Natural Resources Defense Council, Inc. et al v. United States Food and Drug Administration et al

Filing 33

DECLARATION of Jennifer A. Sorenson in Support re: 19 MOTION for Summary Judgment.. Document filed by Center For Science In The Public Interest, Food Animal Concerns Trust, Natural Resources Defense Council, Inc., Public Citizen, Inc., Union Of Concerned Scientists, Inc.. (Attachments: # 1 Exhibit A, # 2 Exhibit B, # 3 Exhibit C, # 4 Exhibit D, # 5 Exhibit E, # 6 Exhibit F, # 7 Exhibit G, # 8 Exhibit H, # 9 Exhibit I, # 10 Exhibit J, # 11 Exhibit K, # 12 Exhibit L, # 13 Exhibit M, # 14 Exhibit N, # 15 Exhibit O, # 16 Exhibit P, # 17 Exhibit Q, # 18 Exhibit R, # 19 Exhibit S, # 20 Exhibit T, # 21 Exhibit U, # 22 Exhibit V, # 23 Exhibit W, # 24 Exhibit X, # 25 Exhibit Y, # 26 Exhibit Z, # 27 Exhibit AA, # 28 Exhibit BB, # 29 Exhibit CC, # 30 Exhibit DD)(Sorenson, Jennifer)

EXHIBIT H
TO DECLARATION OF
JENNIFER A. SORENSON
Excerpt from Institute of Medicine 1988 Report

Ref-, 1S

INSTITUTE
OF
MEDICINE
REPORT OF A STUDY

Human Health Risks with
the Subtherapeutic Use of
Penicillin or Tetracyclines
in Animal Feed

1988

PREFACE

In 1980, at the request of the Food and Drug
Administration (FDA), a committee of the Assembly of Life
Sciences of the National Research Council (NRC) prepared a
report evaluating the effects on human health of the use of
penicillin and two tetracyclines (chlortetracycline and
oxytetracycline) at subtherapeutic concentrations 1 in animal
feed. That committee concluded that the postulated hazards
to human health from such use of antimicrobials had been
neither proved nor disproved. It drew the conclusion largely
because a direct detailed epidemiologic investigation of the
hazards had not been feasible and in part because it was
impossible to ascertain prior antimicrobial exposures of
individual animal sources of meat products for human
consumption. The committee recommended various epidemiologic
studies (especially of human illness due to salmonellae and
pathogenic Escherichia coli) and monitoring and surveillance
of the occurrence of antimicrobial resistance of enteric
bacteria in humans, animals, and foodstuffs.
Several years later, in 1987, FDA asked the Institute of
Medicine to conduct an independent review of the human health
consequences and the risk associated with the use of
penicillin and the tetracyclines at subtherapeutic
concentrations in animal feed. The Institute established a
committee and gave it a tightly drawn charge: specifically,
to perform "a quantitative risk assessment" of those
consequences--to "assess the adequacy of existing human
health data and use such data to arrive at an estimate of
1 The Center for Veterinary Medicine considers any
extended use of antibiotics in feed at 200 g/ton or less
beyond 2 weeks as "subtherapeutic use," whether it is for
growth enhancement or disease prevention. "Use levels are
generally 200 g or less of penicillin or tetracycline per ton
of feed, but dosage units will vary by species. Levels
approved for growth clainLs,andi_sease prophylaxis are
usually lower than those approved for disease treatment;
however, there is some overlap in the claims for dose levels
of 200 g per ton or less." There is more concern in the
agency "with the length of time the antibiotic is used in
feed than in the level of drug." (FDA personal
correspondence, April 26, 1988)
iii

