IPCS INCHEM Home



    INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY



    ENVIRONMENTAL HEALTH CRITERIA 141





    QUALITY MANAGEMENT FOR CHEMICAL SAFETY TESTING








    This report contains the collective views of an international group of
    experts and does not necessarily represent the decisions or the stated
    policy of the United Nations Environment Programme, the International
    Labour Organisation, or the World Health Organization.

    Published under the joint sponsorship of
    the United Nations Environment Programme,
    the International Labour Organisation,
    and the World Health Organization

    World Health Orgnization
    Geneva, 1992


         The International Programme on Chemical Safety (IPCS) is a
    joint venture of the United Nations Environment Programme, the
    International Labour Organisation, and the World Health
    Organization. The main objective of the IPCS is to carry out and
    disseminate evaluations of the effects of chemicals on human health
    and the quality of the environment. Supporting activities include
    the development of epidemiological, experimental laboratory, and
    risk-assessment methods that could produce internationally
    comparable results, and the development of manpower in the field of
    toxicology. Other activities carried out by the IPCS include the
    development of know-how for coping with chemical accidents,
    coordination of laboratory testing and epidemiological studies, and
    promotion of research on the mechanisms of the biological action of
    chemicals.

    WHO Library Cataloguing in Publication Data

    Quality management for chemical safety testing.

        (Environmental health criteria ; 141)

        1.Hazardous substances - toxicity 2.Laboratories - standards 
        3.Quality control 4.Toxicology - methods
        I.Series

        ISBN 92 4 157141 1        (NLM Classification: QV 602)
        ISSN 0250-863X

         The World Health Organization welcomes requests for permission
    to reproduce or translate its publications, in part or in full.
    Applications and enquiries should be addressed to the Office of
    Publications, World Health Organization, Geneva, Switzerland, which
    will be glad to provide the latest information on any changes made
    to the text, plans for new editions, and reprints and translations
    already available.

    (c) World Health Organization 1992

         Publications of the World Health Organization enjoy copyright
    protection in accordance with the provisions of Protocol 2 of the
    Universal Copyright Convention. All rights reserved.

         The designations employed and the presentation of the material
    in this publication do not imply the expression of any opinion
    whatsoever on the part of the Secretariat of the World Health
    Organization concerning the legal status of any country, territory,
    city or area or of its authorities, or concerning the delimitation
    of its frontiers or boundaries.

         The mention of specific companies or of certain manufacturers'
    products does not imply that they are endorsed or recommended by the
    World Health Organization in preference to others of a similar
    nature that are not mentioned. Errors and omissions excepted, the
    names of proprietary products are distinguished by initial capital
    letters.




    CONTENTS

    QUALITY MANAGEMENT FOR CHEMICAL SAFETY TESTING

    INTRODUCTION

    1. GENERAL QUALITY MANAGEMENT APPROACH
         FOR QUALITY ASSURANCE

         1.1. Organization and personnel
                1.1.1. Introduction
                1.1.2. Organization
                1.1.3. Test facility management
                1.1.4. Study director
                1.1.5. Support personnel
                1.1.6. Quality assurance function
                1.1.7. Personnel selection and development
                1.1.8. Orientation and training of new personnel
         1.2. Quality assurance programme
                1.2.1. Introduction
                1.2.2. Quality assurance and quality control
                1.2.3. Organization and personnel
                1.2.4. Inspections and audits
                1.2.5. Records and reports
                1.2.6. Quality assurance SOPs
         1.3. Facilities and equipment
                1.3.1. Introduction
                1.3.2. Facilities for handling test, control
                        and reference substances
                1.3.3. Field study facilities
                1.3.4. Equipment
         1.4. Study plan
                1.4.1. Introduction
                1.4.2. Study plan preparation
                1.4.3. Format and content of study plan
                1.4.4. Use of study plan
         1.5. Standard operating procedures
                1.5.1. Introduction
                1.5.2. Format and content of SOPs
                1.5.3. Preparation of SOPs
                1.5.4. Typical SOPs
                1.5.5. Use and availability of SOPs
                1.5.6. Adequacy of SOPs
                1.5.7. Maintenance of SOPs
         1.6. Test, control and reference substances
                1.6.1. Introduction
                1.6.2. Test, control and reference substance
                        characterization
                1.6.3. Handling
                1.6.4. Storage
                1.6.5. Distribution

                1.6.6. Mixtures of substances with vehicles
                        (carriers)
                1.6.7. Stability
                1.6.8. Labelling
                1.6.9. Facilities and equipment
         1.7. Quality control
                1.7.1. Introduction
                1.7.2. Level of quality control
                1.7.3. Pre-analytical quality control
                1.7.4. Analytical quality control
                1.7.5. Statistical considerations
                1.7.6. Analytical performance evaluation
         1.8. Documentation and record keeping
                1.8.1. Introduction
                1.8.2. Manual data records
                1.8.3. Computer data records
                1.8.4. Indirect computer data records
         1.9. Final report
                1.9.1. Introduction
                1.9.2. Contents
                1.9.3. Indexing
         1.10. Archiving and retention of data
                1.10.1. Introduction
                1.10.2. Facilities
                1.10.3. Responsibilities for an archive
                1.10.4. SOPs for archiving
                1.10.5. Receiving, indexing and identification
                1.10.6. Filing and storage
                1.10.7. Access and security
                1.10.8. Retrieval of data and control of access
                1.10.9. Retention of information

    2. QUALITY MANAGEMENT APPLIED TO TOXICITY
         STUDIES

         2.1. Introduction
         2.2. Procedural requirements
         2.3. Phases of animal use
         2.4. Obtaining animals
         2.5. Shipping and receipt of animals
         2.6. Animal care facilities
         2.7. Animal husbandry supply facilities
         2.8. Facilities for handling test, control and
                reference substances
         2.9. Pre-study evaluation of animals
         2.10. Allocation of animals to a study
         2.11. Exposure of animals to a test or control substance
         2.12. Control of laboratory environment
         2.13. Evaluation of in-life animal responses to test
                and control substances

         2.14. Removal of animals from a study
         2.15. Transfer of animal tissues and specimens to
                archives

    3. QUALITY MANAGEMENT APPLIED TO HUMAN
         AND ENVIRONMENTAL MONITORING STUDIES

         3.1. Introduction
         3.2. Procedural requirements
         3.3. Selection of sampling strategies and study design
         3.4. Sampling procedures and documentation
         3.5. Handling of samples
         3.6. Analytical performance evaluation
         3.7. The regression method
         3.8. Practical application of the regression method
         3.9. Other analytical performance evaluation programmes
         3.10. Analytical performance criteria
         3.11. Quality control samples

    REFERENCES

    APPENDIX I

    WHO TASK GROUP ON QUALITY MANAGEMENT FOR CHEMICAL SAFETY TESTING

     Members

    Professor W. Almeida, Department of Preventive Medicine,
         State University of Campinas, So Paulo, Brazil

    Professor E.A. Bababumni, Department of Biochemistry, 
         University of Ibadan, Ibadan, Nigeria

    Dr A.W. Choudhry, Kenya Medical Research Institute, Nairobi, 
         Kenya  (Chairman)

    Dr C. Morris, International Chemical Consultants, Alexandria, 
         Virginia, USA

    Dr M. Ruchirawat, Chulabhorn Research Institute, Bangkok, 
         Thailand  (Vice-Chairman)

    Dr A. Strik, National Institute of Public Health and 
         Environmental Protection, Bilthoven, The Netherlands 
          (Rapporteur)

    Dr K. Kanagalingam, Office of Compliance Monitoring, 
         United States Environmental Protection Agency, 
         Washington, DC, USA

    Professor M. Vahter, Institute of Environmental Medicine, 
         Karolinska Institute, Stockholm, Sweden

    Dr Z. Xing-Quan, Institute of Environmental Health 
         Monitoring, Beijing, China

    Mr R. Zisa, Office of Compliance Monitoring, United States 
         Environmental Protection Agency, Washington, DC, USA

     Representatives of other organizations

    Dr R.F.M. Herber, Coronel Laboratory, University of 
         Amsterdam, Amsterdam, The Netherlands, representing the 
         International Union of Pure and Applied Chemistry 
         (IUPAC)

    Professor D. de Wied, Rudolf Magnus Institute of 
         Pharmacology, University of Utrecht, Utrecht, The 
         Netherlands, representing the International Union of 
         Pharmacology (IUPHAR)

    Dr H. Knemann, Public Health Inspectorate, Ministry of 
         Welfare, Health and Cultural Affairs, Rijswijk, The 
         Netherlands, representing the Organisation for Economic 
         Co-operation and Development (OECD)

    Dr R. Lnge, Schering A.G., Berlin, Germany, representing the 
         European Chemical Industry Ecology and Toxicology Centre 
         (ECETOC)

     Secretariat

    Dr E.M. Smith, International Programme on Chemical Safety, 
         World Health Organization, Geneva, Switzerland ( Secretary)

    Dr D.T. Mage, Prevention of Environmental Pollution, Division 
         of Environmental Health, World Health Organization, 
         Geneva, Switzerland

    NOTE TO READERS OF THE CRITERIA MONOGRAPHS

         Every effort has been made to present information in the criteria
    monographs as accurately as possible without unduly delaying their
    publication. In the interest of all users of the Environmental Health
    Criteria monographs, readers are kindly requested to communicate any
    errors that may have occurred to the Director of the International
    Programme on Chemical Safety, World Health Organization, Geneva,
    Switzerland, in order that they may be included in corrigenda.

    QUALITY MANAGEMENT FOR CHEMICAL SAFETY TESTING

         The WHO Task Group on Quality Management for Chemical Safety
    Testing met in Bilthoven, The Netherlands, from 28 May to 1 June 1990.
    Dr E. Smith welcomed the participants on behalf of the heads of the
    three IPCS cooperating organizations (UNEP/ILO/WHO). The Task Group
    reviewed and revised the draft monograph and extended its scope.

         The first draft of the part of the monograph dealing with quality
    assurance of toxicological studies was prepared by Mr E.A. Brisson (US
    Food and Drug Agency, Rockville, Maryland, USA) and the part dealing
    with quality control in human health monitoring was prepared by
    Professor M. Vahter (Karolinska Institute, Stockholm, Sweden) with
    contributions from Professor R. Herber (Coronel Laboratory, University
    of Amsterdam, The Netherlands). Additional text on quality assurance
    for environmental monitoring was prepared by Dr R. Lnge (Schering AG,
    Berlin, Germany). Dr H. Knemann (Ministry of Welfare, Health and
    Cultural Affairs, Rijswijk, The Netherlands) reviewed and integrated
    the text. Finally, Dr Knemann, Professor R. Herber (Coronel
    Laboratory, University of Amsterdam, The Netherlands) and Professor M.
    Vahter acted as an  ad hoc editorial group and gave valuable
    assistance in the preparation of the final text. Dr E.M. Smith and Dr
    P.G. Jenkins, both members of the IPCS Central Unit, were responsible
    for the overall scientific content and technical editing,
    respectively, of this monograph.

         Support for the meeting was provided by The Netherlands National
    Institute of Public Health and Environmental Protection (RIVM).

         The efforts of all who helped in the preparation and finalization
    of the monograph are gratefully acknowledged.

    ABBREVIATIONS

    EQC       External Quality Control

    HEAL      Human Exposure Assessment Locations

    IQC       Internal Quality Control

    MAD       Maximum Allowable Deviation

    MRBIS     Mean Running Bias Index Score

    MRVIS     Mean Running Variance Index Score

    NEQUAS    National External Quality Assessment Scheme (UK)

    QA        Quality Assurance

    QAP       Quality Assurance Programme

    QC        Quality Control

    QCP       Quality Control Programme

    SOP       Standard Operating Procedure

    INTRODUCTION

         Chemical safety is a world priority. Considerable effort is being
    devoted by governments and industries to ensure that the manufacture
    and use of chemicals will not have an adverse effect on human health
    or the environment. Many governments have introduced laws,
    regulations, and guidelines designed to prevent human health risks and
    environmental degradation.

         Concerns about the potential of chemicals to have adverse effects
    on human health and the natural environment has led the World Health
    Organization (WHO), together with the International Labour
    Organisation (ILO) and the United Nations Environment Programme
    (UNEP), to cooperate actively in the International Programme on
    Chemical Safety (IPCS). The objectives of the IPCS include the
    evaluation of the effects of chemicals on human health and the
    environment and the development of methodology and testing methods in
    order to produce internationally comparable results. The importance of
    the quality of data in achieving these objectives is self-evident. 
    The need for quality of data generated in laboratory and field
    toxicological studies is parallelled by the need for quality of
    analytical data on tissue concentrations, environmental exposure, and
    exposure surveillance and monitoring studies.  It is important to
    realize that chemical risk assessment utilizes not only data from
    studies carried out for regulatory purposes, such as notification, but
    also data from many types of research studies, both pure and applied. 
    In the process of risk assessment, biological dose-effect and
    dose-response data are integrated with analytical data.  Thus, overall
    quality management of data generation and application of quality
    assurance and quality control are crucial.  Data quality is inherent
    in the IPCS evaluations of the risk to human health and the
    environment of chemicals, published as Environmental Health Criteria
    (EHC) monographs.  The relevance of quality assurance and quality
    control for the generation of sound data is a key element in the
    monographs dealing with principles and methods for the evaluation of
    toxicity, e.g., the monograph on  Principles of Toxicokinetic Studies
    (WHO, 1986a).

         The validity and usefulness of the results from experimental
    studies, whether these relate to basic research or tests carried out
    to meet regulatory requirements, are critically dependent on the way
    in which they are designed, managed and performed.  In risk
    assessment, the quantity and quality of data are both important.
    Limited or inadequate data, even where there are no doubts on their
    quality, cannot result in a balanced evaluation, and the overall
    conclusions on risks to human health and the environment are
    inevitably limited.  An extensive data base that is of poor quality
    also gives rise to poor assessments and, probably, erroneous
    conclusions.

         The quality of data is considered to be an objective matter. Data
    should be meaningful and reliable for use in the assessment of the
    safety of chemicals.  Of course, quality of data depends to a large
    extent on the quality of the scientists and other individuals involved
    in the production of these data, but many aspects of quality can be
    verified, measured or assessed objectively, and therefore, also
    improved systematically.

         Promoting quality is a management responsibility.  Quality
    management is a broad approach, using all possible tools to carry this
    responsibility.  Studies are complex activities involving people,
    facilities, equipment, test systems, chemicals and materials, often
    over a considerable period of time.  Therefore activities must be
    carefully coordinated so that specific events occur when scheduled and
    in the way intended.

         In the field of safety testing of chemicals and preparations,
    quality management was introduced in the 1970s after the discovery of
    some cases of fraudulent tests and increasing concern over the
    careless way in which many tests supporting the registration of drugs
    were being carried out.  The United States Food and Drug
    Administration responded to these concerns by developing regulations
    on Good Laboratory Practices (Lepore, 1979; FDA, 1987a,b).  The
    philosophy and many of the requirements were based on the quality
    management systems developed for industrial production, such as that
    of military equipment.  Experience had already been gained in applying
    similar principles, e.g., the application of Good Manufacturing
    Practices to the production of pharmaceutical products.

         The International Organization for Standardization in ISO 8402
    (ISO, 1986a) defines quality as "the totality of features and
    characteristics of a product or service that bear on its ability to
    satisfy stated or implied needs", quality management as "that aspect
    of the overall management function that determines and implements the
    quality policy", and quality policy as the "overall quality intentions
    and direction of an organization as regards quality, as formally
    expressed by top management".  Quality management focuses on the
    organizational process.  It is directly relevant to the conditions
    under which laboratory studies are planned, performed, monitored,
    recorded, reported and archived.  Quality management principles have
    been laid down in national legislation and in documents from
    international organizations.  Major examples are the Principles of
    Good Laboratory Practice produced by the Organisation for Economic
    Co-operation and Development (OECD, 1982), publications of the
    International Standardization Organization, such as ISO 8402 (ISO,
    1986a) and ISO 9000 and 9004 (ISO, 1987a,b) and quality assurance
    principles for analytical laboratories of the Association of Official
    Analytical Chemists (AOAC, 1984).  The philosophy followed in these
    various documents is broadly similar and has served as guidance for
    developing this monograph.

         A prerequisite for producing data of good quality is the
    availability of adequate facilities, equipment, and personnel who are
    both well educated and well trained.  Studies need to be adequately
    designed and planned.  Routine laboratory procedures need to be well
    defined, so that they can be carried out in the best possible way and
    in a consistent and reproducible manner.  Other key elements in
    performing and reporting studies are the test, control and reference
    substances, the documentation and record keeping, the final report
    and, lastly, the archiving and retention of data.  This monograph
    provides guidance for these activities.

         An important feature of quality is that it should be verifiable
    and indeed be verified.  For that purpose the quality assurance
    approach has been developed (Burck, 1979).  Quality assurance includes
    independent study monitoring that assures laboratory management and
    users of data that facilities, equipment, personnel, methods,
    practices, records and controls conform to accepted quality management
    principles.  An effective quality assurance system provides confidence
    that a study report meets the pre-established quality standards of
    accuracy, integrity, completeness and clarity.  Quality assurance
    should be integrated within the entire study process to ensure that
    the results are valid and that the final report accurately reflects
    these results.  This monograph therefore deals with the organization
    of a quality assurance programme.

         An important quality assurance tool is known as quality control. 
    Quality control is the quality assessment of quantitative, routine
    laboratory determinations.  Many techniques are used for this purpose. 
    Quality control plays a major role in monitoring studies, but in
    addition, it is frequently applied in toxicity studies, for instance,
    to assess the quality of biochemical and haematological data.

         The application of modern quality management approaches to the
    two fields of safety assessment of chemicals is crucial.  These fields
    are a) quality management of laboratory toxicity studies and b)
    studies to monitor the extent of exposure and effects and the presence
    of chemicals in man and the environment.

         Quality has its price.  Considerable investment, especially in
    terms of human resources, has to be made in a laboratory before a
    quality management system is operational.  Furthermore, it requires
    continued effort to keep the system functioning.  On the other hand,
    quality assurance is a major help for management in organizing the
    work, planning it well, ensuring continuity and increasing
    productivity.  In addition, an efficient quality assurance system
    reduces considerably the risk that erroneous data will lead to
    non-acceptance of studies and rejection of reports or the need to
    repeat studies.

         The following guidance is not directly based on legal precepts
    but follows the same philosophy, and it describes quality assurance
    considerations involved in the production of reproducible and reliable
    test data.  It addresses basic elements of a quality management
    programme for chemical safety testing with the aim of promoting the
    quality assurance concept in test facilities throughout the world and
    of facilitating the global sharing of useful information on chemicals. 
    It is not intended to deal with the details of quality management of
    toxicity and ecotoxicity testing and chemical analysis, but chapter 2
    describes quality assurance as applied to biological testing in some
    detail because this is an area that is not familiar to all scientists
    working in this field.  Even though certain countries may not have
    extensive testing facilities, the national and international
    acceptability of their research data can be greatly enhanced by
    utilizing these basic quality assurance and management principles.  In
    addition they can apply the principles of quality assurance and
    quality control to assure the quality and acceptability of safety data
    presented to them for risk assessment purposes.

    1.  GENERAL QUALITY MANAGEMENT APPROACH FOR QUALITY ASSURANCE

    1.1  Organization and personnel

    1.1.1  Introduction

         To ensure that studies are valid and properly conducted, an
    organizational structure should be established that will best serve
    the needs of the testing facility and will employ an adequate number
    of personnel that are well qualified by education, training and/or
    experience to perform their assigned tasks.

    1.1.2  Organization

         Regardless of the specific organizational structure, facility
    personnel include test facility management, study directors, quality
    assurance programme personnel and support personnel. The
    responsibilities for successfully running a laboratory are shared
    among these groups.

         The typical composition and responsibilities of each of these
    groups are discussed below.

    1.1.3  Test facility management

         In general, duties that are more administrative than scientific
    are the direct responsibility of management.  The primary duty of
    management is to ensure that all studies are properly planned,
    conducted and reported, and that sufficient staff and resources are
    made available for the successful conduct of studies.  Before any
    studies are started, management must make various administrative
    decisions concerning the nature of the studies that will be conducted. 
    Management must decide which controls will be in place to ensure the
    successful conduct and completion of studies, and decide the
    organizational coordination within the test facility to carry out the
    studies.  It must establish a quality assurance programme that assures
    compliance with the study plan test facility procedures, and, where
    necessary, government regulations.

         Properly trained personnel are crucial to the conduct of a
    quality study.  It is management's responsibility to see that staff,
    both professional and technical, receive adequate training to ensure
    that they are competent to carry out their duties. Management must
    make available to all personnel the opportunity for external and
    on-the-job training to maintain and improve necessary laboratory
    skills.

         Management must implement organizational and personnel directing
    and coordinating mechanisms that will adequately serve the facility. 
    It must also be certain that an adequate number of qualified employees

    are available who clearly understand the functions they are to
    perform, as well as any health and safety precautions that need to be
    taken during the conduct of a study.

         In addition, management must apply proper managerial skills in
    dealing with the staff in order to promote a high level of job
    performance.  Positive feedback to the personnel for recognition of
    good work performance should be practised continually by management. 
    All personnel should view their work as a positive contribution to a
    high quality end product, the final study report.

         The most fundamental decisions of management concern the types
    and numbers of studies that will be conducted at the testing facility. 
    Taking into account the experience of personnel both within the test
    facility and outside, such as consultants, management must take
    decisions on a number of aspects, e.g., what type of test systems will
    be maintained, whether studies will be of a short-term or long-term
    nature or a combination of both, whether other external test
    facilities will be required to conduct certain portions of a study,
    and how many studies can be properly handled considering the
    personnel, equipment and facilities available in the test facility.

         Management must establish procedural controls that will assure
    that all study activities will be properly conducted.  Typically,
    these include the requirement that written Standard Operating
    Procedures (SOPs) exist for all routine functions conducted within the
    test facility or by other external facilities, and that the test
    facility management is kept informed of study progress or set-backs.

         The designation of job responsibilities and the assignment of
    these responsibilities to the various entities within the organization
    should be documented.  Management must appoint a study director to be
    responsible for overall study management and the single point of study
    control.  A qualified person must be designated to assume the study
    director's responsibilities on a temporary basis in his absence.  If
    a study director needs to be replaced on a permanent basis, this
    should be done promptly.

         Management delegates much of its responsibilities and authorities
    to the appropriate divisions and personnel within the organization. 
    Each division must be held accountable to laboratory management to
    accomplish its assigned tasks.  This requires open communication
    between the various organizational personnel and laboratory
    management.

    1.1.4  Study director

         Experience has shown that unless responsibility for the proper
    conduct of a study is assigned to one person, there is a potential for
    personnel to receive conflicting instructions, which can result in a
    poor study.  Therefore, it is recommended practice, before a study is

    initiated, for management to designate one individual who will serve
    as study director and be considered the chief scientist in charge of
    a study.  It is essential for the study director to have a strong
    scientific background, as well as proven managerial abilities,
    strengths in communication and problem solving, and the ability to
    organize the day-to-day and long-term objectives of a study.

         Specific responsibilities of a study director include assuring
    that:

    *    the study plan is agreed;

    *    the procedures specified in a study plan are followed by  
         personnel engaged in the conduct of that study;

    *    all revisions to a study plan are brought to the attention of  
         appropriate personnel;

    *    test systems are appropriate for the study;

    *    personnel involved in the study clearly understand their  
         functions and are qualified to perform them;

    *    data are accurately and promptly recorded and verified;

    *    health hazards associated with the test system are recognized and
         controlled, including hazards to personnel as well as to the
         integrity and quality of the test system;

    *    studies are conducted in a manner that is safe for personnel;

    *    all raw data, samples and specimens are archived promptly at  
         the completion of the study, as appropriate;

    *    the reported results of the study accurately reflect the raw  
         data;

    *    unforeseen circumstances that may affect the quality and  
         integrity of a study are noted, corrected and documented.

