|
BioSafety Journal
Pontificia Universidad Católica de Valparaíso
ISSN: 1366 0233
Vol. 1, Num. 1, 1995
|
BioSafety
Volume 1, Paper 7 (BY95007), June 24th 1995.
Online Journal, URL -http://bioline.bdt.org.br/by
Biosafety performancs of biotechnology equipment:
Consideration of Performance Criteria and of Equipment
Categorisation.
Paul Hesselink (1), Brian Kirsop (2) and Ria Stoop (1)
(1) TNO Institute of Environmental Sciences,
PO Box 6011, 2600
JA Delft, The Netherlands
Draayer@tno.mw.nl
(2) Biostrategy Associates,
Stainfield House,
Stainfield,
Bourne,
Lincs PE10 0RS, UK.
bio@biostrat.demon.co.uk
Code Number: BY95007
Sizes of Files:
Text: 25K
No associated graphics
SUMMARY
This paper identifies leak tightness, cleanability and
sterilisability as the characteristics of biotechnology equipment
which are of primary importance in biosafety. Performance
criteria which assign equipment to one of three classes in each
of these respects are defined. Leaktightness classes have a
uniform definition over the range of biotechnology processes;
cleanability classification is recognised to depend on the nature
of the soil and of the cleaning protocol, and thus will vary with
the intended use. A distinction is drawn between equipment which
can be sterilised and that which cannot, but in which target
organisms can be destroyed. It is proposed that these criteria
could form the basis of biosafety classification of units of
equipment and of whole plant.
1. INTRODUCTION
Biotechnology equipment plays a vital role in determining process
integrity, product quality and in preventing harm to people and
the environment. Technical performance criteria for equipment are
well established and described in standards and guides to Good
Manufacturing Practice. Criteria for biological safety are not
well established and an approach to developing them is discussed
here. The paper is not concerned with other aspects of equipment
safety, which are well regulated by existing Directives,
regulations, standards and guidelines, nor primarily with the
many aspects of product quality where equipment performance is
important. However, the criteria defined below impinge on product
quality because they affect control over the manufacturing
process, for example by affecting the penetration of contaminant
organisms into production equipment.
Performance criteria standards for biotechnology equipment are
being developed by the European Standardisation Organisation,
CEN, as part of the work of its Technical Committee 233 on
biotechnology (10). This paper presents the views of the authors,
building on the discussions of Working Group 4 of this committee,
which is responsible for drafting the standards for equipment.
Biosafety criteria affect both the new and the traditional
biotechnology industries and care is needed to avoid harming the
traditional activities such as brewing, wine production, waste
water treatment, etc., by producing criteria which are too
stringent or unrelated to the risks presented by these
industries.
2. SIGNIFICANT FACTORS.
Hazardous or potentially hazardous living organisms and/or their
products may cause harm when they are released during:
- primary production;
- downstream processing;
- cleaning and maintenance operations.
The important properties of equipment which affect release
are:
- leak tightness during primary and downstream processing;
- cleanability, which affects whether undesirable organisms
or products are present during servicing/maintenance and
- sterilisability, which determines whether relevant
organisms can be destroyed when necessary.
These properties interact. Thus residues present after
unsatisfactory cleaning can reduce the effectiveness of a
sterilising procedure and unclean equipment may have an increased
tendency to leak.
Because of their importance, criteria for leak tightness,
sterilisability and cleanability need to be defined. Accordingly,
the Commission of the European Union has mandated the European
Standards Organisation, CEN, to do this for 21 units of
equipment. This decision underlines the significance of actual
performance, rather than design. Undue insistence on existing
design criteria may affect biotechnology negatively in the longer
term.
The questions which arise are:
a) how can performance be defined and
b) how can equipment which performs in different ways be
categorised.
3. QUANTIFICATION
We believe that performance should be defined quantitatively
where possible. Operators and regulators will then know what is
required, avoid subjective decisions and will be able to audit
processes from a biosafety viewpoint. If such quantitative
criteria cannot be developed, semi-quantitative or widely
accepted qualitative criteria should be adopted.
The boundary conditions are that a unit of equipment should be
leaktight, sterilisable and cleanable either to a very high
degree or to a very low degree. Thus a fermenter may be
categorised as completely leak tight or as failing to provide
biological containment. It may be cleanable so that soil is
undetectable or it may be impossible to clean. It may be
sterilisable so that all living organisms are destroyed or be
impossible to sterilise.
Some of the wide range of biotechnology processes require the
highest possible performance and others lower performance. Thus
complete leaktightness may be essential for fermenters used with
a hazardous organism but for many traditional biotechnology
processes - brewing, pickling, etc - containment is unimportant
from the biosafety viewpoint.
