🖨️ Personalized Dose Titration via 3D Print Geometry

Why Fixed Tablet Strengths Fail Precision Dosing

Most oral solid drug products are manufactured in a small, discrete set of strengths chosen for manufacturing and regulatory convenience — not for the continuous range of doses that pediatric growth, renal/hepatic impairment, and narrow-therapeutic-index (NTI) titration actually require. When the needed dose falls between strengths, clinicians resort to tablet splitting or extemporaneous compounding, both of which carry substantial, poorly controlled dosing error.

• 12: Levothyroxine strengths (US) (25–300 µg, NTI drug, TSH-guided titration)
• 15–31%: Split-tablet dose deviation (from nominal half-dose, unscored tablets worst)
• ≤10%: Warfarin dose adjustments (typical INR-guided step, below smallest strength gap)
• ~50–75%: Pediatric off-label use (of hospital pediatric prescriptions)

Narrow-therapeutic-index drugs and the titration gap

Fixed dosage forms create a mismatch between what pharmaceutical manufacturing can efficiently produce and what individual patients need:

Levothyroxine (thyroid hormone replacement): • US market: 12 discrete strengths (25, 50, 75, 88, 100, 112, 125, 137, 150, 175, 200, 300 µg) • Narrow therapeutic index: therapeutic window separated from toxicity by <2-fold • Dose titration guided by TSH, typically adjusted in 12.5–25 µg steps • Over/under-replacement risks: atrial fibrillation, osteoporosis (over) vs. fatigue, myxedema (under) • Elderly and cardiac patients often need increments smaller than the 12.5 µg gap between strengths

Warfarin (vitamin K antagonist): • Available as 1, 2, 2.5, 3, 4, 5, 6, 7.5, 10 mg tablets • INR-guided titration: dose changes of 5–20% of total weekly dose are common • Genetic polymorphisms (CYP2C9, VKORC1) create 10-fold inter-patient dose variability • Fractional tablet splitting routinely used off-label to hit intermediate doses

Pediatric pulmonary hypertension — sildenafil: • Weight-based dosing (e.g., 0.5–2 mg/kg three times daily) changes continuously as the child grows • No pediatric-strength tablet exists; extemporaneous suspensions compounded from adult tablets • Compounded suspension stability and potency vary by pharmacy, vehicle, and storage — reported assay deviations of ±10–25% from labeled strength in published stability studies

Propranolol — infantile hemangioma: • First-line therapy since 2008; dosed 1–3 mg/kg/day, split into 2–3 doses • Infant body weight can change >20% within weeks of treatment initiation • Every weight-based recalculation currently requires a new compounded batch or tablet-splitting workaround

Tablet-splitting error data: • Multiple pharmacopeial and clinical studies (van Riet-Nales et al.; Elliott et al.) report splitting deviations of 15–31% from the nominal target weight • Deviation worsens for unscored tablets, small tablets (<8mm), and patients with reduced manual dexterity (elderly, arthritic, pediatric caregivers) • Deviation compounds over serial dose adjustments during active titration

The structural gap: manufacturing discrete strengths is efficient at population scale but poorly suited to the continuous, individualized dose trajectories required during titration — this is exactly the gap 3D-printed dosage forms are designed to close.

From Prescription to Printlet — the Fused-Deposition Dosing Pipeline

A 3D-printed tablet ("printlet," a term coined by the FabRx/UCL group) is built from a drug-loaded filament or paste whose drug concentration never changes. The dose is instead set by how much of that filament is deposited — a function of infill density, shell thickness, and overall tablet geometry — converted automatically from a prescribed mg dose into machine instructions.