         The study director must use the advice, education, experience and
    assistance of other scientists participating in the study.

         Concerning the study director's workload, both the types and
    lengths of the studies have a significant impact on the number of
    studies for which a single study director may be responsible.  The
    maximum number of studies assigned to one study director should be
    very carefully assessed by management.

    1.1.5  Support personnel

         A variety of specialists is required to conduct a study
    adequately and to provide the administrative support to ensure its
    proper conduct.  In addition to clerical and administrative support to
    staff, each study requires the participation of such specialists as
    toxicologists, biochemists, clinical pathologists, veterinarians,
    chemists, histologists, statisticians, equipment and maintenance
    specialists, computer specialists, and animal caretakers.  The
    personnel needs of each facility will be guided by the particular
    studies to be conducted.

    1.1.6  Quality assurance function

         In many laboratories the quality assurance function is run by
    quality assurance programme (QAP) staff carrying out quality assurance
    duties on a full-time basis.  The primary functions of quality
    assurance personnel are to conduct facility inspections and data
    audits.  Through inspections they ensure, in general, that studies are
    conducted in accordance with the quality assurance principles; and in
    particular, that the study plans and SOPs are followed.  Through their
    data audit function they assure that the final reports of studies are
    verifiable from the raw data, and that these data were collected,
    recorded and maintained according the quality assurance principles.

         QAP staff or those exercising a quality assurance function should
    have a scientific or technical background.  They should be well versed
    in:

    *    basic management skills;

    *    communication skills;

    *    understanding of intra-organizational relationships and  
         functions;

    *    good record keeping;

    *    negotiation skills.

         Generally, quality assurance personnel are drawn from experienced
    study personnel and are qualified for their quality assurance
    responsibilities through formal training courses and on-the-job
    training exercises.  It is often helpful for the test facility to
    structure the on-the-job training so that personnel are certified
    proficient as they successfully accomplish each duty listed in the job
    description.  For example, when laboratory staff demonstrate that they
    can review a study plan for adherence to acceptable laboratory
    practices, this would be indicated in their personal records.

    1.1.7  Personnel selection and development

         All personnel engaged in the conduct of a study should have
    sufficient education, experience and training to perform their
    assigned duties.  The careful and systematic recruitment and
    employment of laboratory personnel should be considered an important
    management function.  The job requirements should be delineated in
    written job descriptions established for each position available at
    the laboratory.  Job descriptions should specify the duties and
    responsibilities, the reporting relationships (i.e. to whom the
    employee should routinely report the study status and unexpected
    emergencies) and the reporting requirements expected from the
    employee.

    1.1.8  Orientation and training of new personnel

         The need for adherence to the principles of quality assurance and
    to safety procedures, personal sanitation, and clothing and health
    restrictions should be discussed with each new employee.  Copies of
    the SOPs defining the laboratory's personnel health and safety
    regulations and copies of relevant SOPs for their areas of
    responsibility should also be provided to employees.  Before beginning
    any work, employees should be instructed by their immediate supervisor
    on their particular job duties, responsibilities and working
    conditions.  Employees should be required to review SOPs and quality
    assurance manuals pertinent to their specific jobs.

         This orientation process is followed by a period of on-the-job
    training under careful and direct supervision.  The amount of training
    required for each function will be determined by the nature of the job
    function.  When a supervisor believes that an employee can
    proficiently perform a specific job function, the employee should be
    authorized to perform such functions independently.  Documentation of
    this proficiency and authorization should be maintained by the
    appropriate department (e.g., personnel department and/or division
    where the employee is directly employed).  Supervisors should hold
    additional training sessions as often as needed to emphasize a
    procedure or explain a change in a procedure.  These sessions should
    also be appropriately documented.

         The laboratory should have documentation that explains how
    employees should be oriented and trained for their new positions, as
    well as how their training will be documented.

    1.2  Quality assurance programme

    1.2.1  Introduction

         All testing facilities (e.g., laboratories, field operations),
    regardless of size, that generate data for assessing the impact of
    chemicals (e.g., industrial chemicals, pesticides or pharmaceuticals)

    on human health or the environment should have an efficient management
    system.  The establishment of an independent quality assurance
    programme (QAP) is an essential mechanism for accomplishing these
    goals.

         A QAP does not necessarily require a separate organizational
    entity consisting of personnel permanently assigned to this task. 
    However, it is critical that an organizational separation exist
    between quality assurance inspection and study personnel if a
    laboratory or field operation is to expect an impartial analysis of
    the accomplishments and operations encountered in the conduct of a
    study.

    1.2.2  Quality assurance and quality control

         It is important, at the outset, to distinguish between the
    related concepts of quality assurance and quality control.  Quality
    control is a valuable quality assurance tool.  Assessing study quality
    periodically is an essential aspect of quality assurance.  Thus, the
    quality of an analytical measurement may be validated, for example, by
    comparing analytical results against a known standard taking into
    consideration the sensitivity, accuracy, precision, calibration and
    maintenance of the analytical equipment.  These measurements would be
    part of a quality control system.

    1.2.3  Organization and personnel

         Management assures adherence to the principles of good laboratory
    management practices, authorizes study plans and standard operating
    procedures, and corrects operational defects that are reported by the
    quality assurance programme staff.  A quality assurance programme must
    be established by management so that impartial inspections can be
    conducted and observed defects can be corrected promptly.  Most
    laboratories and field operations generating data for regulatory
    purposes have a quality assurance programme unit as a distinct
    organizational entity reporting to management.  If scientists carrying
    out testing are also to conduct quality assurance inspections and
    audits, they should have no scientific or personal involvement in the
    studies being quality assured and must be impartial.

         It is important for a quality assurance programme to be
    implemented by highly motivated, qualified, and trained individuals
    who possess sufficient knowledge of experimental procedures to permit
    an adequate impartial assessment.

    1.2.4  Inspections and audits

         General inspection of facilities and critical activities and
    auditing final reports are very important tasks in a quality assurance
    programme.  The purpose of inspection is to verify that the study is
    being conducted in accordance with the study plan, the SOPs (see

    chapter 4) and the applicable principles of good laboratory management
    practices.  The goal is to detect and correct systematic or
    unintentional flaws in the study process before the flaws have an
    adverse impact on the quality of the study.  It is very important to
    assure that a study is conducted as directed by the study plan. 
    Auditing has two purposes.  The first is to confirm that the results
    presented in the final report accurately reflect the raw data that
    were collected during the study and the second is to highlight any
    circumstances that would adversely affect the study.  SOPs should be
    developed to cover procedures of study conduct and for the audit of
    the final report.  The QAP must compare these audit findings against
    the study plan and the supporting raw data to ensure that the report
    accurately reflects the results of the study in accordance with the
    study plan.  The SOP should also indicate which raw data are to be
    reviewed during the audit and how they are to be reviewed.

         Prior to initiating an inspection of the facilities and
    equipment, the designated quality assurance person should be familiar
    with the experimental operation to be inspected.  This requires a
    review of the study plan and the SOPs that relate to the laboratory or
    field operations.  Personnel files and equipment logs should be
    reviewed as well as the report of the previous inspection of the same
    operation.  These inspections should be conducted in such a manner as
    to minimize any disruption of normal operations.  The quality
    assurance person should explain the purpose of the inspection to
    laboratory or field staff.  Any deviations should immediately be made
    known to the responsible personnel.  The kinds of items to be examined
    include reagent preparation and labelling, personnel files, presence
    of authorized SOPs, equipment logs, recorded data entries,
    environmental controls, storage for specimens and test substances, and
    general cleanliness and orderliness of the laboratory or field
    operation areas.

         Quality assurance staff may interview laboratory or field
    personnel whenever information on a study-related process is needed. 
    Most importantly, quality assurance staff must witness the actual
    operations to assure that they are being conducted in accordance with
    the study plan and SOPs.  It is not possible to estimate how much time
    should be spent in observing each operation, because this depends on
    the complexity of the operation.  In any event, enough of the
    operation should be observed to conclude that the operations are in
    control and are being conducted in accordance with the SOPs.

         Large amounts of data are often collected during studies, which
    are summarized in the final report.  It is important, therefore, to
    assure that the summarized data in the final reports are accurate and
    complete.  When large amounts of raw data are to be audited, several
    approaches can be used to select data points to evaluate.  One
    approach is a pragmatic sampling plan which is based on the experience
    of the quality assurance staff with certain operational units.  For
    example, activities associated with receipt and handling of the test

    or reference substances generally result in very few recording errors,
    and it may be necessary to sample only 10-20% of the recorded
    information.  In contrast, the recording of test system observation
    may frequently be a problem; 75-100% of these observations should be
    audited and the activity inspected on a regular basis.  Another
    approach used is a random sampling plan.  These plans are developed
    with the aid of statisticians and they provide a statistical
    confidence level for verifying the data in the final report.  In
    accordance with these plans, a certain fraction of the data is
    examined, and if this fraction is found to be accurate, it is assumed
    that the whole report is accurate.  If the fraction examined is shown
    to contain errors, then a larger number of data entries is examined,
    and so on.  One should take into account, however, that errors may not
    occur at random and that statistical sampling plans may give
    misleading results.  Auditing final reports occurs prior to the
    archiving of the study records.  The quality assurance auditor
    assembles the raw data and all other study records for the study to be
    audited.  All of these records are read for completeness and those
    that contain raw data are checked against the data in the final report
    in accordance with a random sampling plan.  It is important to locate
    missing records and it is equally important to confirm the presence of
    all specimens and reserve samples of test and reference substances. 
    All errors in the final report should be brought to the attention of
    the study director for resolution.

    1.2.5  Records and reports

         As with other activities in the laboratory or field operations,
    it is important that timely and accurate reports on quality assurance
    inspections are made and careful records kept.

         There are three kinds of reports:

         a)   A routine, periodic report to management and to the study
              directors of all quality assurance work that was conducted
              during a defined period (e.g., one month).  This report
              should give an overall summary of the inspections and audits
              that were conducted.  It should highlight emerging problem
              areas, if any, and should forecast projected quality
              assurance activities.

         b)   A report made to the study director and to management
              whenever significant problems are found that can have an
              adverse effect on a study's quality.  The report should
              describe completely the observed deficiencies.

         c)   The report that accompanies the final study report.  This
              report lists the dates and phases of inspection conducted in
              the course of the study, and the dates when the inspectional
              findings were reported to management.

         It is necessary for a quality assurance programme to keep copies
    of the following records: the master schedule sheet of all studies
    being conducted by the facility; study plans; periodic reports to
    management; inspection and audit reports; an index of quality
    assurance records and a compendium of all the test facility SOPs; and
    quality assurance programme SOPs.

    1.2.6  Quality assurance standard operating procedures (SOPs)

         In order to accomplish its purpose, a quality assurance programme
    must continually review study records for ongoing and completed
    studies, and inspect ongoing studies and all related facilities and
    equipment.  This requires a number of SOPs specifically for quality
    assurance in order to describe the methods for accomplishing these
    tasks.

         The content of the SOPs will vary depending on the organization,
    but generally the following areas should be covered: structure and
    organization of the quality assurance programme; to whom and how the
    quality assurance staff report findings; procedures for study
    inspection; procedures for auditing the final report; scheduling of
    inspection and audits; and procedures for indexing and maintaining
    records.

         An SOP describes the organization of the quality assurance
    programme and how it relates to the organizational structure at the
    testing facility.  This is important for a number of reasons.  It
    should include the following issues.  Firstly, a quality assurance
    programme must provide reports directly to management to be effective. 
    The SOP should describe the reporting procedure.  It should also
    address the internal organization of a quality assurance programme. 
    Secondly, a quality assurance programme must be composed of an
    adequate number of trained individuals to carry out its
    responsibilities, and the required qualifications for each position
    must be known.  The SOP should include procedures for training
    personnel so that the personnel meet the stated qualifications.

         The SOP needs to include how often reports are made and the
    format in which the reports are to be prepared.  Internal reporting
    within the quality assurance programme should be in writing.  There
    are two important aspects of reporting that should be addressed in the
    SOP: the manner in which reported deficiencies are corrected, and the
    manner in which any corrections are to be documented.

         The SOP covers procedures for the in-process inspection of
    critical laboratory operations and their scheduling.  It is necessary
    for the quality assurance personnel to visit the various areas
    regularly and to observe ongoing study-related activities to assure
    compliance with the study plan and relevant SOPs.

    1.3  Facilities and equipment

    1.3.1  Introduction

         A test site must have suitable facilities and equipment to ensure
    the proper conduct of studies.  These must be of adequate design,
    construction, capacity and location for their intended purpose.  In
    addition, they must be properly used, maintained, and cleaned. 
    Equipment used for generation, measurement or assessment of data must
    be calibrated and/or standardized with maintenance of appropriate
    records.

    1.3.2  Facilities for handling test, control and reference substances

         The facilities must allow separation of areas involved in the
    storage, handling and distribution of test, control and reference
    substances.  This is necessary to prevent contamination of facilities,
    equipment, personnel and test systems, as well as to prevent confusion
    of substances.  To accomplish this, there should be separate areas
    for:

    a)   receipt and storage of test, control and reference substances;

    b)   mixing of test, control and reference substances with a carrier 
         (e.g., feed);

    c)   storage of the test, control and reference substance mixtures.

         In addition, storage areas for the test, control and reference
    substances should be separate from areas housing the test systems. 
    The storage facilities should be adequate to preserve the stability of
    the test, control and reference substances.  For example, adequate
    refrigerator or freezer storage space should be available for test,
    control and reference substances requiring low temperature storage.

         There needs to be separate space for the performance of the
    routine and specialized procedures required by study plans.  This
    would include specialized areas for performing activities such as
    necropsy, histology, X-rays, and analytical and clinical chemistry.

         The extent of space needed and the degree of separation required
    will vary widely depending upon the nature and amount of work being
    conducted.  Larger laboratories may require a complete array of
    facilities capable of carrying out every phase of a study, whereas
    smaller laboratories may hire contract testing laboratories to perform
    some phases of the study that they do not routinely perform, such as
    histopathological slide preparation, clinical chemistry tests, or
    analytical chemistry for test substance characterization.  In either
    case, laboratory operations must be performed under conditions that
    provide an adequate degree of separation to preclude confusion and
    interference between areas performing different aspects of the study.

    1.3.3  Field study facilities

         Adequate facilities should be available for the receipt, handling
    and storage of chemicals, such as herbicides, pesticides, and
    fertilizers, necessary for the conduct of good agricultural practices. 
    These areas should be separated from the areas containing the test,
    control and reference substances.  Adequate space should also be
    available for processing soil and plant residue samples and there
    should be adequate refrigerator and freezer space for interim storage
    before shipment to analytical residue laboratories.

    1.3.4  Equipment

         Along with adequate buildings, a test facility or field operation
    must have equipment of appropriate design and adequate capacity to
    function according to the study plan.  This applies especially to
    equipment used in the generation, application, measurement or
    assessment of data.  It also includes equipment used for laboratory
    environmental control, such as refrigerators, freezers, and air
    conditioning for test system areas and test substance preparation and
    storage areas.  All equipment should be suitably located for ease of
    operation, inspection, cleaning, maintenance and safety.

         Proper maintenance and calibration of equipment is a fundamental
    quality assurance management practice.  Equipment should be inspected,
    cleaned and maintained.  In addition, equipment used for the
    generation, application, measurement or assessment of data should be
    tested, calibrated and/or standardized.

         A testing facility should have SOPs that describe in sufficient
    detail the methods, materials and schedules to be used in the routine
    operation, inspection, cleaning, maintenance, testing, calibration
    and/or standardization of equipment, and there should be appropriate
    documentation.  These procedures should specify remedial action to be
    taken in the event of failure or malfunction and should designate the
    person responsible for performing each operation.  Copies of these
    SOPs should be readily available to personnel in the vicinity of the
    equipment.  In order to assure that all equipment associated with a
    study was operating properly at the time of its use, SOPs must be
    followed and records maintained for all equipment inspection, testing,
    calibrating, standardizing and maintenance.  The records should
    contain the date the operation was performed and, in the case of
    maintenance operation or repairs, the type of function or malfunction,
    how and when it was observed, and, as appropriate, any remedial action
    taken.

    1.4  Study plan

    1.4.1  Introduction

         A clearly written, comprehensive study plan is an essential
    element of all chemical safety studies.  A study plan should state the
    objectives, schedules and all methods for the conduct of a study.  The
    plan, in combination with the SOPs, should provide the complete
    specifications and instructions for the execution of the study.

         There is a practical difference between the role of the SOPs and
    the role of the study plan.  The SOPs describe how specific routine
    operations are to be conducted.  The study plan contains the design
    specifications for a single study or a group of related studies.  Its
    role is to describe what the study objectives are and what methods are
    to be employed to achieve the stated objectives.  In short, the plan
    describes what is to be done and when, while the SOPs address how
    specific study events are to be accomplished.

         As the design specification for a study, the plan has an
    important quality assurance function: it serves as the reference for
    measuring study performance.  If the plan of a well-designed study is
    followed, the study objectives should be achieved.  However, if the
    plan is inadequate or is not followed, the intended objectives of the
    study may not be met.  In this case, the study plan cannot serve as a
    suitable reference against which the quality of the study can be
    assessed.

         The study plan is a useful management tool; it helps the long-
    term planning of, for instance, workload, manpower implications and
    the necessary facilities and equipment.

         Test facilities should have SOPs covering the preparation and
    approval of study plans to help assure an orderly, complete and
    consistent review process.  This is especially important in the
    preparation of the study plan, because errors or omissions can
    jeopardize the entire study.

         Generally, SOPs for study plans should address the persons who
    will be responsible for preparation and review, how the preparation
    process begins and ends, what elements of format and content should be
    included in the plan and to whom it is to be distributed.

    1.4.2  Study plan preparation

         The study director has the primary responsibility for preparing
    the study plan and is generally responsible for designing the plan
    scientifically to assure that the study will achieve its intended
    purpose.  Senior scientists who will participate in various aspects of
    the study should be involved in the preparation and/or review of the
    plan.  This may include writing portions of the plan that involve heir

    area of expertise.  For example, the pathologist for the study may
    prepare the pathology section of the plan.  Involvement of other
    scientists may include reviewing what the study director has prepared,
    correcting errors and omissions, and commenting on its adequacy. 
    Technical personnel should be involved in plan preparation to ensure
    that the functions specified in the plan can be performed.

         Management is also involved in the preparation and review of the
    study plan since it is its responsibility to assure that personnel,
    resources, facilities, equipment, materials and methodologies are
    available as scheduled in the plan.  It is also useful for the study
    plan to be reviewed from the quality assurance perspective to assure
    that it contains the basic elements required in the applicable SOP.

         For studies conducted under contract, the sponsor of a study
    should also be involved in the preparation and review of the study
    plan and must approve it prior to the initiation of the study.  This
    is necessary to permit a clear understanding between the study
    director and the sponsor of the requirements and objectives of the
    study.

    1.4.3  Format and content of study plan

         The specific contents of each study plan will vary according to
    the objectives of a given study and the particular methods and
    procedures that will be employed.  However, there are basic elements
    common to most chemical safety testing studies that should be included
    in the study plan:

    a)   A descriptive title and a statement of the purpose of the study. 
         Stating the title and purpose of the study in the study plan
         clearly establishes the study objectives, and the study
         performance can later be assessed to ensure that the stated
         objectives were achieved.

    b)   The identification of the test, reference and control substances
         by name, chemical abstract number or code number.  If code
         numbers are used to identify the test substance, such numbers
         should be unique to this substance.

    c)   The test system, i.e. any human, animal, plant or microbial
         system, as well as other cellular, subcellular, chemical or
         physical systems, or a combination thereof, which is used for the
         test or control substance.  The selection of the appropriate test
         system is an important feature of study design, and should be
         described fully.  The failure to do so could result in the
         inadvertent use of a test system other than that intended by the
         study director or the sponsor.

    d)   The specific procedure to be used for the identification of
         components of the test system.  This most commonly refers to the
         method of identifying test animals within their respective test
         groups.  Proper identification of the test system is necessary to
         prevent confusion of animals, samples and/or specimens.

    e)   A description of the experimental design, including the methods
         for the control of bias.

    f)   The type and frequency of tests, analyses, and measurements to be
         made.  The plan should specify which parameters are to be
         measured and how often such determinations are to be made.

    g)   The records to be maintained.  Because written records and
         electronically captured data are the tangible results of any
         study, the study plan should identify which records are to be
         kept.  Many test facilities have developed SOPs that describe the
         identity, use, and retention of study-related records.  In
         identifying records to be maintained, reference to the SOPs is
         appropriate.

    h)   A statement of the proposed plan for choosing the appropriate
         methods for statistical analyses of the data set.  Although the
         study plan may assume that a particular test is applicable, a 
         statistical test may show that it is not appropriate and that
         another test (e.g. nonparametric) must be chosen.

         The above elements are not the only items that can be addressed
    in the study plan.  For example, in the case where a study is modelled
    on an established procedure published in the scientific literature,
    the study plan may include literature references.  The study plan may
    also include such items as the proposed starting and completion dates
    of the study and the name and address of the testing facility.

         The study plan should be written in clear, concise language with
    adequate reference to the SOPs involved in the application of the test
    methods described.  The pages should be numbered consecutively and the
    total number of pages should also be given.  The body of the plan may
    be divided into clearly identified sections for ease of reference.

    1.4.4  Use of study plan

         A study plan must be used properly to achieve valid study
    results.  Adherence to the study plan requirements, documenting
    adherence, and controlling and documenting any deviations or changes
    to the study should be monitored by the study director.

         Although all personnel associated with the study refer to the
    study plan, the individuals most involved on a daily basis are the
    personnel performing study operations and the study director. 
    Personnel use the plan to guide the performance of their duties.  The

    study director uses it as a means of informing study personnel of the
    study requirements.  This individual has the overall responsibility
    for assuring that the personnel are aware of the current study plan
    requirements and that the study is carried out as directed by the
    study plan.

         Once the study has started, a quality assurance programme also
    uses the study plan, e.g., to determine that no deviations from
    approved plans or SOPs are made without proper authorization and to
    audit the final report.  To accomplish this, a quality assurance
    programme must maintain a copy of all study plans and current
    amendments.  Quality assurance staff use these to compare actual
    practices against those specified in the plan for a given study.  The
    plan often contains a schedule of study events that is useful to a
    quality assurance programme in scheduling its inspection of a study. 
    In this regard, the quality assurance unit must verify that raw data
    are available to document that the approved study plan was followed.

         To be used effectively, the study plan should be available,
    familiar to personnel and in a convenient format.  However, the
    copying and distribution of the study plan should be controlled to
    prevent unauthorized copying and distribution.  A copy of the study
    plan should be located in the immediate area of the test facility
    where a study event is being performed.

         A final requirement for the effective use of the study plan is
    that it must be current.  Although a well-designed and well-prepared
    plan should require little change after the study has begun, changes
    will almost always be required as the study progresses.  The reasons
    for the planned and unplanned changes are many.  For example, after
    the study has begun, an unusual or unexpected toxic effect may be
    observed that requires additional tests or changes in the frequency of
    observations.  The study plan should be updated to reflect these
    changes in the form of appendices.  The study director must approve,
    by means of a dated signature, the change or revision and document the
    reason for the change.  In addition to dating, it is also desirable to
    number the amendments sequentially so that the most recent amendments
    can be quickly identified and study personnel can be sure they have
    received all amendments.  Also, if study plan amendments are
    themselves revised in subsequent amendments, the current amendment
    should reference the superseded amendment.

         Complete records are important for documenting adherence to the
    study plan.  These include proper recording of study data and of
    records related to study performance.  When the plan directs the
    performance of a study event, records should be kept to document that
    the study plan was followed.

