Certificates of Test - What are the European GMP Requirements?

Published on: 

Pharmaceutical Technology Europe

Pharmaceutical Technology Europe, Pharmaceutical Technology Europe-03-01-2004, Volume 16, Issue 3

This article examines European differences in GMP requirements for the acceptance of Certificates of Tests. The authors look at how pharmaceutical manufacturers can address the issues and suggest a US-compatible framework for GMPs that could be incorporated into EU requirements.

Within the premises of European Union (EU) pharmaceutical manufacturers, there is typically up to 3 weeks' worth of supplied materials inventory waiting to be tested. This is consistent with available US data reported during a benchmarking exercise regarding pharmaceutical manufacturing performance.1 Quality control (QC) waiting times such as these contribute significantly to the typically low stock-turns (5-10 per annum) prevalent for pharmaceutical materials. These high inventories and associated QC costs represent a pressure point as businesses strive to stay competitive.

One well-proven method of containing costs is to expedite the QC testing process for such materials. This may be done by accepting supplier-provided data at the point of receipt into the pharmaceutical facility that confirms the batch's compliance with specifications. The frequently expressed argument "Why is it necessary to test the batch a second time before it is released for use within the pharmaceutical plant?" is a strong one; however, there are many obstacles to overcome to achieve the high level of supplier certification sought by pharmaceutical manufacturers.

Poorly defined GMP rules

One obstacle to efficient QC testing is that requirements associated with this process are poorly defined in the EU good manufacturing practice (GMP) rules,


which do not even contain a specific reference to Certificates of Test (CoTs). Although the UK's Medicines and Healthcare products Regulatory Agency (MHRA) has issued guidance concerning what a certificate should contain,


this only represents one part of a series of otherwise unwritten GMPs on this subject.

This absence has led to differing interpretations by national and regional inspectors, resulting in different economic impacts. This represents a potential quality pitfall - the national GMP requirements are relatively undemanding and are seen as unfair competition in countries where the regulatory authorities apply stricter standards.

This article explains why pharmaceutical manufacturers should actively engage in this process and when they should be careful. Examples are provided of typical GMP requirements for acceptance of CoTs and the article discusses how they vary in different European countries. The authors have developed a recommended process for acceptance of certificates, based on experience of managing supplier development programmes, surveys conducted amongst quality professionals around Europe and interaction from participants on training courses. The article concludes by suggesting a framework for GMPs that are compatible with defined US requirements and could be incorporated into the EU GMP rules for finished pharmaceutical products.

Figure 1: Process for reducing QC testing based on acceptance of certificates of test.

Accepting CoTs

The pharma industry typically purchases the majority of the materials that are used to manufacture and package its products, such as excipients, active pharmaceutical ingredients (APIs) and primary and secondary packaging. Therefore, as 95–100% of raw materials are purchased, the expertise for making and testing these materials resides with the manufacturer and not with the purchasing pharmaceutical company.

However, regulations and registered requirements place responsibility for ensuring the quality of raw materials used with the pharmaceutical manufacturer. Consequently, companies routinely perform chemical, microbiological and physical testing on samples of purchased materials that are often small and possibly not representative of the batch. Similarly, the tests may not be the most suitable to determine the overall quality of the materials because the purchaser does not have the specialized equipment necessary to measure the appropriate parameters.

Manufacturers' products

Producers of materials often take many more representative samples throughout the manufacturing process and use well-established test methods to determine the quality of their products. They also appreciate the implications of subtle variations in test results that the purchaser might not. The majority of manufacturers will also provide the data to their customers with the goods in the form of CoTs, if asked.

The appropriate use of manufacturer's data in the form of a CoT can benefit both the supplier and the purchaser. The buyer can reduce the amount of testing done upon receipt and the supplier knows that the materials are less likely to be rejected because the client has confidence in the data provided.

