Manufacture of a sterile injectable from the protected azetidinone begins with the conversion of the tert-butyldimethylsilyl ether to the free hydroxyl before terminal sterile filtration, because the silyl residue is hydrophobic and not qualified for parenteral administration. Deprotection is conducted in anhydrous tetrahydrofuran at 5–10 °C using a fluoride source at 1.2–1.5 mol/mol; the reaction is quenched with 0.1 M phosphate buffer at pH 6.0 to terminate cleavage without opening the β-lactam ring. Residual fluoride is reduced to ≤ 10 ppm by diafiltration across a 10 kDa polyethersulfone membrane. The resulting active β-lactam solution is passed through a 0.22 μm PVDF capsule filter into 10 mL Type I borosilicate vials. Lyophilization is performed with a shelf ramp from −45 °C to −20 °C at 80–120 μbar, followed by secondary drying at 25 °C until Karl Fischer moisture by USP <921> is ≤ 1.0% w/w. Final container closure is sealed under nitrogen. Subvisible particulates are tested according to USP <788> Method 1, with limits of ≤ 6000 particles per container at ≥ 10 μm and ≤ 600 at ≥ 25 μm. Bacterial endotoxins are controlled to ≤ 0.25 EU/mg; sterility is verified by USP <71>. Unconverted silyl ether is monitored by LC-MS/MS with a reporting threshold of 0.10% area. The reconstituted product is prepared in 0.9% w/v sodium chloride injection at 5–10 mg/mL; the solution is used within 4 h at room temperature because the unprotected β-lactam undergoes pH-dependent degradation. This parenteral route is selected when the final active is not orally bioavailable.

Does Roller Compaction Preserve the Acetoxy Ester in High-Dose Oral Tablets?

When the bulk density of the protected azetidinone falls below 0.35 g/cm³ and the Hausner ratio exceeds 1.45 by USP <616>, direct compression is not used for tablets with a target weight of 400 mg or less because segregation produces weight variation outside USP <905>. The material is pre-blended with microcrystalline cellulose PH-102 and croscarmellose sodium in a bin blender at 12 rpm for 20 min; magnesium stearate is applied by external lubrication at 0.25% w/w during the final 3 min. Roller compaction is conducted on an Alexanderwerk WP 120 with roll gap 2.0–3.0 mm, roll force 50–70 kN, and roll speed 5–8 rpm. Ribbon density is maintained at 0.85–1.05 g/cm³; ribbons are milled through a 1.0 mm screen. The granules are compressed with precompression force 8 kN and main compression force 18–25 kN. Tablet breaking force is 70–110 N by USP <1217>, friability is ≤ 0.8% w/w by USP <1216>, and disintegration is ≤ 15 min in 0.1 M hydrochloric acid at 37 °C by USP <701>. The C4 acetoxy ester is the main stability risk during compaction; localized frictional heating above 40 °C can initiate hydrolysis if residual moisture exceeds 2.0% w/w. Therefore the suite is held at ≤ 30% relative humidity, and product temperature at the nip is monitored with an infrared sensor. Sodium bicarbonate, sodium carbonate, and amine-based buffers are excluded from the blend because alkaline conditions accelerate ester hydrolysis. Published data for degradation rate constants at roll forces above 70 kN are limited, so 70 kN is set as a design-space boundary.

Low-Dose Capsule Filling without Alkaline Excipients Requires a Two-Stage Dilution Train

For hard gelatin or HPMC capsules containing a low-dose β-lactam derived from the protected azetidinone, a direct lactose blend is rejected when potency RSD exceeds 6.0%. A two-stage geometric dilution is required: stage one passes the protected intermediate through a 250 μm sieve with lactose monohydrate at 1:10 w/w in a tumble blender at 10 rpm for 15 min; stage two adds the remaining lactose and pregelatinized starch and blends at 10 rpm for 20 min. The final blend must have a bulk density of 0.45–0.65 g/cm³ and a flow function coefficient ≥ 4.0 measured by ASTM D6773. Capsule filling is performed on an MG2 PLANETA dosator machine with a target fill weight of 350 mg and fill weight RSD ≤ 2.0%. Content uniformity is tested by USP <905> with an acceptance value ≤ 15.0. The silyl-protected intermediate is not filled directly because the TBS ether is cleaved under gastric pH; the final capsule contains the active β-lactam after deprotection. Croscarmellose sodium is limited to 2.0% w/w to minimize moisture uptake.

When Fluid-Bed Ethanol Granulation Is Required for Pediatric Oral Granules

Fluid-bed granulation with anhydrous ethanol is selected when the oral granules must be dispersible and the acetoxy ester cannot tolerate an aqueous granulation step. A binder solution of polyvinylpyrrolidone K30 at 5% w/w in anhydrous ethanol is sprayed at 8–12 g/min per kg of dry substrate. Inlet air temperature is 45–55 °C, product temperature is 25–32 °C, and inlet air dew point is −10 °C or lower. The unit is purged with nitrogen to keep oxygen below 8% v/v and ethanol vapor below 25% LEL. Granulation endpoint is defined by loss on drying 1.5–2.0% w/w, granule d50 180–250 μm, fines below 75 μm at ≤ 15%, and oversize above 500 μm at ≤ 5%. The dried granules are coated with a methacrylic acid copolymer dispersion in acetone-isopropanol to a weight gain of 8–12% w/w. Coated granules are packed in aluminum foil sachets under nitrogen; residual ethanol and acetone are controlled to the ICH Q3C Class 3 limit of ≤ 5000 ppm each. Unconverted silyl ether is limited to ≤ 0.10% area by LC-MS/MS. The final pediatric oral granules are reconstituted to 25 mg/5 mL or 50 mg/5 mL with purified water; the aqueous suspension is stored at 2–8 °C and used within 7 days. The protected intermediate has limited ethanol solubility; undissolved particles can blind the fluid-bed filter bags, so the bags are washed with acetone after every batch. Published data for this specific ethanol granulation configuration is limited; NIR spectroscopy is used to confirm the drying endpoint instead of relying on outlet air temperature alone.

No terminal sterilization by dry heat or gamma irradiation is applied to the protected azetidinone for injectable use, because the tert-butyldimethylsilyl ether is sensitive to radiation-induced free radicals and the acetoxy ester undergoes thermal elimination above 60 °C. Aseptic powder filling is performed in an isolator with EU GMP Grade A particle counts of ≤ 3520 particles per m³ at ≥ 0.5 μm. Sterile powder is filled into 20 mL vials at a target fill weight of 1.0 g with fill weight variation ± 5%. Container closure integrity is verified by vacuum decay according to USP <1207> with a leak size limit of 3 μm. Moisture is maintained below 1.0% w/w; endotoxin is controlled to ≤ 0.25 EU/mg; sterility is tested by USP <71>. The dry powder is reconstituted with 10 mL of water for injection and used immediately because the β-lactam ring hydrolyzes at 25 °C and the pH shifts from 6.0 to 5.5 within 2 h. This aseptic powder-fill route is used when lyophilization of the deprotected active introduces collapse or unacceptable residual moisture; the dry route avoids a prolonged aqueous holding period.

Residual Solvent and Elemental Impurity Limits Across Solid Oral and Parenteral Routes

Route-specific release criteria are compiled in the following matrix; the limits are applied at the release of the protected azetidinone intermediate and the finished dosage form.