United States
                                 Environmental Protection
                                 Agency
                                 Health Effects Research
                                 Laboratory
                                 Research Triangle Park NC 27711
v>EPA
                                 Research and Development
                                 EPA-600/S2-81-054 May 1981
Project  Summary
                                 A  Plan to  Develop and
                                 Implement  a  Quality
                                 Assurance   Program  for the
                                 Ames/Salmonella Test
                                 N. Sexton, L Myers, and T. Hughes
                                  This document discusses a plan to
                                 develop and carry out a quality assur-
                                 ance program for the Ames/Salmo-
                                 nella test. The Ames test is in one of
                                 the 13 categories of tests mentioned
                                 in the Federal Insecticide, Fungicide,
                                 and Rodenticide Act legislation and is
                                 chosen as a model for presenting a QA
                                 program for all test categories.
                                  There are three main sections con-
                                 tained herein. An introductory section
                                 describes the laws requiring this test-
                                 ing and the objectives of this quality
                                 assurance plan.  The second  section
                                 defines   quality assurance  terms,
                                 describes applicable quality assurance
                                 methods, and discusses the mailable
                                 audit materials  approach that  was
                                 chosen for these purposes. The third
                                 section  details  the  mailable audit
                                 materials  program  and discusses
                                 selection of chemicals, preparation of
                                 samples, shipment of audit samples,
                                 pilot-scale audits, scheduling, other
                                 audit programs as  models, data
                                 gathering procedures, data analysis,
                                 reporting of results, and costs.
                                  This Project Summary was develop-
                                 ed by EPA's Health Effects Research
                                 Laboratory, Research  Triangle Park.
                                 NC. to announce key  findings of the
                                 research project that is fully docu-
                                 mented in a separate report of  the
                                 same title (see Project Report ordering
                                information at back).
                                 Introduction
                                  The 1972 Federal Insecticide, Fungi-
                                 cide, and Rodenticide Act (FIFRA) speci-
                                 fied that no pesticide be registered or
                                 reregistered that would cause unrea-
                                 sonable risk to man or to the environ-
                                 ment  After subsequent promulgation
                                 and public comments, tentative guide-
                                 lines were issued containing 13 kinds of
                                 tests  grouped into  three categories.
                                 gene mutations,  chromosomal aberra-
                                 tions, and  primary DNA  damage.
                                 Because of its wide usage, the Ames/
                                 Salmonella test,  a test for gene muta-
                                 tion, was selected as a model for the
                                 preparation of a quality assurance (QA)
                                 plan.


                                 Quality Assurance Strategies
                                  The stated objective of this QA plan is
                                 to present a means of ensuring that
                                 laboratories submitting data to  the
                                 Office of Pesticide Programs concerning
                                 pesticide  registration  can  produce
                                 information that  is reliable and valid
                                 After consideration and discussion of
                                 the various means of implementing a
                                 QA plan, a mailable audit materials
                                 approach was selected because:

                                  1.  Many laboratories are submitting
                                    data, therefore, onsite evaluations
                                    would  be  costly  and time
                                    consuming.

-------
 2   Neither  systems audits nor split-
     sample schemes would provide an
     evaluation of a laboratory's accu-
     racy. In the case of split-sample
     schemes, where there were large
     variances   among   laboratories
     there may be no useful informa-
     tion obtainedforthistask Further,
     there is no practical way to insert
     blind samples  into a laboratory's
     analysis stream.

 3.  There  is   no   widely accepted
     standard method for all aspects of
     performing the Ames/Salmonella
     test  Parts of the systems audit are
     geared  toward   comparing  an
     individual   laboratory's perform-
     ance to well-established  stand-
     ards and protocols This lack of a
     validated  protocol  would  also
     preclude  a formal  certification
     program
The Mailable Audits Program
  Factors considered m  the  mailable
audits   program  were   selection  of
chemicals,  preparation  of   samples,
shipment of samples, pilot-scale audits,
data gathering procedures, analysis of
data,  schedule of  audits, reporting of
results,  and  costs  These  topics  are
discussed in subsequent sections
Selection of Chemicals
  Some  chemicals  giving  a positive
(mutagenic) response and some chem-
icals giving a negative response should
be  included as  audit  materials.  The
physical  characteristics of the chemi-
cals should be known  and should be
compatible with a mailable materials
program.  A  proposed  set of audit
materials would include-

  •  A direct-acting (no metabolic acti-
     vation required) compound in a
     solvent

  •  An indirect-acting compound in a
     solvent

  •  A mixture of direct-acting com-
     pounds in solvent(s) for activating
     only frameshift or only base-pair
     sensitive bacterial strains

  •  A mixture of direct-acting com-
     pounds   that  elicit   a  positive
     response  in  all  five  bacterial
     strains
  •  A mixture of pure compounds that
     require metabolic activation

  •  A  complex  mixture/environ-
     mental sample

  •  A mixture that requires chemical
     fractionation  to  show  mutagen-
     icity.

