Continuous Process Verification (CPV) and Evaluation Templates

Author: Venkiteswaran.T.K

The USFDA’s 2011 guidance on process validation advocates a life cycle approach and defines process validation activities in three stages

1. Design stage: Quality Target Product Profile (QTPP) for the product is defined and the CPPs (Critical Process Parameters), CQAs (Critical Quality Attributes), CMAs (Critical Material Attributes) to achieve the defined product quality are identified
2. Process Performance Qualification (PPQ): The commercial production process is qualified as reproducible and deliver quality products consistently.
3. Continuous process verification (CPV): CPV give assurance that process remain in a state of control during routine commercial production. The principal objective of CPV is to detect variability in process and quality attributes that might not have been evident during the design stage and PPQ stage.

The defined CPPs, CQAs, CMAs are monitored during PPQ and process is validated to show that the process and process parameters performs as intended to give consistent product complying to defined quality requirements. Further during commercial manufacturing, the CPPs, CQAs, CMAs are monitored, trended and statistically evaluated for Continuous Process verification (CPV). Statistical evaluation signal and help detect variations within the specifications and defined process control ranges. CPV also helps to identify continuous improvement opportunities. Deviations from the CQAs, CPPs and CMAs during commercial manufacturing are logged as quality events within the Quality Management System (QMS) and investigated. However, detection of signals during CPV is not a deviation or quality event as the product and process is still within the defined specification. The signals are evaluated for underlying reasons and their potential to cause future non-conformances. 

In this post, Qvents presents a CPV approach and templates to capture Continuous Process Verification, detect, document and evaluate the signals and identify actions.

Defining CPV Parameters and Criteria

The first step in the CPV is to identify the CPP, CQA, CMA to be monitored for continuous process verification. Select parameters that are sensitive to process or material variations. Document the rationale for selection (refer template for CPV Parameter and Criteria Document). For example, the parameters could be

• CPPs: Equipment RPMs, processing temperatures, drying times, compression forces, pH values, etc.
• CQAs: Assay, impurities, residual solvents, dissolution, tablet hardness, disintegration etc.
• CMAs: Particle size, flow properties, pH, water content etc of input materials.

Collect data from PPQ and subsequent commercial batches. Track each batch’s CPV parameters (CPPs, CQAs, CMAs) in an Excel based template (Refer Qvents excel template for CPV Parameter Tracker). Calculate CPV criteria like % Relative standard deviation (%RSD), Process Capability Index (CpK). Generate Control Charts with Upper Control Limits (UCL) and lower control limit (LCL), Upper (USL) and lower (LSL) specification limits and central line marked. Identify appropriate CPV criteria to detect variability in selected parameters. At least 10 data points are recommended for statistically meaningful analysis and calculation of CPV metrics.

CpK or %RSD – Selection of CPV criteria and typical limits:

CPV criteria selected should be appropriate for variation detection in the parameter being evaluated. For example:

• %RSD for parameters and quality attributes which indicate consistency or variance of a process or parameter (e.g. % assay, blend uniformity).
• Process Capability Index CpK for parameters which indicate the process capability to achieve the desired attribute (e.g. % impurities, water content, disintegration time, machine speed, process temperature)

Also note that for parameters where the ratio of mean to standard deviation (SD) is less than 10 [(Mean/SD)<10)] the %RSD criteria become mathematically unstable. In such cases CpK is a more reliable metric (e.g., low-level impurities). Conversely in parameters which has symmetrical two sided specifications with narrow range, even small drifts can affect CpK. In such cases % RSD will be more appropriate (e.g. assay). Also give due consideration to process maneuverability and availability of process levers when choosing the CPV parameter and evaluation criteria.  

CPV Criteria, Process variability and Process control

Typical interpretation of the CPV criteria (%RSD and CpK) with reference to process variability and process control is given below. Evaluate the CPV criteria values in conjunction with signals from control charts for variability detection.