    1.5  Standard operating procedures

    1.5.1  Introduction

         Quality assurance involves the development and use of standard
    operating procedures (SOPs).  SOPs are written procedures which
    describe how to perform certain routine laboratory tests or
    activities, normally not specified in detail in study plans or test
    guidelines.  SOPs set forth in detail how specific routine laboratory
    operations are to be carried out and complement the study plan by
    describing all routine study methods.  SOPs are not described in
    detail in this monograph.  There are no universal SOPs, although those
    detailing the same operation will have many close similarities from
    laboratory to laboratory.  The act of developing SOPs, taking account
    of the operation and how its conduct is directly influenced by
    laboratory organization and structure, is a key function of laboratory
    management.

         The importance of effective SOPs in conducting a study must be
    emphasized.  SOPs are management directives designed to ensure that
    all personnel associated with a laboratory study will be familiar with
    and use the same procedures.  If different individuals perform
    important study functions, such as dosing animals, preparing
    solutions, archiving documents or receiving test articles, these
    operations should be performed in the same manner.  By standardizing
    procedures for the conduct of studies, SOPs have a valuable quality
    assurance function.  They prevent the introduction of possible errors
    in the generation, collection and reporting of data.

         The development of SOPs includes the following aspects: who
    should prepare, review and authorize SOPs; which laboratory operations
    require SOPs; and what information the SOPs should contain.  The
    nature of the laboratory work being done and the training and
    experience of the laboratory personnel at a particular facility will
    determine exactly how extensive the content needs to be.

    1.5.2  Format and content of standard operating procedures

         SOPs should be easy to use.  They should be identified by a
    descriptive title and a number.  The individual pages of an SOP should
    be numbered, and each page should note the total number of pages
    contained in each individual procedure.  This practice is especially
    valuable where SOPs are loosely bound, because it enables individuals
    using the procedure to assure themselves that pages are not missing.

         In addition to the descriptive title an SOP should have a short
    statement of its specific purpose or objective.  This should be
    presented under a separate heading at the very beginning of the SOP. 
    A section listing cross-references to related SOPs or other literature
    is another format element.  For example, an SOP for the operation of
    a piece of laboratory equipment should reference maintenance and use

    of the equipment on the basis of the manufacturer's literature.  An
    SOP for determination of animal body weights or organ weights may
    reference the SOP covering calibration and use of the scales to
    measure the weights.  The appropriate use of cross-references can
    reduce the SOPs volume by eliminating redundancy and can assure that
    users are aware of all other important procedures related to the
    operation that they are performing.

         SOPs should have a section listing all materials and equipment
    required for performance of the indicated operation.  This is
    especially true for highly technical laboratory procedures where the
    unavailability of required equipment or omission of required materials
    may result in the improper performance of the procedure.

         The exact content of SOPs will vary accordingly to the specific
    procedures with which they are concerned.  SOPs concerning the use of
    laboratory equipment should describe the methods, materials and
    schedules to be used for the routine inspection, cleaning,
    maintenance, testing, calibration and/or standardization, and the way
    to document these operations.  They should also specify remedial
    action to be taken in the event of failure or malfunction, and should
    designate who is responsible for the performance of each of the
    operations mentioned above.  Depending upon the complexity of the
    equipment, the SOP should contain sufficient detail to permit trained
    laboratory personnel to operate the instrument.  For example, the use
    of equipment logbooks or forms should be described completely.  SOPs
    should provide procedures for handling emergency situations, such as
    the spillage of chemicals or fires.

    1.5.3  Preparation of standard operating procedures

         Laboratory management is primarily responsible for the
    development of adequate SOPs.

         Scientists, laboratory technicians and quality assurance staff
    play a significant role in the preparation and review of SOPs.  They
    should be written by scientists and laboratory personnel who actually
    perform the studies and should be approved by management.  SOPs that
    are prepared by a single department should be reviewed centrally by
    someone outside that department. This will assure that newly prepared
    SOPs are not in conflict with existing SOPs of other departments. 
    Central review may be performed by a quality assurance programme or by
    representatives from each of the departments in the laboratory (i.e.
    a standard operating procedures review committee).

    1.5.4  Typical standard operating procedures

         It is important to identify which specific routine operations
    require SOPs.  Although specific examples of operations requiring SOPs
    exist, there is no single list that includes all operations that
    require these procedures.  Different laboratories will require the

    existence of different SOPs depending upon such considerations as the
    nature of the work, the kinds of equipment and facilities, and the
    qualifications and training of personnel.  Laboratory management
    should decide which SOPs are required to assure the quality of the
    laboratory's work.  At any rate, there should be SOPs to cover those
    areas and operations that are routinely involved in the conduct of a
    study and that involve or impact on the generation, collection or
    reporting of data.

         To assure proper preparation and maintenance of SOPs, the
    laboratory should have an SOP covering the format, initial
    preparation, review, and approval of SOPs, changes, continuing review,
    update, and distribution.  The SOPs should also identify who is
    responsible for each of the steps in the preparation process.  The SOP
    covering changes to these procedures should identify what types of
    changes will be made (i.e. minor, major, etc.), who will be
    responsible for making changes, and the documentation required to make
    changes.

         In a toxicological laboratory, the following SOPs are typically
    needed:

    *    preparation, review and distribution of SOPs;

    *    receipt, identification, labelling, handling, sampling, usage and 
         storage of test and reference substances;

    *    maintenance, cleaning and calibration of measuring apparatus  
         and environmental control equipment;

    *    preparation of reagents and dosing formulations;

    *    record-keeping, reporting, storage and retrieval of records and 
         reports;

    *    data collection;

    *    preparation and environmental control of areas housing the test
         systems;

    *    receipt, transfer, location, characterization, identification and 
         care of test systems;

    *    handling of the test systems before, during and at the
         termination of the study;

    *    disposal of materials;

    *    use of pest control and cleaning agents;

    *    quality assurance programme operations;

    *    safety and emergency procedures.

    1.5.5  Use and availability of standard operating procedures

         SOPs are used by laboratory management, scientific and technical
    personnel, a quality assurance programme, and other interested parties
    for several different purposes.  Management uses SOPs as a means of
    directing and instructing personnel in procedures that are in
    accordance with the requirements of quality assurance.  Scientific and
    technical personnel use the SOPs directly in the performance of their
    duties and as a means of training new personnel in the conduct of
    study operations.  A quality assurance programme uses SOPs both
    directly and indirectly: directly to carry out its operations and
    indirectly to monitor the operations of every other laboratory
    section.  Part of quality assurance inspections and reviews must
    include a comparison of the procedures that are being used against
    those specified in the SOPs to determine if these are being used and
    are adequate.  Since it is the role of a quality assurance programme
    to assure management that the relevant SOPs exist and are being
    followed, it is desirable for a quality assurance programme to
    maintain a current copy of all of the laboratory's SOPs so they can be
    available for ready reference.  The SOPs may also used by parties
    outside the laboratory, such as inspectors from a regulatory
    authority, and during study audits.

         Every laboratory should have a formal procedure for training
    employees in the use of SOPs, and there should be documentation
    verifying the completion of such training.  Some laboratories
    accomplish this by making the laboratory supervisor responsible for
    providing new personnel with copies of the SOPs and reviewing them
    with the employee.  The employee is then supervised over a period of
    time to ensure that the relevant SOPs are fully understood and the
    procedures are being followed correctly.  This may be documented in
    the employee's training file by a record signed by the laboratory
    supervisor indicating the procedures in which the employee is
    proficient.  Some laboratories also require employees to sign a
    statement that they have received and read the SOPs pertinent to their
    responsibilities.  Of course, employees should be aware of all SOPs
    that may relate to them, not just the ones they use daily in their own
    limited area.  For example, a laboratory animal technician collecting
    specimens of blood for the haematology laboratory needs to be aware of
    the haematology laboratory's SOPs with respect to the labelling,
    storage and processing of specimens.

         Management should identify who is responsible for distributing
    new or revised SOPs and who is responsible for removing obsolete
    procedures from use.  It should also establish a standard distribution
    pattern to assure that all involved personnel and departments receive
    copies of updated SOPs and properly remove obsolete copies.  In

    addition, it should address procedures to prevent unauthorized copying
    of SOPs.  Some laboratories have approached this by the use of an
    official stamp or other mark on each page of the SOPs that will not
    photocopy.  Thus, when there is a quality assurance inspection, any
    copies of the SOPs not bearing this mark will be recognized as
    unauthorized copies and can be removed from use.

         Copies of the SOPs should be located in the immediate area where
    they are to be used, preferably at each work station.  For example,
    SOPs covering the calibration and use of a piece of equipment should
    be located near the equipment, those covering procedures performed in
    animal rooms should be located in the animal rooms, and those covering
    the performance of necropsy or preparation of tissues for
    histopathological evaluation should be located in the laboratory
    performing these operations.

         Copies of all the SOPs should be kept together in a manual, and,
    depending upon the size of the laboratory, there may be one or more
    such manuals.  In laboratories that have SOPs divided into individual
    manuals by department or operation, a general index listing the
    various department's manuals is often useful.  In order for laboratory
    personnel to refer easily to those SOPs that apply to the operations
    for which they are responsible, a table of contents for each
    individual manual is also useful.  In smaller laboratories, where all
    the SOPs are contained in one manual, a single table of contents is
    adequate.

         A complete set of all current SOPs should be kept by a quality
    assurance programme and by management.

    1.5.6  Adequacy of standard operating procedures

         The first evaluation of the adequacy of an SOP is made during
    preparation and review.  Once an SOP has been developed and approved,
    its adequacy still needs to be continually evaluated, because of the
    changing nature of laboratory operations.  The continuous evaluation
    of the adequacy of an SOP is the responsibility of the laboratory
    personnel using it and of quality assurance programme personnel. 
    During its inspections of laboratory operations, the quality assurance
    personnel should examine SOPs and compare them against actual
    operations.  If there are discrepancies, the existing SOP may be
    inadequate and revision warranted.  Also, if there are continuing
    problems in particular operations, this could be an indication of an
    inadequate SOP.  Obsolete SOPs should be kept and not discarded since
    they may be necessary to reconstruct certain aspects of completed
    studies.

    1.5.7  Maintenance of standard operating procedures

         The responsibility for maintaining currently active SOPs is
    shared by laboratory management, technical and scientific personnel, 

    and a quality assurance programme.  Management is responsible for
    reviewing and authorizing new SOPs or significant changes to existing
    ones that are required to keep them current.  Technical and scientific
    personnel who use the SOPs daily are primarily responsible for
    identifying those that have become obsolete.  These individuals should
    identify required changes and seek appropriate action from their
    supervisors to update the SOP.  Laboratory personnel, especially the
    first line supervisors, are also responsible for seeing that current
    SOPs are physically maintained in the laboratory and available for
    use, although a quality assurance programme is primarily responsible
    for verifying that SOPs in use are current, approved, adequate, and
    available.

         The SOPs covering continuing review and update should indicate
    how often SOPs will be reviewed and who is responsible for review and
    update.  There are two approaches to continuing review.  Some
    laboratories have no programme for regular review of SOPs but simply
    rely on the users to initiate updating as needed.  In this case, the
    SOP covering review and update of SOPs would not specify how often
    they are reviewed but should still address how the frequency of review
    and updating is established.  Other facilities have a programme that
    requires a regular scheduled review of all SOPs regardless of apparent
    need.  Either approach is acceptable provided it assures that the
    procedures are maintained up-to-date.

         Maintenance of historical files of SOPs should be a central
    responsibility.  This could be part of a quality assurance programme
    or be located in the unit responsible for maintaining data archives. 
    Alternatively, the various departments could be responsible for their
    respective SOPs.  The historical files of SOPs should include all
    revisions of the SOPs and the dates of the revisions.

         If historical files of SOPs are not maintained by a central unit
    but are instead kept by each laboratory department, the individual
    within each department responsible for maintenance of the historical
    file should be identified.  Also, because there will be a number of
    different individuals involved, there should be a uniform
    laboratory-wide index to facilitate the retrieval of historical SOPs
    from each department by the quality assurance programme or other
    personnel outside of the specific department.  The location(s) of
    historical files should be identified in SOPs covering their
    maintenance.

    1.6  Test, control and reference substances

    1.6.1  Introduction

         The term "test substance" refers to a chemical or a mixture which
    is under investigation for the purpose of evaluating its effects on
    the test system.  A test substance can be any class of product, e.g.,
    drugs, biological products, food additives and pesticides.  Control 

    and reference substances are chemicals that are administered to the
    test system as a positive or negative reference for the purpose of
    establishing a basis for comparison with the test substance.  For
    example, in a mutagenicity assay such as the Ames test, a known
    mutagen is used as a means of comparing the response of the test
    system to the test substance.  The known mutagen is the reference
    substance.  The test, control and reference substances and the test
    system are the major components of a nonclinical safety study.  Unless
    the laboratory has adequate procedures, facilities and equipment for
    proper characterization and handling of these substances and their
    mixtures, a valid safety assessment of the test substance cannot be
    made.

         There should be adequate procedures covering the characterization
    of test, control and reference substances, the handling and
    distribution of test and reference substances, and the preparation and
    handling of mixtures of substances with carriers such as feed. 
    Characterization includes determining the identity, strength, purity
    and composition or any other characteristics that will appropriately
    define the test or reference substances.  Stability determination is
    considered to be a part of the characterization process.  The
    importance of this process in the study is paramount because it is
    essential to know what is being tested before a meaningful evaluation
    can be made of study findings.

         Proper handling of test, control and reference substances is also
    required to preclude the possibility of contamination, deterioration,
    damage or misidentification during the testing process.  Proper
    handling includes adequate study records that document the receipt and
    use of the test and reference substances to establish their
    administration, in known quantities, to the specified test system.

         Laboratories should have the necessary equipment and facilities
    for handling test and reference substances.  They should have separate
    areas for receipt and storage of test and reference substances, mixing
    of test and reference substances with carriers, storage of test and
    reference substance mixtures, and for carrying out experiments, in
    order to prevent contamination and mix-ups.  Laboratories should also
    have necessary equipment to test, prepare, store and administer test
    and reference substances.  Equipment must be maintained, cleaned,
    calibrated and inspected to assure adequate performance.

    1.6.2  Test, control and reference substance characterization

         The identity, strength, purity and composition or other
    characteristics (such as impurities) that will appropriately define
    the test, control or reference substance should be determined and
    documented prior to study initiation.  Methods of synthesis,
    manufacture or derivation of the test and reference substances should
    be documented by the sponsor or the testing facility.  It is necessary
    in any study to characterize fully discrete quantities of these 

    substances.  To do this, specific batches and lots must be obtained
    and used in the test.  The sponsor or the laboratory conducting the
    test must have information such as laboratory analytical reports
    and/or batch records documenting the manufacturing and testing of
    these materials to show that they conform to the specified standards
    of identity, strength and purity.

         The test, control and reference substance containers must carry
    identifying information.  This is facilitated by the use of a batch
    number.  A batch is a specific quantity of substance that has a common
    origin and defined physical/chemical characteristics.  For a study,
    the same batch of test substance should be used for the entire study,
    after being appropriately characterized.  If more than one batch is
    used, each should be fully characterized.  Complete documentation must
    be maintained.  It is not a preferred practice to use different
    batches of the same substance without prior characterization.

         When the test, control and reference substances are received for
    the test, the name of the substances and the lot number must be
    recorded on all study records relating to any analysis of the
    substances.  For example, chromatographs from analysis of the test
    substance for purity or identity should include the chemical or code
    number identifying the compound analysed and the specific batch or lot
    number.

         When marketed products are used as reference substances, full
    characterization testing is not usually necessary because these
    products can be characterized by their labelling.  However, the
    laboratory should have procedures for ensuring adequate label review
    and procedures for accepting shipments of reference substances.  Also,
    label information such as lot numbers, expiration dates, and the name
    of the manufacturer should be documented in the study records so that
    the substances used in the study can be traced.  Again, one batch
    should be used for the entire study, if possible.

         The stability of each test, control or reference substance should
    be determined by the testing facility or by the sponsor before
    initiation of a study.  It is necessary to know the stability of the
    test and reference substances to ensure that the test system is
    exposed to substances of adequate potency and concentration over the
    full course of the study.

         Although laboratory management should assure that the stability
    has been adequately determined, the sponsor may assume this
    responsibility under the same conditions as specified above for
    characterization of the substances; namely, this assumption of
    responsibility should be described in the protocol and/or contractual
    agreement.

         If the stability of the test, control and reference substances
    cannot be determined before initiation of a study, schedules should be

    established and followed to provide for periodic re-analysis of each
    batch.  For test substances whose stability cannot be determined prior
    to study, SOPs or the study plan should specify procedures and
    schedules for sampling and testing the substance for stability.  The
    study plan can be amended as the results of the stability test become
    known.

         There must be ample storage facilities for any stability samples
    retained.  The storage facilities must provide the same environmental
    conditions as those facilities used to store the test and reference
    substances during the study.  The storage conditions, i.e.
    temperature, humidity and other conditions as required, must be
    documented.

         Each storage container for a test, control or reference substance
    must be labelled by name, chemical abstract number or code number,
    batch number and expiry date (if any).  Where appropriate, storage
    conditions necessary to maintain the identity, strength, purity and
    composition as well as safety information, if available, of the test
    or reference substances must be indicated.  If safety information is
    not available, this should also be noted on the label.  The laboratory
    should have procedures for labelling the test substances.  These SOPs
    should require that this information be clearly indicated on the
    label.  They should also address who will be responsible for labelling
    the containers and/or verifying the label of any containers received
    by the laboratory.  The SOPs should also address how labels will be
    applied.  It is generally unacceptable to place the only identifying
    label on the lid of the container.  This practice may lead to
    confusion between containers when lids are removed during use.  SOPs
    regarding labelling may include procedures for assuring that labels
    will remain affixed and legible during conditions of storage and use. 
    For example, test substance containers may be exposed to moisture or
    solvents that could alter the labelling.

         Storage containers should be assigned to a particular test
    substance for the duration of the study.  This is necessary to
    preclude the possibility of cross-contamination, which may occur when
    the same container is used subsequently to store a different test
    substance.  It is good practice to dispose of empty test substance
    containers after a study has been completed rather than attempt to
    decontaminate and reuse them.

         Retention of samples of test, control and reference substances is
    useful should a question arise regarding the quality or identity of
    the test substances actually administered to the animals.

    1.6.3  Handling

         A laboratory should have adequate procedures for a system of
    handling the test, control and reference substances to ensure that
    there is proper storage, that distribution is made in a manner

    designed to preclude the possibility of contamination, deterioration
    or damage, that proper identification is maintained throughout the
    distribution process, and that receipt and distribution of each batch
    is documented.  These procedures should include a description of the
    records to be maintained in order to document that procedures are
    being followed.

    1.6.4  Storage

         Handling procedures should include proper storage.  Storage
    requirements vary with the specific test, control and reference
    substances used.  For example, some test substances may require
    storage at temperatures below 0 C, while others are stable at room
    temperature.  Some may require storage in the dark or in special
    containers and cabinets due to high volatility.  In all cases, the
    laboratory must have procedures that will require each test or
    reference substance to be evaluated upon receipt for special storage
    and handling conditions, and to assure that these conditions are met. 
    For example, there may be a standard form that is filled out upon
    receipt of a substance that requires the recipient to record any
    special storage conditions required, to ensure that these conditions
    are recorded on the label, and that the substances are stored
    accordingly.  Laboratory procedures for receipt of the test and
    reference substances should also require the person receiving them to
    determine, if possible, the storage conditions of the substances
    during shipment and at the time of receipt.  For example, if the
    substances are to remain frozen until used, they should be inspected
    upon receipt for thawing and this should be documented in the study
    records.  The storage requirements for the test, control and reference
    substances should generally be provided by the sponsor or manufacturer
    responsible for initially providing the substances.

         Procedures for ensuring proper storage need, at a minimum, for
    required storage conditions to be specified in the labelling and study
    records, necessary facilities to be identified and available to meet
    required storage conditions, storage conditions during transport and
    use to be monitored and documented, and deviations from the required
    conditions to be investigated and documented.  Where deviations occur,
    the study director should determine their impact on the study and
    record in the final report any that may seriously affect the study.

    1.6.5  Distribution

         Procedures covering test, control and reference substance
    handling should ensure that distribution is made in a manner designed
    to preclude the possibility of contamination, deterioration or damage. 
    Such procedures might include the types of containers used to
    transport articles to the laboratory areas where they are to be used
    and the designation of specific storage areas.  Laboratories dealing
    with large numbers of test substances should have SOPs identifying
    where the substances are to be stored.  Procedures should also

    identify where and how containers of test and reference substances are
    to be opened for dispensing.  This usually takes place in an area
    specially designed for this purpose with an isolated environment. 
    Many laboratories have a specially designated room or laboratory bench
    top under a laminar air flow to prevent the possibility of
    cross-contamination of equipment and facilities, as well as to protect
    employees from potentially hazardous test substances.  Special gowning
    and decontamination procedures are generally required for the test
    substance dispensing areas and are included in the SOPs for test and
    reference substance handling.  Adherence to these procedures must be
    documented.  Such documentation will include records of cleaning the
    areas between use for different compounds.

         A systematic process for receipt and distribution of test,
    control and reference substances should be fully described in
    laboratory SOPs for test, control and reference substance handling. 
    Ideally, there should be a centralized procedure responsible for
    receiving and logging in test, control and reference substances.  This
    should ensure that proper identification is made initially and then
    maintained throughout the distribution process.

         Documentation should specifically include the date and quantity
    of each batch distributed or returned.  An inventory record is often
    used to record the receipt of all test and reference substances. 
    Entries include the name or identifying number of the substance
    received, the batch or lot number, the date of receipt, the amount
    received, and the signature of the individual who received the
    substance and completed the record.  Many laboratories record the
    weight of the container and its contents as received.  When a portion
    of the test substance is subsequently dispensed, the amount is weighed
    and recorded in the inventory record.  The record should also reflect
    to whom and for what purpose the substance was dispensed.  As an
    additional quality assurance measure, many laboratories not only weigh
    out the amount to be dispensed, but also weigh and record the
    remaining substance in the bulk container.  This provides a continual
    inventory of the test substance and allows detection of any dispensing
    or record-keeping errors that may indicate a potential misdosing
    problem.

         It is clear that there should be procedures covering every aspect
    of the handling of test, control and reference substances to ensure
    proper storage, distribution, identification and accountability. 
    Records should be maintained to document that these procedures were
    actually followed.  All study records pertaining to use, testing and
    distribution of the test or reference substances should be directly
    traceable to the specific lot or batch of materials issued, and
    accountability records should document that the quantities and lot
    number used were consistent with study plan requirements.  Periodic
    checks should be made that the amount of test substance in the
    inventory matches with the amount used in the study.

    1.6.6  Mixtures of substances with vehicles (Carriers)

         A vehicle (or carrier) is the material with which the test,
    control or reference substance may be mixed for administration in the
    test system.  It can be feed, water, solvents and/or excipients,
    depending on the form of dosage and route of administration.  A common
    example of a carrier is the rodent feed used to mix a test substance
    for a chronic feeding study.  However, the phrase "mixture of
    substance and vehicle" also refers to solutions and suspensions, e.g.,
    a solution of test substance in distilled water.

         A laboratory should have procedures for the preparation,
    analysis, storage, labelling, distribution and return or disposal of
    mixtures of substances with vehicles.  The procedures should
    accomplish the same goals of maintaining proper identity and ensuring
    proper storage and use as those procedures already discussed for the
    test and reference substances alone.  In addition to the concerns
    already mentioned, mixtures of substances with vehicles present
    special problems.  For example, there must be procedures to assure
    that the mixtures used are uniformly mixed, stable, and provide the
    proper concentration of the test or reference substance.  To
    accomplish this for each test, control or reference substance mixture,
    appropriate analytical methods should be used to determine the
    uniformity of the mixture and to determine, periodically, the
    concentration of the test or reference substance in the mixture.  The
    standard of uniformity should be determined before the start of the
    study.

         Each time a batch of substance carrier mixture is made, it should
    be recorded that the batch was properly formulated and mixed according
    to procedure.  For liquid preparations, uniformity determinations need
    not be made on true solutions, but should be made for suspensions.  In
    laboratories that perform numerous short-term assays (e.g.
    mutagenicity studies) by using liquid dosage forms, it may not be
    practical to analyse dosing solutions from every assay.  In this case,
    it is imperative for the laboratory to have well-documented dilution
    and dosing procedures and SOPs for determining and documenting
    solubility of the test and reference substances.

