the apparent viscosity exhibits shear thinning. A fit to a Herschel-Bulkley model suggests a yield stress of 3–5.5 kPa at storage (5 °C), ambient (22 °C), and physiological (37 °C) temperatures (Fig. 4a and tabulated in Supplementary Table S1). Below these yield stresses, e.g., during storage, the paste exhibits solid-like behavior meaning near infinite viscosity. As experienced during oral extrusion, with apparent shear rates between 1 and 10 s−1 at 37 °C, the apparent viscosity is remarkably high, of magnitude 103 Pa s or 106 cP. The temperature dependence is relatively weak, with the apparent viscosity varying between 1200 and 1900 Pa s between 5 and 37 °C at a shear rate of 5 s−1 (Fig. 4b). The non-monotonic dependence on temperature is largely due to the Poloxamer’s well-known thermo-reversible gelling30. The physical stability, i.e., the absence of sedimentation or phase separation, is evident in Supplementary Fig. S2, showing that the paste withstands centrifugation at 22 °C at 4000g for 2.25 h (9000 g-hr), corresponding to storage at normal gravity for 1 year. With both LD and CD being O2-oxidation-prone catechols, it is their rate of oxidation by dissolved O2 that controls the chemical stability. Being O2 diffusion-dependent, the bimolecular rate is inversely proportional to the viscosity, and is very slow at the near infinite apparent viscosity of the paste stored refrigerated (5 °C) or at ambient temperature (22 °C). Their oxidation is further slowed by the low solubilities of the drugs (1–10 mg/mL in water, nil in the oil), and by the low water concentration (< 7%w/w): less than 5 µmol of either LD or CD is dissolved in 1 g of paste. The stability is additionally improved by packaging in a laminated metal foil pouch. Stability is evident from the less than 20 ppm concentration of the downstream CD oxidation product hydrazine after two years of storage at 5 °C. Supporting chemical and physical stability data is tabulated in Supplementary Table S2. Constant rate delivery At the near-constant temperature of the mouth, the paste is extruded at a constant rate. The rheological data of Fig. 4 is used to characterize the drug paste flow behavior and design the flow controlling tubes of the extruder to provide the intended rate of paste extrusion. Due to the high yield stress of the paste and the slow rate of extrusion, the paste moves by plug flow, and the drug delivery rate is controlled primarily by wall slippage of the paste in the flow controlling tubes of the extruder31,32. Correspondingly, the delivery rate scales approximately with the cube of the internal radii of the flow controlling tubes and inversely with their lengths, as described in Supplementary Note 1. The flow controlling tubes can be tailored accordingly to provide the desired dose-rate by extending or trimming their length or by tuning their internal diameters. In the case of plug flow, the wall slip velocity and delivery rate scale inversely with the formulation-dependent suspending fluid viscosity, as described in Supplementary Note 1, which depends strongly on the water content. The wall slip velocity scales with the slip layer thickness, which scales with the suspended particle size32, where larger particles provide for more wall-lubricating fluid in the particle-excluded volume near the wall. It depends on the material of the flow controlling tube, where, for example, surface roughening decreases the slip length31. In sum, for flow controlling tubes of fixed material and dimensions, the rate of paste slippage is dictated primarily by the wall-lubricating water content and secondarily by the solid particle size distribution. Both the water content and the particle size distribution are tightly controlled and closely monitored during manufacture and storage. For passage through the 0.3–0.4 mm internal diameter flow controlling tubes, the particles are micronized to a mean particle diameter of 2–10 μm and a maximum diameter of 30 µm, such that individual particles are not greater than 1/10th of the flow controlling tube diameter. No change in water content or particle size is observed during storage, as shown in Supplementary Table S2. The photograph of Fig. 5 shows drug paste emerging from the tip of the delivery tube. Notably, the emerging semisolid paste retains its cylindrical shape and cross-sectional profile until it disperses in saliva. The drug delivery extruder in use. A photograph of the semisolid paste emerging from the delivery tube of the extruder; the extruded paste retains the cylindrical shape and the cross-sectional profile of the delivery tube until dispersed in saliva. The paste is extruded at a constant rate by constant-force deformation of the silver foil diaphragm upon volatilization of pressure liquefied 1,1,1,2-tetrafluoroethane confined to the propellant compartment. According to the Gibbs phase rule, the vapor pressure of a single component liquid is about constant at a constant temperature. In the mouth the temperature is about constant. Hot and cold drinks alter the temperature in the mouth only briefly, and as seen in Fig. 4b, the variation in apparent viscosity near 37 °C is mild, such that the amount of extruded paste over any 10-min period is unaffected. The 0.38 mm thick titanium walls withstand the force exerted by the 8.4 bar pressure difference at 37 °C between the interior of the 1,1,1,2-tetrafluoroethane propellant chamber and the external sea-level atmospheric pressure. Each extruder can deliver at a constant rate approximately 80% or 800 mg of its paste, that is, each extruder can deliver about 400 mg LD and 100 mg CD at a constant rate. A patient using 3 extruders per day can therefore receive up to 1200 mg LD and 300 mg CD. Figure 6 shows