risk, the basis of which will be justified." If complete
quantification of human health risks was not possible because
of inadequacies of the available data, the committee was to
evaluate the scientific information that had become available
since the 1980 report and make judgments about the magnitude
of the risks. The committee has not addressed risk
management, nor any aspects related to policymaking because
this was not part of its charge.
In its risk assessment, the committee was to address the
following questions:
o Does the subtherapeutic use of penicillin and the
tetracyclines in animal feed result in an increased frequency
of antimicrobial resistance in pathogens, particularly
foodborne pathogens? If so, can the increase in frequency be
reliably estimated and compared with the increases associated
with other sources of resistance?
o Does antimicrobial resistance increase (or diminish)
the ability of foodborne pathogens to cause disease, change
the number of foodborne pathogens (dose) needed to produce
disease, or alter the severity of disease caused by foodborne
pathogens?
o Does the subtherapeutic use of penicillin and the
tetracyclines in animal feed result in increased prevalence
of pathogens in the animals so fed and in foods derived from
them?
o Does antibiotic resistance attributable to
subtherapeutic use in feed increase the incidence of
foodborne infectious disease in humans or complicate its
medical management?
The current committee is well aware of the longstanding
uncertainty of the benefits of the subtherapeutic use of
antimicrobials in animal feed and its possible restriction in
this country and abroad, and it understands the need for a
risk assessment as a foundation for risk management in FDA's
decision-makin g (rule-making) regarding the use of feed
additives.
It is inherently difficult to relate human morbidity and
mortality associated with a specific antibiotic-resistant
bacterial pathogen directly to that pathogen's origin in
1 ives.tx.ck —tor -poultry)_on _a_farm_or in a feedlot and to
administration of subtherapeutic amounts (as opposed to
treatment amounts) of penicillin and the tetracyclines to the
animals. Unequivocal direct evidence linking mortality to
the postulated initial events is not available--certainly not
in sufficient quantity to establish a cause-and-effect
relationship. For want of direct evidence, the committee has
iv

approached its task indirectly by developing a risk model,
using the most reliable data available for the individual
elements involved, including annual numbers of reported cases
of specified infections, fractions of cases due to bacterial
strains that show antibiotic resistance, mortality rates,
fractions of deaths associated with bacterial strains of farm
origin, and fractions of antibiotic-resistant strains of farm
origin caused by subtherapeutic use of antibiotics in animal
feed. Although some bacterial pathogens (salmonellae,
Campylobacter jejuni, enterohemorrhagic E. coli, and Yersinia
enterocolitica) are commonly foodborne and of animal origin,
salmonella infections are the only ones that have been
reportable for many years and for which incidence figures and
antimicrobial-susceptibility data have been collected.
Salmonellosis has therefore been selected for the risk
assessment model, although we acknowledge that several other
human infections would also be relevant to our charge.
The committee is particularly conscious of the
limitations and inherent weaknesses of the data base used in
the risk assessment model. Where an assumption or estimate
had to be made, we have stated its basis. We are aware that
some estimates used in the model are weaker than others; for
example, the fraction of antibiotic-resistant strains of farm
origin attributable to subtherapeutic use of antibiotics (or
penicillin and the tetracyclines specifically). Because some
data for the model were only estimates, we considered a range
of values (low, mid-range, and high) for each element and
expressed the final risk estimates (deaths per year) as
minimum, median, and maximum.
In addition to the risk assessment, the committee has
reviewed further new information pertinent to human health
that might be related to subtherapeutic use of antibiotics in
animal feed. Some of the new information addresses study
possibilities identified by the former NRC committee on
subtherapeutic antibiotic use in animal feeds. Some of it
deals with the biologic impact of antibiotic resistance in
bacteria and the use of molecular biologic techniques in
identifying clonal features of isolates obtained from farm
animals, from foodstuffs derived from livestock and farm
animals, and from infected humans. Some of it reflects
followup experience in European countries that have, in the

last 10-20 years, by regulatory action prohibited use in
animal feed of subtherapeutic concentrations of antibiotics
that are used in treatment of humans. Much of this
information provides only circumstantial evidence bearing on
the question under consideration. Softe of the facts even
appear to be mutually contradictory.
The committee has not addressed any cost-benefit aspects
of the issues related to this problem, nor has it made any
recommendations regarding regulatory strategies or policies.
It hopes that its report on the subtherapeutic use of

penicillin and the tetracyclines in animal feed will be
useful to FDA in its consideration of the risk involved and
appropriate risk management. The committee stresses the
continuing need for more extensive gathering of detailed
epidemiologic information to define the human health risks
more sharply.

