Chromatography Step Record

A Chromatography Step Record is the executed, reviewable unit of evidence for a purification operation within a batch or lot, typically covering packed-bed column, membrane adsorber, or simulated moving bed steps orchestrated by a skid or DCS. It binds recipe intent (ISA‑88 master/control recipe) to execution actuals, with full traceability to materials, equipment, parameters, and decisions (e.g., pooling rules).

Regulators treat it as part of the batch production and control record (21 CFR 211.188), subject to controls on computerized systems (21 CFR 211.68; EU GMP Annex 11) and electronic records/signatures (21 CFR Part 11). In an MES, it is a first-class object aligned to ISA‑88 unit procedures, providing the bridge from Level 2 control data to Level 3/4 genealogy and release.

The record must capture enough context to reconstruct what happened, prove compliance to the control strategy, and support release/recall decisions. It should be recipe-driven, phase-granular, and complete at the parameter, material, equipment, and decision levels.

• Identity and context: batch/lot, product, stage, site, line/unit, skid/controller IDs, date-time with timezone and synchronized clocks.
• Equipment and column: skid asset ID, column pack/serial ID, resin/membrane vendor, catalog number, lot/exp., column dimensions, bed height, packing verification (e.g., HETP/asymmetry).
• Materials and buffers: buffer/component IDs and lots, preparation records, expiry/use-by, sampling status, WMS issuance and quantities, waste disposition.
• Phases and parameters: equilibrate/load/wash/elute/strip/regenerate/sanitize; setpoints and actuals for flow, pressure, column delta-P, conductivity, pH, temperature, UV (A280/A260), gradient profiles, volumes, contact times.
• In-process controls and criteria: pooling rules (conductivity/UV cutpoints), fraction maps, IPC sample pulls, hold times, acceptance criteria, alarms/interlocks and permissives.
• Results and yields: step yield, purity/enrichment, mass balance (load vs pooled vs waste), deviations/nonconformances, OOS/OOT links.
• Security and review: audit trail entries, versioned attachments (e.g., skid trend exports), electronic signatures, independent review/approval, and record retention metadata.

ISA‑88 provides the grammar to model chromatography consistently across products and skids. A unit procedure (Chromatography) is decomposed into operations (e.g., Column Prep, Load, Elute, Regenerate) and phases (e.g., Equilibrate, Load-to-Breakthrough, Step Elute, Gradient Elute, Strip, CIP, SIP). The control recipe binds equipment modules (skid, pumps, detectors, fraction collector) and parameters (setpoints, limits, methods) to a batch instance.

• Master vs. control recipe: define standard methods, allowable ranges, and criticality; instantiate with batch-specific targets.
• Equipment vs. control modules: map pumps/valves/detectors to phase logic; capture interlocks and permissives.
• Procedure function charts and state models: record transitions, holds, aborts, and exception paths with timestamps.
• Parameter class libraries: standardize units, engineering limits, alarm severities, and calculation rules (e.g., cumulative load mass, HETP).

A chromatography step record aligned to ISA‑88 simplifies recipe governance and validation (traceability between URS–FDS–Test), and enables phase-by-phase release or holds while preserving consistent auditability.

ISA‑95 clarifies boundaries and data exchanges between control (Levels 0–2), MES (Level 3), and business systems (Level 4). A robust chromatography record references real-time data from Level 2 while anchoring material genealogy, quality status, and inventory movements at Level 3/4.

The record should preserve one-up/one-down traceability: resin lot → column pack record → chromatography step → pooled intermediate lot → subsequent steps. ERP/WMS postings must reconcile with MES-recorded consumption/production to maintain accurate inventory and recall readiness.

Chromatography records must be attributable, legible, contemporaneous, original/true, and accurate (ALCOA+) with controlled computerized systems. 21 CFR 211.68 requires appropriate controls over automated equipment, including verification of input/output, invalid data handling, and backup. 21 CFR Part 11 and EU Annex 11 require secure, computer-generated audit trails, user access controls, validated system behavior, and binding electronic signatures for record creation/modification/review.

• Audit trails: parameter changes, method versioning, phase holds/aborts, alarm acknowledgments, pooling overrides, signature events — all time-stamped, user-attributed, and immutable.
• Electronic signatures: role-appropriate signoff on execution and review, linked to the specific record and meaning (per 11.50/11.70).
• Validation: risk-based verification (GAMP 5) of interfaces (skid→MES), calculations (e.g., mass balance), and exception workflows; periodic review and backup/restore tests.
• Security: RBAC, SSO, least privilege, session lock, and time synchronization across skid PLC/HMI, historian, and MES.

