For direct compression of bilirubin pharma grade API into low-load tablet cores with a drug loading of ≤1.0% w/w, the primary process constraint is not chemical instability alone but the particulate behavior of the drug substance as received. The API is first screened through a 250 µm sieve and, when necessary, deagglomerated by a low-shear comil equipped with a 0.039 in round-hole screen. A geometric dilution sequence is used: one part API is triturated with one part lactose monohydrate DC grade, then the mixture is expanded stepwise until the drug substance is distributed through the entire filler mass. Final blending is performed in a bin blender at 10–15 rpm for 15–25 min, with blend uniformity sampling at 10 positions and assay acceptance per USP <905> at an acceptance value ≤15. Magnesium stearate is added last at 0.25–0.5% w/w for 3–5 min because bilirubin is hydrophobic and prolonged lubrication can retard wetting and dissolution. Compression on a rotary tablet press with 8 mm round B tooling at 8–15 kN produces tablets with hardness 30–60 N, friability ≤1.0% per USP <1216>, and disintegration ≤15 min in 900 mL water at 37 °C. Dissolution method development follows ICH Q6A; because bilirubin exhibits pH-dependent aqueous solubility, a two-stage medium consisting of 0.1 N HCl for the first 2 h and pH 6.8 phosphate buffer thereafter is used for preliminary screening. The absence of an official dissolution monograph means that the discriminating power of the method must be demonstrated by deliberate formulation variants, including under-lubricated and over-lubricated cores.

On a rotary tablet press, the main failure modes observed are weight variation at start-up and lamination at the die wall because bilirubin-containing blends have limited plastic deformation. Precompression force is set at 2–4 kN to remove entrapped air, followed by main compression at 8–15 kN. Punch tips and die bores are checked for sticking every 30 min during development; if sticking occurs, press speed is reduced or magnesium stearate is adjusted within the 0.25–0.5% w/w range. Tablet hardness is tested at 10 tablets per sampling point and correlated with disintegration and dissolution to avoid overcompression. The dissolution method uses a paddle apparatus at 50 rpm, and the two-stage medium is degassed by helium sparging because dissolved oxygen can oxidize bilirubin during the test. Samples are protected from light using low-actinic glass and analyzed immediately after collection. The release criterion is not fixed at conventional 80% in 30 min; it is derived from the relationship between in vitro release and the formulation variant under study, using ICH Q8 design-space verification.

The substitution of dry processing with aqueous wet granulation introduces a process conflict: water is needed to build granules, but it also increases the mobility of dissolved oxygen and trace metal ions that accelerate conversion of bilirubin to biliverdin and propentdyopent oxidative fragments. In a high-shear granulator the binder solution, typically hypromellose at 3–5% w/w, is sprayed at 5–10 g/min while the jacket is held at 20–25 °C. The endpoint is controlled by impeller power draw rather than time because the hydrophobic API can produce non-uniform wetting at the same liquid-to-solid ratio. Wet mass is passed through a 1.0 mm sieve and transferred to a fluid-bed drier with inlet air temperature 50–60 °C and product temperature not exceeding 40 °C. Residual moisture is measured by Karl Fischer titration per USP <921>, with an endpoint of 1.0–2.5% w/w. Drying above this temperature window accelerates autoxidation; below it the drying time becomes impractically long. When the process cannot be made oxygen-free, disodium edetate is introduced at 0.005–0.02% w/w because iron and copper ions are known redox catalysts for bilirubin degradation. The dried granules are milled through a 0.8 mm screen, and the oxidation index is monitored by an HPLC method capable of separating biliverdin from bilirubin per USP <621>. Granule densification data are interpreted alongside oxidation index, not in isolation, because dry granule density can remain acceptable while chemical degradation has already occurred.