A high degree of cleanability is essential for the manufacture
of some injectable products; the requirements can usually be much
less stringent for plant in which industrial or agricultural
chemicals are to be manufactured.
High sterilisability is often needed, especially for processes
of long duration using slow growing organisms in rich media and
also when the organism used must be destroyed before the plant
is opened for routine maintenance. Sterilisation may not be
important (or may be undesirable) in traditional processes and
may be impossible for example in domestic waste water treatment
or when wooden vessels are used.
Intermediate situations between the extreme boundary conditions
also exist. A fermenter may not be sterilisable but may be
treatable to remove important organisms. For example, non-sporing
bacteria or yeasts may be the target organism but complete
sterilisation is not required to destroy them. When organisms
rapidly generate inhibitory media, high sterilisability may not
be required; many yeast fermentations are of this type. In some
circumstances in the traditional biotechnologies, such as waste
treatment, the persistence of significant production organisms
from batch to batch is essential. Many processes need plant which
is clean but not to the maximum extent; a degree of leak
tightness less than the maximum is commonly sufficient and
acceptable from a biosafety point of view.
It is concluded that it is necessary to define the extreme
boundary conditions and at least one intermediate condition of
leaktightness, cleanability and sterilisability.
A principal difficulty is that of quantifying the boundary
conditions. That this is difficult is instanced by the use of
terms such as "minimise" or "strongly minimise" by OECD (1) and
NIH (3), and in national regulations (eg 2) designed to implement
the Contained Use (4) and Biological Agents (5) European
Directives , in relation to the release of potentially hazardous
organisms. The possible levels are discussed below for the three
significant parameters.
Leaktightness
Both the European Commission and the National Institutes of
Health have laid down broad requirements for biotechnology
plants, in relation to the degree of perceived hazard.
Thus in Directive 90/219/EEC (4) and the associated guidelines
(5) there is a requirement to "minimise release", to "strongly
minimise release" or to prevent release for different groups of
organisms. In the revised NIH guidelines (3), similar terminology
is used. These terms present difficulties to those who need to
use them, for there is no objective definition of their meaning.
The availability of quantitative criteria would be of great value
to plant operators, machinery manufactures, designers and
regulators.
The regulatory intent is that, in some circumstances, there shall
be no release of organisms from biotechnology plant. This
represents one boundary condition. While this objective is
generally regarded as unattainable in practice (7, 11, 12), it
is reasonable to restate it as "organisms originating in the
biotechnology plant should not be detectable in the work place
and/or the environment". This is the criterion for the highest
category of leaktightness. The opposite boundary is that of no
containment. One or more intermediate conditions may be defined,
when there is some detectable leakage, combined with appreciable
containment. This means that organisms will be detectable in the
workplace/environment surrounding biotechnology equipment but
that the number is substantially reduced compared with the
situation of no containment. This situation may be acceptable
when the risk (being hazard x exposure) is acceptably low. Such
an approach is taken in assessing the performance of safety
cabinets (6,8). The number of such intermediate categories could
be large but in practice it is advantageous to define only one.
The numerical evaluation of this is discussed below.
Sterilisability
Formally, to sterilise means to destroy all organisms present in
the equipment. In practice, because different processes have
different objectives, the term is used not only in this sense but
also with the meaning that unwanted organisms have been
destroyed. Thus it is required that some plants are completely
sterilised with any living organisms present being destroyed;
this is one boundary condition and will be required when release
or survival of any organism present is considered to be a major
hazard to safety or to the product. At the other extreme, some
traditional processes can operate without procedures to destroy
organisms, because there is no organism present which is believed
to be a hazard to people, the environment or the product. This
is the other boundary condition.
An intermediate situation is where plant can be treated to remove
unwanted organisms, though absolute sterilisation cannot be
achieved. Equipment which can be pasteurised but not sterilised
is in this class. The key criterion here is that a defined target
organism can be destroyed.
Again, the interests of practicability indicate that only one
intermediate class should be created.
Cleanability
Absolute cleanliness, defined as the absence of any contaminating
material on the surfaces of the equipment, cannot be achieved.
There are thus parallels with the leaktightness situation, where
absolute containment cannot be expected. As in this latter case,
it seems that one boundary condition for cleanability must be the
inability to detect contaminating material, following the BATNEEK
principle (Best Available Technology Not Entailing Excessive
Cost).
Contaminating soil is always a mixture of substances, which vary
in their ease of removal; cleanability must therefore always be
assessed by measurement of a named substance. This must be
specified with care and again with the concept of BATNEEC
applying. It also seems appropriate to include visual observation
as one factor when assessing cleanability, because of its
sensitivity in some circumstances and because it may be poorly
matched by complex and expensive technology. In addition, the
cleaning methodology must be defined, because this has a major
influence on the extent to which contamination can be removed.
As a result of these considerations, the cleanability of
equipment cannot be regarded as an absolute characteristic, in
that it may be very high for some contaminants and low for
others. Thus equipment made with poor surface finish may be
highly cleanable with one soil and poorly cleanable with another.
QUANTITATIVE EXPRESSION OF LEAKTIGHTNESS
While no detectable leakage may sound a rigorous condition, its
significance is determined by sensitivity of the assay method.
Factors such as the low recovery rate of small particles, the
loss of viability of organisms in air streams, the difficulty of
ensuring growth of stressed organisms on agar media, especially
when they have been genetically constructed to be disadvantaged
in the environment, mean that in some circumstances recovery may
be very low, especially for small bacteria. Recovery of only 1%
of released organisms is not uncommon. In other circumstances (eg
9) more than half of the organisms released have been
recovered.
The significance of these considerations is that the failure to
detect organisms in the workplace may still reflect appreciable
release. Many biotechnology plants have 1000 cubic metres of air
in the workplace. When 1 cubic metre is sampled and the
efficiency of recovery is 1%, at least 100 organisms must be
present in the sampled air to allow detection. This means that
the air in the plant as a whole must contain at least 100,000
target organisms. When air changes result in a fivefold dilution
per hour, then 500,000 must escape per hour to allow detection.
This is leakage per hour of 0.5 ml of medium containing 10^6
organisms per ml. or 10^12 per cubic metre. In this case,
failure to detect organisms in the environment means only that
the leakage is less than 25 ml in a 50 hour fermentation period.
For a fermentation plant containing 25 m^3 of microbial
suspension, this represents less than 0.0001% of the volume over
the whole period, or leakage of <0.00005% per day or
<0.000002% per hour. This may be expressed as a leak factor
or, by analogy with the approach taken in safety cabinet studies,
in terms of a protection factor, defined as the ratio:
Organism concentration in the fermentation medium
------------------------------------------------
Organism concentration in the air
This has a value (per metre cubed) of 10^12/10^2 = 10^10 for the
situation described above.
This clearly represents very substantial containment, even though
release is not completely prevented. There are many factors which
determine the value of the true protection factor in the context
of no detectable release. These include the fact that leaked
organisms will not be evenly distributed and that the volume of
air in the plant will not bear a constant ratio to the volume of
microbial suspension in the contained vessel.
As detection technology improves, for example via pcr technology,
the quantitative meaning of "no detectable leakage" could change,
though it must be emphasized that practical criteria must relate
to methodology which can readily be applied and is not
prohibitively expensive or too complex for routine use.
The intermediate condition of some leakage coupled with
appreciable containment needs to be considered. The use of a
protection factor value seems to be valuable here. The concept
is already used in relation to safety cabinets, where containment
which gives a protection factor of 10^5 is regarded as a
desirable value. If higher protection is needed, the use of glove
boxes, which provide higher containment, is recommended(6).For
biotechnology plant the concentration of organisms in the
fermenter may be high and their absolute concentration in air,
rather then the fraction escaping, may be significant for
worker's health, A higher protection factor than 10^5 thus seems
appropriate for the equipment in the intermediate class; it is
suggested that to be classified in the intermediate class, the
protection factor should be at least 10^8. In the conditions
described above, this would mean that leakage of 0.002% of the
fermenter contents could be tolerated per hour from equipment.
CATEGORIES AND CRITERIA
The degrees of leak tightness, cleanability and sterilisability
have been expressed as three categories, in each case. These are
expressed in Tables below, in two ways. For the boundary
conditions for each parameter, there is no difficulty in defining
the criteria: the ones quoted represent those currently being
considered by Working Group 4 of CEN Technical Committee 233. The
intermediate classes present some difficulty in the case of
leaktightnesss and sterilisability and for these we quote both
the qualitative approach currently being followed by the above
Working Group (Column (a)) and the quantitative values suggested
by the present authors Column (b). We recognised that both may
require modification on the basis of experience but at the time
of writing they appear not only to be reasonable but also to of
value as the first step in a wider discussion of the topic.
TABLE 1. PERFORMANCE CATEGORIES FOR LEAKTIGHTNESS
_____________________________________________________________
Performance Performance criterion
category (a) (b)
_____________________________________________________________
L-A Open equipment or release level not specified
L-B Release tested under Release restricted to
defined conditions give a protection
factor of at least
10^8 (1)
L-C Release not detectable Release not
detectable (2)
_____________________________________________________________
1) Based on comparison of microbial concentrations inside the
fermenter and in the workplace air.
