Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 275
APPENDIX I
INFECTIOUS DISEASE: EFFECT OF ANT IMICROBI~S ON
BACTERIAL POPULATIONS
Thomas F. O'Brienl
EPIDEMIOLOGY OF ANTIBIOTIC RESISTANCE
We have known for a century that germs cause infection.
For a third of a century we have had antibiotics to control
infection and molecular biology to explain, ultimately, how
the process works. We can now see each strain of bacteria as
a specific collection of genes. Some specify the life cycle.
Some code for markers we find convenient for speciation.
Others code for virulence factors or for enzymes that inacti-
vate antibiotics.
The specific linkage of genes in a bacterial strain is not
fixed. We know increasingly more about the ways in which bac-
teria can lose, gain, or exchange genes in experiments and, pre-
sumably, in nature to fill virtually every habitat (Artier, 197 9~.
However, there have been few measurements of actual rates of
bacterial gene reassociation in natural environments (O'Brien et
al., 19773. Such measurements are of critical importance because
cost-benefit estimates of the behavior of bacterial populations
may depend ultimately on their rates of gene reassociation.
We know how most antibiotics work (Kucers and Bennett, 1975~.
Each binds to and inactivates a critical specific target site in
the bacterial cell. The discovery of each new family of antibio-
tics--penicillins, streptomycin, tetracyclines, chloramphenicol,
etc.--proved in retrospect, to have also been the discovery of a
new target site. Unfortunately, few new classes of antibiotics
are being discovered. The last discovery of an antibiotic with a
truly new target site may have been rifampicin, which was first
marketed 15 years ago (Hartmann _ al., 1967~. Most of the newer
antibiotics circumvent resistance mechanisms to reach old target
sites (Christensen et al., 1979~. This declining rate of discovery
of new target sites for antibiotics suggests that these sites are,
like fossil fuels, another unreplenishable resource.
Director, Microbiology Laboratory, Peter Bent Brigham Hospital
Boston, Mass.
275
OCR for page 276
276
Also, it costs increasingly more to reach those target sites.
If it were not for the prevalence of intervening resistance mecha-
nisms, the less expensive, older agents could continue to be used.
These resistance mechanisms force us to use succeeding waves of
newer antibiotics that are expensive to develop and promote. These
costly products now account for most of the antibiotic dollar volume
which is, in turn, a large part of the pharmaceutical dollar volume.
Mutational and "Gained" Resistance
Our understanding of bacterial resistance to antibiotics has
improved over the past two decades. Early concepts were based on
studies conducted with an _ -vitro model of acquired resistance to
streptomycin. These studies showed that a mutation altered the
target site in such a way that it still functioned enzymatically,
but that the antibiotic could no longer bind to it. Thus, suscept-
ibility was lost. We now realize that it is far more common for
resistance to be gained, i.e., the target site remains unchanged,
as vulnerable as ever, but that the bacterial cell acquires some-
thing like an antibiotic-inactivating enzyme or a permeability
barrier that prevents the antibiotic from getting near the target
site (Davies, 1979~.
We can imagine and have begun to observe the implications of
"gained" resistance being more common than lost susceptibility.
Mutational loss of susceptibility to an antibiotic might be expected
to occur at a measurable rate. So should the subsequent overgrowth
replacement of susceptible strains by the mutant-resistant clone in
populations exposed to the antibiotic. Consequently, for the muta-
tional loss-of-susceptibility model, the relationship between number
of bacteria exposed to an antibiotic and the prevalence of antibio-
tic-resistant bacteria seems to be relatively simple with few inter-
vening variables. It would also seem possible that this model could
be used to estimate the prevalence of antibiotic resistance in one
population of bacteria that is ascribable to the administration of
antibiotics to another population, if the rate of interchange of
bacteria between the two populations is known.
Gained resistance is more complex. Usually, one of two types
of genes is acqu~red--a gene for an enzyme, which inactivates the
antibiotic or bypasses a blocked enzyme, or a gene for a protein,
which impairs the antibiotic's entry into the cell (Davies, 1979~.
These are large, specific molecules that are too big to arise often
by chance mutation. Therefore, most strains of bacteria have to
acquire such genes from other bacteria (Artier, 1979; Falkow, 1975~.
OCR for page 277
277
Thus, when relating antibiotic usage to prevalence of resistance
in a bacterial population one would first calculate the probabil-
ity of the arrival of the resistance gene from another population
instead of determining the mutation rate as with loss of suscepti-
bility.
Antibiotic Resistance Genes
Origin and Evolution. The source of antibiotic resistance
genes is not known. Some of them may have developed to protect
antibiotic-producing organisms from their own products (Davies
and Courvalin, 1977~. The 6-lactamases could have evolved if
the active sites of transpeptidases of the bacterial cell wall,
which selectively bind but do not hydrolyze penicillins, had
begun to acquire a hydrolytic function (Crosa et al., 1977;
Tipper and Strominger, 1965~. Techniques for characterizing
products of resistance genes are so new, and surveillance so
sketchy, that there has been little opportunity to distinguish
evolution of a new gene from dissemination of a previously rare
one. However, examples that must involve one or the other have
been observed, e.g., the emergence of the newer aminoglycoside-
inactivating enzymes (Davies and Courvalin, 1977; LeGoffic et al.,
1977~.