vi

X
CONCLUSIONS

The committee has reviewed the extensive and sometimes
conflicting literature pertaining to possible human health
risks associated with the use of subtherapeutic
concentrations of penicillin and the tetracyclines (and other
antimicrobials) in animal feed. It evaluated investigations
of the molecular nature of plasmids, transposons, and other
bacterial antimicrobial-resistance determinants and their
transfer; data on the extent of antimicrobial resistance in
Salmonella species (and in other enteric pathogens) isolated
from humans and farm animals; epidemiologic studies in humans
and farm animals; data on reported cases of human illness and
deaths due to Salmonella transmitted to humans from farm
animals via meat and poultry products; information on the
extent of subtherapeutic use of penicillin, the
tetracyclines, and other antimicrobials in animal feed; and
data from Great Britain on the effects of the restrictions
placed some years ago on the use of antimicrobials in animal
feeds in those countries.
The committee also reviewed the available published
reports dealing with four subjects recommended for further
study in the 1980 report of the National Research Council
Committee to Study the Human Health Effects of Subtherapeutic
Antibiotic Use;in Animal Feeds: the effects of subtherapeutic
and therapeutic doses of antimicrobials on the prevalence of
antimicrobial-resistent enteric bacteria (including
salmonellae) in farm animals; the extent of carriage of
resistance-factor-containing bacteria in vegetarians and
nonvegetarians (to ascertain the extent to which such
carriage is associated with meat consumption); the extent of
carriage of resistance-factor-containing Enterobacteriaceae
in abattoir workers, their families, and neighborhood
controls (to assess the association with occupational
exposure to babteria from animai sources); and the pre-Valence
of urinary tract infections (and urinary tract infections due
to resistance-plasmid-containing Enterobacteriaceae) in
female workers in poultry-processing plants and a control
group of women without contact with farm animals or their
unprocessed meat products.
We consulted with and heard testimony from the
epidemiology staff of the Centers for Disease Control, other
medical epidemiologists, veterinarians, representatives of
193

194
the Animal Health Institute, microbiologists, and
representatives of the pharmaceutical industry.
Using all the resources noted above, we were unable to
find a substantial body of direct evidence that established
the existence of a definite human health hazard in the use of
subtherapeutic concentrations of penicillin and the
tetracyclines in animal feeds. However, we believe that
important--but as yet scant--data indicate the flow of
distinct salmonella strains from farm animals, through the
food processing chain, to humans in whom they cause clinical
salmonellosis. In the one compelling instance of such a
clear link, the multiple-antibiotic-resistant S. newport
originated in farm animals exposed to chloramphenicol, a drug
not approved by the Food and Drug Administration for use in
feed. The committee believes that the molecular
fingerprinting techniques used in this study can provide
(when unique markers are present) the direct evidence needed
to trace the source of antibacterial-resistant bacteria to
human infection. If records of amounts of antibiotic use are
maintained on farms producing food for human consumption,
better evidence can be established for incriminating
subtherapeutic/therapeutic doses in disease outbreaks.
The committee believes that there is indirect evidence
implicating subtherapeutic use of antimicrobials in producing
resistance in infectious bacteria that causes a potential
human health hazard. The evidence is of several kinds:
o There are extensive experimental data on the
properties of R plasmids and their capacity for transfer of
antimicrobial-resistance determinants, both in the test tube
and in the intestinal tract, particularly in the presence of
antimicrobial selective pressure.
o There is evidence of widespread use of subtherapeutic
concentrations of penicillin and the tetracyclines (and other
antimicrobials) on farms and feedlots.
o There is ample evidence of high levels of
antimicrobial resistance among animal isolates of
salmonellae.
o Animal and4Doultry carcasses_in_meat-processing
plants are often contaminated with Escherichia coli and other
enteric pathogens. Few data are available on the frequency
of antimicrobial resistance among such isolates. If the
prevalence of antimicrobial resistance among reported
isolates from diagnostic laboratories is a true
representation of antimicrobial resistance in farm animals
going to slaughter, the frequency of resistance among enteric
pathogens in animal and poultry carcasses would be expected
to be high. However, if the salmonella isolates reported