On production-scale fluid-bed dryers, the main equipment failure is not mechanical but thermal: bilirubin-loaded granules have low thermal conductivity and can remain at above-target temperature near the distributor plate even when the exhaust probe records 40 °C. Therefore, development batches use at least 6 temperature probes placed at different bed depths. The drying curve is monitored by loss-on-drying every 10–15 min until the 1.0–2.5% w/w endpoint is reached. Attempts to replace fluid-bed drying with microwave-assisted drying are rejected because non-uniform energy input creates hot spots that produce biliverdin and other green chromophores. The dried granules are sampled for bulk density, tapped density, and compressibility index per USP <616>, but the release decision is based on assay, related substances, and moisture. The oxidation index is reported as area percent of biliverdin at 450 nm detection using a reversed-phase HPLC method with a C18 column, mobile phase of methanol and ammonium acetate buffer at pH 5.0, and column temperature 25 °C.

Capsule filling is selected when direct compression cannot provide sufficient blend uniformity because the drug substance is present at 0.1–1.0% w/w of the fill weight. The API is micronized in a spiral jet mill using compressed nitrogen at 6–8 bar, and the particle size distribution is checked by laser diffraction per ISO 13320 with a target D90 ≤10 µm. The milled API is blended with lactose monohydrate and pregelatinized starch in a low-shear tumble mixer; colloidal silicon dioxide at 0.1–0.5% w/w is added as a glidant to improve powder flow into the encapsulator. An automatic dosator or tamping-pin capsule filler is used, and fill weight is controlled by weight sorting at 100% for development batches. Shell selection matters because bilirubin absorbs light strongly in the 450–500 nm region; opaque hard gelatin or HPMC shells containing titanium dioxide reduce transmitted light to the powder bed. Capsule shell moisture is maintained at 13–16% w/w for gelatin grades, and the filled capsules are stored under amber glass or foil overwrap after release. Uniformity of dosage units is assessed per USP <905> and mass uniformity per Ph. Eur. 2.9.5. Dissolution testing of capsules uses the same two-stage medium as tablet cores, but the capsule shell introduces a lag time; therefore the method includes a sinker and sampling at 15, 30, 45, and 60 min. Published data for bilirubin-specific capsule dissolution conditions are limited, so the acceptance criterion is established from batch-to-batch data under ICH Q8 design space principles.

Encapsulation of micronized bilirubin API is sensitive to triboelectric charging, which causes powder to stick to the dosing chamber and increases fill weight variability. The encapsulator is grounded and the relative humidity in the filling suite is controlled to 35–45%; outside this band the powder flow becomes erratic. Empty capsule shells are dedusted and filled within 4 h after blending to avoid segregation. Weight sorting is performed at 100% for development batches, and net fill weight data are charted by filling head to identify misaligned tamping pins. Capsule plug hardness is measured with a texture analyzer to confirm that the plug remains intact during ejection but breaks apart within 15 min in the dissolution medium. The capsule shell opacity is quantified by UV-visible spectrophotometry in the 450–500 nm range, and the shell specification includes a transmittance limit at 450 nm of ≤5% at shell thickness 100 µm. Because the drug substance is light-sensitive, the filled capsules are transferred immediately into amber glass jars with screw caps lined with induction-sealed aluminum foil and stored at 2–8 °C until packaging.

