Critical Process Parameters (CPPs) – Batch Control in Validated Manufacturing

This topic is part of the SG Systems Global regulatory & operations glossary.

Updated November 2025 • ICH Q8/Q9/Q10/Q11, 21 CFR 211/111/820, EU GMP • Pharma, Biologics, Devices, Food, Nutrition, Specialty Chemicals

Critical Process Parameters (CPPs) are the process conditions that have a direct, proven impact on Critical Quality Attributes (CQAs) of the product. If a CPP drifts outside its validated range, there is a meaningful risk that the batch will no longer meet specifications for safety, identity, strength, purity or performance. In real operations, CPPs are the handful of “do‑not‑mess‑with” parameters that your entire batch control strategy is built around: temperatures, times, speeds, pressures, flow rates, addition rates and other conditions that must be designed, monitored and responded to in a disciplined way.

“If you can’t point to your CPPs and show how you control them in every batch, you don’t really understand your process.”

1) What Are CPPs and Why They Matter for Batch Control

ICH Q8 defines a Critical Process Parameter as a process parameter whose variability has an impact on a Critical Quality Attribute and therefore should be monitored or controlled to ensure the process produces the desired quality. In practice, CPPs are the parameters that, if wrong, will hurt the patient, customer or product in a way that cannot be easily detected or fixed later. They often sit at the interface between unit operations: granulation conditions that determine tablet dissolution, fermentation conditions that determine potency, or filling conditions that determine sterility assurance.

From a batch‑control perspective, CPPs are where you spend your energy: designing robust setpoints and ranges, specifying automation logic, training operators, writing deviation playbooks and focusing your trending. Non‑critical parameters may move around without much impact, but CPPs must be kept on a short leash. If you treat every parameter as “critical”, you drown in noise; if you treat none as critical, you are gambling with quality. A clear, justified CPP list is how you avoid both extremes and build a defendable, efficient control strategy.

Operationally, CPPs also provide the backbone of your process validation story. When inspectors ask how you know the process remains in a state of control, you will be pointing to CPP trends, alarms, deviations and CAPAs – not just three historical validation batches. CPPs are the through‑line from development to validation, routine operation and lifecycle improvement.

2) CPPs, CQAs and the Control Strategy

CPPs do not exist in isolation; they live in a chain of causality that starts with the patient and ends with the unit operation. At the top are patient and user needs, which are translated into Critical Quality Attributes (CQAs) such as potency, content uniformity, sterility assurance, dissolution, particle‑size distribution or mechanical strength. The control strategy then maps which process parameters materially affect those CQAs, and which of those parameters require tight control to keep the CQA within specification.

In a mature QbD framework, this mapping is captured in design space models and risk assessments. You may have many process parameters, but only a subset will be elevated to CPPs – typically those with a strong, demonstrated impact on CQAs, limited ability for downstream detection or correction, and a realistic likelihood of variation in real manufacturing. These are the parameters that must be explicitly called out in control plans, batch records and automation configurations.

Once identified, CPPs shape your entire control strategy: which parameters require PAT monitoring, which get hard interlocks, which require dual verification or in‑process controls (IPC), and which appear in release justifications and Product Quality Review (PQR/APR). If your CPP list is fuzzy, your control strategy will be fuzzy; if your CPP list is tight and evidence‑based, your whole lifecycle becomes easier to defend.

3) Regulatory Anchors for CPPs

CPPs are not a vendor invention; they sit in the heart of modern regulatory expectations. ICH Q8 introduces CQAs, design space and CPPs as part of a science‑based development approach. ICH Q9 ties that to formal risk management, while ICH Q10 and Q11 embed CPPs into the Pharmaceutical Quality System and drug‑substance development. Regulators now expect sponsors to be able to discuss their processes in CPP/CQA language, not just list generic parameters and tests.

On the execution side, 21 CFR 211, 111 and 820 and EU GMP Chapters 1, 3, 4 and 5 all require processes to be validated, documented and controlled. While they may not always use the term “CPP”, they clearly expect firms to know which process parameters matter most, prove they are controlled, and react when they deviate. When you file or defend a process, you are implicitly presenting your CPP list, even if the slides never show the acronym.