Additional benefits are present that are both intuitive and borne of experience and research. Time invested by the pharmaceutical company to achieve the understanding and assurance levels required pays dividends in terms of reduced errors and fewer suppliers. Overall, these benefits facilitate lower inventories.


Reasons to be careful

Although EU GMP rules are common to all countries, specific GMPs appropriate to the acceptance of CoTs are not clearly defined. Consequently, GMP interpretation can be different in each member state. All countries expect that purchased materials will meet the required standards and most will accept that data from CoTs can be a substitute for additional testing if additional controls are in place. However, research shows that this approach is not universally accepted across the EU. In France, for example, inspectors currently expect full testing to take place for all APIs and excipients, and are not amenable to the use of CoTs. This, however, may change during 2004.

History is important

To use CoTs, a pharmaceutical company must know how the manufacturer makes, samples and tests the products, and what systems are in place for controlling production, distribution and quality. In addition, a good history of the product's supplies must exist. This type of information is usually gathered by questionnaire and/or by audit of the facilities concerned. It is not be advisable to use data from certificates without this minimum level of assurance.

Arranging audits and collecting data is both time and resource intensive. Before embarking on such a course, it is advisable to consider from where the most benefit can be obtained. Materials such as packaging items (for example, labels and cartons) tend to be purchased in large numbers and their acceptance testing is usually repetitive. In this instance, the economic benefits gained by using supplier's data is high. Conversely, materials such as APIs are purchased in smaller amounts; testing can be complex and the benefits are usually less. As an example, a pharmaceutical company that manufactures products that are sensitive to certain parameters in an excipient might want to exclude that particular test from its reduced testing programme. This policy might also extend to materials for sterile applications because of their mode of use.

Current requirements

The current stated GMP requirements for the acceptance of certificated data from suppliers is not extensive. There is no specific mention of the acceptability or otherwise of CoTs within current EU GMP rules for secondary pharmaceutical products. There is a mention in Annex 18, which covers GMPs for APIs, but this Annex has not yet been implemented by regulation. Useful guidance is provided by the MHRA regarding the content of certificates, but this does not enjoy EU-wide acceptance, nor is it supplemented by any guidance covering the nature of the system that should be put in place to support the use of such information.

Table I: Cost/benefit analysis of supplier/material combinations.

It is anticipated that national legislation will be put into effect in France this year that will permit the use of CoTs. The authors understand that the suppliers involved must either be based within the EU or linked by a mutual recognition agreement (MRA) and will also need to submit to a regulatory GMP inspection. Not surprisingly, compliance with the marketing authorization would be expected.

Contractual implications

Importantly, there is an implicit requirement that when data (provided on a certificate) are used by the pharmaceutical company to make decisions relating to the suitability of the batch, then a contract must be in place between the parties involved. This is because the act of accepting a supplier's data represents the contracting out of an analytical service, which is governed by Article 12 of Directive 91/356/EEC.


This requires that "Any . . . operation linked with the manufacture, which is carried out under contract, shall be the subject of a written contract." However, the guidance provided in the appropriate related section of the GMP rules is insufficiently specific for this situation.

CFR more specific

The Code of Federal Regulations (CFR) is much more specific and helpful in this matter because it specifically states in clause 211.84(d) that "a report of a test may be accepted from the supplier of a component . . . provided that the manufacturer establishes the reliability of the supplier's analyses through appropriate validation of the supplier's test results at appropriate intervals."


Although Annex 18 of the EU GMP rules does not yet have a legal basis and applies only to APIs, it does contain some helpful guidance in Section 7.30. "A supplier's CoT can be used in place of performing the tests, provided that the manufacturer has a system in place for evaluating suppliers." This is usefully expanded on in section 7.31 where a requirement to evaluate the reliability of the quality history - and also to validate at least three sets of the supplier's results - is introduced. Full analysis must be performed "at appropriate intervals" and the supplier's certificates checked for reliability.