the constancy over 5 h of the paste extrusion rate for a random sample of n = 9 extruders from two separate product batches intended for 100 mg/hr paste delivery (50 mg/hr LD) and 136 mg/hr paste delivery (68 mg/hr LD). The group mean (± standard deviation) paste flow rates for the two batches are 100.5 ± 4.1 mg/hr and 136.0 ± 3.3 mg/hr, close to the targeted 100 mg/hr and 136 mg/hr, respectively. The chosen 5-hour period represents a typical interval between meals. The system is removed for meals and the extruder is replaced after breakfast, lunch, and dinner. Cumulative drug paste delivery showing constancy and reproducibility. The flow from individual extruders (n = 9) is shown in grey and the group’s average is shown in red circles (100 mg/hr target) or blue squares (136 mg/hr target). Error bars represent standard deviation of the group mean. Discussion This study demonstrates that solid drugs that are intermittently taken as tablets or capsules can be continuously and non-invasively delivered orally. The continuous delivery can be advantageous when the solid drugs have only a short plasma half-life and are short-acting. Their continuous delivery can make the drugs much more efficacious. The study also introduces an oral drug delivery system of solid drugs, formulated as semisolid pastes, such that the solid drugs can be continuously extruded. The semisolid paste extrusion allows for a greater than 100-fold reduction of the size and weight of previous propellant-driven, titanium walled infusion pumps. It demonstrates a miniature, solid drug delivering, constant-rate, propellant-driven, intraoral pump. In this exemplary study, a continuously extruded semisolid paste of 63% by weight 4:1 levodopa-carbidopa was engineered for managing Parkinson’s disease, the levodopa having a plasma half-life of approximately 90 min. This article focuses on the engineering design principles of the drug delivery system consisting of a propellant-driven extruder and semisolid drug paste. In a separate clinical study, continuous oral drug extrusion was demonstrated to safely and effectively reduce the OFF time of advanced Parkinson’s disease patients, validating that the drug delivery system can result in clinically significant improvements in PD treatment1. Outlook With the presented drug delivery system, which can extrude at a constant rate approximately 80% or 800 mg of its paste containing 63% by weight of solid drugs, up to 500 mg of solid drugs can be orally extruded at a constant rate per extruder. To maximize the amount of drug delivered, up to 2.5 g/day of solid drug could be delivered by five daily replaced unilateral extruders residing in one of the cheek pockets, or 5 g from ten bilateral extruders, residing in both cheek pockets. In clinical practice, the 5 g upper limit provides for the non-invasive and continuous delivery of any solid drug. Continuous delivery of such a large daily dose might overcome the failure of ALS management with a drug consisting mostly of sodium phenylbutyrate, which has a 60 min or shorter plasma half-life yet has been administered in recent clinical trials only a few times daily33–35. By tuning the formulation of the drug paste, particularly the water to oil ratio, the surfactant and its weight fraction, and the volume fraction of solids and their particle size distribution, the apparent viscosity can be modulated over a wide range providing not only higher or lower flow rates, but also of solid drugs suspended or dissolved in water or in oil. The 106 cP apparent viscosity semisolid paste of this study is a non-Newtonian fluid flowing by slip-flow, the flow rate scaling with the cube of the internal diameter of the plastic flow controlling tubes of the extruder. A similarly sized and designed 1 mL device could deliver a clear, particle-free, Newtonian fluid of low, e.g., 1–100 cP viscosity drug solution, using a reduced internal diameter flow controller, the flow rate now scaling with the fourth power of the internal diameter. Considering the broad design space enabled by tuning the formulation, extruder design, and dosing strategy, patients suffering of other conditions might also benefit of the continuous oral administration of short-lived drugs. These might include pyridostigmine bromide for managing myasthenia gravis, or baclofen for managing alcoholism or spastic movement disorders. The drug delivery system presented here might also warrant re-examination of drugs that could not previously be maintained in a sufficiently narrow therapeutic window and were therefore rejected as unsafe. Approximately 103 extruders were produced monthly for the study. To scale up the monthly production to over 105 extruders, the design of the delivery tube, the metal of the diaphragm and its bonding, the positions of the drug and propellant filling plugs, and the plastics and their joining were changed. All changes, aimed at mass-manufacturability, enhanced usability, and lower cost, remained consistent with the presented principles of design, and demonstrated the potential viability of continuous oral delivery implemented at a population level. Methods Apparent viscosity The apparent viscosities were measured with a Netzsch Rosand RH2000 capillary rheometer. Its 15 mm diameter barrel was made of nitrided steel. The barrel temperature was controlled using a fluid jacket connected to a Julabo CP-1000F heater/chiller recirculator. The pressure transducers could measure up to 3.5 MPa. The 24 mm long tungsten carbide die had a 1.5 mm diameter capillary and a 180° entrance angle. For shear rate ramps at a constant temperature (Fig. 4a), the shear rate was ramped down from 100 to 0.50 s−1. At each shear rate, a pressure stability endpoint was used