What Are CPPs and CQAs?
Critical Process Parameters (CPPs) are process input variables whose variability has a demonstrable impact on a Critical Quality Attribute. Critical Quality Attributes (CQAs) are physical, chemical, biological, or microbiological properties that must fall within defined limits to ensure product safety and efficacy. The relationship between CPPs and CQAs forms the foundation of Quality by Design (QbD) in biopharmaceutical manufacturing.
In a monoclonal antibody process, for example, upstream CPPs typically include culture pH (6.8–7.2), temperature (35–37°C), dissolved oxygen (30–60% air saturation), and feed rate. These CPPs drive CQAs such as titer (g/L), glycosylation profile (% galactosylation), charge variants (% acidic species), and aggregate levels (<2% by SEC).
Not every process parameter is critical. A typical CHO cell culture unit operation may have 30–50 parameters, but only 5–10 qualify as CPPs after systematic evaluation. The goal of CPP and CQA mapping is to identify which parameters matter, quantify how much they matter, and define the safe operating region — the design space.
| Unit Operation | Typical CPPs | Affected CQAs | Typical Range |
|---|---|---|---|
| Cell culture (upstream) | pH, temperature, DO, feed rate | Titer, glycosylation, charge variants | pH 6.8–7.2, T 33–37°C |
| Protein A capture | Load density, wash pH, elution pH | HCP, yield, aggregates | Load 30–50 g/L resin |
| Low pH viral inactivation | pH, hold time, temperature | Viral clearance (LRV), aggregates | pH 3.4–3.6, 60–120 min |
| Ion exchange polish | Load conductivity, gradient slope, pH | Charge variant removal, yield | Conductivity 5–15 mS/cm |
| UF/DF formulation | TMP, concentration factor, DV count | Concentration accuracy, aggregates | TMP 10–25 psi |
The ICH Q8–Q11 Framework for Design Space
The design space concept is defined in ICH Q8(R2) as the multidimensional combination and interaction of input variables (material attributes and process parameters) that have been demonstrated to provide assurance of quality. Working within the design space is not considered a change, which gives manufacturers operational flexibility without requiring post-approval regulatory filings.
Four ICH guidelines work together to enable QbD:
- ICH Q8(R2) — Pharmaceutical Development: defines the Quality Target Product Profile (QTPP), CQAs, design space, and control strategy.
- ICH Q9 — Quality Risk Management: provides risk assessment tools (FMEA, fault tree analysis, Ishikawa diagrams) for identifying CPPs.
- ICH Q10 — Pharmaceutical Quality System: lifecycle management framework for continuous improvement.
- ICH Q11 — Development and Manufacture of Drug Substances: extends QbD principles to API and biotech-derived active substances.
The QbD workflow follows a logical sequence: define the QTPP, identify CQAs from the QTPP, use risk assessment to link process parameters to CQAs, run DOE studies to quantify relationships, and then define the design space and control strategy. As of 2026, regulatory agencies (FDA, EMA, PMDA) increasingly expect QbD elements in biologics license applications, and companies report roughly 40% fewer batch deviations after adopting QbD approaches.
Step 1: Risk Assessment — Prioritizing Parameters
Risk assessment is the first step in CPP and CQA mapping, reducing a long list of potential parameters to a manageable set for experimental characterization. A typical bioreactor unit operation has 30–50 adjustable parameters, but resource constraints limit DOE studies to 8–15 factors. Risk assessment bridges that gap.
The most common tool is Failure Mode and Effects Analysis (FMEA), which scores each parameter on three dimensions:
- Severity (S) — How badly does the parameter affect a CQA if it deviates? (1–10 scale)
- Probability (P) — How likely is the parameter to deviate from its target? (1–10 scale)
- Detectability (D) — How quickly can you detect the deviation and correct it? (1–10, where 10 = undetectable)
The Risk Priority Number (RPN) = S × P × D. Parameters above an RPN threshold (commonly 80–120) enter process characterization DOE studies. Parameters below the threshold are classified as non-critical and controlled through standard operating procedures.
