A depyrogenated vial, a sterilized metal part, and a product treated in its final container may pass through similar thermal equipment while requiring very different decisions. A sterilization strategy fails when it starts from the available cycle instead of the object to be treated, the microbiological risk, and the necessary protection until use. The operational question is which path allows for achieving and maintaining the intended result with reproducible evidence.
This path mainly concerns moist heat, dry heat, and depyrogenation in pharmaceutical manufacturing. Sterilizing filtration, irradiation, and ethylene oxide can be relevant alternatives in specific situations, but they require their own evaluations and do not constitute interchangeable equivalents to a thermal treatment. Design recommendations are identified as [GUIDEGXP RECOMMENDATION] and must be adapted to the product and the jurisdiction.
1. Formulate objectives separately
Sterilization aims for a defined microbiological result through a validated process. Depyrogenation addresses pyrogens; in thermal processes for pharmaceutical containers, the demonstration normally focuses on bacterial endotoxins. Endotoxins are not viable organisms: inactivating the bacteria that originated them does not prove that they have been eliminated or rendered acceptable.
Therefore, define two linked but distinct evaluations: microbial contamination risk and endotoxin risk. A load may require sterilization without a thermal depyrogenation process, while components intended for a parenteral product may require specific endotoxin control. The objective is not deduced from the machine's name: it must derive from intended use, product contact, materials, and the overall process.
Disinfection, sanitization, and biodecontamination reduce contamination under defined conditions, but they do not automatically authorize a sterility claim. Avoid using effective surface disinfection as a justification for a critical component that should be sterilized. The strategy must also specify motivated exceptions and the necessary approval path.
2. Consider terminal sterilization
[GUIDANCE GMP EU] Annex 1 and the 2019 EMA guideline steer towards terminal sterilization of the product whenever possible. This regulatory preference must guide development, without becoming the claim that one technology is suitable for any formulation. Feasibility requires data on stability, degradation, container performance, and product quality after treatment.
Analyze the product in the intended commercial configuration: concentration, volume, headspace, container-closure system, and load distribution. A favorable result on a small sample does not necessarily demonstrate transferability to the production configuration. Cooling, backpressure, and transition speeds can also affect integrity and quality.
If terminal sterilization is not feasible, document the scientific justifications and the evaluated alternatives. An aseptic path requires an overall contamination control system; it cannot be justified solely by the availability of a filling line. Interfaces with transfers, containers, and barriers are covered in Aseptic Fill-Finish & Barrier Systems.
3. Build the material profile
For each product or component, collect composition, geometry, mass, thermal conductivity, porosity, cavities, and sensitivity to temperature and humidity. For polymers and elastomers, consider deformation, aging, and the consequences of repeated exposures. For metals and glass, evaluate assemblies, contact points, coatings, and characteristics that may hinder heat transfer.
Packaging is part of the treatment system. It must allow for the intended process and subsequently protect the material. Packaging chosen for mechanical resistance may hinder air removal or steam penetration; a barrier suitable for moist heat may not be compatible with a more severe dry heat treatment. Request evidence in the actual configuration.
Specify initial conditions and holding times before treatment. Microbial growth or contamination accumulation during an uncontrolled wait must not be compensated for by a generic increase in temperature. Washing, drying, handling, and storage prepare the load and influence the result; the related controls must be linked to the strategy without confusing cleaning and sterilization.
4. Select moist heat
Moist heat may be appropriate when the material and configuration allow for the necessary heat transfer without compromising quality. For porous loads or packaged parts, residual air, drainage, and surface accessibility are central. For liquids in closed containers, internal thermal response and container protection become determining factors.
Not all moist heat processes use the same saturated steam contact scheme. Systems with steam-air mixtures or superheated water may respond to specific needs, but they must be developed and characterized according to their operation. The simple relationship between total pressure and temperature does not demonstrate saturated steam conditions when non-condensable gases or intentional mixtures are present.
The strategy must indicate how heat distribution, load penetration, repeatability, and any calculated lethality will be demonstrated. F0 is not the time spent in the exposure phase: it represents an equivalent magnitude calculated from the thermal history according to declared reference conditions. The model must be relevant and does not replace the control of other critical parameters.
5. Select dry heat and depyrogenation
Dry heat transfers energy through a system where air circulation, geometry, and the thermal response of the items must be characterized. It can be relevant for materials that tolerate the treatment and for specific sterilization or depyrogenation applications. It is not correct to adopt it solely because steam is unavailable: compatibility and performance must be demonstrated.
Distinguish a batch oven from a continuous tunnel. The former requires control of the load configuration and the full cycle; the latter links exposure, container position, belt speed, air balance, and continuity with the line. Interruptions, startup, and shutdown must be included in the evaluation, as they can produce conditions different from steady-state operation.
[GUIDANCE GMP EU] For thermal depyrogenation processes within its scope, Annex 1, point 8.68, requires a demonstration of suitable Fh and at least a 3 log10 reduction of endotoxins. The reference is specific to that context and does not impose a universal combination of temperature and duration. The challenge must be representative, recoverable, and analytically interpretable.