    1.6.7  Stability

         In addition to analyses for uniformity and concentration, tests
    should be conducted to determine the stability of the test, control
    and reference substances in the substance carrier mixture.  If the
    stability cannot be determined before the study is initiated, the
    study plan should provide for periodic re-analysis of the test and
    reference substances in mixtures.  The laboratory must determine the
    stability of the mixtures over their periods of use.  This would
    include the period between the day the batch is prepared and the day
    the last portion of the batch is used.  Stability should be determined
    under actual conditions of storage and use.  Once the stability of a

    given concentration of a test substance carrier mixture is
    substantiated for one batch, no further stability testing is necessary
    for each subsequent batch of that concentration.  However, periodic
    re-analysis to determine concentration must be carried out, as
    discussed previously.

         A sound practice related to stability testing is the use of
    expiry dates.  Where any of the components for the test or reference
    substance carrier mixture has an expiry date, that date should be
    clearly shown on the container.  If more than one component has an
    expiry date, the earliest date should be shown.  This requirement is
    necessary to assure that outdated or unstable mixtures are not used.

    1.6.8  Labelling

         The importance of proper labelling cannot be overemphasized. 
    This is especially true for the labelling of test, control or
    reference substances, as well as for their mixtures.  Misapplied or
    inadequate labelling can lead to major problems that can invalidate
    study results.  For this reason, every laboratory should have SOPs for
    labelling test, control or reference substances and their mixtures
    with carriers.  These SOPs should address how labelling will be
    applied and what its content will be.

         Labelling for a mixture container should include the expiry date
    of the mixture; the use of expiry dates is intended to preclude the
    use of mixtures that may have deteriorated.  Procedures should exist
    to assure that outdated or deteriorated batches of test and reference
    substance mixtures are not used.

    1.6.9  Facilities and equipment

         There should be a separate area for the receipt and storage of
    test, control and reference substances.  Laboratories may have a
    centrally located facility for this purpose.  These areas are usually
    supplied with storage cabinets, freezers, refrigerators, balances and
    related equipment required for the orderly receipt, storage and
    distribution of the substances received.  These areas are sometimes
    equipped with special environmental equipment, such as laminar air
    flow and biohazardous hoods, depending on the extent to which these
    areas may be used for opening and dispensing test and reference
    substances for use in the mixing areas.  If containers of test and
    reference substances are opened in the receiving/storage area, SOPs
    should prescribe the necessary precautions to be taken to avoid
    contamination of the area and equipment with the various substances.

         Employee safety precautions in handling chemicals are essential. 
    These should include decontamination procedures for spills and other
    emergencies, as well as procedures for routine cleaning.  The nature
    and extent of the facilities and their use will dictate what
    procedures and equipment will be required.  The area should be

    physically secured and only authorized personnel permitted to store or
    remove materials.  In smaller laboratories, there may not be room for
    a separate facility for this purpose.  In this case, test and
    reference substances may be stored in areas used for other purposes;
    however, the same general requirements apply as stated above.  In
    particular the area, such as a cabinet, should be isolated as much as
    possible from other areas and activities and should be made physically
    secure with authorized access only.  It should not be used for any
    other purpose.  Limited access to test substances is required to
    prevent unauthorized use and mix-ups. Accountability records should be
    maintained, documenting receipt and dispersal of test and reference
    substances. Such documentation for all substances is often maintained
    in the receipt and storage area.  The entry or removal of substances
    from the designated areas should be documented.

         In addition to separate areas for receipt and storage of test,
    control and reference substances, the laboratory should provide an
    isolated area for preparing and mixing these substances with carriers. 
    The extent and nature of these facilities will depend upon the scope
    and type of the laboratory.  Larger laboratories may have elaborate
    equipment, facilities and procedures for mixing test, control and
    reference substances with carriers.  In a typical laboratory, engaged
    routinely in conducting safety studies, there will usually be a
    separate area or areas where test substances are stored and weighed. 
    The pure substances are then transferred to the separate mixing and
    dosage preparation areas.

         Another major consideration in the design of a mixing area is
    cleaning and decontamination.  The area should be easily cleanable. 
    The walls, ceiling, and floor should be smooth and impervious to
    cleaning agents used.  Cleaning and decontamination design
    considerations also apply to personnel who work in the area.

         There must be specific SOPs to cover the cleaning and/or
    decontamination of equipment.  This is especially important for
    equipment, such as mixers, that may come in direct contact with test
    substances.  In specifying cleaning agents, the SOP should take into
    account the solubility of materials to be removed from the equipment
    and also the removal of any cleaning agent residues.  In studies where
    the test material may present unusual equipment cleaning problems, the
    procedures for cleaning may be included in the study plan.  It is also
    very important for the cleaning of important equipment to be monitored
    and documented.

         Once the test, control or reference substances have been mixed or
    otherwise prepared, they should be moved to a separate area designated
    for the storage of these mixtures.

         SOPs should also deal with the disposal and/or reuse of
    containers used to store test and reference substances and their
    mixtures.  In laboratories where containers are reused to store these

    substances, there should be strict references and documentation
    covering decontamination and labelling. Procedures for final disposal
    of hazardous waste should be defined by national or local governmental
    regulations.  Laboratories should ensure that they conform to these
    procedures.

    1.7  Quality control

    1.7.1  Introduction

         Quality control is applied to routine laboratory biological,
    chemical and physical analyses in order to assure reliability and
    comparability of test data.  It involves statistical approaches
    designed to demonstrate the constancy or variability, and precision of
    analytical data.

         When the concentration of a chemical is determined repeatedly in
    the same laboratory a certain scatter is always seen.  The scatter is
    due to variation in the analytical step as well as to changes that may
    take place during the collection, preparation and storage of the
    samples.

         Quality control depends, among other factors, on the type of
    compounds involved.  A major issue is "to keep the compound as it is"
    and it is reasonable to make a distinction between inorganic and
    organic compounds. Since over ten million organic compounds exist, a
    further distinction must be made between simple organic compounds,
    such as solvents and metabolites, more complex compounds that may have
    a low vapour pressure, such as some pesticides, compounds that are
    capable of being adsorbed, and very complex biological compounds such
    as proteins and enzymes.

         In the case of inorganic compounds, contamination of the test
    substance with the same compound from an external source is a major
    concern.  Some elements are volatile, e.g., elemental mercury, and
    special care is needed to avoid loss of analyte.  Certain elements are
    readily reduced to a different oxidation state even within the
    container used.  Examples are the reduction from Cr(VI) to Cr(III) and
    the reduction from Hg(II) via Hg(I) to elemental mercury, which
    disappears from the test material.  There may be effects due to
    compounds already present in the test substance, to compounds added
    for preservation or to wall effects of the container.  The latter
    category can be separated into chemical effects, e.g., reduction and
    oxidation, and physical effects, e.g., adsorption to the wall or
    stopper, and precipitation/ coprecipitation of the analyte alone or
    together with other compounds.

         In the case of simple organic compounds, a well-sealed container
    is needed to prevent vaporization loss.  The container or stopper can
    be responsible for adsorption or leakage.  Reduction as well as
    oxidation of metabolites in biological material may occur. 

    Ultraviolet light, as occurs in sunlight, may decompose organic
    compounds by photochemical reactions, and the infrared radiation in
    sunlight may cause thermal degradation of organic compounds.  Thermal
    degradation may also occur due to other infrared sources such as heat
    radiators or laboratory apparatus.  Another factor which may lead to
    thermal degradation is the temperature of the environment of the test
    substance from sampling until final analysis within the apparatus.

         In the case of more complex compounds with a low vapour pressure,
    such as some pesticides, vaporization of the analyte is unlikely. 
    Certain pesticides may undergo reduction or oxidation.  Photochemical
    and thermal degradation may occur in this class of compounds.

         Proteins and enzymes are extremely vulnerable to changes in pH
    and redox potential, heat and exposure to ultraviolet light. 
    Moreover, as many proteins can bind easily to heavy metals, they need
    special care.

         Another important item is the concentration of the compound of
    interest.  Concentrations of compounds in biological materials range
    in general from parts per million, e.g., essential elements in plant
    and animal, to parts per trillion for very toxic compounds such as
    dioxins.  This enormous range of concentrations is also seen in the
    soil, surface water, indoor air, and, to a lesser extent, in sea
    water, drinking-water and outdoor air.  As precision is often
    inversely related to concentration, it is evident that the
    determination of sub-ppm concentrations of compounds is always
    problematic, regardless of the method of choice.

          Analytical variation may be divided into two major categories:
    accuracy and precision.  Accuracy refers to the agreement between the
    measure and the true amounts of the analyte, and precision refers to
    the random variability or reproducibility of the method.

         In order to assure the reliability and comparability of the test
    data, an extensive quality assurance programme should be implemented. 
    This should cover the sampling and sample handling (preanalytical
    quality control), as well as the analytical procedures (analytical
    quality control).  The purpose of quality assurance is to identify
    different types of errors that may invalidate the test data and to
    make sure that the total of all errors is below certain established
    limits.

    1.7.2  Level of quality control

         A basic question in quality control is how accurate and precise
    the different measurements need to be in order to provide reliable
    exposure assessments.  This decision may be based on prior information
    on the inherent variability in the quantity being measured, and this
    is likely to differ for each pollutant and medium.  Existing
    information from measurements already conducted in the locality in

    question, or in similar areas or groups of people, can be used to
    provide a guide.  The variability can then be assessed as the relative
    standard deviation, expressing the standard deviation as a percentage
    of the mean, assuming that a normal distribution basis is adequate for
    the present purpose.

         An important issue in biological test substances is the
    biological variability, which can be separated into the
    intra-individual variability, i.e. the variability  within an
    individual, and the inter-individual variability, i.e. the variability
     between individuals.  The variability may be expressed as a relative
    standard deviation.  The allowed standard deviation for a
    determination used in clinical and environmental chemistry is half the
    standard deviation of the inter-individual variability at a maximum,
    or, alternatively half of the sampling error (Youden, 1967).

         As another example of how to deal with precision, the procedure
    in the UNEP/WHO (1984) and WHO (1986b) studies may be used.  Within
    these studies a relative standard deviation of 20% was set
    arbitrarily.

         The error in precision within a laboratory is thought to be of
    the order of one-half to two-thirds that of the inter-laboratory
    error.

    1.7.3  Pre-analytical quality control

         It is essential to ensure correct sampling, i.e. that the
    collected air, particles, water, food or other samples really
    represent the whole sampling period and the subjects concerned, and
    that contamination of the samples is avoided.  It is important for the
    field personnel and the laboratory staff to be properly trained. 
    Detailed guidelines for the sampling procedures should be prepared
    prior to the sample collection and discussed with all the people
    concerned.

         In the case of many pollutants, it is necessary to check their
    content in containers and other equipment used for sampling and
    storage before sampling in order to avoid contamination.

         Audit procedures are useful to control the reliability of the
    sample collection, transport and storage.  The major steps to be
    audited during sampling would be:

    *    (preventative) maintenance

    *    calibration of sampling and analytical measurement equipment

    *    procedural control checks

    *    tick-off chart

    *    cleanliness during sampling, sample transport and storage

    *    sample deterioration, temperature control and stabilizers

    *    time lapse before analysis

    *    data recording, calculations and record keeping

         Sampling should be representative, and the number of samples
    should be adequate.  Sampling can be random or selective, depending on
    the goal of the study.  If sampling should be random, all subsequent
    steps should also be carried out in a randomized manner.  For example,
    if an "exposed" population of animals or humans is compared with a
    reference group, sampling should be performed randomly.  Also the
    pre-analytical steps, such as the distribution of the samples in the
    containers and the time lapse before determination, should be
    performed carefully to avoid any systematic influence.

         Although sampling errors are often the greatest source of
    variability, especially in environmental measurements, these errors
    are mostly beyond the control of the laboratory, unless the laboratory
    is also responsible for the sampling operation (Horwitz, 1989). 
    Regardless of who collects the samples, this source of variability
    must be considered individually for each lot or population to be
    examined.  The magnitude of this source of variability is assessed by
    taking a number of random samples from the population and examining
    them individually (see also section 1.7.5).

         In occupational toxicology and sometimes in environmental health,
    the timing of specimen collection is important.  Some industrial
    chemicals have a long biological half-life in various body
    compartments, and the time of sampling is not critical  (Herber &
    Schaller, 1986).  For other chemicals the timing of sampling is
    critical because after exposure the compound and/or metabolites may be
    rapidly eliminated from the organism.

         In the case of urine samples, 24-h or 6-h urine samples are
    preferable to spot samples.  Preservation against bacterial growth and
    fungi can be achieved by adding a solution of sodium azide. 
    Determination of enzymes should generally be performed within 24 h. 
    However, some stable enzymes may be determined after a longer time
    interval, but preferably within a week when stored at 4 C.

         In the case of blood samples, clotting is a major source of
    errors and this should be avoided at all costs.  Clotted blood samples
    cannot be used for analysis and must be rejected.  Many urine samples
    have a precipitate but generally this can be redissolved by acids
    without influencing the determination, by carefully heating up to 37
    C, or by ultrasonic treatment.  Another possibility is to remove the
    precipitate by adequate centrifuging.  When a precipitate forms during
    cooling, care must be taken that the samples are stirred so that any

    deposits are dissolved or evenly distributed.  All pre-analytical
    procedures must be checked comprehensively before starting routine
    determinations.

         Storage of samples is optimal at -80 C.  At this temperature
    samples may be stored for years.  A more practical temperature is -20
    C.  At this temperature proteins can be stored for months.  Inorganic
    compounds maybe stored at +4 C, provided that no vaporization takes
    place.

    1.7.4  Analytical quality control

         Validation of analytical procedures with respect to accuracy and
    sensitivity should be accomplished by appropriate quality control
    studies.  The accuracy of the methods used by the different
    laboratories should preferably be established by an external quality
    control programme, in which the reference values of the quality
    control samples are unknown to the laboratories.  If this is not
    possible, comparisons with reference methods or the analysis of
    certified reference samples may be used.  It is also important to
    establish the reproducibility of the routine analytical procedures
    used.  Acceptable limits of variation for control samples should be
    set primarily by considering the data quality requirement rather than
    the analytical characteristics of the procedure.  The stringency of
    the limits depends largely on the purpose for which the test data are
    intended.

         In case of random sampling, the complete analytical procedures
    should be randomized.  Thus, destructions or extractions must be
    performed for an approximately equal number of "exposed" and "control"
    test specimens  within the same run.  This applies to instrumental
    techniques including spectrophotometry, chromatography and atomic
    absorption spectrometry.  Calibration and reference specimens must be
    distributed randomly between the test specimens.

         A major item is the influence of the matrix and, especially in
    urine, the matrix variability; an example is the variable salt
    concentration in urine which may cause problems in the determination
    of metals.

    1.7.5  Statistical considerations

         According to Taylor (1988a), it is essential to define the degree
    of uncertainty of the data.  This is necessary in order to provide
    evidence that samples are "representative" for the environment or
    population to be sampled.  Two factors influence the uncertainty of
    the data, i.e. measurement uncertainty and sample uncertainty.  The
    uncertainty of the measurements can be estimated by replicate
    measurements on a sufficient number of samples.  Taylor proposed that
    the measurement uncertainty should not exceed one-third of the total
    uncertainty tolerance.

         The sample uncertainty is based on the variability of the
    population or environment to be sampled.  Sampling strategies based on
    statistical considerations address this factor.  During sampling a
    number of items can represent components of the sampling uncertainty. 
    These are, for instance, systematic components like the properties of
    sampling equipment and the variability within the actual sampling
    process.  All these components add up to the total sampling
    uncertainty.  It is  practical to define acceptable total uncertainty,
    and then estimate the different components contributing to this
    uncertainty.

         Smith et al. (1988) stated that quality assurance procedures of
    sampling data require the definition of the precision, the bias, the
    representativeness, the completeness and the comparability. Precision
    in this case is the agreement of individual measurements of the same
    property under the same conditions.  It is best expressed in terms of
    standard deviations.  In environmental sampling, it means that the
    total variance of the measurements is the result of the precision of
    the analytical measurements and the precision in sampling.  The
    analytical precision is obtained by determining the standard deviation
    of individual samples, whereas the precision of the sampling as such
    is determined by analysing a number of replicated field samples.  The
    presentation of results from environmental measurements should include
    regression equation coefficients, when appropriate, means and standard
    deviations, and ideally a graph containing the actual data points, the
    best-fit curve and confidence intervals, in order to provide
    information for a quality assessment.

         Bias is the degree of agreement of a measurement or an average of
    measurements having an accepted reference or true value.  It is
    usually expressed as the difference between these two values or as a
    percentage of the true or reference value.  Bias is frequently
    expressed as recovery (100% bias) in environmental monitoring
    programmes, and can be estimated by using spiked field samples for the
    bias in the field sampling phase, and by the repeated analysis of
    field samples or reference material for the analytical phase.

         Representativeness is the degree to which data represent
    precisely and accurately a characteristic of a population or
    environmental condition.  This can be assured by the use of
    appropriate sampling techniques.

         Completeness is a measure of the amount of valid data obtained
    from a measurement system compared to the amount that was expected to
    be obtained to fulfil the objectives.  Various conditions, e.g.,
    meteorological conditions, accidents, during sampling and analysis
    could result in the incompleteness of data, particularly during field
    sampling.  In any case there should be an evaluation of the influence
    of missing or lost data on the overall result of a study.

         Finally, comparability expresses the confidence with which one
    data set can be compared to another.  For this purpose, there should
    be descriptions of the comparabilities of sites elected for sampling,
    the calculation and statistical analysis, and the sampling and
    analysis protocol.

    1.7.6  Analytical performance evaluation

         There are various ways of performing analytical quality
    evaluation.  Since it is impossible to produce errorless analytical
    data or to obtain a measure of the accuracy, it is important to
    estimate the limits of uncertainty of produced data.

         Some general statistical approaches to evaluating the quality of
    chemical measurements have recently been reviewed (Taylor, 1987,
    1988b).  A well-established procedure for comparing the analytical
    performance of several laboratories is the round-robin test using a
    Youden plot (Youden & Steiner, 1975).  Two similar samples, designated
    X and Y, are sent to each collaborating laboratory for analysis.  The
    results provide a pair of coordinates which are plotted in a diagram. 
    If random errors predominate, the results will fall within a circle
    with the center showing the "true" value.  Due to systematic errors
    most of the observations are usually found along a 45 axis.

         Evaluation of the regression line of reported versus reference
    values for a set of quality control samples is a useful method of
    guarding against systematic errors in the likely concentration range
    (Vahter, 1982; UNEP/WHO, 1984; Friberg, 1988).  Further discussion of
    this method is given in section 3.7.

    1.8  Documentation and record keeping

    1.8.1  Introduction

         Any study report must be capable of being validated before it can
    be fully relied upon for accuracy and completeness of findings and
    before any scientific conclusions can be derived from it.  This means
    that the information and conclusions stated in the report must be
    fully supported by raw data documented in the laboratory records
    covering the conduct of the study.  It requires the existence of a
    complete data trail from the initiation of the study to the time when
    the last data point is recorded.  The data trail should be detailed
    enough to allow an independent party to trace every aspect of the
    study.  Validation of a study by an independent party is sometimes
    necessary to assure that all the provisions of the study plan and SOPs
    were followed.

         "Raw data" covers all of the original observations of the study.
    For example, the term "raw data" may mean any laboratory worksheets,
    records, memoranda or notes (or exact copies of these) that are the
    result of original observations and activities of a study and that are

    necessary for the reconstruction and evaluation of the final report. 
    Exact copies of raw data may be valid, provided that they are verified
    as accurate by the dated signature of the person making the copies. 
    Other examples of raw data include photographs, microfilm or
    microfiche copies, computer printouts, magnetic media, including
    dictated observations, and recorded data from automated instruments.

         There are three common processes for capturing data.  These
    processes include manual recording of data, direct computer entry of
    data, and entry of written data into a computer.

    1.8.2  Manual data records

         Many testing facilities find it useful to have standard data
    collection forms for recording items such as health status
    information, body weights, body length, clinical observations, test
    system care, necropsy, clinical chemistry, haematology, histological
    processing, and environmental condition.  These forms should be
    designed with care and should be readily available to personnel
    conducting study functions.  Procedures for proper use of data
    collection forms should be described in SOPs or the study plan. When
    these forms are used during the course of the study, they should be
    identified by study number, when applicable, test substance and
    species, and they should carry entry points not only for the collected
    data but also for dates and initials of the person or persons entering
    the data.

         All data should be recorded in indelible ink, preferably black,
    and corrections in the data should be made so as not to obscure the
    original entry.  The use of white-out, correction tape, erasers,
    overwriting or any other means to obscure the original entry is not
    acceptable.  Changes made in records of original observations should
    ordinarily be made by the individual responsible for the original
    entry.  All corrections should be initialled and dated at the time the
    correction is made and a justification should be given in the record. 
    The record should be corrected in a timely manner.  Some laboratories
    use codes or abbreviations to indicate common reasons for corrections. 
    Such codes should be defined in the record or the associated SOP and
    should be used consistently by all employees making corrections.

         Standard data collection forms should be designed so that they
    are unambiguous and so that they require a notation to indicate
    whether an activity has been performed or an observation has been
    made.  Leaving data entry spaces blank is an unacceptable means of
    documenting an observation.  For example, a standard form to document
    the collection of tissue at necropsy may list all possible tissues and
    provide a space to document the collection of specific tissues.  If a
    technician leaves a blank space to record that no tissue was taken, an
    individual reviewing the record can never be sure if the tissue was,
    indeed, not taken or the technician merely forgot to fill in the space
    when the tissue was collected.  Similarly, a line drawn through all

    data entry spaces on a record can be ambiguous, unless its meaning is
    clearly defined in the record or SOP and is understood by all
    individuals completing the record.

         A study notebook in which narrative notes may be kept of
    significant events, such as animal deaths, changes in batches of test
    substances, clinical examinations, disease outbreaks and animal
    sacrifices, supplements the standard data collection forms.  Actual
    data elements required by the study plan for the test system can be
    recorded in the study notebook.

    1.8.3  Computer data records

         Although there is still a need for written records to document
    important daily events, computers may now be used to save time and
    labour in data collection, validation and report generation.  Direct
    entry of data may be accomplished by use of keyboards, optical bar
    code readers, touch screens, optical scan sheets, and direct digital
    input from laboratory instruments or sensors.  In order to capture and
    correlate effectively diverse data from multiple sources, most
    laboratories collecting data in this manner use a distributed data
    processing system, where remote terminals and microcomputers collect
    data, which are later transmitted to the central computer for
    permanent storage and/or processing.

         In systems such as this, data should be verified at the time of
    collection prior to transfer to the central computer.  This can be
    done by review of the terminal screen at the time of data entry and/or
    review of data print-outs.  As with data collected manually,
    computer-entered data must include the identity of the individual
    making the observation and the date the observation was made.  Any
    correction made later at the local level require that the change be
    verified in the central computer to check that it has been recorded
    correctly.  Experience has revealed cases where legitimate corrections
    were made at a local level but not submitted to the central computer
    or included in the final report.  The same principles apply for
    correcting computer-collected data as for paper data entries, i.e. the
    original entry must not be lost, and the change must be documented to
    show who made the correction, and when and why the correction was
    made.

         A computer system must be designed to meet the requirements of
    quality assurance.  It must be tested prior to initial use and
    whenever significant changes are made to the computer hardware or
    software.  The testing process must include a documented review of the
    system design, development, and acceptance measurements to assure
    proper system performance.  This should be done prior to testing by
    laboratory personnel.  Testing often involves the checking of a
    previously validated array of hand-collected data against the same
    data collected concurrently and reported by computer for one or more

    studies.  For each study, the acceptance testing needs to cover each
    of the data points the system is designed to collect and process.