such that the pressure converged within 0.3% and three sequential measurements were averaged for the reported values. For temperature ramps at a constant shear rate (Fig. 4b), the temperature was ramped down from 37 to 5 °C. A time-based endpoint was used such that every 2 min, three sequential measurements were averaged for the reported values. For all tests, reverse ramp data was collected to confirm a lack of path dependency. Model fits to the Herschel-Bulkley constitutive equation were presented in Fig. 4a and model parameters are presented in Supplementary Table S1. The Herschel-Bulkley model is a generalized non-Newtonian model described by three fit parameters: the yield stress (τ0), the consistency index (k), and the flow index (n). In one-dimension, a constitutive equation for the fluid shear stress (τ) can be written as τ=τ0+kγ˙n for shear stresses exceeding the yield stress (τ≥τ0), where τ is the fluid shear stress [Pa] and γ˙ is the shear rate. For shear stresses below the yield stress (τ<τ0), there is no flow (γ˙=0). The three model parameters were fit using a generalized reduced gradient (GRG) non-linear solver36 to the data plotted in Fig. 4a. Assays by HPLC Assays of the drug product were performed using a Thermo Fisher Scientific Dionex Ultimate 3000 or equivalent HPLC with UV-detection. To assay the LD and CD content, a method based on the USP monograph method for LD and CD tablets was employed. A 30 mg paste sample was dispersed in a 10 mL sample of 80/20 v/v 5% ethanol in phosphate buffer, pH 2.2 mobile phase / 1 N HCl diluent to make a 1.5 mg/mL LD and 0.4 mg/mL CD solution. A 10 µL filtered extracted sample was injected into the HPLC with mobile phase at a flow rate of 1.0–1.2 mL/min through a Phenomenex Luna C18(2) 3.9 × 300 mm, 5 µm column at 32 °C. The UV detector was set for 260 nm (LD) or to 280 nm (CD). In Supplementary Table S2, reported values are the average of two replicate measurements from two samples. Hydrazine was assayed in methanol extracts from 100 mg paste samples. The hydrazine was reacted with p-tolualdehyde in pH 3.5 50/50 v/v 100 mM acetate buffer/methanol, and the product was extracted into hexane. After centrifugation, 50 µL of the filtered hexane extract was injected into the HPLC with a mobile phase of 70/30 v/v hexane/isopropanol at a flow rate of 1.0 mL/min through a Perkin Elmer Spheri Cyano, 220 × 4.6 mm, 5 µm column at 40 °C. The detection wavelength was 310 nm. The values reported in Supplementary Table S2 are averages of four replicate measurements from two samples. Water content The water content was measured using a Mettler-Toledo C10S titrator by Karl-Fischer coulometric titration. A sample of 500 mg paste was extracted in 10 mL of anhydrous methanol. After centrifugation, 1 mL of the extract was coulometrically titrated. The values reported in Supplementary Table S2 are averages of two replicate measurements from two samples. Particle size analysis Particle size distribution measurements of the drug paste were made using a Malvern Mastersizer 3000 or equivalent by laser diffraction. A sample of 25–30 mg paste was dispersed in 25 mL of 1% Span 20 (sorbitan monolaurate) in isopropanol and an adequate amount of the dispersion was added to scatter 5–10% of laser light. The values reported in Supplementary Table S2 are the averages of 5 repeat measurements from a single sample. Flow rate measurements The rates of paste extrusion from extruders were measured by gravimetry using an analytical balance. After initiating the flow by removal of the elastomeric plug from the delivery tube, the extruder was placed in a dry bead bath maintained at 37 °C. The extruder was removed from the bead bath and weighed every hour. After the 5th hour, the propellant was drained from the extruder and the final weight was recorded to verify that the weight loss was due to extruded paste and not due to loss of propellant. If more than 20 mg propellant was lost during testing, the extruder was deemed to have failed, and the data was excluded. In Supplementary Table S2, each reported value is the average flow from a sample group of 10 extruders. Supplementary Information Supplementary Information. Acknowledgements Robert Draper, Carlton King, Jennifer Harmon, Phillip Plante, Taung Lau, Boaz Heller, Mina Truong, and Bruce Rehlaender contributed to the development of the drug delivery system. Robert Draper generated the schematics of Fig. 3. Mina Truong assisted in collecting the data in Fig. 4 and provided the photographs of Supplementary Figs. S1 and S2. Author contributions E.H. and A.H. conceived of the continuous oral extrusion of a paste of levodopa-carbidopa. A.H. and J.S. led the design of the drug delivery system and drug delivery extruder. W.M. and A.H. led the formulation and characterization of the drug paste. A.H, W.M., and E.H. wrote the manuscript. W.M. and A.H. generated the figures, with the help of others. All authors reviewed, commented, and edited the manuscript. Data availability The data presented in this study are available from the corresponding author on reasonable request. Declarations Competing interests At the time of the research and development for this study, all authors were employees and held equity at synAgile Corporation. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information William L. Murch, Email: murch.william@gmail.com Ephraim Heller, Email: eheller01@gmail.com. Supplementary Information The online version contains supplementary material available at 10.1038/s41598-024-78145-4. References Associated Data Supplementary Materials Supplementary Information. Data Availability Statement The data presented in this study are available from the corresponding author on reasonable request.