195
from diagnostic laboratories are principally from animals
that are ill and have received antimicrobials, the figures
would clearly overestimate the frequency of resistant
isolates from meat and poultry carcasses.
o Handling and ingestion of improperly cooked, packaged
frozen or refrigerated meat and poultry contaminated with
bacterial pathogens provides exposure to an infecting
inoculum.
o Experience with antimicrobial drugs in humans over
the last 45 years has revealed the emergence of resistant
strains associated with extensive drug use and the need to
avoid unnecessary and prolonged use, particularly
"prophylactic" use without clear and proven indications.
In addition, the committee has used the results provided
by the risk assessment model presented to estimate
quantitatively the possible risk of mortality associated with
antibiotic-resistant salmonellae due to the subtherapeutic
use of penicillin or the tetracyclines in animals. In the
1980 NRC report, the Committee to Study the Human Health
Effects of Subtherapeutic Antibiotic Use in Animal Feeds
concluded that "the postulated hazards to human health from a
subtherapeutic use of antimicrobials in animal feeds were
neither proven nor disproven." In other words, the risk of
human health as a result of subtherapeutic use of
antimicrobials in feed was not estimated.
We found the available data base on some aspects of the
problem to be limited in quality and quantity; indeed, the
data had not been gathered prospectively for the purpose of
this type of analysis. The committee has used what it
considers the best available information, indicating, where
appropriate, the inherent weaknesses in the data.
Admittedly, in some instances, we used only the best
estimates available in the risk assessment. The assessment
does indicate the presence of risk. Although it does not
provide a distinct numerical "answer" to the question of the
magnitude of the human health risk involved, it does provide
some indication of the probable size of the risk in terms of
numerical estimates or ranges. These are presented below as
numbers of deatho per year-attributable-to the 1.11oth farapeuticuse of antimicrobials (or penicillin and tetracyclines) in
the listing of specific conclusions:
BIOLOGIC IMPACTS
o Use of each new antimicrobial agent over the last
half-century has eventually mobilized genes that encode
resistance to the agent and disseminated them widely through

196
the world's interconnecting bacterial populations. Use of
the antimicrobial agent disseminates the resistance genes in
stages, each of which begins with a rare molecular event that
facilitates further dissemination. Although use of
antimicrobials in a patient or the patient's neighbors might
have triggered overgrowth and clinical manifestation of the
resistant strain, the evolution and delivery of its
resistance genome was the result of prior use in many,
probably distant, bacterial populations.
o Results of surveys of isolates of salmonellae from
animals and humans in the United States and restrictionendonuclease fragment patterns of resistance plasmids from
selected isolates suggest that clones of resistant
salmonellae are endemic in animals and sporadic or
occasionally epidemic in humans.
o Herds of farm animals given subtherapeutic amounts of
antimicrobial agents have more antimicrobial-resistant
intestinal bacteria than herds given no antimicrobials.
o The most important determinant in the selection of
antimicrobial-resistant strains in a bacterial population is
exposure of that population to antimicrobials. Total
duration and concentration of antimicrobial use are important
in selection for resistance. Any measure that fails to
reduce total use appreciably is unlikely to affect the
prevalence of antimicrobial-resistant strains.
o Resistance to antimicrobial drugs among salmonella
strains can interfere with the efficacy of antimicrobial
therapy of human salmonellosis. (Such resistance is usually
R-plasmid-mediated, so it can involve other drugs, such as
trimethoprim-sulfamethoxazole, chloramphenicol, and
ampicillin.) Although such interference with the efficacy of
therapy almost certainly occurs (i.e., patients are treated
with an antimicrobial that is ineffective because of drug
resistance), it is probably quite uncommon in nontyphoidal
salmonellosis.
o The available data are inadequate to conclude that
either o
tions of
antimicrobials are more selective of drug-resistant bacteria.
On theoretical grounds, it is likely that therapeutic and
subtherapeutic dosages exert equal selective pressure for
, •u b. erapeutic dosages
exert more pressure for conjugative transfer of drug
resistance, because of the dosages and the durations of
administration.

197
o Animal and poultry products (including veal, beef,
pork, chicken, eggs, and milk) are the principal sources of
human nontyphoidal salmonellosis. Also, some E. coli
serotypes can also be found in the intestinal flora both of
humans and of farm animals. Thus, there could be an
interconnecting link between these two large pools of enteric
microorganisms, facilitated by the high frequency of
contamination of animal and poultry carcasses in
slaughterhouses. Such a potential link would provide a means
of movement of R plasmids of farm origin to the human
alimentary tract. The interconnection, because of its
nature, would constitute an almost exclusively one-way
passage.
o The overall prevalence of resistance to any of five
commonly used antimicrobials is about 4 times as great in
collections of salmonella isolates from farm animal and
poultry (65%) as those in collections of isolates from humans
(15.5%). This difference suggests that the predominant pool
of resistant salmonellae is in farm animals. Because
ultimately almost all human infections with nontyphoidal
salmonellae result from strains originating in farm animals,
the antimicrobial resistance observed in human isolates most
likely is derived from the animal pool of resistance genes,
rather than from selection due to antimicrobial use in
humans.
EPIDEMIOLOGIC FINDINGS
o Evidence is sparse that directly links the use of
penicillin and tetracycline in subtherapeutic concentrations
in animal feeds to human infections. Several studies have
yielded reliable evidence of spread, from farm animals and
poultry to humans, of E. coli strains in which antimicrobial
resistance had been induced by administration of
subtherapeutic concentrations of antimicrobials as feed
additives. There is evidence from only one study of the
direct spread of multiple-antimicrobial-resistant salmonellae
from farm animals to humans via meat products. However, the
antimicrobial used on the farm was . •
chloramphenicol, a drug
-=e or
! --e _ee
---not—a *
food production. It might be difficult, or impossible, to
provide a total chain of evidence directly relating the
majority of cases of human infection with antimicrobial. •
re . t
to relate the presence of the resistance to the use of
specific antimicrobials in subtherapeutic concentrations in
feed. By the time a detailed investigation of an outbreak of
human salmonellosis occurs, evidence of prior antimicrobial
use patterns might not be available.

198
o It has not been possible to determine whether
antimicrobial resistance of salmonellae caused by the
administration of subtherapeutic concentrations of
antimicrobials in animal feed increases the number of cases
of human salmonellosis.
o Whether the presence of antimicrobial resistance in
salmonellae increases virulence is uncertain; the available
data are limited and conflicting. In special circumstances,
as when R plasmids are linked with virulence genes (e.g.,
those for enterotoxin or hemolysin in E. coli), selection by
antimicrobial agents might promote spread of virulent
strains; however, such an occurrence has only rarely been
reported. It is not clear whether the overall prevalence of
salmonellae in food products is increased by virtue of
antimicrobial resistance. However, the incidence of human
salmonellosis in the United States is increasing, and the
increase is unlikely to be an artifact of better reporting.
As long as most strains of Salmonella are susceptible to the
antimicrobials to which they are exposed, subtherapeutic
administration of antimicrobials might reduce the prevalence
of salmonellae in meat and poultry products that humans
ingest. However, as the prevalence of resistant strains
increases because of repeated and prolonged exposure to
antimicrobials, subtherapeutic administration might actually
favor the increase by suppressing the normal competing flora
and promoting R-plasmid spread. Direct proof of this pattern
in salmonellae in farm animals is lacking.
o The current frequency of R-plasmid-mediated
antimicrobial resistance among isolates of E. coli and
salmonellae in the intestinal contents of farm animals and
poultry is high--much higher than in human isolates. It
would be difficult to predict the period required, after
curtailment of the use of subtherapeutic concentrations of
penicillin and the tetracyclines in animal feed, for Rplasmid-mediated antimicrobial resistance to decrease in any
extent in salmonellae and E. coli strains. Major decreases
might occur only after the passage of years, in view of
(1) the current degree of resistance, (2) the extensive
environmental contamination on farms and feedlots with
resitant organisms, (3) the prolonged prior subtherapeutic

use of antimicrobials, which has allowed extensive permeation
of resistance genes (transposons) throughout the highly
colonization-adapted coliform_flora of farrnanimalsi_and_
(4) the -need—to introduce competing, antimicrobialsusceptible conform bacteria. Results of studies in
confined populations of swine indicate that it could take
many years for major decreases in levels of resistance to
occur.