%RSD
%RSD	<	5%	→	Good control (tight process)
%RSD	–	5-10%	→	Acceptable control
%RSD	>	10%	→	Weak consistency, high variability

CpK
Cpk	>	1.33	→	Good capability, good process control
Cpk	–	1.0–1.33	→	Marginal, acceptable control
Cpk	<	1.0	→	Poor capability

Tracking and evaluating CPV data

Track the control charts updating CPV parameters (CPPs, CQAs, CMAs) after each batch and evaluate the control charts for emerging trends or Signals as below:

	Outlier	:	a parameter data point is outside the control limits
	Shift	:	9 consecutive points on same side of centre line
	Drift	:	6 consecutive points, all in one direction (increasing or decreasing)

Detecting variability

Review the CPV parameters, CPV criteria values (%RSD and CpK) and Signals for identifying any emerging trends of variability within the process and / or quality attributes. If variability is detected, perform additional evaluations to verify if any assignable reasons are contributing to variability and address the same.

A matrix for detecting variability based on CPV parameters and Signals is given below.
Signal	Signal Type	Variability detection and Action
Outlier	A point is outside the control limit (Upper / Lower / Both)	If CpK within control or %RSD is within control: No Action If not, evaluate reasons for the outlier, take action if an assignable reason is found, continue monitoring
Shift	9 consecutive points on same side of center line	If CpK within control or %RSD is within control: No Action  If not, evaluate reasons for the shift, take action if an assignable reason is found, continue monitoring
Drift	6 consecutive points, all increasing or all decreasing	If CpK within control or %RSD is within control: No Action  If not, evaluate reasons for the drift, take action if an assignable reason is found, continue monitoring
Adapted from Reference 1

Evaluation of Variability

It will help to have a set of aspects to be evaluated when variability is detected in each CPV parameter. This should typically be identified during process risk assessment or in the product / process development report. Based on the evaluation if one or a combination of factors are identified as contributing to variability, take action to address the situation or if no action is necessary document the same with justification.

Example Scenario

Variability Aspect	Possible contributing factors (based on Risk analysis or PDR
Increasing process impurity or a shift to higher level in process impurity in API	One or a combination of factors: Higher amount of precursor but within the specifications in raw materials Ineffective washing and purging of impuriti Higher water content in washing solvent Change of composition or quantity of washing solvent due to evaporation
Action, example	Higher evaporation rate is caused by higher spinning speed in a centrifuge resulting in ineffective purging of impurities; Reduce the speed within the design space limits.
No Action with justification, example	Reason for higher impurities in product is relatively higher level of impurities in the raw material used in these batches; however, this is an expected variability in the raw material and process is capable of keeping the resultant process impurity within the limit. No further action is necessary

Conclusion

Continuous Process Verification (CPV) is a powerful tool to proactively detect, evaluate, and manage process variability before it affects product quality, support control over product lifecycle and continuous improvement. The focus of CPV is to identify variation within the specification limits or control strategy even when CPPs, CQAs, CMAs are maintained within the specified ranges and specifications. Since these variations are still within the design space, the risk to quality will be low to nil in batches where variability is detected. But the variations flag future potential failures (alerts). The detection of variations need not trigger a quality event; however the variations should be evaluated.

CPV tracking shall be continuous and evaluation shall be as frequently as possible (e.g. say every 10 batches) for timely detection of variations and initiation of corrective actions.

Important Formulae:

UCL (Upper Control Limit)	=	X̄ + 3σ
LCL (Lower Control Limit)	=	X̄ – 3σ
CpU	=	(USL − X̄) / (3σ)
CpL	=	(X̄ − LSL) / (3σ)
Process Capability Index CpK	=	Min (Cpu, Cpl)
(σ  = Standard Deviation; X̄ = Process mean) Note: If UCL or LCL exceed spec limits, use specification limits for practical control.

Reference:

1. Pharmaceutical Engineering, January / February 2017, Continued Process Verification (CPV) Signal Responses in Biopharma By Syama Adhibhatta, Mark DiMartino, Richard Falcon, Eric Haman, Kevin Legg, Robin Payne, Kevin R. Pipkins, Abdelqader Zamamiri
2. USFDA Guidance for Industry: Process Validation: General Principles and Practices (2011)

Templates: