Benefits and Risks to Human Health
The public has long-standing concerns over potentially harmful drug residues in foods. Many consumers fear that neither the facts regarding the consequences of drug use in food animals are being made available nor are enough animal-derived foods available—or affordable—that allow them to select safe products. The possibility that chemical additives, drugs and their metabolites (drug residues) could cause allergic reactions or disease is not taken lightly by the public or by health care professionals (ERS 1996a). Similarly, the threat of human disease posed by microbial contamination is well documented and increasingly acknowledged and publicized (IOM 1998).
The threat of antibiotic resistance is most commonly associated with the emergence of resistance outbreaks in hospital settings and with improper human applications of antibiotic therapy (CDC 1994; IOM 1998). The cause-and-effect relationship between therapeutic administration of antibiotics and resistance is more readily ascertained—and statistically quantifiable—in hospitals than it is in animal production sites, processing and packaging plants, and transport depots common in animal agriculture. It has been difficult to track and document the link between antibiotic use in farm animals, the development of antibiotic resistance, and disease transference to humans. However, the reporting of such data is increasing with the development of larger and more accessible databases, refined culture and detection methods, and the overall heightened awareness and concern for this potential source of disease. The statistics are more apparent for zoonotic
transfer of overt pathogens that cause specific diseases that must be reported to state or federal health agencies (Lyme disease, rabies, salmonellosis).
The data are increasing (and referenced later in this report) on the transfer of pathogens from farm animals to humans where issues of antibiotic resistance patterns in the invading organism are more frequently tracked. Many of these data come from case studies that followed reported infection and disease in higher risk groups, such as farmworkers (where epidemiological tracking has identified the source). Increased data collection on antibiotic resistance patterns is occurring largely as a result of implementation of newer technologies (developed within the past 5 to 10 years) on a broader, more affordable, and “user friendly” scale and format. In addition, databases on disease occurrence in particular food-animal species are increasing at a rapid rate.
In large part, the appearance of increasing health problems in food animals does not reflect an increase in incidence. Rather, it indicates an increase in documentation of what was probably there all along. The new data arise because of increased vigilance among producers and veterinarians who want to identify problems and provide treatments quickly to maintain productivity. Many of the successes in this effort are the direct result of voluntary implementation of quality-assurance programs and accountability procedures that are expanding throughout the food-animal industry.
The operating premises can be summarized as follows:
Antibiotic resistance is a documented major health threat around the world that has been given high priority by many health agencies (WHO 1997; IOM 1998).
Inappropriate or irresponsible uses of drugs in humans and animals in subtherapeutic and therapeutic regimens contribute to the development of drug resistance (IOM 1998).
There are opportunities in the microbial environment for interconnected ecosystems to allow exchange of DNA, promoting the spread of resistance from one genus to another. The combination of increased bacterial virulence and increased drug resistance creates a potential for increased risk of morbidity and mortality for animals and humans that some have extrapolated to a catastrophic potential. “Catastrophic” and “crisis” are words often applied to this issue, and they evoke emotional, sensational, and oftentimes inflammatory reactions that tend to distract the focus from the goal of factual assessment and hypothesis testing.
Human exposure to pathogens from animal-derived foods has been documented and can result in human disease. The relationship between those diseases and the emergence of antibiotic-resistant disease is less clear, less frequently tracked, and constitutes an area in which there is a fundamental dearth of valid data. Between the farm and the table, the large number of places and opportunities for bacteria to be introduced into the human food chain is an important factor
in the emergence of food-related illness. Irresponsible actions by individuals both before and after harvest of the food (improper storage, poor home sanitary practices, improper cooking techniques) undermine the effort to control microbial proliferation through responsible regulatory compliance, surveillance, and quality assurance. However, sterile packaging and irradiation could substantively alter (eliminate) the capability for even drug-resistant organisms to proliferate in foods prior to cooking and decrease the assessed risk to humans.
Increased international trade, reduced barriers to transport, increased efficiency in processing and delivery, and higher consumption approach or, in some cases, exceed the capacity of current surveillance mechanisms. It is virtually impossible to prevent infectious agents in food from reaching consumers, and efforts toward this end need to be strengthened.
The federally established standards and allowable tolerance levels for many drugs and residues are not zero, and detection of residues should not be equated with adulteration. No assurances can prevent ignorant action, accidents, or breaching of ethical standards in the use of animals that result in animal-derived foods, being adulterated with drug residues. Sophisticated methods for monitoring residues can be used to remove tainted products from the food chain, but every carcass cannot be monitored.
Bacteria are a natural part of the body’s internal and external ecology and environment. Some bacteria are beneficial, most are benign, and their presence is kept in balance through the functions of the immune system, naturally produced antibacterial peptides in skin and epithelial tissues, and microbial populations normally competing with “foreign” bacteria within a stable internal environment. Bacterial infections in any animal, including humans, fall into two categories: subclinical and occult; clinical and overt. Animals and humans can have low levels of pathogens that do not cause detectable disease or illness. A stable internal environment is critical for maintaining health. If environmental, nutritional, or behavioral stresses impinge on an animal or human population, the imbalance in the internal environment (altered adrenal and glucocorticoid hormone concentrations, altered cytokine concentrations, metabolic acidosis, and ruminal disturbances) can trigger the proliferation of bacterial populations that become harmful by spreading infection or release of endotoxins and exotoxins.
