What is Dr. Zubair Khalid's research focus?

Dr. Zubair Khalid specializes in molecular virology, mRNA vaccine development, and computational biology, with a focus on avian pathogens like IBDV and Avian Reovirus.

Where is Dr. Zubair Khalid currently working?

Dr. Zubair Khalid is a Postdoctoral Research Associate at the University of Maryland (UMD), specifically within the Department of Animal and Avian Sciences.

Process Validation in Biotech: Stages, Types, and Best Practices

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Introduction to Process Validation

Process validation is the documented, systematic collection and evaluation of evidence that a manufacturing process, when operated within established parameters, consistently produces a product meeting its predetermined quality attributes. In biotechnology, this is not a one-time exercise but a continuous obligation spanning the entire commercial lifecycle of a biologic. The purpose is to demonstrate, with statistical confidence, that the process is robust, reproducible, and under control before product release to patients.

Definition and Regulatory Framework

The regulatory foundation for process validation rests on several key documents. The FDA's 2011 guidance "Process Validation: General Principles and Practices" replaced the outdated 1987 framework and introduced the lifecycle approach. ICH Q7 (Good Manufacturing Practice for Active Pharmaceutical Ingredients) provides the foundational GMP requirements, while ICH Q8 (Pharmaceutical Development), Q9 (Quality Risk Management), and Q10 (Pharmaceutical Quality System) collectively establish the modern expectation: validation is not a discrete event but a continuous, risk-based activity integrated into product development and commercial manufacturing.

For biologics specifically, the FDA's FDA Approval Process for Biologics requires that process validation data be submitted as part of a Biologics License Application (BLA). The critical distinction from small molecules is that biologics are produced by living systems—cells, microorganisms, or enzymes—which introduce inherent biological variability that must be characterized and controlled. A process that works at 2 L scale in development will not necessarily behave identically at 2,000 L scale, which is why Process Scale-up studies are integral to validation planning.

Lifecycle Approach: Process Design, Qualification, Continued Process Verification

The lifecycle concept, formalized in the 2011 FDA guidance, divides process validation into three interconnected stages that span the product's entire commercial life:

Process Design: Knowledge gathering during development to define the commercial manufacturing process.
Process Qualification: Demonstrating that the process is capable of reproducible commercial manufacturing.
Continued Process Verification: Ongoing assurance that the process remains in a state of control during routine production.

This framework aligns with ICH Q10's pharmaceutical quality system and requires that validation activities be proportionate to risk. A high-risk step—such as viral inactivation or sterile filtration—demands more rigorous validation than a low-risk step like a bulk buffer hold.

The Three Stages of Process Validation

Stage 1: Process Design

Stage 1 begins during development and continues through technology transfer. The objective is to acquire comprehensive process understanding through systematic studies that identify critical process parameters (CPPs) and link them to critical quality attributes (CQAs). This stage relies heavily on Design of Experiments (DoE) principles, scale-down models, and prior knowledge from platform processes.

Key activities include:

Defining the target product profile and CQAs (e.g., purity, potency, glycosylation profile for monoclonal antibodies).
Conducting risk assessments to identify potential failure modes.
Executing DoE studies to characterize parameter ranges and interactions.
Establishing a provisional control strategy, including in-process controls and hold times.

For a typical monoclonal antibody upstream process, this might involve evaluating the impact of dissolved oxygen (20–60% air saturation), pH (6.8–7.2), and temperature (36–37°C) on viable cell density and titer in a scale-down model. The deliverable is a process design that is robust, scalable, and aligned with quality by design (QbD) principles.

Stage 2: Process Qualification (PPQ)

Stage 2, Process Qualification, comprises two elements: facility/utility qualification and process performance qualification (PPQ). Facility qualification confirms that equipment, utilities, and systems are fit for their intended use—this includes installation qualification (IQ), operational qualification (OQ), and performance qualification (PQ) of equipment. PPQ is the heart of Stage 2: it demonstrates that the commercial process, as designed, produces product meeting all CQAs when operated under defined conditions.

PPQ typically involves a minimum of three consecutive successful commercial-scale batches, though the exact number should be justified by risk assessment and statistical rationale. Each PPQ batch must be manufactured using the same procedures, materials, and equipment as routine production. Extensive sampling is required—far more than routine release testing—to provide statistical confidence in process consistency.