There are also opportunities for evolution towards improved
efficacy for existing antibiotic-inactivating enzymes. For exam-
ple, a modification of active site configuration might permit
binding of an increased range of substrate (antibiotic) analogs
as new ones come onto the market (LeGoffic _ al., 1979~. The
R1818 6-lactamase, which produces only low-level resistance, could
increase that resistance if it could evolve to have a higher sub-
strate turnover number (Medeiros et al., 1974~. Evolution of an
improved enzyme environment can also help. The TEM S-lactamase
produces resistance to ampicillin that is much less dependent on
inocula in Escherichia cold than in Haemophilus influenzas because
the outer membrane of the latter is more permeable to ampicillin
(Medeiros and O'Brien, 1975~.
Dissemination. Whatever the origin and rates of evolution of
antibiotic resistance genes, their transmission through bacterial
populations is increasingly well understood. Clearly, there are
multiple levels of genetic mobility and persistence. Many antibio-
tic resistance genes, e.g., a majority of the known 6-lactamases of
Gram-negative bacilli, have never been shown to be plasmid-borne and
are presumably chromosomal (Sykes and Matthew, 1976~. Essentially,
OCR for page 278
278
they can go only where their fixed-host bacteria can go, which
depends on the specific colonizing and competing capabilities
that are built into the other genes on the chromosome.
An entirely different level of mobility is open to antibio-
tic resistance genes that become located on plasmids, but the
extent of the mobility depends greatly on the plasmid on which
the gene is carried. Many plasmids are nonconjugative. They can
transfer between bacteria only if "rescued" by a conjugative plas-
mid (Mitsuhashi, 1979~. Others are conjugative and self-transfer-
able but only at a limited rate or to a limited number of host
bacteria. If transferred, they become unstable in many bacteria
and are readily segregated. Moreover, plasmids cannot transfer to
bacteria that already contain plasmids of the same incompatibility
group (Falkow, 1975~. Thus, a resistance gene on a plasmid of a
group already widely distributed in a bacterial population might
achieve only limited dissemination in that population.
The ultimate dissemination and persistence of a plasmid-
borne resistance gene also depends on the other types of genes
that are on the same plasmid. Many of the plasmids bearing re-
sistance genes are large and have space for many other genes.
Some of these other genes code for resistance to other antibiotic-
The prevalence of any one of these resistance genes, and of the
plasmid itself, is amplified whenever the plasmid's hosts are ex-
posed to one of the antibiotics to which any of the other genes
code resistance. A plasmid is essentially a gene cooperative in
which an advantage accruing to any member gene is shared by all.
In this sense, plasmids can be looked upon as devices that cause
an increased prevalence of resistance to several antibiotics to
result from the use of one antibiotic.
Linkage to Genes for Other Characteristics. Some genes,
which code for functions other than resistance to antibiotics,
also amplify and sustain the prevalence of an antibiotic resist-
ance gene that was associated with them on the same plasmid. Some
plasmid-borne genes code for functions that permit host bacteria to
survive in specific environments such as those containing specific
metallic ions (McHugh et al., 1975; Pinney, 1977~. Others code for
colonization (Silver and Mercer, 1978) or for metabolic functions,
e.g., lactose catabolism (Kopecko et al., 1979), on which host bac-
teria might depend either always or only under certain environmental
conditions. The functions of these other genes, which occupy so
large a part of so many plasmids bearing resistance genes, are just
beginning to be elucidated. The glimpses of them to date, however,
OCR for page 279
279
suggest that they may be major codeterminants of the ultimate
distribution of resistance genes carried with them on the same
plasmids (Novick, in press).
The closeness and stability of the linkages between a
particular antibiotic resistance gene and other resistance or
nonresistance genes on the same plasmid also influence the pre-
valence of that specific antibiotic resistance gene. The in-
creasingly well understood mechanisms for genetic recombination
serve not only to put genes together but also to disrupt and
separate them, ultimately segregating genes that no longer pro-
vide a survival advantage. Conforming to the analogy of the
cooperative venture, genes that never bring any advantage are
eventually dismissed, but the time of dismissal varies greatly.
A gene coding for resistance to a rarely used antibiotic could
persist on a plasmid for long periods during which it offers no
survival advantage if it is closely and stably linked to another
gene that provides an advantage. The close linkage of genes
coding for resistance to streptomycin and sulfonamide on a com-
monly encountered plasmid (van Treeck et al., 1979) could be
part of the reason why the gene for resistance to streptomycin
remains prevalent long after clinical usage of streptomycin in
humans has declined.
Transposons. The recently discovered transposons provide
yet another level of genetic mobility. Transposons are segments
of a DNA chain flanked by special insertion sequences. These
transposons and genes carried on them are able to transpose from
one plasmid to another or from plasmid to chromosome and back
(Heffron _ al., 1975~. Location on a transposon could greatly
enhance the prevalence of any gene.
Genes coding for resistance to a number of commonly used
antibiotics have been found on transposons (Campbell et al.,
1977~. A resistance gene can be limited in its distribution by
its position on the chromosome of, or on a nonconjugative plas-
mid in, a poorly colonizing strain; on a plasmid of limited host
range; or on a plasmid carrying few other genes with survival
advantages. Such a gene could become more prevalent if it could
be transposed to a more mobile plasmid carrying a more advantage-
ous collection of genes, possibly including other resistance
genes, or, better still, to a variety of such plasmids.
OCR for page 280
280
Implications of the Resistance Mechanism
The entire assemblage of these properties for any anti-
biotic resistance gene, including the substrate range and the
efficiency of its gene product, and its genetic location and
mobility, can be regarded as an antibiotic resistance mechanism.
Defined in this way, the mechanism can be viewed as a major var-
iable in determining the prevalence of resistance to antibiotics.