Aqueous injectable development with bilirubin pharma grade API departs from standard small-molecule parenteral formulation because the drug substance has negligible aqueous solubility at physiological pH. Reported aqueous solubility is below 0.1 mg/mL at pH 7.4; therefore a true solution cannot be achieved without a solubilizing carrier or a pH shift. Human serum albumin is the most physiologically relevant carrier because it possesses a high-affinity binding site for bilirubin with an association constant on the order of 10⁸ M⁻¹. In an injectable preparation, the total bilirubin-to-albumin molar ratio is maintained below 1:1 to avoid free, unbound drug substance. The compounding sequence is critical: bilirubin is first dissolved in a small volume of dilute sodium hydroxide or dimethyl sulfoxide under nitrogen, then added slowly to a chilled albumin-containing buffer at 2–8 °C with continuous stirring. The pH is adjusted to 7.8–8.5 using phosphate or tromethamine buffer, and the bulk solution is sterile-filtered through a 0.22 µm polyvinylidene fluoride or polyethersulfone filter. Filter adsorption studies are performed because bilirubin can bind to polymeric membrane surfaces; the loss on filtration is quantified by HPLC per ICH Q2(R1). Terminal sterilization by autoclaving is generally avoided because moist heat accelerates degradation; aseptic processing is used. For lyophilized presentations, mannitol at 2.5–5.0% w/v serves as a bulking agent and the fill is lyophilized with a primary drying shelf temperature not exceeding −25 °C and secondary drying not exceeding 25 °C. Headspace oxygen in the finished vial is controlled to ≤1.0% by nitrogen overlay. Release tests include sterility per USP <71>, bacterial endotoxins per USP <85>, visible particulates per USP <790>, and subvisible particulate matter per USP <788>. The absence of a harmonized solubility monograph for bilirubin injection means that the formulation specification is derived from ICH Q6A and regional pharmacopoeial general chapters.

Analytical control of the injectable solution uses a stability-indicating HPLC method with detection at 450 nm, which distinguishes bilirubin from biliverdin and lumirubin. Forced degradation studies are performed under acid, base, oxidative, thermal, and photolytic conditions per ICH Q1B and ICH Q2(R1). The biggest source of assay variability in parenteral development is not the chromatographic method but sample handling: bilirubin adsorbs to glass and stainless steel surfaces, so all volumetric transfers use polypropylene containers and low-light conditions. The lyophilized cake is reconstituted with water for injection, and the reconstitution time should be ≤3 min without visible aggregates. Subvisible particles are tested per USP <788> with the light obscuration method, and the acceptance limits follow the harmonized pharmacopoeial general chapter. Endotoxin testing uses a rFC or LAL method per USP <85>, with sample pH adjusted to 6.0–8.0 before analysis. Because bilirubin can interfere with chromogenic endpoint readings, a sample-specific interference test is performed on each product dilution.

Uncoated bilirubin tablet cores degrade rapidly under ambient fluorescent and daylight exposure, so the coating system is not decorative but functional. Photostability testing is performed according to ICH Q1B Option 2, with sample exposure to both cool white fluorescent and near-UV lamps. The aqueous film coating dispersion contains titanium dioxide at 1–3% w/w of the tablet weight and iron oxide pigments to attenuate the blue-light region where bilirubin absorption is strongest. A coating pan equipped with a perforated drum is run at 6–12 rpm, inlet air temperature 60–70 °C, exhaust air 40–50 °C, and product bed temperature 38–45 °C. Spray rate for a small development pan is 4–10 g/min at atomization air 1.0–1.5 bar. The target coating weight gain is 3–5% w/w; below this range pinholes allow light transmission, while above this range the disintegration time can exceed the 15 min limit. Cast film opacity is checked at 450 nm with a transmission target ≤5% at 20 µm dry film thickness. Coated cores are then packaged in aluminium/aluminium blisters rather than transparent PVC because the blister material itself must provide a second light barrier. The same coating formulation is used for tablets containing bilirubin and for placebo cores used in photostability trials, allowing separation of drug-related and excipient-related color changes.

The coating dispersion is prepared in a jacketed vessel at 20–25 °C and stirred for 45 min before use. The dispersion contains polyvinyl alcohol or hypromellose as film former, talc as anti-tack agent, and lecithin as plasticizer; the exact composition is adapted to the selected commercial ready-to-use system. The coating pan is preheated until the exhaust air reaches 45 °C before spraying begins. During coating, the bed temperature is kept below 45 °C to avoid oxidation of bilirubin near the tablet surface. Coated tablets are dried for 10–15 min at 40–50 °C in the pan with jogging every 2 min to prevent sticking. The final coated tablets are tested for moisture, hardness, disintegration, and dissolution; light protection is confirmed by exposing coated and uncoated cores to 1.2 million lux hours of cool white light and comparing related substances by HPLC. A package is acceptable only if the coated core and primary packaging together limit total related substances to the approved specification.

