Log in

Cleaning control strategy in the manufacture of Pharmaceutical Active Ingredients in development for clinical use (La Vague 54)

During the production of batches of commercial pharmaceutical active ingredients, validation of the cleaning of dedicated facilities is generally carried out. Systematic cleaning controls are not therefore necessary. In the case of the development of new active ingredients for pharmaceutical use, the problem of cleaning must be taken into account starting from the production of the first batches for use in phase I to III clinical trials. This article describes the strategy applied at ORIL Industrie, Servier Group, during the production of these New Chemical Entities (NCE) for clinical trials.

1The first batches of NCE are generally manufactured in workshops referred to as “multi-product/polyvalent”. In fact, our site simultaneously manufactures batches of active ingredients for several indications, at various stages of development. As the attrition rate of products in development is very significant, no installation can be dedicated before the end of phase III. Numerous active ingredients and intermediates therefore follow each other in the installations. As the number of patients varies considerably depending on the progress of trials and indications, the batch sizes, types of reactors, dryers, mills, etc. are very different from one batch to the other for the same project. Because of this diversity, crossover between products cannot be predicted but the risk of cross-contamination must be considered in keeping with Good Manufacturing Practices.
No cleaning validation can be planned and therefore testing must be systematic.

A progressive approach

Visual inspections are required at all stages of synthesis. These are carried out by operators and tracked in the process control forms for difficult-to-access areas. Equipment is partially or completely disassembled in order to apply effective visual checks. As needed, wiping may assist the visual examination of difficult-to-access areas.

A progressive approach is proposed for analytical cleaning control based on a risk analysis.(1) Analytical cleaning control is conducted systematically on the equipment used for the last chemical step and the last physical step of production of the active ingredient. For the intermediate steps, analytical controls are run for projects at most advanced stage of development (phase II and III).

Acceptance criteria

The acceptance criteria are determined using all the data then available for these products. Online monitoring of the results of toxicology and pharmacology studies is necessary to ensure their inclusion in the calculation of acceptance criteria. Computer tools have been developed in order to increase the reliability of storage of this information. Several calculation methods are systematically considered:

• Calculation on the basis of the daily dose administered.
• Calculation on the basis of toxicological data: the PDE is included when it is available. In the absence of this data, the LD50 values can be considered. The genotoxic nature of the compounds is also taken into account and an acceptance criterion can be calculated following the ICH M7 Guideline.(2)
• Use of a chemical criterion of 10 ppm of residual contaminant in the following batch.
• After calculation of all the possible criteria from the data available, the lowest value is selected for acceptance of the equipment.

The cleaning methodology

Before the first transfer to the production facility, laboratory studies are run in order to acquire solubility data. These are used to develop the first cleaning protocols.
Subsequently, depending on the experiences gained during synthesis, cleaning conditions can be optimized where necessary.

Sampling for cleaning control analyses

Most analytical cleaning tests are performed by rinsing especially for reactors, filters and isolators. Whenever technically possible, heating under reflux is carried out.
A risk analysis conducted on the site authorizes the exemption from testing of some small items that are easy to inspect visually and which have a surface area much smaller than that of the main equipment.
The equipment is placed in quarantine before rinsing or wiping is carried out. Quarantine is only lifted after a compliant analysis has been obtained. For the last step of physical processing of the active ingredient, the testing includes the equipment train in its totality.

Analytical techniques for the testing of cleaning

Most of our NCE have chromophore group,s which makes UV the detection technique of choice. The method developed around ten years ago was non-specific and consisted in producing a UV spectrum of the rinse solvent and subtracting it from the rinsate blank. Quantification was then performed using a reference spectrum of the product to be investigated. The feedback on this non-specific method of analysis in our context is negative as many false positives were obtained. This was linked in particular to the topology of the facility which has a system of solvent batchers to supply the reactors. Thus contamination of the rinse solvent by 10 ppm of another solvent very likely containing a chromophore with industrial solvent batchers resulted in a non-compliance and in repeated cleaning of the installation when the contaminant was not present.

Example of cleaning control using UV spectrometry

To improve the relevancy of our tests we have developed a specific separating method. Ultra-high-performance liquid chromatography is most often used because of its accuracy and speed. UV detection at an appropriate wavelength allows for the measurement of contaminants potentially contained in the equipment. This technique preserves analytical speed since the result is obtained in around 15 minutes after preparation of the sample (each run lasts for 3,5 minutes). The stationary phase was chosen to optimally retain a maximum number of compounds. The contaminant is quantified by external standard calibration or via boundary testing. Technique sensitivity is adjusted to the acceptance criteria of our cleaning controls.

Example of chromatogram run for a cleaning test

The other compounds potentially present are also considered and quantified above a certain threshold. The cleaning test can be rejected in the event that excessive content of an unknown compound is present. The eluants were chosen for their compatibility with a mass detector. Thus in the event of persistent non-compliance, this type of detection can help to understand the source of the problem.

Here is an example of mass spectrometry input: this cleaning test was non-compliant after several cleaning and rinsing operations because of the presence of an unknown product.


Through mass detection, a structural hypothesis was generated for this contaminant then verified against a standard. The quantification of this residue coupled with the risk assessment performed by the operators and quality assurance then demonstrated the low impact on the next phase of synthesis of the active ingredient. The quarantine on the equipment was lifted while patient safety was ensured.
For each new investigation of a contaminant, validation measures were applied to the choice of rinse solvent to check the appropriateness of the method. In this way, more than 100 substances have been successfully investigated since the method was developed.

Other alternative methods such as Thin Layer Chromatography (TLC) or Gas Chromatography (GC) can also be used where there is no chromophore grouping. The methodology remains the same.


The development of a specific analytical method allows our test laboratory to run more than 150 cleaning control analyses a year for the production of several dozen batches of active ingredients for clinical use and to ensure controlled risk of cross contamination.


[1] Guidance on aspects of Cleaning Validation in Active Pharmaceutical Ingredient Plants Active Pharmaceutical Ingredients Committee (APIC) Sept 2016
[2] ICH Guideline M 7 on assessment and control of DNA reactive (mutagenic) impurities in pharmaceuticals to limit potential carcinogenic risk Step 4

By Solenn JANVIER - Servier


PDF version