When a population of bacteria is exposed to an antibiotic, the
extent and distribution of the resulting antibiotic resistance
will depend largely on the properties of~the mechanisms of re-
sistance to that antibiotic in that population.
This can be best illustrated by considering some examples.
If no mechanism for resistance to the antibiotic is available
to the exposed population of bacteria, there will be no resist-
ance. This situation prevailed for a number of years in popu-
lations of Enterobacteriaceae exposed to gentamicin in the United
States (O'Brien et al., 1978~. If the only mechanism available
is a gene coding for an enzyme with a narrow specificity includ-
ing only that antibiotic and if that gene is the only resistance
gene on the chromosome of a bacterial species with poor colonizing
and competing abilities, the ensuing resistance would be directed
only against the antibiotic used and would be unlikely to persist.
However, resistance would become widespread and persistent if the
population contains a resistance mechanism consisting of a gene
coding for an enzyme with a broad substrate range and if the gene
is located on a transposon in a plasmid with a broad host range
and is rich with closely linked genes for resistance to other
antibiotics and for other survival advantages.
Consider a bacterial population in a chemostat or in nature
that includes one cell with a resistance mechanism and a similar
population that includes one cell with a different mechanism.
Each population could be exposed for the same period to an anti-
biotic to which its mechanism produces resistance. Each popula-
tion could then be scored for the total number of resistant cells
times the total number of antibiotics to which each is resistant
times the number of generations during which resistance persists.
The values could differ greatly and would constitute for each
mechanism an index of its efficiency at converting selection
pressure into resistance prevalence.
A different kind of index could acknowledge the heterogeneity
of bacterial populations in nature. Such populations contain many
different strains and species of bacteria coexisting, competing,
OCR for page 281
281
and cooperating to occupy all possible niches in the ecosystem.
For a resistance mechanism introduced in one strain in such a
population, it would be possible to score the number of different
strains and specimens into which the mechanism became incorporated
after a controlled period of antibiotic exposure. Some mechanisms
could remain only in the strain in which they were introduced.
Such a strain would presumably overgrow its susceptible neighbors
during exposure to the antibiotic but might not by itself be able
to occupy all of the niches very fully. After the exposure to the
antibiotic, the strain might promptly be displaced and revert to
its natural domain in the population, which could be nonexistent.
In the same population exposed for the same period to the
same antibiotic, a different mechanism of resistance on a plasmid
with a broad host range could be incorporated into many strains
and species. Or a resistance gene on a transposon in a plasmid
of limited host range might be transposed to one or more plasmids
with a broader range of hosts and enter many strains and species
by this route. In each strain and species the resistance gene
would, in effect, be associated with a different assemblage of
other genes. Therefore, this score would be an index of gene re-
association and, to the extent that the selection pressure pro-
motes it, antibiotic-induced gene reassociation.
Two consequences of this process are worth examining. If
many of the strains and species constituting the original bacter-
ial population came to possess the resistance mechanism enabling
them to survive exposure to an antibiotic, they would then pre-
sumably be better able to fill all niches during the exposure than
would a single strain. They would also be less rapidly displaced
by susceptible species after the exposure ended. In the hypothe-
tically perfect situation, all strains and species in the original
population would receive the resistance mechanism so that exposure
to antibiotics would not alter their ecological balance. In such
a case there would be no tendency after the exposure had ended for
reemerging or reentering susceptible species to displace resistant
strains in the process of regaining their old niches. In this
population, the prevalence of resistance after exposure would be
diminished primarily by the segregation rates of the resistance
mechanism from each of the strains, which may also be a variable
characteristic of the resistance mechanism (Levin, in press) or,
possibly, the normal influx of new strains colonizing and dis-
placing the older members, although these new strains could also
be susceptible to infection by a mobile plasmid widely resident
in the population.
OCR for page 282
282
Secondly, if a resistance mechanism has evolved to gain a
great increment in prevalence per episode of exposure to antibio-
tics and to have a high degree of interspecies mobility, it has
a greater chance of ultimately getting into a pathogen. The bac-
teria that actually infect humans or animals are only a miniscule
part of the total bacterial flora of the world and, usually, of
the infected individual as well. But they do all the damage. The
arrival of a resistance mechanism in a pathogen can be a signifi-
cant and tragic event. The recent arrival of the TEM S-lactamase
(Medeiros and O'Brien, 1975) into group B Haemophilus influenzas,
the commonest cause of bacterial meningitis in children,has led to
treatment failure and to the need to treat all children thought to
have the disease with a potentially toxic antibiotic, chlorampheni-
col (Howard et al., 1978~.
All of these considerations lead to one essential question:
What is the source of the more efficient and more efficiently
distributing resistance mechanisms? Everything suggests that
the mechanisms themselves are the ultimate evolutionary products
of the cumulative application of antibiotic selection pressures
to bacterial populations. Exposure to antibiotics is the only
imaginable stimulus that would provide evolutionary survival ad-
vantage to antibiotic-inactivating enzymes or to other resistance
genes prevalent enough to overcame the odds against their forming
favorable linkages with one another or with other advantageous
genes or eventually becoming located on plasmids or transposons.
Summary
The antibiotic resistance observed in a population of bac-
teria after exposure to an antibiotic can be attributed to:
1. The net effect of all prior antibiotic exposure on the
bacterial flora of this planet in shaping and delivering to this
particular population the specific resistance mechanisms it has
acquired.
2. The effect of those specific resistance mechanisms in
determining the number and types of rests tent bacteria that will
be generated by the present episode of exposure to the antibiotic.
OCR for page 283
283
INTERCHANGE BETWEEN DIFFERENT BACTERIAL FLORAS
Compartmentalization of Floras
The preceding considerations have a very practical bearing
on how we view the relationship of antibiotic usage in different
compartments of the world's bacterial flora. These compartments
could be defined as domains within which bacteria can be imagined
to circulate more readily than they do between compartments. For
example, the flora of each human being, the flora of an intensive
care unit as opposed to the rest of the hospital, the flora of an
entire hospital (O'Brien et al., 1975) as opposed to the community,
or, on a larger scale, the flora of humans in the United States
as opposed to that in humans in some other part of the world.
Although compartmentalized, the bacterial flora of the world un-
doubtedly circulates between all of compartments, but generally
less than within them.
1
If the only concern with antibiotic resistance were quan-
titative, the resistance in compartment A would change the re-
sistance in compartment B only to the extent that bacteria from
A moved into B. If A had 100% resistance and B by itself would
have had only 10% resistance, yet 1Z of B's strains actually came
from A, then they would add 1% resistance for a total of 11%
resistance in B.
If there are qualitative differences among resistance mecha-
nisms, as discussed in the preceding section, then the influence
of resistance in one compartment on that of another may be quite
out of proportion to the circulation between them. Suppose com-
partment A were to contain a mechanism of resistance still absent
in B but ideally suited for circulation there because of its plas-
mid incompatibility grouping, its special ability to inactivate an
antibiotic widely used there, or its linkage to some other genes
admirably suited for survival in B. Even if there were very little
circulation between the compartments, a few bacteria with the
mechanism could transfer from A to B. Once there, the mechanisms
could become prevalent.
Recent Observations
-
Recently, an apparent example of this has been observed with
the use of gentamicin, which had been widely used since 1970 in
OCR for page 284
284
the United States to treat hospitalized patients infected with
Enterobacteriaceae. Through 1975, however, exceedingly few
strains of Enterobacteriaceae isolated in the United States had
any ability to resist that antibiotic (O'Brien et al., 1978~.
For purposes of bacterial ecology, there was a great need for
some mechanism of resistance, but apparently none, or none with
appreciable genetic mobility, had become available to strains in
the United States. In other parts of the world--in France, for
example--there were genes for two aminoglycoside-inactivating
enzymes (O'Brien _ al., 1978; Witchitz and Chabbert, 1972~.
The 2 -aminoglycoside nucleotidyl transferase (AND 2 ~ and the
3-aminoglycoside acetyl transferase (AAC-3) were known to be
widely distributed in nosocomial strains.
In early 1976 a strain of Klebsiella pneumonias of serotype
-
2 with an unusual biotype was isolated from a patient in one of
the U.S. hospitals in our international collaborative study
(O'Brien _ al., in press). This strain had a distinctive conju-
gative plasmid of 56 Mdal belonging to incompatibility group M,
which carried genes for resistance to sulfonamides as well as to
chloramphenicol acetyl transferase, the TEM 1 6-lactamase, and
the AND 2 enzyme, one of the two gentamicin resistance enzymes
that are common in other parts of the world.
Over the next 4 months this distinctive strain of K. pneu-
moniae was isolated from 40 patients in the hospital. Thereafter,
this strain was isolated only infrequently, but by then, the plas-
mid was being found in isolates of other strains of K. penumoniae
and in other species of Enterobacteriaceae as well (O'Brien et al.,
in press). Ultimately, this one plasmid was found in 49 different
biotypes of six different bacterial strains. Restriction endonu-
clease digestion analyses conducted on plasmids from Escherichia
cold K-12 transconjugants of eight different strains of six
different species confirmed the identity of the plasmid through-
out the outbreak (O'Brien et al., in press).
In this instance, a restrictive mechanism not previously
circulating in one compartment (the hospital flora) and probably
not circulating in a much larger one (flora of humans in the United
States), but which circulated widely in another compartment for
several years (European hospital flora) appeared in one hospital
compartment and rapidly became prevalent. An almost identical se-
quence in one other U.S. hospital has since been reported (Gerding
_ al., 1979; Sadowski et al., 1979~. We are beginning to search
files in microbiological laboratories in other U.S. hospitals for
evidence of further spread.
OCR for page 285
285
There is no proof that the mechanism was imported from out-
side of the flora compartment of humans in the United States,
although it had been prevalent there (O'Brien et al., 1978) and
might have been expected to be imported if there had been any
appreciable exchange between these compartments. Another item
circumstantially suggesting such importation was the plasmid's
carriage of the chloramphenicol resistance gene, which tends to
be relatively rare in the flora of humans in the United States
but common in flora outside that compartment. Proof of its gene-
alogy will have to come from studies identifying plasmids on a
larger scale, which are beginning to become practical.
This experience illustrates that if there are qualitative
r differences between resistance mechanisms, the influence of re-
sistance in one bacterial flora compartment on that of another may
be completely out of proportion to the rate of interchange between
them. The compartment in which this particular complex mechanism
had evolved could not have had great interchange with the compart-
ment of flora in humans in the United States because only traces
of it could be found there prior to the outbreak. Apparently,
however, this particular mechanism was qualitatively so "fit" for
this hospital flora compartment that once introduced it circulated
widely and became prevalent.
The consequences of the entry and circulation of this one
plasmid in a hospital where it had previously been absent were
substantial. The resistance of all Gram-negative bacillary iso-
lates at the hospital to a set of six antibacterial agents (APAR
index) (O'Brien et al., 1978) increased approximately 50% because
of this plasmid. The number of isolates with resistance to more
than six antibiotics increased nearly 10-fold. A number of cases
of Gram-negative sepsis were virtually untreatable. Most of the
resistant isolates came, as usual, from patients being treated or
having recently been treated with antibiotics. Yet the level of
antibiotic usage had not appeared to change appreciably. As dis-
cussed previously, the entry and dissemination of a qualitatively
different resistance mechanism had reset the equilibrium so that
a given amount of antibiotic usage now generated a greater preva-
lence of resistance to antibiotics.
IMPLICATIONS TO HUMAN HEALTH FROM ANTIBIOTIC USE IN ANIMALS
Exchange Between Compartments
The considerations discussed in the preceding paragraphs
generate concern about the veterinary use of antimicrobial agents.
OCR for page 290
290
This problem is a very general obstacle to tracing the
spread of resistance mechanisms in naturally occurring bacter-
ial populations. Most common phenotypic resistance markers
have been so widely distributed in most bacterial populations
that their further exchange is extremely difficult to detect,
let alone measure overall. Generally, we are able to detect
and measure exchange only when a new marker enters a population
as it did in the previously mentioned outbreaks, which involved
a previously rare aminoglycos ice-inactivating enzyme (O'Brien et
al., in press). Similarly, incursions of resistance without gene
exchange is usually observed only when a previously rare type
of resistance is involved, as with penicillinase-producing gono-
cocci, ampicillin-resistant Haemophilus influenzas, methicillin-
resistant staphylococci, and penicillin-resistant pneumococci.
With regard to the dissemination of antibiotic resistance, it
seems likely that only the tip of the iceberg has been observed.
A corrolary to this is that improved discrimination of any
component in this system--levels of antibiotic resistance, identi-
fication of inactivating enzymes, identification of specific
transposons or plasmids, or identification of specific biotypes or
serotypes of host bacteria--will greatly increase our ability to
find distinctive combinations that can be traced through bacterial
populations. Every strain of bacteria would be distinctive and
traceable if we knew enough about it.
Despite these obstacles, some investigators have observed
the spread of antibiotic resistance from the flora of animals to
the flora of humans. For example, farm workers who work with anti-
biotic-fed animals have in their flora a higher prevalence of anti-
biotic-resistant, plasmid-containing bacteria than do other humans
(Levy, 1978; Wiedemann and Knothe, 1971~. Moreover, multiresistant
strains of salmonellae have been traced from infected animals to
humans on many occasions (Anderson, 1968; Anonymous, 1979; Center
for Disease Control, 1977~.
These documented episodes of the flow of antibiotic resistance
from the flora of animals to the flora of humans are important
because they confirm, in spite of the obstacles, what molecular
biologists have implied.
In the near future, it should be possible to overcome some of
these obstacles and to improve our ability to observe the spread
of resistance, including that between compartments of flora, e.g.,
between animals and humans. Two new techniques should enable us to
observe antibiotic resistance as an interrelated system throughout
the world.
OCR for page 291
291
One technique is computerized analysis of results of tests
for antibiotic susceptibility. Every day, thousands of labora-
tories in all parts of the world test the susceptibility of tens
of thousands of clinical bacterial isolates. In the aggregate,
these data, which have already been paid for, represent the work-
ings of the antibiotic resistance exchange system. The problems
of quality control, accurate sampling, data management, and ana-
lytic methods have been examined (O'Brien et al., 1977, 1978),
and an international collaborating group now representing 30
centers is pooling these types of data for continuing analysis
(International Antibiotic Resistance Survey Group and O'Brien,
1978 and in press). In addition, other sets of susceptibility
testing results, such as data observed from isolates of salmo-
nellae from humans and animals (Figure 2) and a collection of
test files from several hundred U.S. hospitals are being put
into computers and similarly analyzed.
These analyses provide a detailed view of the distribution of
antibiotic resistance and its trends in time. Much of the contro-
versy over any aspect of antibiotic use concerns the magnitude of
the resistance problem and the direction it is taking. Information
on both of these issues should be extractable from the data in the
computer files and could be observed continually thereafter if mon-
itoring is maintained. Moreover, this type of reconnaissance might
serve as a starting point in the development of strategies to mini-
mize resistance resulting from the use of antibiotics in humans or
animals.
Furthermore, these analyses should also provide unprecedented
detail about the distribution of resistance mechanisms in the
flora of humans in the United States. This is extremely impor-
tant to this report because it should enhance the detection of the
incursion of individual or a combination of resistance mechanisms
into the flora of humans in the United States from another compart-
ment. Previously noted incursions of this kind, e.g., penicilli-
nase-producing gonococci from the Far East, have been observed,
usually because they were extremely conspicuous. The more sensitive
and selective are one's techniques for identifying resistance mech-
anisms, the greater is one's ability to detect incursion of more
subtly distinctive groups from the flora of humans in the United
States or from the flora of animals. For example, it was possible
to trace the entry and dissemination in the flora of a patient in a
U.S. hospital of the AND 2 resistance gene, previously more common
in isolates from humans in other countries, because in its early
circulation it was linked with a grouping of other resistance genes
virtually unique for that hospital (O'Brien et al., in press).
OCR for page 292
OCR for page 295
OCR for page 296
OCR for page 297
OCR for page 298
OCR for page 299
OCR for page 300
Representative terms from entire chapter:
resistance mechanisms
mO ~ gO ~ =~1 ! 1 1 1 i ~' 0~ _ _ 1 1 1 1 1 1 =~) m_ ~
z ]: ~ i - -lo ~ :~, ~
C=c ~ ~ ,~, ~ it. ~
I
A_ I I I I 11 1 d gO - - - - - - ~ ' =~1 ~ ~ I I I I ~ FUJI I I I I I 'I 1°OI I I I I I I I I '-I Icn I I I I L I I I ~ '
I :t ~ ~ 2_~ _~- ~ ~7
=~ air' m:' m:' it's
' W~:L:; - L:
= ~:'~
~ ~:l:
G ~ ~ IU~ 'A ~ ~
~
I On
~ ~ ~ ~ ~o ~
'a
rl I ~ 0.) Ed ~ ~ d
Lo ~ ~ ~ ~ o
0 .~ ~ a) On ~ ~ ~ ~^ 0
.~ ,~ ~ ~ . ~ ~ ~ ~^=
= ~ ~ ~ ~ ~o ~ ~¢
.- ~ ~ ~m ~ ~ O ~Zen
usado ~ ~ ~ ~ o ~¢
= ~ ~ ~ ~ o ^
o ~ o Cal
,1 ~ ~ ~ ~ ~lo, -
~ ~ ~ ~ ~ ~ ~ ~ ~ o ~ -
o CO ~ o ~ ~o ~ ~ ~0
· ~ ~ ·
u' O o
,1 ~ ~
294
Surveillance of the prevalence of antibiotic resistance
mechanisms in the compartment of flora of humans outside of the
United States Will continue to have an important bearing upon
the issue surrounding the use of antibiotics in animal feed for
two reasons. First, this is the major source other than animal
flora of the importation of distinctive resistance mechanism-
host combinations into flora of humans in the United States,
and it has thus far provided the most conspicuous examples of
such importation, including penicillin-resistant gonococci,
penicillin-resistant pneumococci, chloramphenicol-resistant
typhoid, methicillin-resistant staphylococci, etc. Therefore,
knowledge of what is prevalent in human flora abroad could help
greatly in deciding the likelihood that a particular mechanism
observed in the flora of humans came from animals, rather than
from humans from another country.
Figure 2 shows that resistance to sulfonamide in salmonella
isolates from humans in the United States could have come from
animal sources or from foreign human sources, in both of which
it appears to be highly prevalent. Examination of the serotype
distribution of sulfonamide resistance in the three compartments
might provide further evidence for exchange. In contrast, the
low-level resistance to tetracycline (zone diameters of approxi-
mately 10 mm) observed in several of the isolates of salmonella
from humans in the United States appears more likely to be related
to strains from animals, which have a high proportion of this type
of resistance, than to isolates of salmonella from humans outside
the United States, in which only a very small fraction of resist-
ant isolates appear to be of this type.
The second reason for the pertinence of foreign data is
that different countries, in effect, have been testing grounds
for a variety of antibiotic uses and food procurement and distri-
bution practices, which have developed in them. The prevalence
of various resistance mechanisms in the flora of humans from
these different countries are presumably largely the consequences
of these practices. Observation of these results could, therefore,
be an indispensable guide to the development of optimal antibiotic
usage policy. For example, the prevalence of resistance to chlor-
amphenicol, presumably due to the chloramphenicol acetyl transfer-
ase gene, was found to be more than 10 times more prevalent among
isolates from humans in France than in the United States (O'Brien
_ al., 1978~. The French investigator collaborating in this study
believes that the use of chloramphenicol in humans is not excessive
in France but that chloramphenicol is the antibiotic most widely
used in food-source animals in that country. If this could be
295
shown to be true, it could constitute a powerful argument for the
influence of antibiotic use in food animals on the prevalence of
antibiotic resistance in human flora.
Similarly, resistance to streptomycin remains relatively com-
mon in isolates of Enterobacteriaceae from humans in-the United
States long after its use in treating humans has declined. This
could represent fortuitous linkages of the streptomycin resistance
gene, as discussed previously, but it might also reflect continuing
use of streptomycin in livestock in the United States.
The second resource for improved future observation of the
antibiotic resistance system is the newer techniques for character-
izing resistance mechanisms, ranging from the detailed characteri-
zation of inactivating enzymes to analyses conducted by incubating
the plasmid~s) with restriction endonucl~ases to establish the
identity, diversity of, or to map, plasmids. The increasing avail-
ability of these methods should improve discrimination of resistance
mechanisms, which, as mentioned previously, should greatly enhance
our ability to trace their spread in bacterial populations. These
techniques may be particularly useful when coupled with computer
survey of the large data bases. Discriminant analysis of these
data bases may result in the identification of specific antibiotic-
inactivating enzymes (Guzman, and O'Brien, 1978; Kent et al., in
press). The computer analyses could indicate isolates of special
interest for these laboratory studies and then project the results
of the studies back to a large population.
296
REFERENCES
Anderson, E. S. 1968. The ecology of transferable drug resist-
ance in the enterobacteria. Ann. Rev. Microbial. 22:131-180.
Anonymous. 1979. Salmonellosis--an unhappy turn of events.
Lancet 1:1009-1010 (Editorial).
Arber, W. 1979. Promotion and limitation of genetic exchange.
Science 205:361-365.
Babcock, G. F., D. L. Berryhill, and D. H. Marsh. 1973. R-fac-
tors of Escherichia cold frog dressed beef and humans.
Appl. Microbiol. 25:21-23.
Campbell, A., D. Berg, E. Lederberg, P. Starlinger, D. Botstein,
R. Novick, and W. Szyblaski. 1977. Nomenclature of trans-
posable elements in prokaryotes. Pp. 15-22 in A. I. Bukhari,
J. A. Shapiro, and S. L. Adhya, eds. DNA Insertion Elements,
Plasmids, and Episomes. Cold Spring Harbor Laboratory,
Cold Spring Harbor, N.Y.
Center for Disease Control. 1977. An Outbreak of Multiple-Drug
Resistant Salmonella Heidelberg, Connecticut. Reported by
M. L. Cohen, J. G. Wells, C. L. Samples, P. A. Blake, J. L.
Conrad, and E. J. Gangarosa. Report Number EPI-77-13-2.
Center for Disease Control, Atlanta, Ga. 8 pp.
Christensen, B. G., L. J. Ruswinkle, and L. D. Cama. 1979. The
design of new drugs that resist microbial inactivation.
Rev. Infect. Dis. 1:64-72.
Crosa, Je H., J. Olarte, L. J. Mata, L. K. Luttropp, and M. E.
Penaranda. 1977. Characterization of an R-plasmid associ-
ated with ampicillin resistance in Shigella dysenteriae
type 1 isolated from epidemics. Antimicrob. Agents Chemo-
ther. 11:553-558.
Davies, J. 1979. General mechanisms of antimicrobial resistance.
Rev. Infect. Dis. 1:23-27.
Davies, J., and P. Courvalin. 1977. Mechanisms of resistance to
aminoglycosides. Am. J. Med. 62:868-872.
297
Elwell, L. P., J. De Greaff, D. Seibert, and S. Falkow. 1975.
Plasmid-linked ampicillin resistance in Haemophilus influen-
zae type b. Infect. Immun. 12:404-410.
Falkow, S. 1975. Infectious Multiple Drug Resistance. Pion
Limited, London, England. 300 pp.
Gerding, D. N., A. E. Buxton, R. A. Hughes, P. P. Cleary, J.
Arbaczawski, and W. E. Stamm. 1979. Nosocomial multiply
resistant Klebsiella pneumonias: Epidemiology of an out-
break of apparent index case origin. Antimicrob. Agents
Chemother. 15:608-615.
Guzman, M. A., and T. F. O'Brien. 1978. Mechanisms of resist-
ance to aminoglycoside antibiotics in clinical isolates of
Serratia marcescens. Abstract 288 in Program and Abstracts.
Eighteenth Interscience Conference on Antimicrobial Agents
and Chemotherapy. Sponsored by the American Society for
Microbiology, Washington, D.C., 1-4 October 1979, The
Atlanta Hilton, Atlanta, Ga.
Hartmann, G., K. O. Honikel, F. Knusel, and J. Muesch. 1967.
The specific inhibition of the DNA-directed RNA synthesis
by rifamycin. Biochem. Biophys. Act. 145:843-844.
Heffron, F., C. Rubens, and S. Falkow. 1975. Translocation of
a plasmid DNA sequence which mediates ampicillin resistance:
Molecular nature and specificity of insertion. Proc. Nat.
Acad. Sci. (USA) 72:3623-3627.
Howard, A. J., C. J. Hince, and J. D. Williams. 1978. Antibio-
tic resistance in Streptococcus pneumoniae and~Haemophilus
influenzae. Br. Med. J. 1:1657-1660.
International Antibiotic Resistance Survey Group and T. F. O'Brien.
1978. Multicenter sensitivity studies. International colla-
borative antibiotic resistance survey. Pp. 534-536 in W.
Siegenthal and R. Luthy, eds. Current Chemotherapy. Pro-
ceedings of the 10th International Congress of Chemotherapy,
Volume l, Zurich/Switzerland, 18-23 September 1977.
International Survey of Antibiotic Resistance Group and T. F.
O'Brien (coordinator). In press. Average percent antibiotic
resistance in isolates from multiple centers. In J. D. Nelson
and C. Grassi, eds. Current Chemotherapy and Infectious Dis-
ease. American Society for Microbiology, Washington, D.C.
298
Kent, R. L., T. F. O'Brien, A. A. Medeiros, J. J. Farrell, and
M. A. Guzman. In press. Identification of antibiotic-
inactivating enzymes by stepwise discriminant analysis of
susceptibility test results. In J. D. Nelson and C. Grassi,
eds. Current Chemotherapy and Infectious Disease. American
Society for Microbiology, Washington, D.C.
Kopecko, D. J., J. Vickroy, E. M. Johnson, J. A. Wohlhieter,
and L. S. Baron. 1979. Molecular and genetic analyses of
plasmids responsible for lactose catabolism in salmonellae
isolated from diseased humans. In Collected Abstracts from
the IV Symposium on Antibiotic Resistance, Smolenice,
Czechoslovakia.
Kucers, A., and N. McK. Bennett. 1975. The Use of Antibiotics.
A Comprehensive Review with Clinical Emphasis. Second Edi-
tion. J. B. Lippincott Company, Philadelphia, Pa. 679 pp.
Le Goffic, F., M. L. Capmau, D. Bonnet, C. Cerceau, C. Soussy,
A. Doblanchet, and J. Duval. 1977. Plasmid-mediated pris-
tinamycin resistance PAC IIA: A new enzyme which modifies
pristinamycin IIA. J. Antibiot. (Tokyo) 30:665-669.
Le Goffic, F., N. Moreau, A. Martel, and M. Masson. 1979. Could
a single en7-yme inactivate aminoglycoside antibiotics by two
different mechanisms? In Collected Abstracts from the IV
Symposium on Antibiotic Resistance, Smolenice, Czechoslovakia.
Levin, B. R. 1979. The conditions for the existence of plasmids
in bacterial populations. In Collected Abstracts from the
IV Symposium on Antibiotic Resistance, Smolenice, Czechoslo-
vakia.
Levy, S. B. 1978. Emergence of antibiotic-res~stant bacteria in
the intestinal flora of farm inhabitants. J. Infect. Dis.
137:688-690.
McHugh, G. L., C. C. Hopkins, R. C. Moellering, and M. N. Swartz.
1975. Salmonella typhimurium resistant to silver nitrate,
chloramphenicol, and ampicillin. A new threat in burn units?
Lancet 1:236-239.
Medeiros, A. A., and T. F. O'Brien. 1975. Ampicillin-resistant
Haemophilus influenzae type B possessing a TEM-type ~ -lacta-
mase but little permeability barrier to ampicillin. Lancet
1:716-718.
299
Medeiros, A. A., R. L. Kent, and T. F. O'Br' en. 1974. Charac-
terization and prevalence of the different mechanisms of
resistance to beta-lactam antibiotics in clinical isolates
of E scherichia colt. Antimicrob. Agents Chemother. 6: 7 91-
801.
Mitsuhashi, S. 1979. Nonconjugative resistance plasmids. In
Collected Abstracts from the IV Symposium on Antibiotic Re-
sistance, Smolenice, Czechoslovakia.
Novick, R. In pres s. The development and spread of antibiotic
resistant bacteria as a consequence of feeding antibiotics
to livestock. Ann. N. Y. Acad. Sci.
O'Brien, T. F., R. L. Kent, and A. A. Ceded ros. 1975. Computer
surveillance of shif ts in the gross patient flora during
hospitalization. J. Infect. Dis. 131:88-96.
O'Brien, T. F., R. Ae Norton, R. L. Kent, and A. A. Medeiros.
1977. International surveillance of prevalence of antibio-
tic resistance. J. Antimicrob. Chemother. 3(suppl C):59-66.
O'Brien, T. F., J. F. Acar, A. A. Medeiros, R. A. Norton, F.
Goldstein, and R. L. Kent. 1978. International comparison
of prevalence of resistance to antibiotics. J. Am. Med.
Assoc. 239:1518-1523.
O'Brien, T. F. ~ D. G. Ross, M. A. Guzman, A. A. Medeiros, R. W.
Hedges, and D. Botstein. In press. Dissemination of an
antibiotic resistance plasmid in hospital patient flora.
Antimicrob. Agents Chemother.
Petrocheilou, V., M. H. Richmond, and P. M. Bennett. 1979.
Persistence of plasmid-carrying tetracycline-resistant
Escherichia cold in a married couple, one of whom was
receiving antibiotics. Antimicrob. Agents Chemother. 16:
225-230.
Pinney, R. J. 1977. Pharmaceutical significance of plasmid-
mediated mercury resistance. J. Pharm. Pharmacol. 2 9
(suppl.~:70P.
Riley, M., and A. Anilionis. 1978. Evolution of the bacterial
genome. Annul Rev. Microbial. 32: 519-560.
300
Sadowski, P. L., B. C. Peterson, D. N. Gerding, and P. P. Cleary.
1979. Physical characterization of ten R plasmids obtained
from an outbreak of nosocomial Klebsiella pneumonias infec-
tions. Antimicrob. Agents Chemother. 15:616-624.
Shapiro, M., T. R. Townsend, B. Rosner, and E. H. Kass. 1979.
Use of antimicrobial drugs in general hospitals. II. Analy-
sis of patterns of use. J. Infect. Dis. 139:698-706.
Shooter, R. A., S. A. Rousseau, E. M. Cooke, and A. L. Breaden.
1970. Animal sources of common serotypes of Escherichia cold
in the food of hospital patients. Lancet 2:226-228.
Silver, R. P., and H. D. Mercer. 1978. Antibiotics in animal
feeds: An assessment of the animal and public health aspects.
Pp. 649-664 in J. N. Hathcock and J. Coon, eds. Nutrition and
Drug Interrelations. Academic Press, New York, San Francisco,
London.
Sykes, R. B., and M. Matthew. 1976. The ~ -lactamases of Gram-
negative bacteria and their role in resistance to 6-lactam
antibiotics. J. Antimicrob. Chemother. 2:115-157.
Threlfall, E. J., L. R. Ward, and B. Rowe. 1978. Spread of multi-
resistant strains of Salmonella typhimurium phase types 204
and 193 in Britain. Br. Med. J. 2:997.
Tipper, D. J., and J. L. Strominger. 1965. Mechanism of action
of penicilllins: A proposal based on their structural simi-
larity to acyld-alanyl-d-alanine. Proc. Nat. Acad. Sci.
(USA) 54:1133-1141.
van Treeck, U., B. Wiedemann, and W. Kalthofen. 1979. Character-
ization of a frequently occurring plasmid/rPBl/ in clinical
isolates of E. colt. In Collected Abstracts from the IV Sym
-
posium on Antibiotic Resistance, Smolenice, Czechoslovakia.
Wiedemann, B., and H. Knothe. 1971. Epidemiological investiga-
tions of R factor-bearing enterobacteria in man and animal in
Germany. Ann. N.Y. Acad. Sci. 182:380-382.
Witchitz, J. L., and Y. A. Chabbert. 1972. Resistance transfer-
able a la gentamicine. II. Transmission et liaisons du
caractere de resistance. Ann. de L'Institut Pasteur 122:24,
368-378.