An Ishikawa (fishbone) diagram is useful before FMEA scoring to ensure no parameters are overlooked. Organize branches by category: equipment (agitation, aeration, vessel geometry), raw materials (media lot, buffer components), environment (temperature, humidity), process (pH, DO, feed timing), and measurement (probe calibration, sampling frequency).
Step 2: Process Characterization with DOE
Process characterization uses designed experiments (DOE) to quantify how CPPs affect CQAs, including main effects, interactions, and curvature. This is the most resource-intensive step in CPP and CQA mapping, typically requiring 40–80 bioreactor runs across screening and optimization phases for a single upstream unit operation.
A two-stage DOE strategy is standard in industry:
Stage 1: Screening (identify which parameters matter)
- Design type: Fractional factorial (Resolution IV+), Plackett-Burman, or Definitive Screening Design (DSD)
- Factors: 8–15 potential CPPs from risk assessment
- Runs: 20–30 experiments
- Goal: Identify statistically significant main effects (p < 0.05) and flag potential interactions
- Outcome: Narrow to 4–6 confirmed CPPs
Stage 2: Optimization (quantify relationships and define design space)
- Design type: Central Composite Design (CCD), Box-Behnken Design (BBD), or face-centered design
- Factors: 4–6 confirmed CPPs
- Runs: 25–50 experiments (including center points and axial points)
- Goal: Build response surface models with interaction terms and quadratic effects
- Outcome: Regression models predicting each CQA as a function of CPPs
| Design Type | Factors | Runs (typical) | Detects Interactions? | Detects Curvature? | Best For |
|---|---|---|---|---|---|
| Plackett-Burman | 8–15 | 12–20 | No | No | Quick screening, many factors |
| Fractional Factorial (Res IV) | 5–8 | 16–32 | Partially | No | Screening with some interaction info |
| Definitive Screening (DSD) | 5–12 | 2k+1 runs | Yes | Yes | Efficient screening + optimization hybrid |
| Central Composite (CCD) | 3–6 | 25–50 | Yes | Yes | Full response surface modeling |
| Box-Behnken (BBD) | 3–5 | 15–30 | Yes | Yes | Optimization when extremes are impractical |
Scale-Up Calculator
Translate your design space parameters across bioreactor scales. Compare P/V, tip speed, kLa, and mixing time.
Understanding CPP–CQA Interactions
Parameter interactions are the reason multivariate DOE is essential for CPP and CQA mapping — univariate one-factor-at-a-time (OFAT) studies systematically miss them. An interaction means the effect of one CPP on a CQA depends on the level of another CPP. In mAb production, the pH × temperature interaction on glycosylation is one of the most well-documented examples.
Consider culture pH and temperature effects on galactosylation (a CQA for monoclonal antibodies). At pH 7.0 and 37°C, galactosylation might be 45%. Dropping temperature to 33°C at pH 7.0 reduces it to 30%. But dropping temperature to 33°C at pH 6.8 reduces galactosylation to only 35% — the pH partially offsets the temperature effect. This interaction is invisible to OFAT experiments.
The interaction plot below shows how pH and temperature jointly affect galactosylation in a typical CHO mAb process. Non-parallel lines confirm a significant interaction — the temperature effect depends on pH level.
Common CPP–CQA interactions in mAb manufacturing include:
- pH × temperature → glycosylation: Lower temperature reduces enzyme activity; pH modulates intracellular nucleotide sugar pools.
- DO × agitation → oxidation variants: High DO with high shear increases methionine oxidation (target: <5% oxidized species).
- Feed rate × temperature → titer: Temperature shift to 33°C extends culture viability, but requires adjusted feeding to maintain productivity.
- pH × CO2 → charge variants: Elevated pCO2 at low pH increases acidic charge variants through non-enzymatic deamidation.
Step 3: Defining Design Space Boundaries
The design space is defined by overlaying acceptance criteria for all CQAs onto the response surface models from DOE, then identifying the region where every CQA simultaneously meets specification. Three nested boundaries describe the operational hierarchy from strictest to broadest.