• 1.75 mm: Filament diameter (standard desktop FDM feedstock)
• 5% w/w: Typical drug loading (API dispersed in PVA/HPMC matrix via HME)
• 160–210 °C: Nozzle temperature (polymer-dependent (PVA higher, HPMC lower))
• 40–60 mm/s: Print speed (per-layer head travel speed)

Hot-melt extrusion, slicing, and dose-to-geometry conversion

Step 1 — Drug-loaded filament manufacture (hot-melt extrusion, HME): • API physically dispersed (dissolved or suspended) into a thermoplastic polymer carrier — polyvinyl alcohol (PVA), hydroxypropyl methylcellulose (HPMC), or Eudragit grades • Twin-screw extruder blends API + polymer + plasticizer at 120–180°C, cools into 1.75mm filament wound on a spool • Drug loading fixed at manufacture (commonly 1–30% w/w depending on API potency and target dose range) — this loading is the one formulation variable that stays constant across every dose strength printed from that spool

Step 2 — Dose prescription and parameter calculation: • Pharmacist/software enters target dose in mg • Required printlet volume = target_dose / (drug_loading × polymer_density) • Given a fixed external tablet shape (e.g., 10mm biconvex disc), required volume maps to an infill percentage via calibration curve established in Stage 3 • Software (Cura-derived slicers, or bespoke pharmaceutical control software) auto-generates: infill %, number of shell perimeters (typically 2–3), top/bottom solid layers, layer height

Step 3 — Slicing to G-code: • 3D model (STL) sliced into sequential horizontal layers at chosen layer height (100–300 µm) • Infill pattern selectable: grid, rectilinear, honeycomb, gyroid — each with different mechanical and disintegration properties • G-code specifies nozzle XYZ path, extrusion rate, temperature, and speed per layer

Step 4 — Printing: • FDM: filament fed through heated nozzle (0.4mm typical diameter), melted, deposited layer by layer, building the tablet bottom-up • Semi-solid extrusion (SSE) / pressure-assisted microsyringe (PAM): drug-loaded paste or gel extruded at room or mild temperature — gentler on heat-labile APIs, no HME step required • Drop-on-demand / binder jetting: inkjet-style deposition of liquid binder onto a powder bed (the ZipDose approach used in Spritam, see Stage 4) • A standard 200mg printlet at 200 µm layer height and 45% infill typically completes in 12–17 minutes on a single-nozzle desktop-class pharmaceutical printer

Step 5 — Post-processing: • Optional drying/curing step for SSE prints • Visual and dimensional inspection before release to content-uniformity testing (Stage 3)

Proving Volume Predicts Content — Linearity, Range, and Content Uniformity

The entire personalized-dosing premise rests on one empirical claim: printed geometry predicts delivered drug content with pharmaceutical-grade precision. This is established by printing a calibration series across the infill/geometry range, assaying each printlet by HPLC, and regressing measured content against nominal design dose — then confirming batch-to-batch consistency against USP <905> content-uniformity acceptance criteria.

• R² > 0.99: Reported linearity (infill % / volume vs. HPLC-assayed dose)
• 5–400 mg: Demonstrated dose range (single filament, geometry-only adjustment)
• ≤ 15: USP <905> AV limit (acceptance value, 10-unit content uniformity)
• 1–3%: Typical assay RSD (HPLC-UV, n=10 printlets per level)

Calibration curves, achievable dose range, and USP <905> compliance

Calibration protocol: 1. A fixed external tablet geometry is chosen (e.g., 10mm diameter biconvex disc, or a caplet) 2. A series of printlets is manufactured spanning the practical infill range (e.g., 10%, 25%, 33%, 50%, 75%, 100%) or, equivalently, varying tablet height/volume at fixed infill 3. Each printlet is individually assayed by reverse-phase HPLC-UV (or HPLC-MS for low-dose/high-potency APIs) after dissolution in a validated diluent 4. Measured drug mass (mg) is plotted against nominal design volume or infill % → linear regression

Representative published results (academic FDM printlet literature): • Early proof-of-concept (Skowyra et al. 2015, prednisolone printlets): dose range 2.0–3.9 mg achieved via infill 10–100%, linear fit R² = 0.9983 • Paracetamol / 5-fluorouracil / captopril printlet series (FabRx/UCL body of work): linearity consistently R² = 0.99–0.999 across infill 10–90% • Achievable absolute dose range depends on drug potency and target tablet size: reported spans from single-digit mg (potent APIs, low loading) up to 400+ mg (bulk excipient-heavy tablets, e.g. isoniazid, high loading) • Because loading % is fixed per filament spool, one spool typically covers roughly a 5–8-fold dose range before a new higher/lower-loading spool is needed for extreme ends