In newer areas such as Pharma 4.0, advanced therapy medicinal products (ATMPs) and continuous manufacturing, regulators lean even harder on CPP language. Hybrid batch‑continuous processes, short hold times and limited end‑testing options make live control of CPPs absolutely central, with PAT, CPV and Real‑Time Release Testing (RTRT) built around them.

4) How CPPs Are Identified in Development

CPPs should emerge from development and scale‑up work, not be invented after the plant is built. Typical inputs include risk assessments (FMEA/HAZOP), small‑scale experiments, Design of Experiments (DoE), mechanistic models and historical manufacturing data from similar products. The aim is to understand which parameters significantly affect CQAs, within what ranges, and with what sensitivity.

At lab and pilot scale, teams deliberately stress parameters – mixing speeds, times, temperatures, pH, gas flows, feed rates – to see how CQAs respond. Parameters showing strong, non‑linear or poorly understood influence become candidates for CPP status. Others may be shown to be robust and left as normal operating parameters (NOPs) with wider allowable ranges. This classification should be documented, reviewed and updated as more data arrive.

During tech‑transfer and commercial‑scale validation, the CPP hypothesis is tested: can you hold these parameters within proposed ranges across batches and equipment, and do CQAs remain acceptable when they vary within that space? CPPs that prove too fragile may require tighter ranges, improved equipment capability or changes in process design. Conversely, parameters initially flagged as critical may be downgraded if data and mechanistic understanding consistently show low impact on CQAs. The key is traceability from data and risk arguments to the final CPP list.

5) CPPs vs NOPs, KPPs and Other Parameters

Not every “important” parameter is a CPP. Many organisations use a hierarchy: Critical Process Parameters (CPPs), Key Process Parameters (KPPs) and normal operating parameters (NOPs). CPPs have a direct, demonstrated impact on CQAs and demand formal control. KPPs affect efficiency, robustness or yield, but do not usually threaten product quality within expected variation. NOPs may matter for practicality or economics, but not directly for quality.

For example, in a granulation step, binder addition rate and impeller speed might be CPPs because they directly impact granule size and dissolution. Jacket temperature might be a KPP because it influences drying time and energy use. Mixer lighting might be a NOP. The discipline is to keep the “critical” label for parameters that genuinely carry quality risk; otherwise SOPs, MES configuration and QRM become bloated and unmanageable.

This hierarchy also helps during investigations. A deviation affecting a CPP demands a different level of scrutiny and potential impact on batch release than one affecting a non‑critical parameter. If everything is labelled “critical”, you end up either rejecting too many batches or rationalising too many deviations. A clear CPP vs KPP vs NOP structure prevents that drift and keeps your batch control decisions consistent and risk‑based.

6) Typical Types of CPPs Across Unit Operations

Although every process is unique, CPPs tend to cluster into familiar categories. For thermal operations (sterilisation, drying, lyophilisation, cooking), time‑temperature profiles and ramp/hold conditions are classic CPPs. For mixing and blending, impeller speed, power input, mixing time and addition sequence often become critical. In fermentation and cell culture, dissolved oxygen, pH, temperature, feed rates and agitation form the usual CPP set.

In filling and packaging, CPPs might include fill‑weight or volume control parameters, stopper or cap placement conditions, crimping force, seal‑bar temperature, dwell time and line speed. In coating, spray rate, inlet/outlet temperature, pan speed and atomising air pressure frequently show up as CPPs. Environmental CPPs, such as room differential pressure or humidity, may also be defined where they affect microbial or physical stability.

Recognising these patterns helps teams avoid starting from a blank page and instead perform targeted risk assessment. However, “typical” does not mean “automatic”: a parameter should only be declared critical when your specific product, process and equipment show that it matters. Copy‑pasting CPP lists from other products without evidence is as weak as having no CPP list at all.