199
o Although the extent of antimicrobial resistance among
salmonella strains isolated from humans is probably growing,
it is still low enough for suitable intervention to forestall
possible further increases and eventually to lower the
overall extent of antimicrobial resistance.
ANTIBIOTIC USE PATTERNS
o The use of subtherapeutic dosages of penicillin and
the tetracyclines in animal feeds is extensive in the United
States. Such use is for the purpose of either growth
promotion or disease prevention and often continues for a
substantial portion of the growth cycle of farm animals. The
specific rationale for use in a given herd at a given time is
not always clear. Of over 31 million pounds of
antimicrobials produced each year in the United States, about
42-48% is designated for addition to animal feeds or other
unspecified (minor) uses. The best estimates (they are only
estimates) indicate that penicillin and the tetracyclines
account for almost 60% of the antimicrobials sold to the feed
trade (and presumably ultimately used on farms and in
feedlots). Of the total amount of tetracyclines produced in
this country, for use in both humans and animals,
approximately 70% is sold for use in livestock and poultry
feeds. An estimated 88% of all antimicrobial use in
livestock and poultry is in subtherapeutic concentrations.
Thus, subtherapeutic use of penicillin, the tetracyclines,
and other antimicrobials in animal feeds--which accounts for
some 40% of antimicrobial production in the United States-constitutes a sizable segment of the total antimicrobial
selective pressure (for resistant enteric microorganisms)
exerted on the combined human and farm-animal intestinal
bacterial populations.
o Interpretation of the results in Great Britain after
banning the subtherapeutic use of penicillin and the
tetracyclines in animal feed is difficult, in part because
total farm use of these antimicrobials might not have
decreased because use could have taken the form of
therapeutic or prophylactic doses in feed for disease
prevention as prescribed by a veterinarian. The
-treatment—er
appearance of new epidemic strains of antimicrobial-resistant
salmonella serotypes during the period of interdiction of
subtherapeutic use further confounds interpretation. It

might-take-years-for-dilution of unlimierobial-resistan
strains of Salmonella and E. coli in the farm animal
population before any substantial changes might be
observable.

200
RISK ANALYSIS
o The committee has been unable to find substantial
direct evidence that bacterial resistance resulting from the
use of subtherapeutic concentrations of penicillin or the
tetracyclines in animal feed causes an excess risk to human
health as a result of consumption of food products derived
from the treated animals, as a result of contact with such
animals, or as a result of exposure to an environment
contaminated by resistant enteric bacteria from such animals.
Lacking this direct evidence, the committee turned to the
tools of risk assessment to develop some quantitative
estimate of the probable risk to human health associated with
this form of the subtherapeutic use of these antimicrobials.
o Use of penicillin and the tetracyclines in
subtherapeutic concentrations in animal feed has led to
increased antimicrobial resistance in foodborne commensals
and pathogens. The risk analysis in this report focused only
on human infection with salmonella serotypes, because
available data on other species were insufficient. The
committee has not assessed the potential risk to human health
associated with drug resistance in other gram-negative
bacillary species (Campylobacter jejuni, Yersinia
enterocolitica, and enterohemorrhagic E. coli) of animal
origin, because the data on human cases are too limited and
because antimicrobial susceptibility data on those bacteria
are not routinely obtained.
o Because the committee's risk assessments are based on
estimates using sparse data, these estimates should be
interpreted and used with caution. Such estimates are best
seen as scientific hypotheses about the possible extent of a
problem. This does not mean that they are "hypothetical" in
the weak sense of being speculative. Rather, they are
hypotheses that are consistent with all available information
and scientific understanding, but they have not been tested
by traditional scientific methods. All the estimates
presented in this report should be viewed in that
perspective.
ID -Annucti 1unthers5r deattff-from salmonellosis

attributable to subtherapeutic uses of any antimicrobials for
prophylaxis and growth promotion have been estimated. The
likeliest estimate is 70 deaths per year.

o The likeliest estimate of mortality from
salmonellosis attributable to subtherapeutic uses of
penicillin/ampicillin and/or tetracycline for prophylaxis and
growth promotion is 40 deaths per year. Caveat--these are
not necessarily "excess deaths," but rather estimates of the