Pooling criteria often combine process signals (e.g., UV/cond cutpoints) with in-process testing (e.g., purity, host-cell protein) from LIMS. The chromatography record must show the decision basis and trace sample IDs → results → pool inclusion/exclusion, including any overrides with scientific justification and QA approval.

• Sample management: auto-generate sample pulls tied to phases (load, peak, tail), barcoded containers, chain-of-custody, and analytical methods.
• Decision logic: document deterministic rules (e.g., include fractions where conductivity 4–8 mS/cm and UV > X mAU) and any confirmatory test gates.
• Result integration: import certified results with analyst/review signatures, method version, and uncertainty; lock pool decisions until required results are present.
• Exception handling: OOS/OOT links to QMS, hold states on intermediate lots, and defined rework/discard paths captured in the record.

The chromatography step record is a key artifact for process validation and ongoing lifecycle control. It should demonstrate that critical process parameters (CPPs) stayed within validated ranges, cleaning/sanitization met acceptance criteria, and column performance remained fit-for-use across cycles.

• Alignment to control strategy (ICH Q11/Q10): define CPPs (e.g., delta-P, flow, gradient profile) and CQA surrogates (purity, HCP/DNA) with acceptable ranges and proven acceptable operating ranges (PARs).
• Cleaning verification: capture CIP/SIP recipes, contact times, temperatures, agent lots, rinse conductivity/pH, bioburden/endotoxin where applicable, and carryover swab/rinse results.
• Column performance: HETP/asymmetry checks, dynamic binding capacity (e.g., DBC10%), breakthrough curves, maximum cycles, and repacking criteria with trend charts.
• Hold times: document resin/column storage conditions, sanitization state, and requalification triggers after extended holds.

Chromatography step records feed Continued Process Verification (CPV), enabling statistical monitoring, early-warning indicators, and lifecycle assessments. Define standard metrics, units, and calculation provenance in the record so trends are reproducible and reviewable.

• Process metrics: step yield (%), recovery (% of load), pool purity/enrichment, delta-P at reference flow, number of cycles since pack, binding capacity, gradient accuracy.
• Data quality metrics: missing samples/trends, time sync offsets, alarm rates, override frequency, review-by-exception rates.
• Capability and control: CP/CPK of CPPs, run charts of delta-P drift, EWMA of UV tailing, excursion Pareto.

• Disconnected data: skid trends not referenced or preserved with integrity metadata; inability to reconstruct pooling decisions.
• Weak audit trails: parameter changes, holds, and alarm acknowledgments not recorded or not human-readable; shared accounts.
• Transcription risks: manual re-entry of material usage or results without verification; lack of independent review.
• Inadequate equipment linkage: missing column pack IDs/resin lots; no lifecycle checks (HETP/DBC) before use.
• Ambiguous criteria: pooling rules not versioned with the recipe; overrides without scientific justification and QA approval.
• Poor time discipline: unsynchronized clocks across PLC/HMI/MES; timezone ambiguity; daylight saving time anomalies.

V5 models chromatography as an ISA‑88 unit procedure with operation/phase granularity, enforcing master/control recipe ranges and capturing setpoint/actuals, alarms, and events in real time from the skid or historian. Material consumption/production is reconciled with WMS issuance and inventory, while LIMS-integrated samples and results gate pooling and release. QMS workflows (deviation, CAPA, change control) are natively linked; electronic signatures, audit trails, and review-by-exception fulfill Part 11/Annex 11 expectations. Maintenance and calibration status for critical instruments are checked at phase start to prevent use of out-of-tolerance assets.

1. Define chromatography master recipes with parameter class libraries, ranges, and criticality tags; align with control strategy (ICH Q10/Q11).
2. Map ISA‑88 structure (unit procedure → operations → phases) and configure exception handlers (holds, aborts, retries) with audit trail coverage.
3. Engineer interfaces (skid/PLC/Historian → MES) with accurate time sync, data integrity (hashes), and buffering for loss scenarios; validate per GAMP 5.
4. Model materials and equipment: resin/column IDs, buffer lots, skid assets, calibration status checks, and lineage (pack → use → clean → store).
5. Configure pooling logic and sample plans; integrate LIMS with result-based gates and override governance (justification + QA e-sign).
6. Design review-by-exception dashboards with anchor evidence (trends, alarms, calculations) and enforce independent QA review and release signatures.
7. Harden security and data retention: RBAC, SSO, backup/restore tests, disaster recovery RTO/RPO, and retention per product lifecycle/regulatory needs.
8. Operationalize CPV: define metrics, control charts, and automated alerts; schedule periodic assessments of column fitness and cleaning effectiveness.