For a downstream process, PPQ would include verification of column performance (e.g., Protein A chromatography binding capacity ≥ 40 mg/mL resin), viral clearance studies (e.g., ≥ 4 log reduction for low-pH inactivation), and filtration integrity testing. The PPQ protocol must be approved before execution, and deviations must be thoroughly investigated.

Stage 3: Continued Process Verification (CPV)

Stage 3, Continued Process Verification, is the ongoing monitoring of the commercial process to ensure it remains in a state of control. This is not a static activity but a dynamic program that uses statistical process control (SPC) to detect trends, shifts, or drifts before they result in out-of-specification (OOS) results.

CPV programs typically include:

Real-time monitoring of CPPs and CQAs using control charts (e.g., X-bar and R charts).
Periodic review of batch data, including yield, purity, and potency trends.
Annual product quality reviews (APQRs) that synthesize data across batches.
Trigger-based investigations when trends approach alert or action limits.

The intensity of monitoring should be risk-based: a step with a narrow operating range or high failure impact warrants tighter monitoring. For example, a viral filtration step with a validated log reduction value (LRV) of ≥ 4 for minute virus of mice (MVM) would require continuous pressure and flow monitoring, while a less critical buffer hold step might only require periodic confirmation.

Types of Process Validation

Prospective Validation

Prospective validation is performed before commercial distribution of a new product or after a significant process change. It is the standard approach for new biologics and involves executing the PPQ protocol with three (or more) consecutive batches. All data must demonstrate that the process performs as intended before any product is released to the market. This is the only acceptable approach for a novel biologic where no historical data exist.

Concurrent Validation

Concurrent validation is conducted while the product is being commercially distributed. It is reserved for exceptional circumstances—for example, when a validated process has undergone a minor change that does not affect product quality, or when there is an urgent patient need. The rationale must be documented, and the risk must be carefully assessed. In practice, concurrent validation is rare in biologics due to the high cost of a product recall and the regulatory scrutiny involved.

Retrospective Validation

Retrospective validation uses historical batch data to validate a process that has been in commercial use without formal validation. This approach is generally not acceptable for biologics under current regulatory expectations, as it contradicts the lifecycle philosophy. It may be considered only for legacy products with extensive, consistent historical data, but even then, regulators expect a transition to prospective or concurrent validation.

Revalidation

Revalidation is required when changes occur that could affect product quality, safety, or efficacy. Triggers include:

Changes in raw material suppliers (e.g., a new cell culture media vendor).
Equipment modifications or replacement.
Process changes (e.g., a new chromatography resin).
Facility changes (e.g., a new manufacturing suite).
Deviations or trends indicating loss of control.

The scope of revalidation should be determined by risk assessment. A change to a single buffer pH might only require a focused revalidation of that step, while a change to the cell line would require full process revalidation.

Process Validation Protocol and Report

Protocol Elements

The validation protocol is the master document that defines the entire validation exercise before execution. It must be detailed enough to be executed without ambiguity and must be approved by quality assurance (QA) and relevant subject matter experts. Essential elements include:

Objective and Scope: A clear statement of what is being validated and the boundaries (e.g., "Validation of the 2,000 L fed-batch cell culture process for Product X").
Responsibilities: Identification of all parties involved—manufacturing, QA, regulatory affairs, and validation.
Process Description: A detailed description of the process, including a flow diagram, equipment list, and critical parameters.
Acceptance Criteria: Predefined, measurable criteria for each CQA and CPP. For example, "Product titer ≥ 2.5 g/L at harvest" or "Host cell protein (HCP) ≤ 100 ppm in the drug substance."
Sampling Plan: A detailed plan specifying what samples are taken, at which points, in what quantities, and by whom. This must be more extensive than routine release testing.
Analytical Methods: Identification of all test methods, including their qualification status (see Assay Method Validation for method qualification requirements).
Statistical Methods: Predefined statistical analyses, such as confidence intervals, tolerance intervals, or capability indices.
Deviation Handling: Procedures for documenting and investigating any deviations during the validation runs.
Change Control: How changes to the protocol will be managed and documented.