Mannitol is preferred over lactose as a diluent in bilirubin-containing oral granules when reducing sugar excipients are intentionally avoided and when the downstream sachet will be exposed to humidity. In a fluid-bed top-spray granulator, bilirubin is first dispersed in a mannitol/pregelatinized starch blend and granulated with a binder solution of povidone K30 or hypromellose at 3–5% w/w. The inlet air is set at 50–60 °C and the product temperature is maintained at 32–38 °C; atomization air pressure is 1.5–2.0 bar and spray rate is 8–15 g/min for a 2 kg development batch. The granulation endpoint is not defined by temperature alone; it is confirmed when the D50 particle size reaches 150–250 µm and fines below 75 µm remain ≤20% by sieve analysis per USP <786>. Residual moisture after drying is controlled to ≤2.0% w/w by Karl Fischer titration. The granules are filled into stick packs made from aluminium foil laminate with a water vapor transmission rate ≤0.1 g/m²/day at 38 °C and 90% RH. Each stick pack contains a silica gel sachet of 1 g for development batches. The oral granule presentation requires reconstitution or direct administration from the sachet, and dissolution testing is adapted from the tablet method with a reduced particle-size artifact because granules disperse more rapidly than compressed cores. No official bilirubin granule monograph exists, so release testing is aligned to Ph. Eur. 2.9.3 for dissolution and Ph. Eur. 2.9.5 for uniformity of mass of single-dose preparations.

The fluid-bed granulation process is sensitive to nozzle blockage when the binder solution is not filtered through a 150 µm screen before use. Blockage changes the droplet size distribution and produces overwetted pockets that dry into hard agglomerates. These agglomerates are removed by milling through a 0.8 mm screen, but the milled fraction can increase fines and reduce granule flow. Granule flow is measured by angle of repose and compressibility index per USP <616>, with a target angle of repose ≤40° for stick-pack filling. The stick-pack filler is set to operate with a jaw temperature of 140–160 °C for foil laminate sealing; dwell time is adjusted to achieve a seal strength ≥30 N/15 mm when measured by a tensile tester at 300 mm/min. Seal integrity is tested by vacuum leak method with methylene blue dye ingress, and the pass criterion is no dye penetration through the seal. Filled stick packs are pouched in aluminium foil over-wrap with desiccant, and residual oxygen in the pouch is purged with nitrogen to ≤2.0%.

Container-closure selection for bilirubin pharma grade API and its oral solid dosage forms is driven by the same two degradation vectors that control manufacturing: oxygen ingress and light exposure below 500 nm. Cold-form aluminium/aluminium blister cavities are used instead of PVC/PVDC when the coated tablet or capsule will be stored for more than 6 months in climate zone IVb conditions. The foil laminate provides a moisture barrier below 0.1 g/m²/day and eliminates direct light transmission, but it increases packaging cost and requires higher sealing temperatures. For bottling, high-density polyethylene containers with amber coloration are used with a heat-sealed foil induction liner and a child-resistant closure. Desiccant quantity is based on the sorption isotherm of the finished dosage form; a 1 g silica gel canister is often sufficient for a 30-count bottle when the tablet or capsule moisture specification is ≤2.0% w/w. Container performance testing follows USP <671> for moisture permeation and light transmission, and photostability of the packaged product is confirmed under ICH Q1B conditions. The packaging validation includes seal integrity by vacuum decay or dye ingress, and the acceptance criterion is no leak path greater than 15 µm. Batch-to-batch variance in aluminium foil thickness, which is typically 20–25 µm for the base foil and 60–90 µm for the lidding foil, must be recorded because thinning below the specification creates pinholes that compromise the light barrier. Published data on the long-term stability of bilirubin-containing pharmaceutical products in specific packaging configurations are limited, so the packaging system is selected through accelerated and stress studies with assay, related substances, and moisture as the primary responses.