         For the purposes of retaining raw data, the magnetic media
    containing the raw data (i.e. disc, tape) and/or a verified printout
    should be kept.  However, manufacturers of magnetic media advise that
    data retention is not guaranteed beyond a limited time, and, if discs
    are erased, revised periodically or reused, some errors in the discs
    can arise.  Thus, a paper print-out of the data for data review and
    storage as original data is essential.  However, any print-out
    retained as original raw data must be verified.  This means it must be
    reviewed for accuracy and this review must be documented by the
    signature and date of the individual(s) performing it.  Proper
    verification would include review by the supervisor and/or study
    director and quality assurance personnel.

         Any computerized data collection system must include SOPs to
    describe operations, specification, security, validation and
    maintenance of both the local data acquisition systems and the central
    computer system.  These documents should be in the appropriate work
    stations and should contain such information as how the data are to be
    entered and what remedial action is to be taken in the event of system
    failure.

    1.8.4  Indirect computer data records

         The principles outlined above also apply to manual data that are
    subsequently entered into a computer.  In addition, evidence must
    exist that the manual data were correctly and completely entered. 
    Frequently, this is a quality assurance programme function.  The
    extent to which the "raw data" are compared with those of the initial
    paper print-out (or screen display) needs to be defined in SOPs.

    1.9  Final report

    1.9.1  Introduction

         The final report is the end-product of a carefully planned and
    conducted study.  The report must be well organized and reflect
    accurately all the experimental data.  It must contain a detailed
    account of the study including, where relevant, unexpected deviations
    in the controls, study plan and environmental conditions.

    1.9.2  Contents

         The final report should include the following information:

         a)   a descriptive title;

         b)   clearly defined objective(s) and study plan;

         c)   an informative summary of the results of the study;

         d)   an identification of the test and reference substances by
              chemical name, code or chemical abstracts number;

         e)   a description of the characterization of the test and 
              reference substance including purity, stability and
              homogeneity;

         f)   name and address of the test facility, and the name and
              address of any facility that may have performed parts of the
              study;

         g)   the name of the study director and any other principal
              scientists who contributed reports included in the final 
              report;

         h)   the dates on which the study was initiated and completed;

         i)   a description of the methods, procedures and materials used,
              highlighting any changes in the study plan;

         j)   all information and data called for in the study plan,
              including "outliers";

         k)   a description of the test system, and the procedure used for
              identification;

         l)   exposure conditions;

         m)   an accurate presentation of results, including calculations, 
              and a description of the statistical methods used to analyse
              the data;

         n)   an evaluation and discussion of test results and any
              conclusions drawn from the results;

         o)   the dated signature of the study director.

    1.9.3  Indexing

         An index is essential for each volume of data comprising the
    final report.  Appropriate references to appended tables and figures
    are necessary to facilitate the peer review of the validity of the
    conclusions drawn from the study.  Appendices containing graphs,
    tables and statistical evaluations of all the data and observations
    are as important as the narrative report.  A glossary of terms and
    abbreviations used in the report should also be included, since this
    will facilitate an understanding of the findings of the report.

    1.10  Archiving and retention of data

    1.10.1  Introduction

         Records, specimens and reports constitute lasting proof of the
    validity of a study; facilities and procedures for their archiving
    should be available.  The following basic points should be observed:

    *    all raw data, documentation, study plans, appropriate specimens
         and final reports generated as a result of a study should be
         retained;

    *    there should be archives for the orderly storage and retrieval of
         all raw data, documentation (e.g., training and qualification
         records), equipment maintenance, master schedule of studies and
         quality assurance reports, study plans, specimens, and interim
         and final reports;

    *    material retained or referred to in the archives should be
         indexed by test substance, date of study, test system and nature
         of study;

    *    an individual should be responsible for the archives;

    *    only authorized personnel should be permitted to enter the
         archives.

    1.10.2  Facilities

         Storage facilities required for archiving data and specimens must
    be adequate to preserve specimen quality and to control access.  The
    storage conditions in the archives should be monitored to prevent
    accelerated deterioration of data and/or specimens.  Archives should
    be designed to have a controlled access area, i.e. with access limited
    by doors that lock, and the area should be isolated from operational
    laboratory areas but still conveniently accessible.  Environmental
    conditions should avoid extremes of temperature and humidity.  For
    example, tissue blocks should be kept cool and dry enough to prevent
    melting, sticking or mould growth.  Paper records should be protected
    from fire, water, rodents, etc.  Environmental conditions in the
    archives should be monitored to assure that appropriate conditions are
    maintained.  Finally, proper storage cabinets are needed to organize
    and store data and specimens for easy retrieval.

    1.10.3  Responsibilities for an archive

         Normally a testing facility has the responsibility for providing
    an archive facility and authorized archivist.  However, when there is
    a sponsor/contractor relationship involved, there can be problems over
    the allocation of responsibility for archiving material between the
    laboratory and the sponsor.  During the course of the study, the

    original data remains with the test facility conducting the study. 
    When the study has been completed, there are several ways that data
    may be archived, depending on the contract between the sponsor and the
    test facility.  Firstly, the test facility may archive the data. 
    Secondly, it may transfer all archival material (e.g., data, slides,
    tissues) to the sponsor for archiving.  In this case, there should be
    a record of transfer, and documentation identifying the storage
    location should be maintained in the contract facility's archive.  The
    third option is for the test facility to send the originals to the
    sponsor and maintain copies in its archive.

    1.10.4  Standard operating procedures for archiving

         There must be SOPs for each activity performed in operating an
    archive.  Some of the relevant SOPs are:

    *    receiving, indexing and identification

    *    filing and storage

    *    access to archives

    *    security

    *    data retrieval

    *    retention of material

    *    removal and return of records and raw data (as appropriate)

    1.10.5  Receiving, indexing and identification

         Generally, the central archive of a test facility is utilized for
    the paper and computer data from a study.  This includes all raw data
    in notebooks, forms and/or computer print-outs, approved study plans,
    amendments, and final reports generated.  These materials should be
    received and accepted by an authorized person (the designated
    archivist or his replacement).

         Prior to accepting the data, the archivist should be given an
    index, signed by the study director, of the submitted data listing the
    items being archived.  When the items have been checked against the
    index and the archivist is satisfied that the package is complete, the
    index should be signed and dated by the archivist.  After acceptance,
    the total study package should be archived using a classification
    system (e.g., a unique number).

    1.10.6  Filing and storage

         Material files should be filed and cross-referenced for easy
    access and retrieval of data.  Index cards are often used for
    cross-referencing, but data are increasingly being indexed on
    computers.

         Archived material should be paginated and indexed by test
    substance, date of study, test system and nature of study.  Other
    titles used for cross-referencing are unique study numbers, the study
    title and the name of the study director.

         In the case of specimens, slides, blocks, tissues, test
    substances and other materials that are not in the central archive
    with the raw data and final report, there should be a method for
    indicating where these articles are archived.  One method is the use
    of a form to be filled out by the person maintaining the items.  For
    instance, if the teratologist maintains an archive for fetal specimens
    and the pathologist maintains a separate archive for wet tissue,
    blocks and slides, they should complete and sign a form certifying the
    location and storage of the specimens.  This information would be
    maintained in the central archive with the other paperwork generated
    as a result of the study.

    1.10.7  Access and security

         There should be control of access to any archive where data,
    specimens and/or slides from studies are stored.  The door to the
    archive should be locked at all times when it is unattended.  Only the
    archivist should have keys for access to the archive records. 
    Ideally, the archive should be designed with a separate area for
    people to wait while the archivist is retrieving the requested data in
    order to help maintain the limited access environment.

    1.10.8  Retrieval of data and control of access

         Although stored data and specimens should be easily identified
    and retrievable, their removal from the archive should be discouraged.

         For the retrieval of data specimens, there must be a standard
    operating procedure.  The approval required may vary, but in many
    testing facilities, the authority to release data is given to the
    archivist.  For records management and to maintain a tracking system,
    a "Data and Specimen Request Form" is useful because it becomes a
    permanent record for the study file and helps make the person
    requesting data aware of the need to ensure the security and integrity
    of the items removed from the archive.

         The title of the study, name of the borrower, purpose,
    organization and items requested are entered on the form.  When the
    request is completed, the archivist will sign and date the form and

    the borrower will sign, indicating that the items were received and
    that the security and integrity of the data will be ensured.  This
    form remains in the archive.

         When the borrower returns the data, the archivist must check the
    items returned against those listed on the request form.  If all items
    are present and the archivist is satisfied that the integrity of the
    study has not been compromised, the form is signed by both the
    borrower and archivist and becomes a permanent record for the study
    file.

    1.10.9  Retention of information

         The length of time that study records must be retained will vary,
    this depending mainly on legal considerations.  It is the
    responsibility of test facility management and/or the study sponsor to
    ensure that records, specimens and reports are retained for the
    required time.

         For archived material, such as wet specimens, plates for
    mutagenicity testing, test and reference substances, haematology
    slides and histochemical and other specimens that are relatively
    fragile and that vary in stability and quality during storage, the
    retention period can be only as long as the period during which the
    material is of a quality that affords a valid and meaningful
    evaluation.  Management should develop an SOP establishing the
    retention time for these types of specimens.

    2.  QUALITY MANAGEMENT APPLIED TO TOXICITY STUDIES

    2.1  Introduction

         Laboratory animals are the test system for many studies.  In
    order to provide meaningful safety data, the animals must be properly
    selected and cared for to ensure that any response observed in the
    animals is caused by controlled exposure to the test, control or
    reference substances and not by uncontrolled variables such as disease
    and adverse environmental conditions.  Appropriate laboratory practice
    needed to assure proper care and use of animals includes providing
    proper facilities and employing necessary procedures to eliminate or
    control factors that could interfere with the response of animals to
    the test and reference substances.  The care, housing and treatment of
    animals used in research and testing is regulated by law in many
    countries.

         Animal care facilities can comprise a large portion of the
    physical plant of a laboratory because sufficient animal rooms or
    areas are needed to ensure separation of species or test systems, and
    provide for quarantine requirements and routine or specialized housing
    of animals.

    2.2  Procedural requirements

    *    There should be SOPs for housing, environmental conditions,
         feeding, handling and care of animals.

    *    Animals recently received from outside sources should be
         appropriately isolated and their health status should be
         evaluated before they are used in a study.

    *    At the beginning of a study, animals should be free of diseases
         that might interfere with the purpose or conduct of the study.

    *    Animals that become diseased while under study should be
         appropriately isolated to prevent infection of other animals.

    *    Such animals may be treated if treatment is authorized and
         documented, and if the treatment will not interfere with the
         interpretation of the test.

    *    Animals should be identified individually as appropriate (e.g.,
         tattoo, ear tag) to assure that they can be identified with their
         respective dose regimens and the in-life and postmortem
         observations.  Their individual housing units (e.g., cages)
         should be labelled with all the information needed to identify
         specifically each animal within the unit.

    *    Animals of different species and/or projects should be
         appropriately separated to preclude inadvertent exposure to test
         or reference substances, disease transmission or other
         uncontrolled stress to the animals.

    *    Animal cages, racks and accessory equipment should be cleaned at
         appropriate intervals.

    *    Food and water used for animals should be analysed periodically
         for potential interfering contaminants, as specified in the study
         plan.

    *    Bedding used in animal cages and pens should not interfere with
         the purpose or conduct of the study, and should be changed as
         needed to keep animals clean and dry.

    *    The use of any pest control materials should be approved by the
         veterinarian/study director and be documented.  Cleaning and pest
         control materials that interfere with the study should not be
         used.

    2.3  Phases of animal use

         The procedures listed in section 2.2 must be applied in the
    various phases of animal use, which include:

    *    obtaining animals

    *    shipment and receipt of animals

    *    pre-study evaluation of animals

    *    allocation of animals to the study

    *    exposure of animals to the test substance

    *    evaluation of in-life animal response to test and reference
         substances

    *    removal of animals from a study

    *    transfer of animal tissues and specimens to archives

    2.4  Obtaining animals

         Some laboratories obtain their animals from commercial breeders,
    while others maintain their own breeding colonies.  In either case,
    the laboratory should have SOPs for animal procurement.  These SOPs
    should address the source of supply as well as the responsibility for
    initiating and authorizing animal orders.

         With respect to the source of animals, the SOP should
    specifically identify acceptable suppliers for each of the different
    species of animals used at the laboratory.  Ideally, the list of
    suppliers should be developed by a laboratory veterinarian or other
    animal health care professional with responsibility for directing the
    laboratory's overall animal health programme.  Animal suppliers should
    be evaluated for their ability to provide healthy animals of known
    genetic and health background.  The laboratory should keep records of
    its animal health evaluations or should ask to review such records at
    the animal supplier in order to establish which suppliers can provide
    suitable animals consistently.  Some laboratories require that their
    veterinarian or health care professional inspect the facilities of
    animal suppliers as part of the evaluation of the supplier's
    suitability.  Once a list of acceptable suppliers has been developed,
    only animals from this source should be used.  This practice, coupled
    with an effective programme at the laboratory for evaluating the
    health status of incoming animals, will help assure the quality of
    incoming animals.

         If animals are from the laboratory's own breeding colony, there
    should be SOPs for the maintenance and testing of the colony to assure
    the production of animals of consistent quality.  Breeding colonies
    should be kept separate from the areas housing animals being tested. 
    This helps to preclude inadvertent exposure of the breeding colony to
    test substances and disease.  Records of health examinations,
    treatments and breeding should be kept for review by a health care
    professional responsible for assessing the suitability of animals for
    use in tests.

         In addition to addressing the source of supply, the SOPs for
    ordering animals should define the person(s) with responsibility for
    ordering animals, the information required to order animals, and the
    documentation to be maintained.  A single individual or department
    should be assigned responsibility for ordering animals.  It is good
    practice to have a single authority for ordering animals to assure
    that the animals are obtained only from an acceptable source.  This is
    often the animal husbandry department, which is headed by the
    laboratory's veterinarian.  However, the individual or department
    charged with ordering animals cannot assume full responsibility for
    this function alone.  The study director is responsible for assuring
    that the test system (i.e. the laboratory animals) used is the one
    specified in the study plan.  One means of ensuring this is to require
    that the study director provides a written request to the animal
    husbandry department specifying the species, strain and number of
    animals required for a given study.

         The SOP for ordering animals should specify the information to be
    included in the request.  Normally, this will be the same as the
    information on the test system in the study plan.  For example, when
    requesting animals, the study director should specify at least the
    following: the species, strain or substrain; number and sex of

    animals; weight range and/or age of animals; and any special
    requirements such as timed-pregnant animals or surgically altered
    animals.  The request should also state the proposed starting date of
    the study so that animals will be available on time and within the
    specified age/weight range.  The number of animals ordered should also
    take into account the loss of animals due to shipping.  The written
    request for animals should be signed by the study director indicating
    that the authorization and animal specifications conform to the study
    plan.  The written request for animals should be retained with the raw
    data to document the involvement of the study director in this phase
    of the study.  The SOP for ordering animals should also specify what
    documentation is to be retained.  In addition to the request for
    animals, some laboratories retain purchase orders, shipping tickets
    and/or other records to document the date the order was made, that the
    order conformed to the request for animals and that it was placed with
    an acceptable animal supplier.

    2.5  Shipping and receipt of animals

         When animals are shipped from an outside source, the receiving
    laboratory may not have much influence over the handling of animals
    during shipment.  Most reputable animal suppliers will select the
    shipper and provide adequate housing, feed and water to accommodate
    the animals' needs during shipment.  However, the testing laboratory
    should be familiar with the supplier's normal means of shipment and
    inform the supplier of any special shipping requirements.  For
    example, a supplier may normally make deliveries to a laboratory on
    certain days of the week.  However, if for some reason, there will be
    no one at the laboratory to receive the animals on the normal day, the
    supplier/shipper should be advised.  Otherwise, animals may be left
    unattended in an unsuitable environment.

         In laboratories that breed their own animals, transport of the
    animals from the breeding colony to the areas where the animals are
    received and used should be defined in an SOP for transfer of animals. 
    The SOP should include a description of the housing to be used during
    transport as well as the precautions to assure that animals are
    properly fed and watered from the time they leave the breeding
    facilities until the time they arrive at their point of use.  The
    procedures should clearly identify the individuals responsible for
    transfer of the animals and should provide for direct communication
    between the breeding colony and the department that is to receive the
    animals.  The SOP should include any procedures required to preclude
    unnecessary stress to the animal during transport.  Animals must not
    be allowed to remain in an uncontrolled or a stressful environment any
    longer than necessary.  It is also important that animals transported
    in-house be clearly identified on their housing as to their origin and
    destination in order to assure proper routing and use.

         Unlike shipping, the receipt of animals ordered by the laboratory
    from its own breeding colony is under its full control.  The

    laboratory should have SOPs covering the receipt of animals, which
    define, at a minimum, who will be responsible for receiving animals,
    where the animals are to be received, which tests should be done
    immediately upon receipt, and what documentation should be kept.

         At the time of receipt, animals are usually counted and assigned
    temporary numbers or other means of individual identification.  They
    are placed in standard housing units for transfer to the area where
    their health status is to be checked (i.e. quarantine).  There should
    be an SOP that covers identification and housing of animals during
    this pre-study evaluation period.  Animals should be identified so
    that observations of their health can be clearly documented for the
    individual animal.

         Documentation of animal receipt is important information and
    should be kept with the other study records.  It should cover the
    following details: where the animals were shipped from; the date and
    time they were received at the laboratory; who received the shipment;
    the general condition of the shipment, including a description of the
    number of containers; and the number of live, dead and moribund
    animals in each container.  The assignment of animals to cages for
    transfer to the pre-study evaluation area should be documented and
    should include the actual number of animals transferred.  Any other
    information obtained for each animal during the receipt phase should
    also be recorded.  For example, animals may be weighed or, in the case
    of some large animals, vaccinated or otherwise treated.  This
    treatment should be recorded.  It is useful to retain a copy of the
    shipping tickets and animal container labels in the study records in
    order to document the receipt of the animals.

    2.6  Animal care facilities

         Depending on the nature of a laboratory's work, animal care
    facilities may comprise a large portion of the physical plant.  The
    laboratory should have a sufficient number of animal rooms or areas to
    assure proper separation of species or test systems, isolation of
    individual projects, quarantine of animals, and routine or specialized
    housing of animals.

         The need for separation of species should be evaluated by a
    veterinarian or animal health professional to assess where separation
    by species is required to prevent disease or stress in the animals. 
    Although not ideal, it is possible to house different species together
    where it has been determined that such housing arrangements will not
    potentially affect the health of the animals or otherwise adversely
    affect the study.  This is not a desirable practice, however, and
    should be avoided.

         Individual projects should be isolated to avoid potential
    confusion and cross-contamination of animals.  As a general rule,
    animals being used in separate projects should be housed in different

    rooms.  This is especially true in feeding studies where environmental
    contamination may be present due to dust from handling dosed feed and
    spillage of feed by animals.  Although it may be permissible to house
    two different projects together under some conditions (e.g., same test
    substance, same species), it is generally inadvisable to house
    projects together that use different test substances.  In any event,
    precautions should be taken to minimize or eliminate exposure of
    laboratory personnel to the test substance.

         Facilities for quarantine or isolation of animals are essential. 
    Laboratories should have space available to isolate animals
    effectively, when necessary, to avoid adverse effects on the conduct
    of studies.  For example, many laboratories quarantine incoming
    animals in the room where the study will be conducted.  They are held
    in that room until their health status is evaluated and then they are
    placed on study in the same room.  Thus, they are effectively
    quarantined from other animals at the laboratory, even though there is
    not an animal room strictly dedicated for quarantine.

         Areas for routine or specialized housing are required.  Routine
    housing of animals may include the maintenance of a breeding colony or
    the housing of animals in a routine study, such as a chronic feeding
    study.  Specialized housing should be available where required.  For
    example, studies requiring collection of metabolism or behavioral data
    may require special caging or environmental controls.  Also, animals
    used in inhalation studies may be housed in special inhalation
    chambers.  The laboratory management should assure that these and all
    other facilities are available as required.

         In addition to the animal facilities discussed above, a
    toxicological laboratory should have a number of animal rooms or areas
    separate from these to ensure isolation of studies being done with
    test systems or test and control substances known to be biohazardous,
    including volatile substances, aerosols, radioactive materials, and
    infectious agents.  These areas should include facilities and
    equipment for disposal of waste and other contaminated material from
    this area, and special equipment or clothing to protect laboratory
    personnel.

         Separate areas should also be provided, where appropriate, for
    the diagnoses, treatment and control of laboratory animal diseases.
    These areas should provide effective isolation for the housing of
    animals either known or suspected of being diseased or of being
    carriers of disease from other animals.  The extent of, and the need
    for, these facilities will depend upon the testing laboratory's policy
    regarding the handling of diseased animals.  Many laboratories using
    rodents have a policy of immediately removing and discarding any
    animal that develops disease.  Such a policy would essentially
    eliminate the need for any facilities dedicated to diagnosis and
    treatment of disease in these animals.  However, when primates, dogs
    or other larger animals are used, they frequently are treated for

    disease.  In this case, special areas should be provided for such
    diagnosis, treatment and isolation.  The extent to which diseased
    animals need to be isolated to prevent adverse effects on other test
    animals will require the judgment of the laboratory veterinarian or
    animal health care professional.  However, should isolation be
    required, facilities must be available.  The need for dedicated
    treatment facilities will also vary depending on the nature of the
    treatment.  Treatment that is highly stressful to the animal or which
    includes aseptic procedures, such as surgery, should not be performed
    in the animal room.  On the other hand, oral administration of
    medication and monitoring of body temperature are examples of
    diagnostic and treatment procedures that may be carried out in the
    animal room.

         Adequate facilities for animal care must include facilities for
    collection and disposal of animal waste.  In testing facilities that
    do not dispose of the waste directly, there must be facilities for
    safe sanitary storage of waste until it can be removed from the
    laboratory for disposal.  Disposal facilities should be designed and
    operated to minimize vermin infestation, odours, disease hazards and
    environmental contamination.  Facilities for storage and disposal of
    waste should be adequately separated from animal areas to preclude
    contamination of the animals, and schedules must be established,
    followed and documented to assure the timely performance of waste
    removal activities (e.g., animal cage changing).

         A schedule for routine maintenance of buildings should be
    established.  Although many laboratories have such schedules for
    equipment, many do not have building maintenance schedules.  The need
    for such schedules is most easily illustrated in the area of animal
    facilities.  These facilities receive much wear from both animals and
    personnel.  Walls, floors and ceilings may crack, making these
    surfaces uncleanable or exposing underlying surfaces that may produce
    dust.  Paint may also crumble.  Light fixtures and doors wear out. 
    All of these conditions may have an adverse effect on the animals
    housed in the facility.  It is difficult to repair some of these
    conditions after a study is underway without adversely affecting the
    animals' environment.  Consequently, maintenance should be scheduled
    to eliminate the need for repairs while studies are underway.  This is
    especially true for rooms used to house animals in chronic studies. 
    Of course, consideration must always be given to the potential impact
    on the study of chemical agents used in building maintenance and
    cleaning.  Similarly, attention must be paid to the possible
    environmental contamination of the building during construction,
    renovation or repair of laboratory facilities, e.g., the use of
    certain plastics that leach out toxic chemicals or asbestos used in
    ceilings or to protect pipes.  A schedule of routine preventive
    maintenance for buildings will ensure proper cleaning and maintenance
    of all areas.

    2.7  Animal husbandry supply facilities

         Storage areas should be supplied, as needed, for feed, bedding,
    supplies and equipment related to animal care and use.  Storage areas
    for feed and bedding should be separated from areas housing the test
    systems and should be protected against infestation or contamination. 
    There should also be facilities, as needed, for the preservation of
    perishable supplies, e.g., certain feeds requiring refrigeration.

