When 2.5 mg and 5 mg dose strengths are processed by direct compression, the milled apixaban intermediate is first geometrically pre-blended with a lactose monohydrate carrier and microcrystalline cellulose. The purpose of this step is not only to satisfy USP <905> uniformity criteria but also to prevent segregation of a low-dose active with a D90 below 20 µm. A 600 L bin blender filled to 60% of gross volume and operated at 12 rpm typically reaches an in-process assay RSD below 5.0% after 15–20 min for a 1–3% w/w active premix. Production-scale lines encounter two competing constraints. Micronized intermediate improves dissolution but reduces bulk flow; a flow function coefficient below 2.0 can produce weight variability and sticking on 9 mm round tooling. Fumed silica at 0.25–0.5 wt% of the total blend mass is introduced to raise the flow function coefficient above 4.0 without retarding release. Magnesium stearate is limited to 0.5–1.0 wt% and total final blending time to 3–5 min; over-lubrication creates a hydrophobic film that delays dissolution and lowers tablet tensile strength. Compression on a rotary press at 15–25 kN yields tablets with tensile strength above 1.5 MPa and friability below 0.8% when measured by USP <1216>. Dissolution is assessed by USP <711>; for immediate-release cores, release above 80% within 30 min is a routine development target in surfactant-containing media. Segregation during hopper transfer remains the main batch-to-batch defect: press speed is limited to 60–80 tablets/min when the blend contains more than 15% fines below 150 µm. If direct compression is not viable, the same intermediate is re-routed to dry granulation.

Can Low-Fill-Weight Hard Capsules Maintain Content Uniformity Without Over-Lubrication?

Capsule filling of a 2.5 mg apixaban intermediate blend presents a different failure mode than tablet compression. The target fill weight for a size 3 hard capsule is often below 150 mg; at this scale, a tamping pin filler requires a powder bed with Carr index below 25 and Hausner ratio below 1.25. Without these flow properties, fill weight RSD exceeds 4.0% and content uniformity testing under USP <905> becomes the primary release risk. The intermediate is pre-dispersed in a dry blend of lactose monohydrate, microcrystalline cellulose, and croscarmellose sodium. Sodium stearyl fumarate at 0.5–1.0 wt% is used in place of magnesium stearate when dissolution lag is observed; the former generates less hydrophobic surface coverage on the active. Hard gelatin shells equilibrated at 13–15% moisture transfer water into a dry fill blend and can initiate sticking inside the dosing chamber. HPMC shells at 4–6% moisture offer lower water activity and are better matched to amide-containing actives. Disintegration is verified by USP <701> at 37 ± 2 °C; immediate-release capsules in development are expected to disintegrate within 15 min in aqueous media. Dissolution using USP <711> apparatus II at 50 rpm in 900 mL surfactant-containing medium is run on 12 units; a release plateau below 80% at 30 min typically traces back to over-blending with magnesium stearate or excessive tamping pin pressure. Production-scale capsule fillers with 150,000 capsules/hour capacity must be validated for pin lubrication and dust extraction; dust generation from low-fill-weight blends without adequate fines control can shift fill weight drift by more than 3% across an 8-hour batch. Capsule shell moisture and fill blend water activity are logged at line start, mid-batch, and end-batch to control shell brittleness and cross-linking potential.

Granule Flow and Residual Moisture Boundaries for Sachet and Sprinkle Formats

High-shear wet granulation introduces a moisture-driven degradation threshold that is not present in direct compression or roller compaction. Apixaban intermediate contains amide linkages; exposure to water at 40–60 °C during wet massing can increase hydrolysis-related impurities if drying is delayed. The granulating fluid is therefore formulated as a 15–25% w/w aqueous binder solution added over 3–5 min at rotor speed 200–400 rpm, or as an isopropanol-water mixture to reduce water activity. Wet mass is discharged through a 1.0 mm screen and dried in a fluid bed at inlet air 50–60 °C until loss on drying is below 2.0% by USP <921> Karl Fischer or LOD method. Final granules are screened to retain a 180–850 µm fraction with D50 between 250 µm and 400 µm. Fines below 150 µm increase segregation risk in form-fill-seal sachet filling; oversized material above 850 µm causes incomplete dispersion when the dose is sprinkled into water or soft food. The granule flow function coefficient should exceed 4.0 before a vertical sachet filling line can hold weight variation below 5% RSD. A dispersion test in 50 mL water at 25 °C is used to reject granule batches that retain agglomerates; the suspension is passed through a 500 µm sieve and visually inspected for undispersed particles. Published data for a specific apixaban sachet formulation are limited; the above thresholds derive from general scale-up data for low-dose poorly soluble anticoagulant granulations. If the sachet format includes a desiccant pouch, the granule moisture specification may need to be tightened below 1.5% to offset moisture ingress through laminate edges. Stability under 40 °C/75% RH conditions per ICH Q1A is used to define the final blend packaging configuration.