2) Based on tests following the principles of BATNEEC (best
available techniques not entailing excessive costs)
Table 2: PERFORMANCE CRITERIA RELATED TO
CLEANABILITY
_____________________________________________________________
Performance Performance criterion
category (a) (b)
_____________________________________________________________
C-A Visible soil or cleanability level not
specified
C-B cleanability level 95% of defined soil
specified in terms of removed by defined
defined soil and cleaning protocol
cleaning protocol
C-C defined soil not detectable (1) after
application of a defined cleaning protocol
____________________________________________________________
(1) assessed by BATNEEC
_____________________________________________________________
Table 3. PERFORMANCE CRITERIA RELATED TO
STERILIZABILITY
_____________________________________________________________
Performance Performance criterion
category
_____________________________________________________________
SI-A Sterilizability of equipment not
specified
SI-B Equipment cannot be sterilised but
defined target organisms can be
destroyed
SI-C Equipment can be sterilised
_____________________________________________________________
DISCUSSION
In biotechnology, as in other industries, some processes are held
to be hazardous to a greater or lesser degree. Where there is
risk of harm to people or to the environment, measures are taken
to minimise the risk by the selection of suitable operating
conditions and appropriate equipment. Regulatory authorities have
the responsibility of defining the degree of risk; plant
designers and managers have the responsibility of constructing
and operating the plant appropriately. Equipment characteristics
dominate plant performance but, until recently, little attention
has been paid to defining the biosafety performance criteria
which make equipment suitable for particular operations. In this
paper, criteria have been defined for leaktightness, cleanability
and sterilisability of equipment. Three classes have been
specified in each case. They are differentiated by possessing the
property to a high degree, to an intermediate extent or not at
all. It is hoped that the classification of equipment in this way
will allow suppliers to define their products more precisely with
regard to biosafety. Equally, users of equipment will be able to
select for the intended use on the basis of performance and
monitor performance in practice against agreed criteria.
The application of the classification scheme requires appropriate
test methods and, while these exist to some extent, it is clear
that more research is need before there is a fully adequate
methodology. The insensitivity of existing testing methods is
discussed in relation to leak tightness; here it is clear that
failure to detect escaped organisms can coexist with appreciable
lack of containment.
ACKNOWLEDGEMENT
We have valued the opportunity to discuss performance criteria
for equipment with colleagues in Working group 4 of Technical
Committee 233 of the European Standardisation Organisation, CEN.
REFERENCES
1 Anon., 1996. Recombinant DNA Safety Considerations. OECD,
Paris, 1986
2 Anon., 1987 Guidelines for the large scale use of
genetically manipulated organisms. AGCM/HSE Note 6,
Health and Safety Executive, London
3 Anon., 1994.NIH Guidelines for research involving
recombinant DNA molecules. Federal Register, July 5, 1994
Separate part IV, 1-73
4 Anon EEC, Council Directive on the contained use of
genetically manipulated organisms. Official Journal of the
European Communities, L117/1-14, 1990.
5 Anon EEC, Council Directive on the protection of workers from
risks related to exposure to biological agents at work.
Official Journal of the European Communities, L 374, 1, 1990
6 Anon., British Standard BS 5726. British Standards
Institution, 387 Chiswick High Road, London W4 4AL
7 Berg, P, Baltimore, D., Brenner, S., Roblin, R.O. and
Singer, M. Asilomar conference on recombinant DNA
molecules. Science, 188, 991-995, 1975
8 Clark, R., 1995. Appraisal of the practical effectiveness
of Biosafety Controls in Biotechnology, with special
reference to Safety Cabinets, Biosafety (an online
journal), Paper 3, BY95003.
9 Ferris, L. E., Keshavarz-Moore, E., Leaver, G., Noble, M.
and Turner, M.K. 1995. Microbial population balancing as a
quantitative aid for evaluating release from a high
pressure homogeniser. J. Inst. Chem. Eng., in the press.
10 Kirsop, B.H., Trends in Biotechnology, 1,1 375-378, 1993
11 Turner, M.K., 1989.Categories of Large-Scale Containment
for Manufacturing processes with Recombinant Organisms.
Biotechnology and Genetic Engineering Reviews, 7, 1-43,
12 Walker, P.D., Narendranathan, T.J., Brown, D.C.,
Woolhouse, F., and Vranch, S.P.,1987. Containment of
microorganisms during fermentation and downstream
processing. In Separations for Biotechnology, M.S. Verrall
and M.J. Hudson, Eds., 469-482, Ellis Horwood, Chichester,
1987.
Published by Bioline Publications Ltd and Science and Technology
Letters.
Copyright held by the authors.
Editorial Office: biosafe@biostrat.demon.co.uk
|