         Most large testing facilities maintain central storage areas for
    bulk storage of feed and bedding.  In the case of feed, the individual
    responsible for running the bulk feed storage facility will dispense
    small lots of feed to the various animal areas for short-term storage
    and immediate use.  These small lots are often stored in separate
    rooms in the immediate animal area under the responsibility of animal
    husbandry personnel working in the area.  Feed may then be further
    dispensed to relatively small storage containers in the actual animal
    rooms.  Feed is dispensed from these containers into feeders for the
    animals.  Despite the difference in size and location of feed storage
    facilities at the testing laboratory, they must be protected against
    infestation or contamination.  This is generally not a problem where
    feed is stored in designated areas outside of the animal rooms.  In
    these designated areas, feeds are bagged and stored to facilitate
    stock rotation and to preclude contamination with cleaning agents. 
    The chance of vermin infestation is also reduced.  However, if feed is
    to be stored in animal rooms, special precautions must be taken. 
    Firstly, only small amounts of feed should be permitted to be stored
    in the room, i.e. amounts of feed necessary for two weeks or less. 
    Secondly, the feed must be stored in sealed vermin-proof containers
    that can protect feed from contamination during cleaning of the room. 
    Thirdly, the feed must be clearly identified and stored in an area
    separate from the animal housing unit and designated for no other
    purpose.  Finally, there should be procedures in place to assure the
    proper rotation of feed.  Often, new feed is dumped into a container
    in the room without emptying it of the previous feed.  This practice,
    if continued, may lead to contamination of the feed with mould and
    accompanying mycotoxins.

         There should be facilities for the storage and handling of
    bedding so as to prevent contamination and vermin infestation.  Some
    laboratories store substantial amounts of bedding in the area where
    cages are cleaned.  This bedding is sometimes placed in the cages
    prior to their transport to the animal room.  In this case, the
    facilities must be adequate to protect the bedding from contamination
    by equipment cleaning agents and dirty equipment brought to this area
    for cleaning.

         Finally, animal supply facilities should provide areas for the
    storage of clean and dirty equipment, as well as facilities for
    cleaning it.  Such equipment would include cages, racks, water
    bottles, feeders, and similar items.  Some facilities have corridors

    and rooms dedicated strictly to either clean or dirty equipment.  The
    laboratory and flow of work are designed to prevent the mingling
    and/or mixing of clean and dirty equipment.  Although designated
    clean/dirty facilities are ideal, an adequate degree of separation can
    be maintained by scheduling work so that corridors and cleaning areas
    are not occupied by clean and dirty equipment at the same time.

    2.8  Facilities for handling test, control and reference substances

         The facilities must allow separation of areas involved in the
    storage, handling and distribution of test, control and reference
    substances and of mixtures containing them.  This is necessary to
    prevent contamination of facilities, equipment, personnel and test
    systems, as well as to prevent confusion of substances.  To accomplish
    this, there should be separate areas for:

    a)   receipt and storage of test, control and reference substances;

    b)   mixing of test, control and reference substances with a carrier 
          (e.g., feed);

    c)   storage of the test, control and reference substance mixtures.

         In addition, storage areas for the test and control substances
    and test and control mixtures should be separate from areas housing
    the test systems.  The storage facilities should be adequate to
    preserve the stability of the substance and mixtures.  For example,
    adequate refrigerator or freezer storage space should be available for
    substances requiring low temperature storage.

    2.9  Pre-study evaluation of animals

         A fundamental laboratory practice is to place all newly received
    animals in quarantine until their health status has been evaluated. 
    The procedures for doing this vary from laboratory to laboratory. 
    Some maintain animal colonies under barrier conditions with extensive
    programmes for the evaluation of animals entering the barrier.  These
    include thorough screening for bacterial, viral and fungal pathogens. 
    Sentinel animals may also be maintained in the room housing the
    animals under test.  These sentinel animals are evaluated in the same
    manner as the animals sampled during the pre-study health evaluation.

         Whatever pre-study evaluation is employed, a laboratory should
    have SOPs to assure that the animals' health status is fully evaluated
    prior to the test.  This evaluation may be performed by the laboratory
    veterinarian, animal care employees or the study director.  These
    individuals should have documented training and experience that
    qualifies them to recognize symptoms and to diagnose disease problems
    for the animal species involved.

         A laboratory should have facilities that can adequately separate
    newly received animals from those already housed, and a designated
    quarantine area is desirable.  Laboratories may quarantine newly
    received animals in the room where they will be housed during the
    study.  This is acceptable provided that employees follow adequate
    procedures to prevent cross-exposure of other animal rooms with
    equipment, waste or other materials from the rooms where the animals
    are housed for pre-study evaluation.  The use of the same room for
    pre-study health evaluation and conduct of the study also permits the
    acclimatization of animals to the actual environmental conditions in
    the room prior to the beginning of the study.  The facilities required
    for housing animals during quarantine are generally the same as for
    housing animals being tested.  There should be an adequate number of
    cages of appropriate design, construction and size for the species to
    be housed.  There should also be facilities for the removal and
    disposal of animal wastes, and adequate environmental controls for
    temperature, humidity and lighting.  Technical guidance regarding
    facilities, equipment and proper animal husbandry procedures may form
    part of government regulations for animal welfare.

         SOPs for the pre-study health evaluation of animals should define
    the parameters to be monitored for each species during the pre-study
    evaluation period and the frequency of observations.  For example, the
    SOP might require rodents to be observed daily for mortality and
    clinical signs of illness such as rough hair coat, diarrhoea, laboured
    breathing and weight loss.  A very important SOP topic is the length
    of the pre-study evaluation period.  Animals should be isolated long
    enough to provide reasonable assurance that the most common diseases
    affecting that species will have ample time to manifest themselves
    clinically.  This time varies from species to species and should be
    based on acceptable veterinary medical practice.  The SOP should
    identify who will make the observations, define the circumstances
    under which a veterinarian or the study director should be called, and
    what, if any, treatment is to be given to the animals.  Any signs of
    illness observed or treatment rendered should be documented in study
    records as described in the SOP or study plan.  Records of clinical
    signs of illness should include entries specifically identifying
    individual animal(s) involved, with a detailed description of the
    clinical features, the date and time of onset and duration.  The
    records should identify who made the observation.

         Records of treatment should include the name of any medications
    used and the complete details of their administration, including the
    date of each administration and the amount given.  Records should
    clearly indicate when treatment began and ended for each animal and
    who authorized treatment; the laboratory veterinarian, in consultation
    with the study director, should authorize all treatment.  These basic
    requirements apply whether treatment is administered during the
    pre-study evaluation period or during the study itself.  In any case,
    no treatment should be administered that will interfere with the
    purpose or conduct of the study.  All records of animal observations

    and treatment made during the pre-study evaluation period should be
    retained with the study records.

         The SOP covering pre-study evaluation must reference the SOP for
    housing, feeding, handling and care of the animals.  It should specify
    the types and sizes of cages to be used, feeds and feeding schedules,
    bedding schedule and procedures for changing and cleaning cages and
    equipment, procedures for watering animals, and procedures for
    monitoring and documenting environmental conditions such as lighting,
    temperature and humidity.  The SOP should specify acceptable ranges
    for these parameters and remedial action to be taken when readings are
    outside these ranges.  Like the SOP for treatment of animals, the SOPs
    for the housing, feeding, handling and care of animals is general and
    required for all animals whether the animals are in quarantine, on
    test or in the breeding colony.

    2.10  Allocation of animals to a study

         Most laboratories document the suitability of the test animals at
    the end of the quarantine period by means of a form signed by the
    veterinarian in charge of quarantine, which authorizes the release of
    the animals for use in the study.  The study director is notified and
    the animals are transferred to the room where the study will be
    conducted if they are not already housed there.  At this point,
    animals are ready for allocation to the study.

         Three important events occur during the allocation to study
    phase.  These are:

    *    Animals are evaluated to determine if they meet study plan
         requirements such as age, weight and physiological condition.

    *    Animals are permanently identified so that they can be
         specifically related to any data or specimens generated during
         the test.  Animal housing units are also identified so that all
         information needed to identify each animal within the housing
         unit is located on that unit.

    *    Animals are allocated to the respective test groups in a manner
         that precludes bias (i.e. they are randomized).

         Related procedures and documentation should exist to assure that
    each one of these vital study functions is properly performed.  In the
    animal evaluation, some overlap may occur with the quarantine phase of
    the study.  For example, it may be necessary to use animals that have
    specific blood chemistry values.  To establish these values, several
    measurements may be required over a period of weeks.  Blood samples
    may have already been drawn and analysed for several weeks prior to
    release of animals from quarantine.  Now, however, the results of
    those blood analysis must be compared against the study plan
    requirements, and the study director must decide which animals qualify

    for study.  Also, weight requirements and other parameters that must
    be met at the time of study initiation should now be evaluated for
    conformance to the study plan.  These are some examples of pre-study
    evaluations that are made during the allocation phase of the study. 
    They are dictated by the study plan, and the study director has the
    final say as to whether the criteria are met.  The most important
    record-keeping requirement at this point includes all records of
    laboratory reports and other observations made to determine if the
    animals conform to study plan requirements.  There should also be a
    record of the individual reviewing this documentation and authorizing
    the animals for use in the study.

         At this point, animals should be individually identified in such
    a way that they can be traced for any in-life or postmortem
    observation.  This is often done by the use of tattoos or ear notches. 
    Whatever method is chosen, it should be thoroughly described in the
    SOP for identifying animals.  The method selected should not interfere
    with the study, should not separate from the animal, and should not be
    obliterated over the course of the study.  It should also be easily
    read by animal technicians and all other laboratory personnel handling
    animals.  It is especially important to provide adequate instruction
    to animal technicians on the proper reading of the identification code
    used.  For example, a system of ear notches may be used to represent
    animal numbers.

         Such a code should be fully explained in an SOP maintained in the
    animal room for ready reference.  The SOP should also cover how
    numbers are assigned for a given study to assure their traceability. 
    Study records must very clearly document the assignment of permanent
    animal numbers in such a way that they can be accurately
    cross-referenced with any temporary number assigned during quarantine. 
    This is necessary to maintain data integrity for a given animal from
    receipt to disposal.

         In addition to identifying individual animals, housing units
    should be marked to indicate which animals are housed in the unit.
    Most laboratories utilize cage cards or labels for this purpose.  The
    cage identification should include the number, sex, dose group and
    study number for the animals housed in the cage.  This practice helps
    assure that the animals are selected from and returned to their proper
    cage during cage-changing operations and during removal of the animals
    for required study observations.

         The allocation of animals to their assigned test groups is
    another important study event.  Once animals have been appropriately
    evaluated and approved for use on the study, they must be divided into
    treatment groups.  In a typical study, there are four treatment groups
    of animals for each sex: control, low-dose, mid-dose and high-dose. 
    Additional groups may be required as dictated by the study plan. 
    Animals should be placed in these groups in a manner that is designed
    to eliminate bias.  This is accomplished by using a procedure that

    randomly assigns animals to a given treatment group.  The procedure
    for randomizing animals should be described in a written SOP or the
    study plan.  The allocation of animals using the specified procedure
    should be fully documented.

         In addition to randomizing animals, many laboratories employ
    procedures to eliminate bias in the housing of animals.  This is
    necessary because, even in the best animal facilities, environmental
    conditions within the room will vary.  For example, there may be areas
    of lower air circulation or uneven heating and cooling.

         The microenvironment of the animal cages in a single rack housing
    multiple cages may also vary.  For example, animals in higher cages
    may be exposed to more intense light and higher temperatures than
    animals in lower cages.  To overcome this, most laboratories have
    established SOPs for the placement and rotation of animal housing
    units within a rack and racks within a room.  Commonly, in rodent
    studies, cages housing animals of a single dose group are placed
    together in a structured fashion to facilitate easy identification of
    the group during observation, feeding and dosing.  The cages are then
    moved systematically as a group through all locations within the rack
    during the course of the study.  The racks themselves are
    systematically moved through all locations in the room.  This assures
    uniform exposure of the test animals to the full range of
    environmental variations within the room.  The moves are often
    performed with the same frequency as other animal room procedures,
    such as cage changing.  The moves should be documented by recording
    the date of the move.  Laboratories may use diagrams of racks and
    rooms to document the movements and specific location of racks and
    cages.

    2.11  Exposure of animals to a test or control substance

         There are two major concerns during this phase of the study. 
    Firstly, known quantities of defined substances should be administered
    to specifically identified test animals over a given period of time. 
    Secondly, test animals must not be exposed to uncontrolled stress that
    may affect their response to the test substance.  Both of these
    concerns can be met by appropriate care and use of laboratory animals.

         The administration of substances should be carried out in
    accordance with the SOP and/or the study plan requirements to assure
    proper treatment of the test animals.  A laboratory should have SOPs
    to cover fully every aspect of receipt, testing, preparation,
    distribution and use of test and control substances.  Laboratory
    procedures should ensure that properly defined and known quantities of
    correctly labelled test and control substances are delivered to the
    laboratory personnel for administering to the test animals.  Personnel
    should have their SOPs and study plan readily available to assure
    proper administration of the materials.  The study plan should specify
    the route of administration and the dosage levels.  However, there

    should also be SOPs to cover the technical aspects of administering
    the substances, as well as specific procedures to prevent confusion of
    animals or test and control substances during administration.

         Examples of SOPs for dosing animals might include instructions on
    how to insert a stomach tube for administration of materials by oral
    gavage or how to administer substances intravenously.  Examples of
    SOPs designed to prevent confusion would include procedures that
    address the order in which the animals are to be dosed, how many
    animals can be removed for dosing at one time, and procedures for
    verifying animal and test substance identification.  The SOP or study
    plan directions for the administration of test substances should
    provide directions on how to document the administration of the test
    and reference substances.  This is crucial to the subsequent
    validation of study findings.  Such documentation usually consists of
    a dosing record kept by means of a standard form.  Documentation
    should include the following information: the name, lot number and
    expiration date of the test and reference substances, as applicable;
    the exact quantity of substance administered; the specific identity of
    the individual animal receiving it; the date, time and duration, as
    applicable, of administration; and the names of the individuals
    administering the dose.  The individuals involved should sign or
    initial the record.  During test and reference substance
    administration, the fact that each animal was dosed and the specific
    weight or volume of dose that each animal received must be recorded. 
    If the test substances must be prepared or manipulated by technicians,
    this should also be documented.  It is equally important to document
    positively that control animals were treated according to study plan
    requirements.

    2.12  Control of laboratory environment

         The second major concern during the dosing phase is to prevent
    the exposure of test animals to uncontrolled stress.  This is
    important throughout the in-life portion of the study.  Uncontrolled
    stress can be caused by numerous factors: these include inadequately
    controlled animal room environment; disease; contaminants in feed,
    bedding and water; exposure to cleaning agents, pesticides and
    uncontrolled test substances; poor husbandry practices; improper
    handling; and inadequate facilities.

         To prevent interference from environmental factors, the
    laboratory must have animal facilities that will provide a controlled
    environment appropriate for the species housed.  This generally means
    that temperature, humidity, lighting, ventilation, noise and personnel
    activity in the animal rooms should be monitored and/or controlled. 
    The monitoring and control activities should be recorded.

         Laboratories housing rodents should have the capability to
    control temperature and humidity.  There should be SOPs to define
    acceptable ranges for environmental parameters and to describe actions

    to be taken if acceptable ranges are exceeded.  Some laboratories
    continuously monitor and record the temperature and humidity (by means
    of hygrothermographs).  The recording charts from such devices should
    be regularly checked and retained as study records.  The charts should
    be initialled and dated by the individual checking them and there
    should be explanations for any reading that is abnormal.  Such
    readings should be brought to the attention of the study director so
    that their impact on the study can be assessed.

         Although they are desirable, hygrothermographs or other elaborate
    temperature and humidity recording devices are not absolutely
    essential.  Many laboratories monitor these parameters with
    nonautomatic instruments on a twice daily basis to coincide with other
    activities that are being performed.  All that is required is an
    accurate thermometer and a hygrometer or sling psychrometer.  If these
    are used there should be SOPs for their use and the collected records
    should be reviewed and retained as raw data.  If temperature is to be
    observed only once a day, it is desirable to have a thermometer that
    will show the minimum and maximum temperatures reached over the 24-h
    period.  It is also good practice to have any temperature monitoring
    device periodically calibrated against an official standard to assure
    accuracy.  This also applies to thermocouples or other electronic
    sensors used to monitor temperature and humidity.

         Other environmental factors that affect animals include lighting,
    noise and ventilation.  Lighting is often controlled by the use of
    timers to cycle lights on and off on a regular schedule.  For small
    animals, it is common practice to provide equal periods of light and
    dark over a 24-h period (i.e. 12 h on, 12 h off).  Lighting cycles
    should be monitored periodically to assure that they meet study plan
    requirements.  Such monitoring should be recorded.  This sometimes
    presents a problem and may be overlooked when the light cycles occur
    after normal working hours.  To overcome this problem, some
    laboratories have SOPs that direct the security or building
    maintenance personnel to monitor and document the light cycles after
    working hours.  This is an acceptable practice, provided accurate
    documentation is maintained.  It is also good practice to have animal
    room doors identified with the name and telephone number of the study
    director or other responsible individual so that they may be contacted
    should a problem with the animal room be detected after working hours.

         Noise in animal facilities is generally controlled by facility
    design and proper placement of animals.  In most laboratories, noise
    usually comes from three sources: animals, equipment and personnel. 
    To control animal noise, noisy animals such as primates and dogs
    should be housed separately from other animals.  Their housing should
    have doors and windows designed to reduce noise transmission between
    rooms and provide for arrangement of animals to accommodate their
    social behaviour.  For example, animals housed so that they can see
    each other are often quieter.  Limiting the entrance of unnecessary
    personnel into these areas will greatly reduce noise levels.

         The most stressful noise probably originates from equipment and
    personnel.  For example, cage washing operations often generate sharp
    irregular and/or continuous noise.  Noise from these sources can be
    controlled by locating these areas out of hearing range of animal
    housing areas.  Employees slamming doors, moving racks and cages, and
    playing radios can also create uncontrolled stress in the animals.

         Ventilation of animal rooms should always be controlled and
    monitored.  Proper ventilation with clear, filtered air is necessary
    to maintain acceptable temperature and humidity levels and to remove
    offensive odours or other airborne contaminants.  SOPs for animal care
    and housing should specify acceptable air exchange rates for animal
    rooms.  These should be monitored and documented periodically.

         Another factor that may produce uncontrolled stress in test
    animals is disease.  In cases where animals become diseased during a
    study, the disease may be treated.  However, such treatment must be
    fully authorized and documented as discussed in section 2.6, and the
    study director must determine that treatment will not interfere with
    the purpose or conduct of the study.

         Contaminants in feed, water and bedding can adversely affect the
    animals.  To preclude this, feed and water used for animals should be
    analysed periodically to ensure that contaminants known to be capable
    of interfering with the study, and reasonably expected to be present
    in such feed and water, are not present at levels above those
    specified in the study plan.  Documentation of such analysis should be
    maintained as raw data.  The study director should identify any such
    contaminants in the study plan.  Likewise, bedding should be free of
    any interfering contaminants or naturally occurring constituents that
    could interfere with the test.  For example, pine and cedar shavings
    may contain naturally occurring aromatic hydrocarbons that can affect
    the liver metabolism of laboratory animals.

         Other potential contaminants may be introduced into the animals'
    environment in the form of cleaning agents, pesticides and test or
    control substances from different studies.  There should be SOPs
    covering the identification, use and monitoring of these agents.  For
    example, the facility should have an SOP detailing the cleaning agents
    acceptable for use in animal facilities or on equipment coming into
    contact with animals.  This would include detergents and sanitizers
    for cleaning floors, walls, ceilings and counter tops in animal rooms,
    as well as these same agents used to clean cages, pans, water bottles,
    racks and feed mixers.  The SOP for cleaning must also include
    procedures for documenting that cleaning was performed and that the
    agents used were in accordance with the SOP.  A similar SOP should
    exist for any pest control materials that are used.  The SOP should
    identify which chemical pesticides are approved for use, the method of
    application and the specific locations where they can be used.  The
    SOP should include the provision that responsible laboratory personnel
    should accompany any contract pest control personnel when they apply

    pesticides to assure that they are applied in accordance with the SOP. 
    This should be documented in writing.  Many laboratories forbid the
    use of any pesticides and instead rely on mechanical traps and
    barriers.  The use of these should also be described in the SOP.

         Other sources of uncontrolled stress to animals include
    laboratory personnel and poor husbandry practices.  There should be an
    SOP regarding the admittance of personnel to the animal facilities;
    this should identify which employees are permitted in the animal rooms
    and under what circumstances.  It should also describe any special
    health or safety precautions necessary to protect employees and
    animals.  For example, the SOP should describe any protective clothing
    to be worn or decontamination procedures to be followed by personnel
    entering or leaving animal areas.  Many laboratories have procedures
    designed to control disease and environmental contamination of the
    animal colony by prohibiting movement of employees between different
    animal facilities or laboratories.  SOPs should also exclude from
    direct contact with animals an employee with an illness that may
    adversely affect animal health.  Such employees should be reassigned
    until their health recovers.

         Poor husbandry practices can be avoided by the use of adequate
    SOPs covering housing, feeding, handling and care of animals.  This
    should include procedures to assure that animal cages, racks and
    accessory equipment are cleaned and sanitized at appropriate
    intervals.  SOPs should also specify minimum acceptable cage sizes and
    room capacities to prevent animals being overcrowded.  Every activity
    performed in the animal room during a study should be documented.  To
    do this, some laboratories keep room activity logs.  These records are
    maintained in the animal rooms during a study.  Each time anyone
    enters the room they must sign the log documenting who they are, when
    they entered the room, what activity they were performing, and when
    they finished and left the room.  This is a good practice for
    controlling and monitoring the level of activity in an animal room. 
    Excessive activity can increase stress in the animals.

         These animal care and use procedures are not all inclusive but
    they do cover the major factors.  The procedures apply not only during
    the exposure of the animals to the test substances, but also to every
    phase of studies involving the maintenance and use of animals in
    laboratories.

    2.13  Evaluation of in-life animal responses to test and control
          substances

         This phase of animal use covers the period from the first
    administration of the test substances until the animals are submitted
    for postmortem evaluation.  It is during this phase that crucial data
    are obtained by observing the response of test animals to exposure of
    test and reference substances.

         In a typical chronic rodent study, animals are observed as
    specified in the study plan and/or SOPs.  Examples of observations to
    be made include the following: general physical examination; daily
    observations for mortality and toxicological or pharmacological
    effects, body weight and feed consumption measurements; palpation for
    tissue masses; and special procedures such as ophthalmoscopy.  The
    study plan must specify the frequency of these observations.  Also,
    each of these procedures should be covered by the study plan or an
    SOP; SOPs often complement the study plan by providing specific
    guidance.  For example, the study plan may simply state that animals
    are to be observed for toxicological effects.  The SOP may define such
    effects as lethargy, hyperactivity or convulsions.

         SOPs can standardize observations by prescribing the use of
    standard terminology and methods.  For example, in palpating animals
    for tissue masses, the SOP should require that the location and size
    of masses be reported.  It should define the limits of specific
    location description, such as dorsal front right, to distinguish
    masses in this area from adjoining areas.  The SOP should also require
    that masses reported in a previous observation be positively accounted
    for during each subsequent observation session from the time they
    appear until the time they regress or are observed at postmortem
    observation.  SOPs should also cover how and where observations will
    be recorded.

         Proper identification of animals is important because all
    observations made must be related to specific animals.  There should
    be SOPs that assure verification of the accuracy of animal
    identification during dosing, observation and transfer of animals. 
    All documentation should clearly identify who made an observation,
    when it was made and to which animal it pertained.

         An important consideration is the training of personnel.  It is
    crucial to have employees who are trained and experienced in working
    with animals to make in-life observations.  All personnel should have
    documentation of their training and experience.  Complete familiarity
    with a species and its characteristics is essential to detect the
    often subtle clinical changes that may be indicative of toxic effects
    of the test substances.

    2.14  Removal of animals from a study

         Animals may be removed from a study because they die while the
    study is in progress, because they become sick or because they are
    sacrificed at the times specified in the study plan.  In these cases,
    it is important that proper identification of the animals and related
    specimens is maintained and that the animals are submitted promptly
    for postmortem evaluation to preclude the loss of data due to
    autolytic changes in tissue.  This is accomplished by requiring
    multiple daily checks for mortality and morbidity and specifying
    after-hours procedures for storage of dead animals.

         Documentation should be maintained for unscheduled removal of a
    dead or moribund animal and include, at a minimum, the identity of the
    animal, the date and time of removal, the reason for removal, and the
    identity of the individual who removed the animal.  For moribund
    animals, the reason for removal should include specific observations
    of animal behaviour and physical condition.  The study records should
    clearly indicate whether the animal was sacrificed or found dead. 
    Animals that are killed accidentally during the study should be fully
    reported.

         The same general requirements, with respect to the required
    documentation, apply to scheduled deaths.  An important requirement in
    the scheduled sacrifice of animals is that documentation should be
    maintained to demonstrate the method of sacrifice used.  This may be
    specified in the study plan or SOP, but should include a method that
    will not interfere with postmortem evaluation of animal tissues and
    specimens.

         Postmortem evaluation includes gross necropsy, histological
    preparation of tissues and organs, and histopathological evaluation. 
    A major concern is integrity of animal identification.  There should
    be procedures covering the identification of animals and their
    respective tissues and specimens from the time they are submitted for
    postmortem evaluation until they are sent to storage.

         There should always be positive identification attached to or
    accompanying each animal.  Many laboratories have a form that
    accompanies the animal to necropsy, identifying it by number, sex,
    dose group and study.  For animals that are individually numbered with
    ear tags, tattoos or other permanent identification, SOPs generally
    require the necropsy personnel to compare the animal received against
    the form to confirm proper submission.  Animals are then necropsied
    and tissues, organs and specimens are collected and processed as
    directed by the study plan.  Each of the items collected should be
    identified in a manner that relates it back to the animal it came
    from.  This is often accomplished by placing unique identifying
    information on containers used in the storage or processing of these
    tissues, organs or specimens.  A unique sequential accession number
    can be assigned to containers that relate to the animal, or the unique
    animal identification number may be utilized.  Whatever method is
    used, it must be traceable back to a specific animal in a specific
    test.  The procedures used must be fully described in the SOP.  Not
    only must tissues be identified, but all documentation for the
    receipt, preparation and evaluation of tissues must be identified to
    relate them to a specific animal from a specific study.  The records
    of gross necropsy findings, microscopic findings and in-life
    observations must account for all gross observations.  For example, if
    an animal technician reports four tissue masses during the last
    in-life observation period of an animal prior to submission for
    necropsy, the necropsy records must confirm the presence or absence of
    each of the masses reported.

         To achieve this, the last in-life observations must be available
    to the individual making necropsy observations.  Each mass accounted
    for at necropsy should be uniquely reported and identified in the
    gross necropsy findings for the animals.  Trimming and tissue
    processing records must also account for masses collected at necropsy
    for processing.  Finally, necropsy findings should be provided to the
    pathologist making microscopic evaluation of the tissues.  Records of
    this evaluation, likewise, must account for all tissues, masses and
    lesions reported.  To facilitate tissue accountability for animals
    with multiple masses, number or letter designations should be assigned
    to each mass at necropsy.  The method used to assure identification
    and accountability of animals and tissues during this phase of the
    study must be fully described in the SOP.

    2.15  Transfer of animal tissues and specimens to archives

         Safety studies generally produce large amounts of data and
    specimens.  To accommodate their storage and retrieval, a laboratory
    should have space designated as an archive for all raw data and
    specimens from completed studies.  This area should have access
    limited to authorized personnel only.  Although most laboratories have
    a centrally located archive to support the whole testing facility,
    some have a number of archives.  For example, some support
    departments, such as histology and chemistry, may maintain space for
    storage of data and specimens they generated or analysed.  This is
    acceptable, provided access to these materials is limited to
    authorized personnel and the data and specimens are properly
    identified and indexed as to their location.  The best situation,
    however, is to have archive facilities available to provide central
    storage of all data and specimens.  These should be under the control
    of a designated individual.

         After evaluation of animal tissues and specimens, the final phase
    is the transfer of these materials to the archives.  At this point,
    the gross remains of animals are placed in sealed containers while the
    tissues are contained in blocks and slides.  There should be an SOP to
    assure that all these items are properly identified, accounted for,
    inventoried and placed in the archives.  It is the study director's
    responsibility to assure that all specimens are transferred to the
    archives at the conclusion of the study.

    3.  QUALITY MANAGEMENT APPLIED TO HUMAN AND ENVIRONMENTAL MONITORING
        STUDIES

    3.1  Introduction

         The manufacture and use of chemicals can lead to the deliberate
    or unintentional exposure of natural ecosystems to potentially
    hazardous substances.  In this case, humans may be affected by contact
    with the chemicals through the consumption of food and drinking-water,
    and air inhalation.  The natural environment can change its structures
    due to effects on the complex processes governing the functions of
    ecosystems.

         In recent years, the awareness of risks of adverse health effects
    due to exposure to various environmental factors has increased
    substantially.  In many countries major health problems and nuisances
    are still related to particulate matter, sulfur dioxide and
    pesticides.  In the more developed industrialized countries the
    interest has shifted to effects such as cancer, allergy and disturbed
    function of the central nervous system, caused by exposure to
    pollutants such as hydrocarbons, nitrogen oxides and toxic metals.  In
    the future, interest is likely to focus on environmental health
    effects, e.g., effects on the nervous and immune systems, that are
    more difficult to detect and often appear after long periods of
    exposure to relatively low doses of specific environmental factors or
    to combinations of different factors.

         One aim for continuous health surveillance is to follow the
    development of morbidity and mortality patterns within various
    population groups.  This requires access to extensive descriptive
    data.  In order to relate observed health effects to certain
    environmental factors, reliable exposure data are needed.  It is
    important that observed changes in the health effect pattern are
    validated on a local level.  Many observed changes have been shown to
    be caused by artefacts.

         The function of an ecosystem is determined by the physical and
    chemical processes, as well as by the relationships and interactions
    of the living organisms within the system.   With the overall aim of
    the safe manufacture and use of chemicals, several objectives are
    covered by monitoring studies in natural environments.  These can be
    described, for instance, as the monitoring of the levels of chemicals
    in various compartments within ecosystems, i.e. a  descriptive study
    of situations and trends.  Furthermore, programmes monitoring the fate
    and effects of chemical substances in the field are carried out in
    order to compare the results with existing information from similar
    studies carried out in the laboratory, e.g., the investigation of
    exposure-response relationships for risk assessment.  In addition,
    monitoring of contamination levels in a selected environment is

    performed in order to evaluate compliance with specific environmental
    quality criteria or targets, for instance, from a regulatory point of
    view.

         Consequently, in environmental monitoring studies of natural
    ecosystems, it is predominantly the concentrations of chemicals,
    transformation processes and products which are investigated, as well
    as accumulation and ecological effects in wildlife, and the abundance
    and distribution of species.  For this purpose, environmental
    monitoring studies are carried out in different compartments of
    various ecosystems, e.g., sediments and soils, air and water, and
    terrestrial and aquatic biota.

         The incorporation of quality management approaches in
    environmental monitoring studies should ensure that the resulting data
    are reliable and reconstructible.  Quality management should cover all
    phases of an environmental monitoring study, i.e. the planning of the
    study, selection and preparation of sampling equipment, sampling
    procedures, analyses, measurements and observations, and the
    reporting.

         Reliable exposure data are essential for the establishment of
    dose-response relationships for toxic effects of environmental
    pollutants or other chemicals in human subjects, in epidemiological
    studies and for the assessment of risks of adverse health effects upon
    contact with such substances.  Traditionally, pollution monitoring has
    been concerned primarily with the movement of pollutants through the
    environment and the concentrations in various environmental media such
    as air, food and water.  Such monitoring provides little information
    on the amounts of pollutants actually coming in contact with people.

         Direct measurements of human exposure levels may involve
    determination of the chemical under study in air, food and water,
    preferably through personal monitoring, or determination of the
    chemical or its metabolites in tissues or body fluids (biological
    monitoring).  Ideally, the concentration of a pollutant, or its
    metabolites, in an indicator medium will give information on the
    degree of exposure, the dose at the critical organ and the risk of
    adverse health effects.  Some information concerning the type and
    degree of exposure may also be obtained from determinations of
    pollutant concentrations in micro-environments combined with studies
    of human activity patterns.

         The World Health Organization (WHO) and United Nations
    Environment Programme (UNEP) have been involved in human exposure
    monitoring since 1977, when global studies on the biological
    monitoring of lead and cadmium (Braux et al., 1979; Vahter, 1982;
    Friberg & Vahter, 1983; Bruaux & Svartengren, 1985; Vahter & Slorach,
    1989) and organochlorine compounds (Slorach & Vaz, 1983) were
    initiated.  On the basis of this work the WHO health-related
    programme, which involves monitoring of pollutants in air, water and

    food, has included a new component, the Human Exposure Assessment
    Locations (HEAL) programme.  Initially, the HEAL programme has focused
    on methods for exposure monitoring, including methods for quality
    assurance.  General principles and procedures for the development of
    quality assurance in relation to exposure monitoring have been
    prepared (UNEP/WHO, 1984; WHO, 1986b).  A comprehensive document on
    quality assurance in biological monitoring of metals has been
    published (Friberg, 1988).

    3.2  Procedural requirements

         A crucial point in exposure monitoring is to ensure correct
    sampling, e.g., that the collected air particles, food, blood and
    other samples really cover a representative period and that the
    correct sampling procedures have been used.  As in other types of
    studies, it is important that the field personnel and the laboratory
    staff are properly trained.  In many exposure monitoring activities
    the subjects under study collect the samples themselves, e.g., 
    duplicate diets.  In this case it is necessary to train the subjects
    properly, and to have staff available for assistance or advice during
    the entire sampling period.  A study plan and detailed SOPs for the
    sampling procedures should be prepared and discussed with all the
    people concerned. It is also important to motivate the subjects and to
    give them adequate information about the aim and the design of the
    project before starting collecting the samples.

         The SOPs should cover all aspects of the monitoring exercise,
    particularly the sampling procedure, use of equipment and materials,
    transport and processing of samples, and analytical and observation
    methods.  For example, a description of the equipment to be used for
    field sampling should take into account that outdoor application
    requires robust constructions, simple handling and easy-to-read
    instructions.  Careful transportation is necessary to maintain the
    functioning of equipment.  Furthermore, calibration procedures should
    be established which appropriately consider that the performance of
    many instruments, not particularly designed for outdoor use, may alter
    under field conditions; the laboratory calibrations may not be valid.
    The SOP should consequently define how the functions of the
    instruments for field measurements are to be tested.

         The correct, unambiguous and non-erodible labelling of samples
    and specimens is particularly essential, since the samples taken in
    the field are usually transported and analysed or measured at other
    locations and at a later date.  Thus the conditions under which the
    samples have to be transported and stored must be established in an
    SOP.  The correct recovery of a chemical from an environmental sample,
    for instance, is very much dependent on its chemical and biological
    stability in the medium.  Considerable care has, therefore, to be
    taken to ensure that the content and often the structure of a sample
    remain unchanged until it is analysed, or until observations or
    measurements have been carried out.

         In the study plan, the personnel responsible for different parts
    of the study should be clearly identified, since the field work may be
    carried out by different institutions to those that carry out the
    laboratory analyses and measurements.  Therefore, the study plan is
    the most important document for providing information to all
    participants on all aspects of the study, including sampling dates,
    amounts and character of samples.

         Deviations from the study plan, particularly concerning changes
    in the selected sampling stations, sampled populations and sampling
    tissues, necessitate a detailed amendment to the study plan.  This is
    of great importance, since in many monitoring studies sampling occurs
    on a regular basis, and the data obtained may be worthless if the
    preset sampling scheme is not followed.  Therefore, changes need
    authorization of the study director.  Acceptable levels of deviations
    in respect of time or space of sampling should also be stated.

         If a change in a given parameter or the impact of a certain
    process has to be monitored over time or space, a control site or
    population often has to be selected.  The selection of the control
    must ensure that conditions are similar to those of the site under
    investigation but that the influencing factor is lacking.  In
    contrast, background sites (or populations) that are not affected by
    human activities (Black, 1988) can be treated as "blanks" in
    environmental monitoring programmes.

    3.3  Selection of sampling strategies and study design

         The study plan for a monitoring study must state specifically the
    selected sampling area, sampling station or population to be sampled,
    medium to be sampled, frequency and time of sampling, numbers of
    samples and the parameters for evaluation.  For instance, when the
    contamination level in an aquatic ecosystem is monitored over a long
    period, samples from the water column, sediments and living organisms
    are commonly taken on a regular basis.  The number of samples is
    selected according to the size and character of the investigation
    zone.  The volume of the sampled medium is chosen according to the
    expected concentration of the chemical and the sensitivity of the
    analytical method.  The wildlife species to be sampled should be
    representative for the biota and known to accumulate the substance of
    concern sufficiently.

         Different objectives of environmental monitoring programmes
    require various kinds of sampling techniques.  In general, it has to
    be decided if random or non-random sampling should be applied to
    achieve the objectives.

         For example, if the fate and effect of a contaminant distributed
    by a diffuse source has to be monitored, a random sampling technique
    is appropriate.  On the other hand, if the dilution of an industrial
    effluent by natural surface waters is monitored, the sampling stations

    need to be selected according to various factors, e.g. water flow,
    while sampling times may be randomly chosen.

         In cases where exposure data for a group of individuals are used
    for estimating exposure in the population, the selection of
    individuals for monitoring is a critical step.  The sample of
    individuals has to be selected so that it will provide valid
    inferences for the target population.  Thus, the selected individuals
    must be representative of the population under study.  Quality
    assurance related to selection of study groups and individuals is
    described in the guidance provided by WHO on sample selection and data
    analysis for HEAL studies (WHO, 1992).

         Three major random sampling techniques exist; these are commonly
    described as "stratified random sampling", "simple random sampling",
    and "systematic sampling" (Cochran, 1963; Kelley, 1976; Sokal & Rohlf,
    1981).  All three techniques have a common purpose, namely that the
    samples are representative for the population or compartment to be
    sampled.

         In simple random sampling the selection process is totally,
    unconditionally random.  The disadvantage of this method lies in the
    possibility that the samples may unintentionally be clumped together,
    while other parts of the population or compartment may not be sampled
    at all.  This may be particularly unfavourable when samples show a
    large variation.

         Distribution problems can be solved by dividing the population or
    compartment into either equal segments, where the investigator selects
    the interval systematically and in which segment a sample is to be
    taken, or into segments that are unequal in size or number ("strata"),
    and where at least one sample is taken in each segment.  The first
    situation (systematic sampling) is unfavourable in cases where, during
    repeated sampling, the functions to be monitored vary, perhaps
    periodically.  In the second case (stratified random sampling), some
    information on functions that are monitored must be available before
    the study starts and has to be used in designing the strata.  Within
    each stratum the samples are selected randomly.  The sampling size can
    be adjusted for each segment of the population or compartment.

         Thus, the quality and value of the data obtained by the
    monitoring exercise is governed to a great extent by the selected
    sampling technique.  For example, if the effects of the so-called acid
    rain on forest ecosystems should be monitored, simple random sampling
    might not lead to valid results, since the effects might vary
    considerably between, for instance, deciduous and coniferous forest
    systems.

         All these aspects demonstrate the extent to which a well-
    designed study influences the quality of the results.

    3.4  Sampling procedures and documentation

         The parameters to be measured or observed must be specified by
    SOPs and/or the study plan.  Where and by which means the samples are
    taken and how measurements and observations are carried out should be
    described.  For instance, if concentrations of a substance are
    monitored in natural aquatic ecosystems, it is necessary to establish
    the volume of water to be sampled, from which depth the samples must
    be taken, and which device should be used.  For blood samples the time
    of sampling can be of great importance.

         For collecting human samples, new syringes and containers should
    preferably be used, but for environmental samples this may not always
    be feasible.  In any event, clean sampling devices and containers
    should always be used.  In order to avoid contamination of samples,
    the necessary cleaning procedures for sampling devices and containers
    must be specified.  During sample collection the sampling devices can
    contribute to chemical contamination due to the use of improper or
    unstable material as well as improper cleaning procedures.  The SOPs
    should state clearly the material to be used for sampling and storage
    of samples, as well as the cleaning procedures.  The
    cross-contamination of biological material can also occur, e.g., a net
    sample with planktonic organisms, where organisms from a previous
    sample are introduced into another because the net was not properly
    rinsed.  In this example, if the species composition is monitored,
    biased results can be expected.  Thorough rinsing would avoid this. 
    In the case of chemical contamination, cleaning procedures such as
    acid washing may be useful.

         Containers and chemicals may be potential sources of
    contamination, for instance with trace metals.  Special cleaning
    procedures are needed to minimize this problem.

    3.5  Handling of samples

         The samples obtained during a study are normally shipped from 
    the field into the laboratory where they are to be evaluated.  Often
    there is a considerable delay between sampling and arrival at the
    laboratory where the samples and specimens are to be stored.

         For many pollutants there is a great risk of introducing errors
    during sampling, sample handling and chemical analysis.  The risk of
    contamination of the sample during sampling and sample handling is
    particularly great for samples with low concentrations of pollutants
    ubiquitous in the environment or present in materials and tools coming
    into contact with the sample.  For example, a blood sample of 1 ml
    normally contains as little as 0.1-0.5 ng cadmium, and it is obvious
    that considerable measures have to be taken in order to avoid
    contamination.  Tobacco smoke often contains cadmium at levels that
    may seriously contaminate the blood samples.  Therefore, all sample

    handling must be carried out in rooms where smoking is prohibited, and
    preferably by non-smokers.

         Concentration changes during storage may be due to precipitation
    or evaporation of the analyte or to evaporation of the solvent. 
    Adsorption of the analyte to the container may also occur.  Changes in
    the sample matrix, e.g., clotting of blood, may change the
    concentration of the chemical to be measured.  Often the samples have
    to undergo a series of preparation steps before the determination of
    the concentration of a pollutant is carried out. Mistakes in the
    dilution, weight determination and calculations may appear.  Chemicals
    added to the samples may be contaminated with the substance under
    study.

         The use of blanks is the most common analytical tool for
    controlling contamination.  Blanks can be used in the laboratory in
    the form of instrument blanks, calibration blanks and reagent blanks. 
    In the field, matched-matrix blanks (determination of contamination
    during sample collection, handling, storage and transport) are used to
    simulate the sample matrix and are carried through all processes. 
    Blanks of the media used in sampling (like filters, nets, traps and
    containers) and of the sampling devices (by collecting the rinsing
    media) should also be used.

         Appropriate provision must be made and measures taken in order to
    keep the samples under the conditions required to maintain the
    character of the samples.  The stability of the chemical to be
    analysed, in terms of, for instance, physical, chemical or biological
    degradation, must be known.  If there are special storage conditions
    required, adherence to these conditions must be controlled and
    documented.

         One means of controlling the stability of the samples for
    chemical analysis is the use of spiked samples.  These samples contain
    the same matrix as the environmental sample but without the respective
    analyte, which is then added in a known quantity.  The spike samples
    are then handled and stored in the same way and under the same
    conditions as the environmental samples and are concomitantly
    analysed.  It should be noted that the integrity of volatile fractions
    in samples is particularly difficult to maintain.

         The stability of biological material is typically ensured by the
    use of preservation materials or processes, but proper and appropriate
    preservation methods must be used so that very fragile structures,
    such as single cells in plankton samples, are not damaged or
    destroyed.  This is necessary to ensure that measurements or
    observations of the samples are representative.

    3.6  Analytical performance evaluation

         The main objective of the analytical performance evaluation is to
    assess the accuracy of data.  This evaluation should be published
    along with the data so that users can assess the quality.  The
    following techniques are available for analytical performance
    evaluation:

    *    external quality control programmes;

    *    comparison with results using an independent technique;

    *    comparison with results of a reference laboratory;

    *    analysis of commercial standard reference materials;

    *    analysis of spiked samples prepared at the laboratory.

         Ideally, the analytical performance evaluation should be
    coordinated by an external quality assurance coordinating centre or by
    the quality assurance unit of the laboratory.  The coordinators
    provide the laboratory with external quality control (EQC) samples,
    the concentrations of the substance being unknown to the analyst.  The
    EQC samples should be analysed together with the collected samples,
    and the results should be evaluated by the quality assurance
    personnel.

         If it is not possible to have the analytical performance
    evaluated by an external quality assurance personnel, duplicate
    samples should be analysed using other analytical techniques,
    preferably at a reference laboratory.  If other techniques are not
    available, duplicate samples should be analysed at a reference
    laboratory.

         Since the certified concentrations of standard reference
    materials are published, and this is known to the analyst, analysis of
    such samples alone is often not enough for independent evaluation of
    the accuracy of the resulting data.  Correct results for the internal
    quality control (IQC) samples do not always guarantee accurate results
    for the EQC samples.  Also, standard reference materials are often
    only available at one or two concentrations.  The quality control
    samples should cover the range of concentrations likely to be found in
    the monitoring samples.  Good analytical performance at one
    concentration is no guarantee of good performance at other
    concentrations.

         The analytical performance evaluation should be carried out at
    the time of the analysis of the monitoring samples.  Reference to
    previous participation in interlaboratory comparison programmes cannot
    be used for evaluating the accuracy of the data produced.

    3.7  The regression method

         Unfortunately it is not possible to measure of the accuracy of
    data produced, but the limits of uncertainty can be estimated.
    Principles for analytical quality assessment in human exposure
    monitoring, giving limits for the uncertainty, have been recommended
    by WHO (1986b).  The method is based on the quality control programmes
    developed in the UNEP Biological Monitoring of Lead and Cadmium
    (Vahter, 1982; Friberg & Vahter, 1983; Vahter & Slorach, 1990). 
    Basically, it involves the analysis of sets of quality control samples
    and evaluation of the results using linear regression analysis.  In
    the WHO/UNEP Human Exposure Assessment Locations (HEAL) monitoring
    programme for lead and cadmium, a quality control set consisted of 3-6
    external EQC samples, the metal concentrations of which were not known
    to the laboratories, as well as 1-2 IQC samples, the metal
    concentrations of which were given.  The IQC samples are used for the
    analyst's own control of the analytical conditions, while the EQC
    samples are used for the analytical performance evaluation.  One or
    more sets of quality control samples were analysed together with the
    monitoring samples, and the results were evaluated by a coordinating
    centre.

         The regression method, i.e. the evaluation of the regression line
    of reported versus "true" values for a set of quality control samples
    analysed together with the monitoring samples, is a useful method of
    guarding against systematic errors in the whole range of
    concentrations likely to occur (Vahter, 1982; UNEP/WHO, 1984; Friberg,
    1988).  The regression line represents the average analytical
    performance.

         In order to obtain limits for the uncertainty of the data
    produced, the maximum allowable deviation (MAD) of the empirical
    regression line from the ideal line  y =  x is defined.  The MAD
    criteria have to be decided separately for each pollutant and for each
    medium.  In the UNEP/WHO HEAL project on lead and cadmium, the MAD was
    generally set to  (5-10%  sigma), where sigma was the estimated
    error of the method, based on several quality control runs.

         Since the regression line, based on the results of a set of
    quality control samples, has an operating error, the decision on
    acceptance or rejection of the regression line must be based on
    statistical criteria, i.e. the probability of making right or wrong
    decisions.  A laboratory's results may be erroneously rejected when in
    fact the laboratory performance is satisfactory, or erroneously
    accepted when the performance is bad.  Table 1 illustrates the
    different decisions that can be made on the basis of the results of
    the quality control analyses, and the associated probabilities
    (UNEP/WHO, 1984).

        Table 1.  Decision-making on the basis of quality control resultsa
                                                                             
    True condition                        Decision

                             Acceptance               Rejection
                                                                             
    Methodology is           correct decision         wrong decision
    satisfactory             (1-alpha)                Type I error (alpha)

    Methodology is not       wrong decision           correct decision
    satisfactory             Type II error ()        Power (1-)
                                                                             

    a  From: UNEP/WHO (1984)
    
         In the HEAL project, a total power of 90% was employed, which
    means that the probability of accepting an unsatisfactory performance
    (the true regression line falling outside the MAD interval) was not
    more than 10%.  For acceptance of data the empirical regression lines
    had to fall not only inside the MAD interval, but also inside an
    acceptance interval.

         Fig. 1 shows an example of a regression line of the results of
    six quality control samples, the MAD interval (solid lines) and the
    acceptance interval (broken lines).  The distance between the MAD
    lines and the acceptance lines is 1.645 times the operating error,
    sigma, calculated according to the formula:

    sigma2 = sigma2y/x (1 +      d2      )
                        n   (n-1).sigma2x

    where

    n = number of observations

    d = difference between x value and x mean

    sigmax = standard deviation of x values

    sigmay/x = error of method or residual deviation (estimated from
    previous analyses)

    FIGURE 1

         It is obvious from the formula that the acceptance interval (AI)
    lines will get closer to the MAD lines with increasing numbers of data
    points (quality control samples).  Also, a smaller error of method
    will decrease the difference between the MAD lines and the AI lines.

         The random error of method may be calculated from the results of
    each quality control set.  However, with quality control sets
    consisting of only 4-6 samples, the empirical error of method may be
    largely influenced by one or two occasional gross errors.  Therefore,
    it is good practice to estimate the error of method based on previous
    experience.  It is nevertheless important to calculate the empirical
    error of method, since it may serve as a supplementary guidance in the
    evaluation of the analytical performance.  When the current random
    error of method deviates too greatly from the estimated one, this is
    an indicator of bad performance, and the evaluation should not be
    based on the estimated error of method.

         For many types of chemical measurements, the error variance tends
    to vary with the true concentration.  Variance-stabilizing
    transformations, e.g., logarithmic ( z = ln  x ) or square-root
    ( z =  x ) transformation, of the data may make the error variance
    independent of the true concentration (Starks, 1989).  In the HEAL
    nitrogen dioxide project, where many quality control samples were
    used, the quality control results were considered satisfactory if the
    regression line, the 90% confidence interval and all data points were
    inside the MAD interval (Matsushita & Tanabe, 1991).

         The regression method may give valuable information concerning
    the type of error.  For example, a regression line parallel to the
    ideal line ( y =  x ) indicates an absolute error caused by, for
    instance, a false blank value.  A slope deviating from 1.0 indicates
    a constant relative error due to, for instance, incorrect standards or
    errors in the concentrations of the standards.

    3.8  Practical application of the regression method

         The regression method for analytical performance evaluation was
    developed for use in a WHO study on the assessment of human exposure
    to lead and cadmium through biological monitoring.  Since then, the
    method has been used in a number of studies, involving both metals
    (Lind et al., 1987, 1988a,b; Zheng & Ji, 1987; Vahter & Slorach, 1990)
    and nitrogen dioxide (Matsushita & Tanabe, 1991).

    3.9  Other analytical performance evaluation programmes

         There are several other methods for external analytical
    performance assessment (ISO, 1986b; Horwitz, 1988).  Some examples are
    given below.  Most of these methods are descriptive and do not give
    acceptability limits for possible errors.

         The WHO European Regional Study on Health Effects of Exposure to
    Cadmium, coordinated by the Coronal Laboratory for Occupational and
    Environmental Health, was specially developed for countries
    participating in the United Nations Development Programme.  A quality
    control programme was developed for the determination of lead and
    cadmium in blood, as well as cadmium, 2-microglobulin and retinol
    binding protein in urine (Herber, 1990).  In the metal programme,
    30-40 laboratories each analysed six samples of human blood and six
    samples of human urine.  The samples contained cadmium and lead
    nitrate at levels up to about 20 g/litre for cadmium and up to 800
    g/litre for lead.

         Regarding the proteins in urine, there were only six or seven
    participating laboratories, and another, more simple, evaluation
    procedure was used.  Urine from patients with kidney disease was
    diluted with urine with normal physiological protein concentrations.

    The acceptance criteria were  20% of the median for albumin and  40%
    of the median for 2-microglobulin and retinol binding protein.

         The Guildford Trace Element Quality Assessment Scheme,
    coordinated by Robens Institute of Industrial and Environmental Health
    and Safety and St. Luke's Hospital, Guildford, United Kingdom, is an
    external quality assessment scheme for trace elements in human
    biological fluids (Taylor et al., 1985).  More than 100 laboratories
    from over 15 countries participate in the scheme.  The programme
    includes several combinations of analytes and biological samples. 
    Every month during a six-month cycle, three specimens of each matrix
    are sent to the participants.  Monthly reports include consensus mean,
    standard deviation, relative standard deviation, a histogram of
    distribution and a tabulation of the results.  At the end of the
    six-month cycle, the 18 results for each sample-matrix analyte are
    summarized and the analytical performance is assessed based on the
    proximity to the consensus mean, the difference between results
    analysed on two occasions and the recovery of added analyte.  A
    "performance score" is calculated for each individual laboratory. 
    Targets or markers of satisfactory performance have been established
    based on what is necessary for clinical purposes and what can be
    achieved with available analytical techniques.

         The Guildford external quality assessment (EQA) programme for
    aluminium in serum has been in operation since 1981 (Taylor, 1988). 
    Samples of horse serum, supplemented with known amounts of aluminium,
    are prepared for a series of 6-monthly cycles.  During a cycle, nine
    different specimens are distributed in duplicate, three samples each
    month.  The results are evaluated as described above.  The programme
    has shown poor performance in large numbers of laboratories and
    excellent performance in a few laboratories.  In an attempt to explain
    this, an evaluation of instrumentation and methodologies was carried
    out.  It showed that the quality of the results was not influenced by,
    for example, specific features of equipment or instrumentation but
    rather by good, careful analysts, who were able to ensure that
    everything was properly set up and that contamination was avoided.

         The National External Quality Assessment Scheme (NEQAS),
    coordinated by the Wolfson Research Laboratory in Birmingham, is
    another external quality assessment programme in the United Kingdom
    (Bullock & Wilde, 1985).  Samples are distributed to the participating
    laboratories for analysis, and based on the results obtained, a
    variance index (VI) is calculated.  The VI is the difference between
    the result obtained and a trimmed mean, divided by a chosen relative
    standard deviation for the analyte and expressed as a percentage.  A
    Mean Running Variance Index Score (MRVIS) and a Mean Running Bias
    Index Score (MRBIS) are also calculated.  It is believed that a MRVIS
    below 33 should be the target for the participants, and that any MRVIS
    over 66 should stimulate investigation of the laboratory's method.

    3.10  Analytical performance criteria

         The limits of the uncertainty in the produced data should be
    based on the accuracy requirements for the monitoring data and what
    can be achieved with available analytical techniques.  The accuracy
    and precision of routine and reference methods should be evaluated in
    order to determine the feasibility of the proposed criteria.

         Often the criteria for the analytical performance are determined
    by the sensitivity and accuracy of the analytical method.  In the
    WHO/UNEP monitoring project on cadmium and lead in blood, the limits
    for the MAD lines for lead in blood were set to
     y =  x  (0.1 x + 20).  This means that if the average
    concentration of lead in blood in a group of people was found to be,
    for example, 100 g/litre, the criteria guaranteed with a probability
    of 90% that the true average concentration was somewhere between 70
    and 130 g/litre.  If the average blood lead concentration found was
    50 g/litre, the criteria guaranteed that the true average was between
    25 and 75 g/litre.  With such a great "uncertainty" in the monitoring
    data, it is of course difficult to detect differences in lead exposure
    between various groups in the general population.

         The experience from the quality control activities in the HEAL
    pilot project on lead and cadmium shows that, for lead concentrations
    in blood (g/litre), experienced laboratories can meet the MAD
    criteria  y =  x  (0.05 x + 10), which guarantees that an obtained
    mean blood lead concentration of 50 g/litre lies with 90% probability
    between 37.5 and 62.5 g/litre. For concentrations of cadmium in
    blood, well-experienced and well-equipped laboratories met the MAD
    criteria  y =  x  (0.05 x + 0.2), which guarantees that an
    obtained mean blood cadmium concentration of, for example, 0.5
    g/litre lies with 90% probability between 0.28 and 0.72 g/litre.

    3.11  Quality control samples

         The analytical procedures and performance may vary considerably
    between various types of samples.  Good analytical performance for a
    pollutant in one type of medium is no guarantee of good performance
    with other media.  Thus, it is important for quality control samples
    to have a matrix similar to that of the monitoring samples.

         As already mentioned above, the quality control sample should
    cover the range of concentrations likely to be found in the monitoring
    samples.  It must be emphasized that good analytical performance at
    one concentration is no guarantee for good performance at other
    concentrations.

         Various types of reference samples are commercially available
    (Muntau et al., 1983; Belliardo & Wagstaffe, 1988; Klich & Caliste,
    1988; Okamoto, 1988; Parr et. al., 1988; Rasberry, 1988). However,
    many of the commercially available reference materials are certified

    for a limited number of substances, and usually only for one or two
    different concentrations.  Commercially available reference materials
    are suitable for internal quality control purposes but cannot, as a
    rule, be used for published data.

         If possible, specimens of the samples should finally be archived
    in order to prove again the validity of an environmental monitoring
    study, if it should be necessary.  This may be appropriate where the
    samples are still integral after evaluation.

         Biological tissue is typically contained in slides or blocks,
    while specimens are stored in containers with appropriate fixation. 
    Aliquots from the various media for chemical analysis may be stored
    after appropriate preservation, carrying a label with an expiry date
    based on stability analysis.  All written documentation including the
    raw data is also transferred to the archive at the conclusion of the
    study.

    REFERENCES

    AOAC (1984) Quality assurance principles for analytical laboratories.
    Washington, DC, Association of Official Analytical Chemists, 224 pp.

    Belliardo JJ & Wagstaff PJ (1988) BCR reference materials for food and
    agricultural analysis: An overview. Fresenius Z Anal Chem, 332(6):
    533-538.

    Black SC (1988) Defining control sites and blank sample needs. In:
    Keith LH ed. Principles in environmental samples. Washington, DC,
    American Chemical Society, pp 109-118.

    Bruaux P & Svartengren M ed. (1985) Global Environmental Monitoring
    System (GEMS). Assessment of human exposure to lead: Comparison
    between Belgium, Malta, Mexico and Sweden. Stockholm, National Swedish
    Institute of Environmental Medicine and Department of Environmental
    Hygiene, Karolinska Institute, and Brussels, Institute of Hygiene and
    Epidemiology, 57 pp (Prepared for United Nations Environment Programme
    and World Health Organization).

    Bruaux P, Claeys-Thoreau F, Ducoffre G, & Lafontaine A (1979)
    Exposition saturnine de la population bruxelloise. Choix des groupes
    de population. Choix des indicateurs.  Premiers rsultats. Arch Belg
    Md Soc, 37: 488-501.

    Bullock DG & Wilde CE (1985) External quality assessment of urinary
    pregnancy oestrogen assay: further experience in the United Kingdom.
    Ann Clin Biochem, 22: 273-282.

    Burck K (1979) The role of quality assurance in good laboratory
    practices. Clin Toxicol, 15: 627-640.

    Cochran WG (1963) Sampling techniques. New York, John Wiley & Sons, 13
    pp.

    FDA (1987a) Good laboratory practice regulations. Fed Reg, 52(172):
    33768 (21 CFR Part 58) (Docket No. 83N-0142).

    FDA (1987b) Part 58 - Good laboratory practice for non-clinical
    laboratory studies. Code Fed Regul, 21(Ch1): 228-242.

    Friberg L (1988) Quality assurance. In: Clarkson TW, Friberg L,
    Nordberg GF, & Sager PR ed. Biological monitoring of toxic metals. New
    York, Plenum Press, pp 103-126.

    Friberg L & Vahter M (1983) Assessment of exposure to lead and cadmium
    through biological monitoring. Results of a UNEP/WHO global study.
    Environ Res, 30: 95-128.

    Herber RFM (1990) The World Health Organization study on health
    effects of exposure to cadmium. Toxicol Environ Chem, 27: 1-10.

    Herber RFM & Schaller KH (1986) Analytical variability of biological
    parameters of exposure and early effects. In: Notten WRF, Herber RFM,
    Hunter WJ, Monster AC, & Zielhuis RL ed. Health surveillance of
    individual workers exposed to chemical agents. Berlin, Springer
    Verlag, pp 102-109.

    Horwitz W (1988) Protocol for the design, conduct and interpretation
    of collaborative studies: IUPAC Workshop on the Harmonization of
    Collaborative Analytical Studies, Geneva, Switzerland, 4-5 May 1987.
    Pure Appl Chem, 60: 855-864.

    Horwitz W (1989) Harmonization of methodology: international
    perspective. In: Cowgill UM & Williams LR ed. Aquatic toxicology and
    hazard assessment. Philadelphia, American Society for Testing and
    Materials, pp 5-10 (Special Technical Publication No. 1027).

    ISO (1986a) International Standard 8402. Quality - Vocabulary. Geneva,
    International Organization for Standardization, 15 pp.

    ISO (1986b) International Standard 5725: Precision of test methods -
    Determination of repeatability and reproducibility for a standard test
    method by inter-laboratory tests: 2nd edition. Geneva, International
    Organization for Standardization, 49 pp.

    ISO (1987a) International Standard 9000. Quality management and
    quality assurance standards: guidelines for selection and use. Geneva,
    International Organization for Standardization, 6 pp.

    ISO (1987b) International Standard 9004. Quality management and
    quality system elements - Guidelines. Geneva, International
    Organization for Standardization, 19 pp.

    Kelley JC (1976) Sampling the sea In: Cushing DH & Walsh JJ ed. The
    ecology of the seas. Oxford, Blackwell Scientific Publications, pp
    361-387.

    Klich H & Caliste JP (1988) COMAR - database for certified reference
    materials.  Fresenius Z Anal Chem, 332(6): 552-555.

    Lepore PD (1979) FDA's good laboratory practice regulations. Pharm
    Technol, 3: 71-74.

    Lind B, Elinder C-G, Friberg L, Nilsson B, Svartengren M, & Vahter M
    (1987) Quality control in the analysis of lead and cadmium in blood.
    Fresenius Z Anal Chem, 326: 647-655.

    Lind B, Vahter M, Rahnster B, & Bjors U (1988a) Quality control
    samples for the determination of lead and cadmium in blood, feces, air
    filters and dust. Fresenius Z Anal Chem, 332(6): 741-743.

    Lind B, Bigras L, Cernichiari E, Clarkson TW, Friberg L, Hellman M,
    Kennedy P, Kirkbride J, Kjellstrom T, & Ohlin B (1988b) Quality
    control of analyses of mercury in hair. Fresenius Z Anal Chem, 332(6):
    620-622.

    Matsushita H & Tanabe K (1991) Earthwatch - Global Environmental
    Monitoring System (GEMS). Exposure monitoring of nitrogen dioxide. An
    international pilot study within the WHO/UNEP Human Exposure
    Assessment Location (HEAL) Programme. Nairobi, United Nations
    Environment Programme, 109 pp.

    Muntau H, Schramel P, Brtter P, & Knapp G (1983) Trace element
    investigations on some new biological test materials. In: Brtter P &
    Schramel P ed. Trace element - Analytical chemistry in medicine and
    biology. Berlin, New York, Walter de Gruyter & Co., vol 2, pp 819-831.

    OECD (1982) Good laboratory practice in the testing of chemicals.
    Paris, Organisation for Economic Co-operation and Development, 62 pp.

    Okamoto K (1988) Biological reference materials from the National
    Institute of Environmental Studies (Japan). Fresenius Z Anal Chem,
    332(6): 524-527.

    Parr RM, Schelenz R, & Ballestra S (1988) IAEA biological reference
    materials. Fresenius Z Anal Chem, 332(6): 518-523.

    Rasberry SD (1988) NBS activities in biological reference materials.
    Fresenius Z Anal Chem, 332(6): 528-532.

    Slorach SA & Vaz R ed. (1983) Global Environmental Monitoring System
    (GEMS). Assessment of human exposure to selected organochlorine
    compounds through biological monitoring. Uppsala, Swedish National
    Food Administration, 134 pp (Prepared for United Nations Environment
    Programme and the World Health Organization).

    Smith F, Kulkarni S, Myers LE, & Messner MJ (1988) Evaluating and
    presenting quality assurance sampling data. In: Keith LH ed.
    Principles in environmental samples. Washington, DC, American Chemical
    Society, pp 157-168.

    Sokal RR & Rohlf FJ (1981) Biometry: The principles and practice of
    statistics in biological research, 2nd ed. New York, Oxford, W.H.
    Freeman, 859 pp.

    Starks TN (1989) Global Environmental Monitoring System (GEMS).
    Statistical guidelines for quality assurance in human exposure
    assessment location studies. Geneva, World Health Organization, 48 pp
    (Document PEP/89.9).

    Taylor A (1988) Reference materials for measurement of aluminium in
    biological samples.  Fresenius Z Anal Chem, 332(6): 616-619.

    Taylor JK (1987) Quality assurance of chemical measurements. Chelsea,
    Michigan, Lewis Publishers, 328 pp.

    Taylor JK (1988a) Defining the accuracy, precision and confidence
    limits of sample data. In: Keith LH ed. Principles in environmental
    samples. Washington, DC, American Chemical Society, pp 101-108.

    Taylor JK (1988b) The role of statistics in quality assurance.
    Fresenius Z Anal Chem, 332(6): 722-725.

    Taylor A, Starkey BJ, & Walker AW (1985) Determination of aluminium in
    serum:  findings of an external quality assessment scheme. Ann Clin
    Biochem, 22: 351-358.

    UNEP/WHO (1984) Principles and procedures for quality assurance in
    environmental pollution exposure monitoring. Geneva, World Health
    Organization, 31 pp (EFP/HEAL/84.4).

    Vahter M (1982) Global Environmental Monitoring System (GEMS).
    Assessment of human exposure to lead and cadmium through biological
    monitoring. Stockholm, National Swedish Institute of Environmental
    Medicine and Department of Environmental Hygiene, Karolinska
    Institute, 136 pp (Prepared for United Nations Environment Programme
    and World Health Organization).

    Vahter M & Slorach S ed. (1989) Global Environmental Monitoring System
    (GEMS). Exposure monitoring of lead and cadmium: An international
    pilot study within the UNEP/WHO Human Exposure Assessment Location
    (HEAL) Programme. Nairobi, United Nations Environment Programme, 82
    pp.

    Vahter M & Slorach S ed. (1990) Global Environmental Monitoring System
    (GEMS). Exposure monitoring of lead and cadmium: An international
    pilot study within the WHO/UNEP Human Exposure Assessment Location
    (HEAL) Programme. Nairobi, United Nations Environment Programme, 82
    pp.

    WHO (1986a) Environmental Health Criteria 57: Principles of
    toxicokinetic studies. Geneva, World Health Organization, pp 20-24.

    WHO (1986b) UNEP/WHO Global Environmental Monitoring System (GEMS).
    Human Exposure Assessment Location (HEAL) Project: Guidelines for
    integrated air, water, food and biological exposure monitoring.
    Geneva, World Health Organization, 389 pp (Document PEP/86.6).

    WHO (1992) Earthwatch - Global Environmental Monitoring System (GEMS).
    Guidance on survey design for Human Exposure Assessment Locations
    (HEAL) Studies. Geneva, World Health Organization (Document
    WHO/PEP/92.6).

    Youden WJ (1967) The role of statistics in regulatory work. J Assoc
    Off Anal Chem, 50: 1007-1013.

    Youden WJ & Steiner EH (1975) Statistical manual of the Association of
    Official Analytical Chemists. Washington, DC, Association of Official
    Analytical Chemists, 88 pp.

    Zheng XQ & Ji RD (1987) Assessment of lead contamination of the
    general population through blood lead levels. Environ Monit Assess, 9:
    169-177.

    APPENDIX I

    GLOSSARY

    Batch: A specific quantity or lot of a test or reference substance
    produced during a defined cycle of manufacture, in such a way that it
    could be expected to be of a uniform character and should be
    designated as such.

    Carrier (vehicle): Any agent that serves as a vehicle used to mix,
    disperse, or solubilize the test or reference substance to facilitate
    the administration to the test system.

    Calibration: Assigning values to the output of an instrument or
    other device.  Calibration can apply to instruments, glassware or any
    other measuring device.

    Calibration verification (calibration check): Determining if the
    output of an instrument or device is within some established
    tolerance.

    Code number: A unique number assigned to each packaged unit of the
    test substance.

    Final report: A comprehensive report that includes all information
    relative to the evaluation of the results of a specific nonclinical
    laboratory study.

    Laboratory technician: A laboratory worker who performs an operation
    covered by an SOP, under the supervision of a professional scientist.

    Master schedule sheet: A document that contains essential
    information on completed and in-progress studies in the laboratory,
    i.e. a list of studies, initiation and completion dates, test systems,
    route of administration and name of study director.

    Quality assurance programme (QAP): An internal control system of
    inspections and study audits designed to ascertain that a study is in
    compliance with defined guidance principles.  It assures laboratory
    management that facilities, equipment, personnel, methods, practices,
    records and controls conform with these principles.

    Quality assurance unit (QAU): Any person or organizational element,
    designated by testing facility management to perform inspections of
    laboratory operations and study audits relating to quality assurance
    of studies.

    Quality control system: A system used to assess product quality
    whenever the quality attributes of the final product can be expressed
    as discretely measured parameters.

    Raw data: All original laboratory records and documentation, or
    verified copies thereof, which are the result of the original
    observations and activities in a study.

    Sample: Any quantity of the test or reference substance.

    Specimen: Any material derived from a test system for examination,
    analysis or storage.

    Sponsor: A person or entity who commissions and/or supports a study. 
    A testing facility can also be a sponsor if it both initiates and
    actually conducts the study.

    Standardization: Determining the response of an instrument to known
    quantities of a test agent according to a specific method; accurately
    determining the concentration of a solution.

    Standardization verification (standardization check): Determining if
    the response of an instrument to a particular standard is within some
    established tolerance (e.g., verifying that a previously prepared
    standard curve (or other procedure) is still valid).

    Standard operating procedure (SOP): A written procedure that
    describes how to perform certain routine laboratory tests or
    activities that are normally not specified in detail in study plans or
    test guidelines.

    Study: An experiment or set of experiments in which a test substance
    is examined to obtain data on its properties and/or its safety with
    respect to human health and the environment.

    Study audit: A comparison of the raw data and associated records
    with the interim or final report in order to determine whether the raw
    data was accurately reported, and whether testing was carried out in
    accordance with the study plan and SOPs to obtain additional
    information not provided in the report, and to establish whether
    practices were employed in the development of data that would impair
    their validity.

    Study director: The individual responsible for the overall conduct
    of the study.

    Study plan: A protocol which defines the entire scope of the study.

    Test facility (testing facility): The persons, premises and
    operational unit(s) that are necessary for conducting the study.

    Test facility management: Consists of the executive level of
    management charged with the ultimate responsibility of the test
    facility and studies.

    Test substance: A chemical substance or a mixture that is under
    investigation.

    Test system: Any human, other animal, plant, microbial, as well as
    other cellular, subcellular, chemical or physical system or
    combination of these, that is exposed to a test, control or reference
    substance.

    Validation: Establishing documented evidence that provides a high
    degree of assurance that the intended use of such things as a
    procedure, test system, test substance or control substance is
    accomplished.

    Validation procedure(s): A written procedure stating how validation
    will be conducted.  Validation procedures should exist for all
    elements of the test system and should specify the procedures and
    tests that will be conducted and the data to be collected.  Examples
    include the validation procedure used for animal supply vendors,
    facility equipment, test substances control substances, and SOPs.


    See Also:
       Toxicological Abbreviations