  •  Nonmutagemc materials

  The importance  of preaudit verifica-
tion of audit materials was emphasized.
Preparation of Samples
  American  Society  of  Testing  and
Materials (ASTM) methods could be
used to ensure the homogeneity of solid
samples  Choice of  solvents would be
dependent  on  solubility  of  the  test
materials, stability,  and toxicity of the
solvents   to   bacteria   Laboratories
preparing samples should adhere to the
Guides for  Quality Assurance in Envi-
ronmental  Health Research  and the
proposed "good  laboratory practices"
(Federal Register, 11/19/76, 1/7/77,
and 1/28/77)
Shipment of Audit Samples
  Title 49, Code of Federal Regulations,
covers all modes of transport of mate-
rials within the United States, and Parts
100-177 deal with the  transport of
hazardous materials In view of Title 49
regulations, three modes of transporta-
tion were considered: Federal Express,
United Parcel Service  (UPS), and U.S.
Mail  Details of the additional  regula-
tions covering each of the three modes
of transportation, as well as costs, are
discussed in an appendix
Pilot-Scale Audits
  A  pilot-scale  audit of seven labora-
tories was proposed to test clarity  of
instructions,  stability  of  materials
during transport, time required for ship-
ment, applicability of data  entry forms,
and  schemes for data  interpretation  A
consensus among  eight  laboratories
(the  verification laboratory plus seven
others) would constitute a statistically
significant probability (at the 1  percent
level) that the  answers were, in  fact,
correct and  not purely  chance occur-
rences.
Scheduling of Audits
  It  was  proposed  that  laboratone
wishing  to  enter  the  QA  prograr
receive  samples  when they  mitiall
submitted data or when they asked t
participate  Otherwise, all participatin
laboratories would receive a set of audi
samples each year  For laboratories tha
did not show acceptable performance,
second, then a third set of audit sample
would be  provided. After failing thre
attempts,  a  laboratory would  have t
wait one  year before reapplying  fo
evaluation.
Data  Gathering Procedures
  Of the several data systems for bio
assay  information that are currentl
available, the system designed and n
current  use  at  the  Environments
Protection  Agency's  Health   Effect
Research  Laboratory  in   Researcl
Triangle  Park (HERL-RTP) was selectei
as most applicable for this QA program
Data entry forms and instructions an
presented as an  appendix.  A supple
mentary  questionnaire, also presentee
in the  appendixes, was recommende<
for initial evaluation of a laboratory.
Data Analysis
  The HERL-RTP pesticide QA prograrr
served as a model  In this QA prograrr
for the Ames test, five major points wil
be used for evaluation

  1   The   mutagenic/nonmutagenic
     (+/-) determination by the testing
     laboratory as compared with the
     expected  results. The expectec
     results would be based on  pub-
     lished  response  values  for the
     chemicals chosen as audit mate-
     rials and on the preliminary anc
     pilot audit analyses.

  2   The   mutagenic/nonmutagenic
     (+/-) determination by the testing
     laboratory as compared with the
     +/- determination by a reviewer ol
     the  raw  data.  Mathematical
     models have been developed that
     consider all raw data (plate counts
     on controls and samples), factor in
     toxicity, and give an assessment of
     the mutagenic or nonmutagenic
     nature of the sample.

  3   The   mutagenic/nonmutagenic
     (+/-) determination by the testing
     laboratory as compared to the +/-
                                   1

-------
     determination by other laborator-
     ies submitting data

 4   Positive  and  negative  control
     values as reported  by a testing
     laboratory and compared with the
     appropriate  normal  ranges
     Normal ranges for controls will be
     deter mined from values published
     in a  consensus report by NIEHS,
     from valuesdetermmed in the pre-
     liminary  tests  and  pilot  audits
     program,  and  from  a  statistical
     analysis of the values reported by
     all laboratories submitting data

 5   Ranking by the testing  laboratory
     of the set of audit samples, from
     the most  mutagemc to the least
     mutagemc within eachstram/S-9
     combination of bacteria, with and
     without   metabolic   activation,
     compared to a  similar  ranking
     according  to  expected  results.
     Again, expected  results will be
     based  on  determination  from
     preliminary and pilot studies and
     values reported by all laboratories
     submitting data.

  It  is  proposed that the  following
weighting scheme be applied for evalu-
ation of a laboratory's capabilities in the
Ames/Salmonella test.
     Criteria

        1
        2
        3
        4
        5
Points

  50
  25
   5
  10
  10

 100
  The data interpretation  model men-
tioned in point 1 is an extension of the
Poisson  model  developed  by  Stead,
Hasselblad,  Creason, and Claxton of
HERL-RTP. These models are described
in an appendix
Reporting of Results
  The QA program would generate five
types of reports. First,  a report of a
laboratory's raw data, as encoded for
interpretation, would be mailed to that
laboratory. Then an executive summary
of a laboratory's performance, based on
the  previously mentioned  evaluative
criteria,  would  be prepared  for the
Office of Pesticide Programs. Third, an
                annual  report  on all  aspects  of  the
                program would be produced. A fourth
                report would be the evaluative report
                mailed back to submitting laboratories
                Fifth, it  is anticipated that results of
                such a QA program would be of general
                interest to the scientific community and
                could be  published in the appropriate
                technical and scientific journals
                Costs
                  Costs were considered for  prelimi-
                nary . planning  ($27,000),  the  pilot
                program  (2,180  man-hours;  sample
                preparation,  sample  testing,  small
                scale audits),  and the full-scale audit
                program (1,100 man-hours plus $15,000
                computer time, packaging and mailing,
                data compilation and evaluation, report-
                ing,  mailing second and  third sets). For
                270 sample sets and given  $3000 per
                man-hour,  the cost per sample set is
                $520.00.
                   N Sexton, L Myers, and T  Hughes are with the Research Triangle Institute,
                    Research Triangle Park, NC 27709.
                   Larry Claxton is the EPA Project Officer (see below).
                   The complete report, entitled "A Plan to Develop and Implement a Quality
                   Assurance Program for the /Imes/Salmonella  Test," (Order No. PB 81-182 065;
                    Cost. $9.50, subject to change) will be available only from.
                          National Technical Information Service
                          5285 Port Royal Road
                          Springfield,  VA 22161
                          Telephone  703-487-4650
                   The EPA Project Officer can be contacted at:
                          Health Effects Research Laboratory
                          U S. Environmental Protection Agency
                          Research Triangle Park, NC 27711
                                                                                     ! US GOVERNMENT PRINTING OFFICE. 19(1.757-012/7094

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
                      ".

-------