The WHO standard

The World Health Organization (WHO) draft guidance on

Good Trade and Distribution Practice of Pharmaceutical Starting Materials

outlines the standards to be adopted by suppliers concerning the content and provision of Certificates of Analysis, skip lot testing by the manufacturer, traceability and the controls required should batches become mixed.


These should be incorporated into the evaluation process employed.

Acceptance of supplier's data

The authors' recommended approach for the acceptance of supplier's data is based on experience and surveys conducted amongst companies throughout Europe. This core process, which is outlined in blue in Figure 1, defines the necessary sequence of activities.

This five phase process is interlinked with the three groups that have a stake in a successful outcome: regulatory bodies requiring compliance with appropriate regulations; suppliers whose data is destined to be used on trust; and the pharmaceutical company's management who are seeking to improve efficiency and maintain the confidence of regulators and patients. This is a complex mix that requires a high level of understanding and focus, but a mix that can enhance both the efficiency and quality of supplied materials.

The five phases of activity and the important principles that underpin each one are given in detail below:

Phase 1:Understanding the current performance of each material/supplier combination received and the relevant regulatory requirements. This phase ensures that

  • The current QC system reflects the registered requirements precisely and gives an understanding of whether any specific constraints on reducing testing are defined in the marketing authorization. For example, there may be a specific requirement defined to perform QC testing on each batch for a sensitive parameter.

  • Measurement processes are in place for both the supplier's quality system and the quality of the material supplied. These must be considered separately as both are important when deciding whether or not to proceed. The measurement for the quality system would be qualitative and should address the question "Is this system well enough developed to initiate or continue supply?" The measurement for the material quality should be quantitative and should incorporate batch reject rates, concession rates and complaints from external and internal customers (for example, production). It should also enable comparison between different companies by normalizing the number of batches (or volume) received and allocating a specific score for each type of deficiency. Available process capability data should also be reviewed.

Phase 2: Identify potentially cost-effective material/supplier combinations where the acceptance of supplier's data would be acceptable to regulators and suppliers. This phase involves identifying suitable candidate supplier/material combinations such as

  • Understanding the constraints placed by national/state inspectors regarding the acceptability of reducing QC testing. These constraints represent the unwritten GMP requirements and the next section shows how these can vary within the EU.

  • Performing a cost/benefit analysis for each supplier/material combination. This should consider factors such as the cost of QC and holding stocks of materials, together with the increased costs of higher level evaluation of a supplier's system and performance measurement process maintenance. The matrix shown in Table I illustrates that supplier/material combinations in Box 2 are ideal for reducing QC testing; it also shows appropriate strategies for other scenarios.

  • Evaluation of the supplier's quality system and management attitude to understand whether it has the capability to develop its system to the level required. This demands a listening approach and checks for coherence between declared intentions and delivered performance.

Phase 3: Develop and validate the acceptance process in conjunction with the supplier and internal customers. This phase requires

  • The discussion of a draft contract with the supplier to define parameters such as sampling frequencies and locations, test methods, format of the certificates and change notification.

  • Obtaining draft contract agreement with internal stakeholders such as production management and the qualified person. Note that this will probably require progression through the change control process.

  • Validation of the supplier's results using a conventional approach when the accuracy required for the test results is defined in advance of the validation study. This might highlight that the supplier is achieving more accurate results than the QC laboratory in the pharmaceutical company.

Phase 4: Introduce the new process. This phase requires that

  • A procedure is written to define how the system will work. The monitoring frequency and process should be defined, for example, as should the situations wherein full testing should be reinstated. It is particularly important that such 'alarms' are clearly defined at this stage because once QC testing has been reduced and analysts redeployed it is often difficult to revert back. An alarm could be, for instance, a significant difference between the supplier's reported result and a monitoring result. Another example is a drop in the supplier/material score to below a preset value.

  • Analysts are trained in the new monitoring process and also made aware that this change is completely separate from any reduction in identity testing that may or may nor be implemented in accordance with Annex 8 of the EU GMP rules.