7) From CPP Lists to Practical Batch Control

A CPP list in a development report is useless if it does not translate into concrete batch control. That translation happens through recipes, control logic and documentation. For each CPP, you must define setpoints, allowable ranges, measurement methods, sampling frequency, alarm and interlock logic, and actions to take when values approach or exceed limits. These rules then need to be embedded into batch records and automation, not merely described in a slide deck.

In a modern plant, that means configuring CPP limits, warnings and hard stops in the MES, DCS/PLC and eBR. Manual steps may require operator prompts and enforced data entry; automated steps may require validated interfaces from sensors and controllers. For portable equipment, CPP‑relevant settings (such as mixer speed ranges) must be locked or verified each batch, not left to tribal knowledge.

Where batch control is still largely paper‑based, the same logic applies but the risk is higher. CPP‑related instructions and limits must be clearly visible, not buried in long paragraphs. Sampling, verification and sign‑off must be strongly emphasised and reviewed. Over time, most companies find that sustained control of CPPs is practically impossible without at least partial digitalisation – which is exactly why CPPs are a natural anchor for MES and automation investment decisions.

8) PAT, Sensors and Real‑Time CPP Monitoring

CPPs are only as good as your ability to measure and react to them. For some parameters, classical instrumentation (temperature probes, pressure sensors, flowmeters, load cells) is enough. For others, PAT tools such as NIR, Raman, particle‑size analysers, in‑line titration or spectrophotometry become essential. These instruments allow you to monitor attributes like blend uniformity, concentration or crystallinity in real time, rather than using slow off‑line tests.

Once PAT and sensor data are available, they should feed into a central manufacturing data historian and, increasingly, a GxP data lake and analytics platform. This enables trend analysis, multivariate monitoring, soft sensors and early‑warning systems that detect CPP drift before it hits specification limits. It also creates the data backbone required for RTRT, APC and MPC strategies that actively adjust inputs to keep CPPs stable.

However, adding sensors and PAT without clear CPP logic can actually backfire: you generate streams of data with no plan for how to interpret or act on them. The discipline is to start from CPPs and CQAs, then decide what to measure, at what frequency, with what accuracy, and how that measurement feeds into alarms, interlocks and decision trees. Technology follows the CPPs, not the other way around.

9) CPP Excursions, Deviations and Batch Disposition

So what happens when a CPP goes out of range? This is where the theory meets regulatory reality. A true CPP excursion is not a minor nuisance – it is a formal deviation or non‑conformance that must be investigated, documented and considered in batch‑release decisions. The predefined response may include automatic process stops, segregation or rejection of affected material, enhanced sampling or reprocessing (where allowed) and an escalated risk assessment.

Investigations into CPP failures should link back to your original risk analysis and validation work. Did the excursion take you outside the proven acceptable range or design space? Were CQAs still met, and if so, why? Is there mechanistic or empirical evidence that the specific deviation window is acceptable, or are you simply rationalising after the fact? Over time, recurrent CPP excursions may indicate that your ranges are unrealistic, your equipment cannot meet capability requirements, or your understanding of the process is incomplete.

A common failure pattern is to treat CPP excursions no differently from minor housekeeping deviations. That might get a batch out of the door in the short term, but it erodes the credibility of your CPP framework and invites regulatory challenge. If you declare a parameter critical, your behaviour when it goes wrong must reflect that label – otherwise the entire criticality hierarchy collapses.

10) CPP Trending, SPC and Continued Process Verification

CPPs are not just pass/fail gates; they are continuous signals about process capability and health. Applying Statistical Process Control (SPC) charts, capability indices (Cp, Cpk) and multivariate trending to CPP data is central to Continued Process Verification. Instead of only asking “did we stay within the limits?”, you ask “how comfortably and consistently are we staying within them?”

Drift in CPP averages, creeping increase in variability, frequent near‑misses against limits – all are early warnings that the process may be sliding out of its validated state. These signals should feed into root cause analysis, maintenance decisions, training and CAPA. In well‑run organisations, CPP trending is a standing agenda item in PQR/APR and management reviews, not hidden in a data scientist’s notebook.