Report and Data Review

The validation report is the companion document that summarizes the executed protocol, presents all data, and draws conclusions. It must include:

A summary of all results against acceptance criteria.
A detailed discussion of any deviations and their impact.
Statistical analysis results.
A final statement on whether the process is validated and suitable for commercial use.
Recommendations for ongoing monitoring (Stage 3).

The report must be reviewed and approved by the same functions that approved the protocol. Any failure to meet acceptance criteria must be investigated, and the root cause must be addressed before proceeding. In some cases, additional PPQ runs may be required.

Key Steps in Conducting Process Validation

Risk Assessment (FMEA)

Risk assessment is the starting point for any validation activity. Failure Mode and Effects Analysis (FMEA) is the most commonly used tool in biotech. The process involves:

Identifying each unit operation (e.g., cell thaw, seed expansion, production bioreactor, harvest, Protein A chromatography, viral inactivation, polishing chromatography, viral filtration, ultrafiltration/diafiltration).
Listing potential failure modes for each operation (e.g., "Bioreactor pH controller fails high").
Assessing the severity (S), occurrence (O), and detectability (D) of each failure mode on a scale of 1–10.
Calculating the Risk Priority Number (RPN = S × O × D) and prioritizing actions.

For example, in a Protein A chromatography step, a failure mode might be "resin fouling leading to reduced binding capacity." Severity might be 8 (affects yield and purity), occurrence 3 (moderate), and detectability 4 (detected by pressure rise). The RPN of 96 would trigger a mitigation action, such as implementing a pre-use column performance check.

The output of the FMEA directly informs the validation strategy: high-risk steps receive more extensive sampling and tighter acceptance criteria.

Defining CQAs and CPPs

CQAs are physical, chemical, biological, or microbiological properties that must be within appropriate limits to ensure product quality. For a monoclonal antibody, typical CQAs include:

Aggregates (e.g., high molecular weight species ≤ 5% by size exclusion chromatography)
Charge variants (e.g., acidic variants ≤ 30% by ion-exchange chromatography)
Glycosylation (e.g., afucosylation ≤ 10% for enhanced ADCC activity)
Potency (e.g., relative binding affinity 80–125% of reference standard)
Host cell proteins (e.g., ≤ 100 ppm by ELISA)
Residual Protein A (e.g., ≤ 10 ppm)

CPPs are process parameters whose variability has an impact on a CQA and therefore must be monitored and controlled. For upstream processing, these might include seeding density, temperature, pH, dissolved oxygen, and feed strategy. For downstream processing, examples include column loading density, residence time, buffer pH and conductivity, and membrane pressure.

The relationship between CPPs and CQAs is established through DoE studies during Stage 1. For instance, a design space might show that Protein A elution pH between 3.4 and 3.8 yields aggregate levels below 5%, with a target of 3.6.

Executing PPQ Runs

PPQ execution follows the approved protocol precisely. Key considerations include:

Batch Number: Typically three consecutive batches, but the number should be justified statistically. For a process with high inherent variability, more batches may be needed.
Scale: PPQ must be performed at commercial scale. Scale-down models cannot substitute for PPQ.
Materials: All raw materials must be from commercial sources and meet specifications.
Personnel: Operators must be trained and qualified.
Environment: Manufacturing must occur in a qualified facility under appropriate environmental conditions.

During PPQ, extensive sampling is performed. For a cell culture process, this might include daily cell counts, viability, metabolite analysis (glucose, lactate, glutamine), and off-line pH/gas analysis. For downstream, samples are taken at every column fraction, every pool, and every intermediate hold point.

Statistical Analysis and Acceptance Criteria

Statistical analysis is the backbone of PPQ. The goal is to demonstrate that the process is capable and consistent. Common approaches include:

Confidence Intervals: Calculate 95% confidence intervals for key CQAs and verify they fall within specification limits.
Tolerance Intervals: Determine the range that contains a specified proportion (e.g., 95%) of the population with a specified confidence (e.g., 95%). This is more rigorous than simple confidence intervals.
Process Capability Indices: Calculate Cp and Cpk values. A Cpk ≥ 1.33 is often considered acceptable, indicating the process is centered and capable.

For example, if the aggregate specification is ≤ 5% and the three PPQ batches show aggregate levels of 2.1%, 2.4%, and 2.2%, the 95% confidence interval might be 1.9–2.7%, which is well within specification. The process capability index would be high, supporting the conclusion that the process is robust.

Methods and Tools for Process Validation

Design of Experiments (DoE)

DoE is a systematic approach to understanding the relationship between inputs (factors) and outputs (responses). It is essential during Stage 1 for defining the design space and identifying CPPs. Common designs include:

Full Factorial: All combinations of all factors at all levels. Useful for 2–4 factors.
Fractional Factorial: A subset of the full factorial, used when many factors are screened.
Central Composite Design (CCD): Adds center and axial points to estimate curvature and interactions.
Box-Behnken Design: An alternative to CCD with fewer runs, useful for 3–7 factors.

For a chromatography step, a DoE might evaluate the effect of pH (3.0–4.0), conductivity (5–15 mS/cm), and residence time (3–6 min) on yield and aggregate removal. The resulting model would define the design space and identify the robust operating region.

Process Analytical Technology (PAT)

PAT is a framework for real-time monitoring and control of critical quality attributes and process parameters. It is not a specific technology but a philosophy that includes:

In-line sensors: Probes placed directly in the process stream (e.g., Raman spectroscopy for glucose and lactate monitoring in bioreactors).
At-line analyzers: Instruments that analyze samples taken from the process (e.g., HPLC for product titer).
On-line analyzers: Automated sampling and analysis systems (e.g., automated cell counters).

In a fed-batch process, Raman spectroscopy can provide real-time measurements of glucose, lactate, and viable cell density, enabling automated feed control. This reduces variability and provides data for CPV. PAT is particularly valuable for processes with narrow operating windows or where off-line testing introduces significant lag time.

Statistical Process Control (SPC)

SPC is the primary tool for Stage 3 (CPV). It involves:

Control Charts: X-bar and R charts for continuous variables, p-charts for proportions. Control limits are typically set at ±3 standard deviations from the mean.
Run Rules: Western Electric rules detect non-random patterns, such as 7 consecutive points on one side of the mean.
Capability Analysis: Periodic calculation of Cp and Cpk to ensure the process remains capable.

For example, a control chart of bioreactor harvest titer across 50 batches might show a mean of 3.2 g/L with control limits of 2.8–3.6 g/L. If the last 5 batches trend upward toward 3.5 g/L, the run rules would trigger an investigation before the process exceeds the upper control limit.

Common Pitfalls and How to Avoid Them

Inadequate Sampling Plans

A common failure is designing a sampling plan that is too sparse to provide statistical confidence. For example, taking a single sample per chromatography pool when the pool is 200 L may not capture heterogeneity. The solution is to use a risk-based approach: high-risk steps require more samples, and samples should be taken at multiple time points and locations. For a 2,000 L bioreactor, this might mean sampling at 24, 48, 72, 96, 120, and 144 hours, plus daily metabolite analysis.

Insufficient Number of Batches

Three batches is the minimum, not the default. If the process has high inherent variability—for example, a cell culture process with a historical titer coefficient of variation (CV) of 15%—three batches may not provide adequate confidence. In such cases, a statistical justification for a larger number (e.g., 5–10 batches) is required. Conversely, a highly robust process with a CV of 3% might justify three batches with a rigorous statistical rationale.

Poor Documentation Practices

Validation is a documentary exercise. Common failures include:

Incomplete or illegible batch records.
Missing signatures or dates.
Deviations documented after the fact rather than in real time.
Data recorded in pencil or with correction fluid.

The solution is a robust data integrity program aligned with ALCOA+ principles (Attributable, Legible, Contemporaneous, Original, Accurate, Complete, Consistent, Enduring, Available). Electronic batch records and laboratory information management systems (LIMS) can significantly reduce documentation errors.

Neglecting Continued Process Verification

Many organizations treat Stage 3 as an afterthought, only reviewing data at the annual product quality review. This is insufficient. CPV requires real-time or near-real-time monitoring with predefined alert and action limits. A process that drifts gradually—for example, a slow decline in cell viability due to gradual media degradation—may not be detected until it results in an OOS result. The solution is to implement automated SPC with alerts and to review control charts on a batch-by-batch basis.