Phase 5: Ensure that the monitoring process is effective. This phase requires

  • Monitoring the supplier's data at intervals defined in the procedure. Typically this is initially set at every fifth batch and, when confidence develops, can move to every tenth or twentieth batch for high volume materials. At least one batch should be tested every year.

  • Monitoring the supplier's quality system at intervals defined in the procedure.

  • Annually monitoring/reviewing the supplier's performance history.

  • Providing feedback to the supplier on the outcome of this review at least annually.

  • Maintaining the supplier/material history and making comparisons with agreed internal acceptance standards.

Unwritten GMP requirements

Although specific EU GMPs appropriate to the acceptance of CoTs are not clearly defined, research has shown that despite wide variations, several national inspectorates express a similar set of expectations, which typically include

  • a contract to be in place

  • a good history of previous supply

  • a satisfactory audit (note that the audit standard varies)

  • a procedure that includes a defined monitoring process.

The absence of a specifically defined set of GMPs means that the expectations of a particular inspectorate cannot be prejudged and must be ascertained before implementing any reduction in materials QC testing. For example, the French authorities do not currently permit the use of certificates for APIs and excipients, and one German state is reported not to accept certificates from outside the EU unless an MRA is in place. The subject of which standard should be applied to the audit is also not well defined. Whereas some authorities require that ISO 9001 is applied, this is not consistently expected. The basic quality system standard ISO 9001 is inadequate in that necessary GMPs appropriate to suppliers are omitted. The Pharmaceutical Quality Group's GMP application standards for suppliers (PS 9000/PS 9100), which build on ISO 9001:2000, provide a much better basis.5,8

Reduced incoming QC testing for conformance to the specification should not be confused with reduced identity testing, for which the GMPs are well defined in Annex 8. This concern derives from disasters such as the Haitian diethylene glycol incident.9


The authors' research has shown that where regulatory authorities specify requirements for the acceptance of certificated data, their requirements are consistent with those specified in sections 7.30 and 7.31 of Annex 18 for APIs and clause 211.84(d) of the CFR. Additionally, they often ask for a contract to be in place between the two parties. However, some national regulatory authorities insist on additional specific expectations that must be understood before any reduction in QC testing is implemented. The inclusion of a specific baseline set of GMPs within Annex 8 of the EU GMP rules would represent a valuable clarification of regulatory expectations for this economically important subject.

Authors' recommendations

The authors have provided the basic requirements and reasons for reducing incoming QC testing. Ultimately, it is the manager of the quality unit who must decide how to approach the use of other companies' data and to assess whether it will be beneficial to his/her organization. To make that decision and gain substantial benefits, the manager must spend time on the project and set up some measurement indices that will show where to concentrate the effort.

Initially, the authors recommend that experience be gained by working with one or two suppliers of packaging materials. Other suppliers can be included as experience of the balance between the benefits and the management implications becomes apparent.


1. G.K. Raju and C. Cooney "Benchmarking Pharmaceutical Manufacturing Performance," Working Paper No. 11-94, MIT Programme on the Pharmaceutical Industry (1995).

2. MCA Rules and Guidance for Pharmaceutical Manufacturers and Distributors 2002, Part 1, Chapter 4, 6th Edition (Medicines and Healthcare Products Regulatory Agency, London, UK, 2002).

3. MCA Rules and Guidance for Pharmaceutical Manufacturers and Distributors 2002, Part 1, Chapter 2, 6th Edition (Medicines and Healthcare Products Regulatory Agency, London, UK, 2002).

4. http://pharmacos.eudra.org/F2/eudralex/vol-1/new_v1/dir91-356en.pdf

5. www.iqa.org/information\d2-10.shtml

6. www.fda.gov/cder/dmpq/cgmpregs.htm

7. www.who.int/en/

8. www.pqg.org

9. www.essentialdrugs.org/edrug/archive/199708/msg00024.php

This article is reproduced with kind permission from GMP Review 2(4), 6-12 (2004).