Content uniformity — USP <905>: • Acceptance Value (AV) calculated from 10 dosage units: AV = |M − X̄| + k·s • AV ≤ 15 required for release (L1 stage); tighter L2 stage (30 units) triggered if L1 fails • Printlet studies typically report RSD of 1–3% across a printed batch — comparable to or better than conventional direct-compression tableting for low-dose actives • Sources of variability: extrusion consistency (nozzle clogging, filament diameter tolerance ±0.05mm), moisture uptake by hygroscopic polymers (HPMC) between print and assay, cooling-rate effects on crystallinity

Why this matters clinically: • A validated linear model lets a pharmacist type a prescribed mg dose and receive an automatically generated, assay-traceable print file — no separate formulation development needed per strength • The same regulatory logic used for conventional tablet strength bridging (dissolution similarity, content uniformity, stability) applies, just parameterized by geometry instead of formulation

Spritam, FabRx, and the Regulatory Path for Printed Dosage Forms

3D-printed medicines already have an FDA approval on record — Spritam (levetiracetam), approved August 2015 — establishing that the regulatory pathway for printed solid dosage forms is real and navigable. Its mechanism, however, is porosity-for-disintegration, not geometry-for-dose, and it is worth being precise about that distinction before extrapolating to personalized titration.

• Aug 2015: Spritam FDA approval (Aprecia Pharmaceuticals, ZipDose)
• <10 s: Spritam disintegration (with a sip of liquid, high-porosity matrix)
• CE-marked 2020: FabRx M3DIMAKER (first pharma-grade 3D printer, Class I device)
• 100–400 µm: FDM print resolution (XY ~50–100 µm; layer height 100–300 µm)

Distinguishing porosity-engineered disintegration from geometry-tuned dosing

Spritam (levetiracetam) — what it actually is: • Manufactured by Aprecia Pharmaceuticals using ZipDose technology, licensed from MIT's original binder-jetting 3DP patents • Process: powder bed of levetiracetam + excipients, layers bound by an inkjet-deposited liquid binder — a drop-on-demand additive process, not FDM • Result: a highly porous, friable matrix that disintegrates in under 10 seconds with a small sip of liquid — designed for patients with dysphagia or seizure disorders who struggle to swallow conventional tablets • Manufactured in bulk, in fixed strengths (250, 500, 750, 1000 mg) — every tablet of a given strength is nominally identical • Critical distinction: the printing process here is exploited for porosity/disintegration engineering, not for continuously varying dose by geometry — Spritam is not personalized-dose 3D printing, but it is the regulatory proof that additively manufactured tablets can clear FDA review

Academic precedent — FabRx and the UCL School of Pharmacy group: • Founded 2014 as a UCL spin-out; scientific founders include Abdul W. Basit and Simon Gaisford, with Alvaro Goyanes as a lead researcher driving much of the dose-linearity body of work • Coined "printlet" for a 3D-printed tablet; published the first FDM-printed dose-varying tablets (prednisolone, 2015) demonstrating the infill-controls-dose principle used throughout this pipeline • M3DIMAKER: first pharmaceutical-grade 3D printer purpose-built for medicines manufacture, CE-marked as a Class I medical device in 2020 — designed for hospital and community pharmacy point-of-care use, not industrial-scale production • Broader portfolio: polypills (multiple APIs in one printlet with independent release profiles), chewable pediatric printlets, orodispersible films

Resolution and process parameters across the literature: • FDM print (XY/nozzle) resolution: roughly 100–400 µm depending on nozzle diameter (commonly 0.4mm) and extrusion consistency • Layer height: 100–300 µm is the practical range — finer layers improve dose resolution and surface finish but increase print time roughly proportionally • SSE/semi-solid extrusion offers comparable resolution with lower thermal stress on the API, useful for heat-labile drugs (peptides, some biologics-adjacent small molecules) • Binder-jetting (Spritam-style) trades dose-geometry precision for very high porosity and very fast disintegration — a different point on the same design space