6. Use a transparent decision matrix
| Issue | Moist Heat | Dry Heat | Thermal Depyrogenation |
|---|---|---|---|
| Objective | Defined sterilization process | Sterilization of compatible materials | Demonstrated endotoxin reduction |
| Constraint to verify | Compatibility with heat and humidity | Thermal tolerance and heat transfer | Component tolerance and initial risk |
| Critical configuration | Air, packaging, drainage, and load | Spacing, shielding, and circulation | Format, trajectory, exposure, and flows |
| Core evidence | Physical measurements and appropriate microbiological support | Distribution and penetration with microbiological strategy | Thermal profile and suitable endotoxin challenge |
| Subsequent protection | Intact barrier and controlled transfer | Controlled cooling and maintenance | Prevention of recontamination until filling |
The matrix does not assign absolute scores. Before comparing costs and capacities, exclude alternatives incapable of meeting critical product requirements. Subsequently, compare robustness, ease of management, maintenance, available evidence, and future expansion. Making missing information explicit prevents an estimate from being presented as a conclusion.
7. Define families and worst-case conditions
A family brings together loads for which a scientific justification allows for the use of representative evidence. The criterion may involve material, dimensions, mass, packaging, filling, or geometry, but it must be consistent with the mechanism that limits the process. Two outwardly similar items can have very different thermal behaviors.
The worst-case condition may differ for heating, air removal, drying, cooling, and risk of damage. Avoid a single "worst case" label without explaining which performance is being challenged. The strategy should provide a matrix of conditions and indicate which tests cover each risk, including minimum and maximum loads when relevant.
Define rules for admitting new configurations: necessary information, multidisciplinary review, equivalence criteria, and management of differences. An approved load photograph is useful for execution, but must be supported by the logic that justifies its use. Otherwise, every change becomes a discussion lacking shared criteria.
8. Link development, validation, and routine control
Development explores system behavior and identifies a robust operating window. Validation demonstrates that the defined process meets pre-established criteria under the relevant conditions. Using exploratory tests as if they were already a confirmatory protocol makes it difficult to distinguish technical choices, deviations, and criteria modified after seeing the results.
Establish in advance which measurements must support routine acceptance. A validation rich in independent probes must explain how the measurements available in routine represent critical conditions. If the relationship is not demonstrated, the operator might accept a cycle based on data poorly representative of the load.
Biological indicators, when appropriate, supplement physical characterization. Their selection requires relevance to the process, documented performance, controlled management, and consistent interpretation. Chemical indicators provide information related to the conditions to which they were exposed; they do not replace the body of evidence necessary to conclude that the process is acceptable.
9. Example: components for an aseptic line
A line must prepare packaged metal parts and glass vials. Treating them as a single problem of "sterile materials" hides the different objective for vials, the continuity required at filling, and the methods of protecting parts. The first decision is to separate the families and describe the risk of each interface.
For parts, evaluate air removal, steam contact, drainage, drying, and packaging integrity. For vials, evaluate washing, endotoxin control, tunnel profile, challenge, and protection at the exit. The two technologies can be complementary in the same strategy; the choice does not derive from an abstract ranking between moist heat and dry heat.
The final document links each path to its load, acceptance criteria, and data demonstrating its control. It also includes material management after an interruption and the responsibility for authorizing line restart. These decisions must be agreed upon before the equipment is purchased.
10. Make trade-offs and exclusion conditions explicit
A technically feasible technology may prove to be not very robust in site routine. Consider utility availability, operator skills, laboratory capabilities, maintenance accessibility, and time needed to re-examine data. These factors do not authorize a reduction in quality requirements, but help choose between solutions that satisfy them. A process requiring constant exceptional interventions deserves a design review.
Establish exclusion conditions before supplier selection. Examples include demonstrated material incompatibility, inability to protect the treated component, insufficient data on medium penetration, or the absence of a credible strategy for managing interruptions. A high commercial score must not compensate for the failure to meet these conditions. Preliminary tests can resolve uncertainty, but one must define who performs them and how they will be evaluated.
Also evaluate consumption across the entire process, including non-productive time and waste. An apparently shorter cycle may increase instantaneous utility demand or make cooling more sensitive; a higher nominal capacity may produce waits incompatible with material preparation. Compare realistic scenarios, keeping measured results, supplier estimates, and project assumptions separate.
The transition to operational development should produce a precise package: list of families, objectives, variables to study, measurement methods, preliminary criteria, responsibilities, and decision points. Record open questions with a deadline and expected evidence. In this way, an approved strategy is not interpreted as authorization to adopt any cycle available on the machine.
11. Approve and maintain the strategy
[QRM] The risk assessment must be proportional to the complexity and uncertainty of the decision. A table with scores does not replace failure mechanism analysis. For each relevant risk, indicate preventive controls, detectability, consequences on the material, and information to collect when the control fails.
Assign owners to product, load, cycle, utilities, automation, and data. Define which changes impose a strategy review: new materials, packaging changes, transfer to another machine, updated recipes, changes to flows, or initial conditions. The strategy remains useful if it accompanies the lifecycle and guides subsequent decisions.
Before approval, verify four elements: distinct objective, compatible technology, planned evidence, and subsequent protection. If one is missing, the availability of a nominal cycle does not solve the problem. Keeping discarded alternatives and the reason for the choice allows for re-evaluating the process when product or production constraints change.
References and further reading
Sources verified on September 23, 2026: EudraLex Volume 4, Annex 1; EMA, guideline on the sterilization of the medicinal product, active substance, excipient and primary container, effective from 2019; ICH Q9(R1), Quality Risk Management, 2023. The sources have different scopes and natures and must be applied to the authorized context.
Consult the Sterilization & Depyrogenation Systems hub and the URS checklist. For steam as a utility and SIP systems, continue in Critical Utilities and Cleaning, CIP & SIP, respectively.