From a regulatory perspective, CPV based on CPPs is also your best defence against the “three‑batch validation” mentality. Being able to show several years of stable CPP performance with clear responses to excursions is far more persuasive than a stack of old protocols and batch records. It demonstrates that your process is monitored, understood and actively managed – the essence of lifecycle validation.

11) CPPs in MES, eBR and ISA‑88/ISA‑95 Structures

Standards such as ISA‑88 (batch control) and ISA‑95 (enterprise‑control integration) provide the structural language for implementing CPPs in recipes and systems. At the master‑recipe level, CPPs appear as parameters on phases and operations, with defined limits, units, default values and mode of control. At the control‑module and equipment‑module level, they translate into control loops, setpoints, alarms and interlocks implemented in PLC/DCS logic.

In the MES and eBR, CPPs appear as structured fields, data‑entry prompts and checks, not free‑text notes. MES can enforce that certain CPP‑related steps are completed, readings are within range and evidence (such as attachments from instruments) is attached before the recipe can advance. This is where concepts like Batch Review by Exception (BRBE) depend heavily on CPP logic; the system must know which parameters are truly critical to decide what is an “exception”.

If your digital infrastructure is built without explicit CPP fields, you end up with thousands of raw tags and values that are difficult to manage and interpret. Elevating CPPs as named, first‑class objects in recipes, data models and dashboards is what allows senior stakeholders to see, at a glance, how well the process is controlled where it matters most.

12) CPPs, APC, MPC and Digital Twins

Once CPPs and CQAs are well understood and instrumented, they become natural inputs to more advanced control strategies. Advanced Process Control (APC) layers, such as model‑based controllers, can adjust non‑critical parameters in real time to keep CPPs stable. Model Predictive Control (MPC) goes further, predicting future CPP trajectories based on current states and planned moves, then optimising inputs within constraints to keep the process in its sweet spot.

In parallel, digital twins of processes use mechanistic and data‑driven models to simulate how CPPs will respond to recipe changes, raw‑material variability or equipment differences. This allows organisations to test “what‑if” scenarios for new products, equipment upgrades or control‑strategy changes without risking live batches. Again, the twin is only as good as its foundation: clear definitions of CPPs, CQAs and the relationships between them.

These capabilities are often marketed under the umbrella of Pharma 4.0 or Industry 4.0. But underneath the buzzwords, they all depend on basic hygiene: agreed CPP lists, good instrumentation, reliable data and well‑designed recipes. Skipping that groundwork and jumping straight to “AI optimisation” is a short path to unconvincing models and fragile controls.

13) Governance, Change Control and CPP Lifecycle

CPPs are not static; they have their own lifecycle. As more data accumulate, equipment is upgraded, raw materials change and new markets are added, previously assumed relationships can shift. A robust governance process is needed to add, remove or reclassify CPPs using formal change control. This includes updating risk assessments, validation documents, batch records, automation logic, training and regulatory filings where relevant.

Ownership is key. Someone – often a cross‑functional process owner or technology‑transfer lead – must be accountable for the CPP list and its consistency across sites. Without named ownership, CPP definitions diverge quietly between plants, CMOs and documents, and investigations become painful: different teams use the same word “critical” to mean different things.

Periodic review of CPPs should be a defined activity, not a side‑effect of crises. Annual PQR/APR cycles, major equipment changes and recurring deviations are natural triggers to revisit classifications and ranges. When firms can show a clear audit trail of how and why CPPs evolved over time – grounded in data and risk, not opinion – regulators are much more comfortable with their lifecycle control story.

14) Multi‑Site, CMO and Supplier Perspectives on CPPs

In multi‑site networks, inconsistent CPP definitions are a silent killer of standardisation. One site may treat granulation end‑point as a CPP with tight PAT‑based control; another may treat it as a vague “operator judgement” parameter. Harmonising CPP lists and control strategies across sites allows meaningful comparison of performance, easier tech transfer and clearer communication with regulators.