Practical Summary and Decision Framework

Decision Tree for Validation Approach

The choice of validation type and the intensity of validation activities should be guided by risk and product lifecycle stage:

Is this a new product? → Prospective validation with full Stage 1, 2, and 3 activities.
Is this a significant process change? → Revalidation with scope determined by risk assessment. A change to the cell line requires full revalidation; a change to a buffer pH may require only a focused study.
Is this a minor change (e.g., new supplier of the same material)? → Risk-based revalidation, possibly including a single engineering run and a reduced PPQ.
Is this a legacy product without formal validation? → Retrospective validation may be considered, but transition to prospective or concurrent validation is expected.
Is the process in routine production? → Stage 3 CPV with SPC and annual product quality reviews.

Key Takeaways for Industry Scientists

Process validation is a lifecycle activity, not a one-time event. It begins at process design and continues through commercial manufacturing.
The three stages—Process Design, Process Qualification, and Continued Process Verification—are interconnected and must be planned holistically.
Risk assessment (FMEA) should drive all validation decisions, including sampling intensity, batch numbers, and acceptance criteria.
Statistical rigor is non-negotiable. Use confidence intervals, tolerance intervals, and capability indices to support conclusions.
Documentation must be contemporaneous, accurate, and complete. Data integrity is a regulatory expectation, not an option.
CPV is not optional. A process that is not monitored is a process that is not controlled.
Engage QA early and often. Validation is a cross-functional activity that requires alignment across manufacturing, quality, and regulatory affairs.

Frequently Asked Questions

What are the types of process validation?

The four types are prospective validation (performed before commercial distribution), concurrent validation (performed during commercial distribution under exceptional circumstances), retrospective validation (using historical data, generally not acceptable for biologics), and revalidation (triggered by changes or deviations).

What are the steps of process validation?

The key steps are: (1) risk assessment using FMEA, (2) defining CQAs and CPPs, (3) developing a control strategy, (4) writing and approving the validation protocol, (5) executing PPQ runs at commercial scale, (6) performing statistical analysis of results, (7) writing the validation report, and (8) implementing Stage 3 continued process verification.

What are the stages of process validation?

The three stages are: Stage 1 (Process Design), where process understanding is developed; Stage 2 (Process Qualification), where the process is demonstrated to be capable of reproducible commercial manufacturing; and Stage 3 (Continued Process Verification), where the process is monitored to ensure it remains in control.

What is process validation?

Process validation is the documented evidence that a manufacturing process, when operated within established parameters, consistently produces a product meeting its predetermined quality attributes. It is a regulatory requirement and a lifecycle activity spanning process design, qualification, and ongoing verification.

What is a process validation protocol?

A process validation protocol is the approved master document that defines the objective, scope, responsibilities, process description, acceptance criteria, sampling plan, analytical methods, statistical methods, and deviation handling procedures for a validation exercise. It must be approved before execution.

Can you give an example of process validation?

For a monoclonal antibody, Stage 1 might involve DoE studies to define the Protein A chromatography elution pH range (3.4–3.8). Stage 2 would involve three PPQ batches at 2,000 L scale, with extensive sampling to confirm aggregate levels ≤ 5%, HCP ≤ 100 ppm, and yield ≥ 80%. Stage 3 would involve SPC monitoring of these parameters across all commercial batches.

Why is process validation important in biotech?

Biologics are produced by living systems with inherent variability. Process validation provides the statistical evidence that the process is robust and reproducible, ensuring patient safety and product efficacy. It is also a regulatory requirement for FDA Approval Process for Biologics, and it protects the manufacturer from costly recalls and supply disruptions.

Key Takeaways

Process validation is a lifecycle obligation spanning process design, qualification, and continued verification—not a one-time regulatory checkbox.
Risk assessment (FMEA) must drive all validation decisions, from sampling plans to batch numbers and acceptance criteria.
PPQ requires commercial-scale runs with extensive sampling and predefined statistical acceptance criteria; three batches is a minimum, not a default.
Continued Process Verification (Stage 3) is mandatory and requires real-time SPC monitoring with predefined alert and action limits.
Documentation must be contemporaneous, accurate, and complete, adhering to ALCOA+ data integrity principles.
Revalidation is triggered by any change that could affect product quality; the scope must be justified by risk assessment.
Statistical tools—confidence intervals, tolerance intervals, and capability indices—are essential for drawing defensible conclusions from validation data.