Because a sterile injectable presentation of apixaban intermediate has no monograph-recognized formulation, development begins with a solubility screen rather than a fixed manufacturing route. Published data specific to apixaban injectable solutions are limited. The neutral amide core can require pH adjustment, co-solvents such as propylene glycol and polyethylene glycol 300 at 20–40% total organic content, or complexation with a cyclodextrin to reach a therapeutically relevant concentration. Co-solvent levels above 40% create pain on injection and raise osmolality above 320 mOsm/kg; the target isotonic range is 280–320 mOsm/kg. Terminal sterilization at 121 °C for 15 min may stress amide-containing APIs. Degradation pathways must be mapped before a terminal sterilization cycle is selected; otherwise, aseptic filtration through a 0.22 µm PVDF membrane with a validated bubble point test is used. Fill volume for a single-dose injection typically ranges from 1 mL to 10 mL. Subvisible particulate matter is controlled by USP <788>; for small-volume injections, the light obscuration method limits are not more than 6000 particles per container at ≥10 µm and not more than 600 particles per container at ≥25 µm. Visible particulates are controlled by USP <790> under inspection lighting. Bacterial endotoxin testing by USP <85> uses a product-specific limit calculated from the maximum bolus dose in mL/kg and the endotoxin limit of 5 USP EU/kg per hour for intravenous administration, not from a universal input value for the raw API. Sterile filling is executed in an ISO 14644-1 Class 5 environment with unidirectional airflow and a restricted-access barrier system or isolator. If solution stability data show unacceptable impurity growth after 24 h at 25 °C, lyophilization is evaluated; the glass transition temperature of the frozen matrix is not predicted from the API alone, so collapse temperature is measured by freeze-drying microscopy before cycle design. No injection route should be locked without spike-and-recovery data for the filter membrane and compatibility data for silicone tubing and elastomeric closures.

Roller compaction offers a dry granulation route for apixaban intermediate when direct compression blends fail flow function criteria or when a higher drug load is targeted in a fixed-dose combination. Ribbon density is controlled between 1.1 g/cm³ and 1.3 g/cm³ by adjusting hydraulic roll pressure from 30 kN to 60 kN at roll speed 5–15 rpm. Milling through a 0.8–1.0 mm screen produces granules with D50 between 200 µm and 400 µm. The dominant failure mode is fines generation: if granules below 150 µm exceed 30% of the batch, segregation during bin transfer and compression hopper hold-up causes content uniformity drift exceeding 5% RSD. In that event, the granulation is re-screened and blended with 0.25 wt% colloidal silicon dioxide before the final lubrication step. The granule flow function coefficient after roller compaction typically increases from below 2.0 to above 4.5, enabling rotary press speeds of 60–80 tablets/min for 9 mm round tooling without weight variation beyond 3% RSD. Post-granulation mixing with magnesium stearate is limited to 3–5 min; longer times redistribute hydrophobic lubricant onto granule surfaces and depress early dissolution. Ribbon density variability across the roller width is measured at the left, center, and right positions because edge leakage produces low-density ribbon that mills into excess fines. If ribbon density drops below 1.0 g/cm³, the roll side seals must be inspected for wear and the feeding screw speed adjusted to maintain a stable nip angle. Dissolution is tested by USP <711>; any release plateau shift after 4 weeks at 40 °C/75% RH should trigger a review of the lubricant level and granule porosity rather than automatic reformulation.

What Changes When an Immediate-Release Film Coat Is Applied to Low-Dose Cores?

Film coating of apixaban immediate-release cores introduces heat and moisture stress that can shift dissolution and mechanical integrity. A 24-inch side-vented pan with pan speed 6–12 rpm, inlet air 60–70 °C, exhaust 40–50 °C, and spray rate 8–15 g/min applies an aqueous Opadry II dispersion at 10–12% solids to a 2–4% weight gain. If the core contains croscarmellose sodium, water uptake during the first coating phase can induce swelling and edge chipping; lowering the spray rate and raising inlet air temperature reduces moisture penetration into the core. Hardness should not fall below 1.5 MPa and friability should remain below 0.8% after 2–4% coating weight gain. Dissolution is retested by USP <711> because excessive coating weight or uneven pan distribution can create a release-lag artifact that is unrelated to the API. Coating pan load size is selected to keep tablet bed surface speed constant; a batch that fills 60–70% of pan volume is used for reproducible mixing, while lower fill levels can result in spray drying and higher fill levels can cause tablet breakage at the bed surface. Core preheating is monitored with an infrared sensor; exhaust temperature above 50 °C for more than 10 min can overdry the core and increase friability. Final visual inspection under USP <790> lighting conditions is used to reject coated tablets with film peeling, color nonuniformity, or edge exposure. Coating suspension preparation must follow a validated hold time because Opadry II dispersions beyond 12 h can undergo viscosity drift and deliver variable spray droplet size. Final release testing under FDA 21 CFR 211.165 links the coating step back to core uniformity, dissolution, and impurity data before the batch is dispositioned.