Normal Operating Range (NOR) is the tightest region — day-to-day operating targets with narrow control limits. This is where routine production runs. Proven Acceptable Range (PAR) is the region demonstrated by DOE data to produce acceptable quality for each parameter individually. Design Space is the multivariate region where all CQAs meet acceptance criteria simultaneously, accounting for parameter interactions. The design space can extend beyond individual PARs when interaction effects are favorable.
The design space is typically represented as an overlay plot (for 2 CPPs) or a response surface contour plot (for 2–3 CPPs). For processes with 4+ CPPs, Monte Carlo simulation or Bayesian approaches generate the multidimensional design space and verify that a proposed operating point has a high probability (>95%) of meeting all CQA specifications simultaneously.
Building Your Control Strategy
The control strategy translates the design space into operational reality by defining how CPPs are monitored, controlled, and maintained within acceptable ranges during routine manufacturing. A robust control strategy integrates three layers of control.
- Level 1 — Real-time process control: Automated feedback loops for CPPs with fast dynamics. Examples: pH controlled by CO2/base addition (±0.05 units), DO controlled by cascade (agitation → air → O2 enrichment), temperature controlled by jacket (±0.2°C).
- Level 2 — In-process testing: Off-line or at-line measurements during the run. Examples: daily VCD and viability, metabolite profiles (glucose, lactate, ammonia), osmolality checks.
- Level 3 — End-of-batch release testing: CQA measurements on the final product. Examples: titer by Protein A HPLC, glycan profile by HILIC, aggregates by SEC, charge variants by icIEF.
As of 2026, the trend is moving toward Process Analytical Technology (PAT) — real-time CQA monitoring using Raman spectroscopy, capacitance probes, or soft sensors that predict CQAs from online CPP data. PAT enables a shift from Level 3 to Level 1 control, enabling real-time release testing (RTRT) and reducing batch release timelines from weeks to hours.
Clone Selection Scorecard
Score and rank clones by titer, growth rate, product quality, and stability. Weighted multi-criteria analysis with radar charts.
Worked Example: mAb Upstream Design Space
This worked example walks through CPP and CQA mapping for a CHO-based mAb upstream process, from risk assessment through design space definition. The scenario is a Phase III mAb with three key CQAs: titer (≥3.0 g/L), galactosylation (25–55% G1F+G2F), and aggregates (≤2.0% by SEC).
Worked Example — CHO mAb Upstream Design Space
Step 1: Risk assessment. An FMEA scored 35 process parameters across severity, probability, and detectability. Eight parameters exceeded the RPN threshold of 100:
- pH set point (RPN = 240)
- Temperature shift timing (RPN = 210)
- Temperature shift magnitude (RPN = 200)
- Feed rate (RPN = 180)
- Dissolved oxygen set point (RPN = 160)
- Seed density (RPN = 140)
- Osmolality (RPN = 120)
- pCO2 (RPN = 110)
Step 2: Screening DOE. A Definitive Screening Design with 8 factors required 17 runs (2k+1 = 17). After analysis (p < 0.05):
- Significant main effects: pH, temperature shift, feed rate, DO
- Significant interaction: pH × temperature
- Non-significant: seed density, osmolality, pCO2 (moved to SOP control)
Step 3: Optimization DOE. A face-centered Central Composite Design with 4 factors and 5 center points required 29 runs. Response surface models were fitted for each CQA:
Titer (g/L) = 2.1 + 0.42×pH* + 0.31×T* − 0.15×Feed* + 0.22×pH*×T*
R² = 0.91, p < 0.001
Galactosylation (%) = 38.2 + 8.1×pH* − 5.4×T* + 3.2×pH*×T*
R² = 0.88, p < 0.001
Aggregates (%) = 1.1 − 0.18×T* + 0.25×Feed* + 0.09×pH*²
R² = 0.85, p < 0.001
(* = coded factors: −1 to +1)
Step 4: Design space. Overlaying all three CQA acceptance criteria on the response surfaces defines the design space. The viable region where titer ≥ 3.0 g/L AND galactosylation 25–55% AND aggregates ≤ 2.0% spans:
- pH: 6.85–7.15 (NOR: 6.95–7.05)
- Temperature: 33.5–36.5°C (NOR: 34.5–35.5°C)
- Feed rate: 0.8–1.2× target (NOR: 0.95–1.05×)
- DO: 35–55% (NOR: 40–50%)
Step 5: Verification. Three confirmation runs at the NOR center point and two at design space edges confirmed all CQAs within specification (titer: 3.4–3.8 g/L, galactosylation: 38–48%, aggregates: 0.8–1.5%).