The regulatory takeaway: Spritam shows FDA will approve a 3D-printed drug product built around a validated, fixed manufacturing process. Personalized dose-by-geometry systems will need to additionally demonstrate that point-of-care or pharmacy-level manufacturing, with per-unit geometry varying by prescription, meets the same content-uniformity, stability, and process-control expectations — a harder but increasingly precedented bar, discussed next.

Closing the Loop — From Prescription to Bedside-Manufactured Dose

The clinical payoff of geometry-tuned dosing is a workflow where the prescribed dose and the delivered dose become the same number, verified at the point of dispensing rather than approximated by splitting or compounding. Pilot programs in hospital and community pharmacies are now testing exactly this: prescription in, exact-dose printlet out, same visit.

• 5–20 min: Point-of-care print time (per unit dose or small batch)
• NIR / Raman: In-line QC methods (plus gravimetric weight check per unit)
• Pediatric, geriatric, NTI: Target population (levothyroxine, warfarin, propranolol titration)
• ↑ reported: Adherence signal (flavored/shaped pediatric printlet acceptability studies)

The point-of-care pharmacy workflow and quality-control requirements

End-to-end workflow being piloted: 1. Prescriber enters an exact mg dose into an e-prescribing system, informed by weight, renal/hepatic function, or lab-guided titration target (TSH, INR) 2. Dosing software (validated against the Stage-3 calibration curve for that drug/filament combination) converts the mg target into infill %, layer count, and shell thickness 3. Local pharmacy 3D printer (e.g., FabRx M3DIMAKER-class device) manufactures the printlet on-site — typically 5–20 minutes for a single dose or small batch, well within a same-visit dispensing window 4. In-line quality control: gravimetric weight check against expected mass, plus near-infrared (NIR) or Raman spectroscopy for rapid non-destructive content verification — replacing destructive HPLC assay for routine release 5. Dispensed to the patient with full batch/print-parameter traceability tied to the original prescription

Quality-control requirements unique to distributed point-of-care manufacture: • Process analytical technology (PAT) must run in real time at each pharmacy node, not just centrally — NIR calibration models must be validated per drug/polymer combination • Filament/paste lot traceability: drug loading % must be confirmed per spool/batch before any dose calculation is trusted • Printer qualification (IQ/OQ/PQ) and preventive maintenance become pharmacy-level responsibilities, not just central-manufacturing ones • Environmental control: humidity affects hygroscopic polymers (HPMC) both during printing and short-term storage before dispensing

Patient-facing outcomes reported or targeted in pilot work: • Pediatric formulations: flavored, colored, and shape-customized printlets (stars, toy-like shapes) reported to improve palatability and caregiver-reported acceptability versus bitter extemporaneous suspensions • Titration precision: geometry-defined dosing removes the ±15–31% error band inherent to tablet splitting, directly relevant to levothyroxine, warfarin, and weight-based pediatric regimens where small errors compound over serial adjustments • Polypills: multiple APIs printed into a single dosage form with independently tunable release profiles per compartment, targeted at reducing pill burden in polypharmacy (elderly, cardiovascular) patients • Turnaround: because print time is minutes rather than the days required for external compounding pharmacy dispatch, a titration adjustment can in principle be actioned the same day a lab result (TSH, INR) returns

The strategic insight of this pipeline is that dose personalization is decoupled from drug reformulation. The API concentration in the filament or paste never changes — only the printed volume does. That means a single FDA/EMA-reviewed formulation (one drug loading %, one polymer carrier, one validated calibration curve) can in principle cover an entire clinically relevant dose range, turning "personalized medicine" from a bespoke-manufacturing problem into a software-and-geometry problem — the same regulatory logic as adjusting an infusion pump rate, applied to a solid oral dosage form.