For CMOs and CDMOs, CPP definitions must be baked into Quality Agreements and tech‑transfer packages. Sponsors should be explicit about which parameters they consider critical, how they expect them to be controlled, and how excursions should be reported. In turn, CMOs need the flexibility to adapt CPP control to their own equipment and systems, while maintaining the underlying risk rationale.

Upstream, suppliers may not talk in CPP language, but their raw‑material variability and documentation directly affect your ability to keep CPPs stable. Linking CPP excursions to supplier lots in your lot traceability and Supplier Quality Management (SQM) processes helps close the loop: critical parameters are not only about what happens inside the plant, but also about what you bring into it.

15) Implementation Roadmap for a CPP‑Centric Control Strategy

A practical CPP roadmap usually starts with a reality check: collect existing development reports, validation protocols, risk assessments and batch records for a representative product family. Extract every parameter currently labelled “critical”, “key” or “essential”, and compare that against actual control logic, instrumentation and deviation data. You will typically find a gap between the theory (what documents say is critical) and practice (what systems and people actually control and react to).

Next, run a focused QRM exercise for each unit operation to confirm or revise CPP candidates based on evidence. Capture the result in a simple but rigorous CPP register: parameter name, unit operation, CQA linkage, rationale, ranges, measurement method, alarm limits and predefined responses. Use that register as a single source of truth for updating URS, recipes, SOPs, MES configurations and validation documentation.

Finally, operationalise CPPs through digitalisation and culture. Embed them into eBR workflows, historian tags, dashboards and SPC charts. Train operators, engineers and QA reviewers to think in CPP/CQA terms when making decisions. Feed CPP trends into CPV, PQR and management reviews. Over time, the goal is simple: when anyone asks “is this process under control?”, the first thing people reach for is CPP evidence, not guesswork.

FAQ

Q1. Are CPPs and CQAs the same thing?
No. CQAs are product attributes (such as potency, content uniformity, sterility assurance or dissolution) that must be controlled to ensure product quality. CPPs are process parameters (such as temperature, time or feed rate) that have a proven impact on those CQAs. CQAs live in specifications and analytical methods; CPPs live in recipes, equipment settings and control strategies. They are linked but not interchangeable.

Q2. How many CPPs should a typical process have?
There is no magic number, but most robust commercial processes have a relatively small, focused set of CPPs per unit operation – often between a handful and a dozen. If every parameter is labelled critical, the term loses meaning and control becomes unmanageable. The right number is the one justified by risk, data and mechanistic understanding, not by habit or copying from old documents.

Q3. Can a parameter stop being critical over time?
Yes. As process knowledge grows and more data are collected, some parameters initially treated as critical may be downgraded if evidence consistently shows low impact on CQAs within realistic variation. Conversely, previously ignored parameters can be upgraded to CPPs if deviations or CPV trends reveal hidden sensitivity. The key is to manage such changes through formal change control, update filings where necessary and keep a clear rationale.

Q4. What if a CPP goes out of range but all finished‑product tests pass?
This is one of the hardest judgement calls in batch disposition. Passing end‑product tests does not automatically make a CPP excursion irrelevant, especially if the parameter went outside the proven acceptable range. You need a documented, science‑based justification that considers the magnitude and duration of the excursion, mechanistic understanding, historical data and any additional testing performed. Repeated “pass but out‑of‑range” events are a signal that your ranges, models or risk assessments need re‑work.

Q5. Are CPPs only relevant for highly regulated products?
No. Even in less‑regulated sectors such as cosmetics, speciality chemicals or some food categories, knowing which parameters are truly critical is essential for quality, safety and cost control. Regulatory pressure forces the discipline in pharma and devices, but the underlying logic – focus your control energy where it matters most – is universal. Many “non‑regulated” plants quietly adopt CPP thinking because it simply makes operations more predictable and defensible.

Related Reading
• Development & Validation: QbD | Process Validation | CPV | QRM
• Control & Monitoring: IPC | PAT | SPC | RTRT
• Digital & Systems: MES | eBR | Process Historian | GxP Data Lake | Pharma 4.0