201
yearly mortality attributable to salmonellosis of the
indicated origin. The deaths might to some extent replace
deaths (in the same patients or others) that occur from
infections due to salmonellae susceptible to penicillin/
ampicillin and tetracycline if subtherapeutic dosages of
these antimicrobials had not been used in animal feed.
Estimation of such "replacement" of deaths is not possible
with the evidence at hand.
o The likeliest estimate of mortality from
salmonellosis attributable to subtherapeutic uses of any
antimicrobial for growth promotion only is 20 deaths per
year. As in the preceding (and following) estimates, the
caveat regarding "excess deaths" applies.
o The likeliest estimate of mortality from
salmonellosis attributable to subtherapeutic uses of
penicillin/ampicillin and/or tetracycline only for growth
promotion is 15 deaths per year.
o The likeliest estimate of mortality from
salmonellosis in the "etiologic fraction" attributable to
subtherapeutic uses of any antimicrobial for prophylaxis and
growth promotion is 6 deaths per year. The "etiologic
fraction" is the proportion of persons exposed to an
antimicrobial-resistant salmonella strain who are at
increased risk of illness by virtue of recent use of
antimicrobial drugs for whatever reason. Therefore, such
deaths can be considered as "excess deaths"; i.e., they would
not occur if the infecting salmonella strain were not
antimicrobial-resistant and if its multiplication were not
promoted, presumably, by suppression of growth of the
competing normal antimicrobial-susceptible normal flora. In
the same way, the number of foodborne pathogens (inoculum
size) needed to precipitate disease might have been
decreased. Whether a similar effect can be produced by prior
antimicrobial use in persons infected with antimicrobialsusceptible salmonellae (due to possible differential
antimicrobial susceptibility between susceptible salmonellae
and normal components of the intestinal flora) is unknown,
and the committee has not been able to find data bearing on
o The likeliest estimate of mortality from
salmonellosis in the "etiologic fraction" attributable to
subtherapeutic uses of peniciIIIntampiciIIin-andior
tetracycline for prophylaxis and growth promotion is 6 deaths
per year.
o The likeliest estimate of mortality from
salmonellosis in the "etiologic faction" attributable to

202
subtherapeutic uses of any antimicrobial only for growth
promotion is 2 deaths per year.
o The likeliest estimate of mortality from
salmonellosis in the "etiologic fraction" attributable to
subtherapeutic uses of penicillin/ampicillin and/or
tetracycline only for growth promotion is 2 deaths per year.
o Infections with antimicrobial-resistant strains of
Salmonella are more often fatal than infections with
susceptible Salmonella. Therefore, the increased difficulty
of providing effective therapy for human disease can be
estimated. The increased difficulty in providing effective
treatment may be due to increased virulence of antimicrobialresistant strains, to the presence of resistance to one of
the antimicrobials ordinarily used to treat such infections
when they are severe or when they occur in particularly
vulnerable persons, or to some other factor. The likeliest
estimate of mortality from salmonellosis arising because of
increased difficulty of treatment attributable to
subtherapeutic uses of any antimicrobial for prophylaxis and
growth promotion is 40 deaths per year.
o The likeliest estimate of mortality from
salmonellosis arising because of increased difficulty of
treatment attributable to subtherapeutic uses of
penicillin/ampicillin and/or tetracycline for prophylaxis and
growth promotion is 20 deaths per year.
o The likeliest estimate of mortality from
salmonellosis arising because of increased difficulty of
treatment attributable to subtherapeutic uses of any
antimicrobial only for growth promotion is 8 deaths per year.
o The likeliest estimate of mortality form
salmonellosis arising because of increased difficulty of
treatment attributable to subtherapeutic uses of
penicillin/ampicillin and/or the tetracyclines only for
growth promotion is 8 deaths per year.
o Evaluation of the foregoing estimates of mortality
—from aalmonellosis attributable to subtherapreutic u5eb -of

antimicrobials in animal feed requires consideration in a
broader context. What possible benefits accrue from such
subtherapeutic use of antimicrobials in food production?
---Woul-d---human-deaths—from salmonellosis be reduced Mr the
discontinuation of subtherapeutic use of penicillin/
ampicillin and/or the tetracyclines? The committee's thesis
is that, although some deaths due to antimicrobial-resistant
strains might be "replaced" by deaths due to susceptible
strains, the total number of deaths would decrease, however,

203
this cannot now be proved. The committee offers no
recommendations regarding policy-making because that was not
part of its mandate.