CHO Troubleshooter
Diagnose CHO cell culture problems — viability drops, lactate spikes, glycosylation shifts. Interactive decision-tree logic.
Understanding the interaction between pH and temperature is essential for setting the design space boundaries. The chart below shows how galactosylation varies across the design space as both pH and temperature change, overlaid with the CQA acceptance window.
Frequently Asked Questions
What is the difference between CPP and CQA in bioprocessing?
A Critical Process Parameter (CPP) is an input variable whose variability impacts product quality, such as pH, temperature, or dissolved oxygen. A Critical Quality Attribute (CQA) is a product output property that must meet defined limits to ensure safety and efficacy, such as potency, glycosylation profile, or aggregate levels. CPPs drive CQAs through cause-and-effect relationships established during process characterization.
How do you define a design space for a biologics process?
A design space is defined through three steps: (1) risk assessment to identify potential CPPs and CQAs, (2) DOE-based process characterization studies to quantify CPP–CQA relationships and interactions, and (3) multivariate analysis to establish the multidimensional region where all CQAs meet acceptance criteria simultaneously. ICH Q8(R2) defines the design space as the combination of input variables demonstrated to provide assurance of quality.
How many DOE runs are needed for process characterization?
The number depends on the factor count and design type. A screening design for 8–12 parameters typically requires 20–30 runs using a fractional factorial or DSD. An optimization study for 4–6 confirmed CPPs using CCD or Box-Behnken requires 25–50 runs. Most biopharmaceutical companies run 40–80 total experiments across screening and optimization phases per unit operation.
What is the difference between design space and proven acceptable range (PAR)?
The PAR is the univariate range for a single parameter demonstrated to produce acceptable quality when all others are held at target. The design space is the multivariate region accounting for parameter interactions. The PAR is always a subset of the design space because it does not capture how parameters interact. Operating within the design space but outside individual PARs is permissible when interaction effects are favorable.
Does moving within the design space require a regulatory filing?
No. ICH Q8(R2) states that movement within an approved design space is not considered a change and does not require a regulatory post-approval change filing. This operational flexibility is a key advantage of QbD. Movement outside the design space requires a filing (Prior Approval Supplement in the US, Type II Variation in the EU).
How do you assess criticality of process parameters?
Criticality is assessed on a continuum using a risk-based approach. Start with an Ishikawa diagram to list all parameters, then score each using FMEA (severity, probability, detectability). Parameters exceeding an RPN threshold enter DOE characterization studies where statistical significance and effect size on CQAs confirm or refute criticality.
Related Tools
- Scale-Up Calculator — Translate CPP set points across bioreactor scales using P/V, tip speed, kLa, and mixing time criteria.
- Clone Selection Scorecard — Multi-criteria weighted scoring for clone ranking by CQAs including titer, growth, and product quality.
- CHO Troubleshooter — Diagnose CQA deviations (glycosylation shifts, aggregation, low titer) with root cause analysis.
References
- ICH Q8(R2): Pharmaceutical Development. International Council for Harmonisation. ich.org/Q8_R2
- ICH Q9(R1): Quality Risk Management. International Council for Harmonisation, 2023.
- A-Mab Case Study: A Case Study in Bioprocess Development. CMC Biotech Working Group, ISPE. ispe.org/a-mab
- Wohlenberg OJ et al. Optimization of a mAb production process with regard to robustness and product quality using quality by design principles. Eng Life Sci. 2022;22(7):484-497. doi:10.1002/elsc.202100172
- FDA. Process Validation: General Principles and Practices. Guidance for Industry, January 2011 (Revision 1).