Antibiotics are used to treat infections, but maintaining the animal’s internal environment (the gastrointestinal tract and absorptive processes) is another use in animal production. This involves giving antibiotics for longer periods of time and at concentrations lower than those administered for therapeutic treatment (Fagerberg and Quarles 1979).
Antibiotics can be applied in three ways. In one, a single antibiotic is administered at subtherapeutic concentrations for an extended period to maintain the
normal population of gastrointestinal microorganisms and prevent emergence of any that could be pathogenic. The second is the use of rotating classes of multiple antibiotics at low, subtherapeutic concentrations. Again, the aim is to eliminate the development of opportunistic bacteria that could emerge as pathogenic or be passed from one animal to another. This strategy is used when animals are transported from one location to another, where the surroundings and feeding methods are different and the animals are reared with more-intensive management practices. The potential for antibiotic-resistant populations of organisms to develop still persists. Therefore, a third application strategy, involving a gradient subtherapeutic regimen, is introduced. Antibiotic concentrations are gradually increased, so that the effective dose for bactericidal action is greater, at least in theory, than a concentration of antibiotic to which microorganisms might have resistance. This strategy is effective because both the efficacy of a drug in controlling disease and the development of resistance are dose dependent. The benefits to animals and humans associated with overall therapeutic antibiotic use in food animals outweigh the risks of use because the development and spread of pathogenic organisms are held in check (CAST 1981).
In assessing the risk–benefit ratio of antibiotic use in food-producing animals, the nature of the applications for which antibiotics are either prescribed or administered must be known. The exercise of ranking risks and benefits to animals and humans of antibiotic use in food animals might change dramatically according to who assesses the risk and how the availability of related facts strengthens or weakens hypotheses derived from conceptual possibilities. A significant threat to humans exists in the form of zoonotic transmission of diseases. Zoonotic infection results from an animal pathogen that is transmitted directly to humans causing a similar infection. Examples of potentially life-threatening zoonotic infections are tuberculosis, leptospirosis, toxoplasmosis, brucellosis, salmonellosis (DT-104), hemorrhagic Escherichia coli O157:H7 (colisepti-cemia), and rabies, to name a few. Treatment is the first response when microbial disease is diagnosed in any animal. For a clinically infected animal, the choices are to treat it with therapeutic concentrations of antibiotics for a defined course of administration or not to treat it at all. If the animal is not treated, the organisms can spread throughout the environment to infect other animals and humans and possibly to decrease the animal’s productive lifetime (Fagerberg and Quarles 1979).
If the animal is treated, there is a small chance that some microorganisms could become resistant to the class of antibiotics administered. In some cases, the bacteria developing resistance might, in fact, not even be the species causing the disease (CAST 1981). The risk in antibiotic use in food animals (that is, giving antibiotics to cure or prevent disease) is seen by some as a human health benefit,
because treating a sick animal directly maintains the health of other animals and humans (Carneval, R. 1997. Animal Health Institute, Alexandria, VA, personal communication). Some risk is involved in the practice of giving antibiotics to animals, but the ranking of risks and benefits cannot be accomplished easily because of the lack of validated data and controlled studies.
BENEFITS OF ANTIBIOTIC USE
Antibiotics are used in food-animal production for the primary benefit of (1) the health and welfare of the animal (Gustafson 1986; Ziv 1986), (2) carcass quality and overall efficiency of growth and production (Langlois et al. 1986; Mackinnon 1993), (3) economics (CAST 1981; Walton 1986), and (4) human public health. The benefit to human health in the proper use of antibiotics in food animals is related to the ability of these drugs to combat infectious bacteria that can be transferred to humans through direct contact with the sick animal, through consumption of food contaminated with pathogens, or through proliferation in the environment. The advantages of antibiotic use in animals are related to the prevention of overt bacterial disease and improvement in animal performance through reducing the physiological costs of limiting growth that are incurred in the process of fighting low-level and overt disease (Hays 1986; Espinasse 1993). Those limitations need to be minimized to permit better nutrient use, enhanced growth rate, and feed efficiency (Elsasser et al. 1995, 1997; Beisel 1988; Roura et al. 1992; see earlier discussion in Chapter 2). However, because of the controversy surrounding the development of antibiotic drug resistance in animal and human populations, and because of the consequences for human health and clinical practices, use of antibiotic drugs in food-producing animals has been questioned by the Food and Drug Administration (FDA), policy makers, health care professionals, and consumer organizations, among others, and has been studied regularly since the 1960s (see IOM 1989; OTA 1995) as directed by several federal agencies. Some groups have argued for a substantial reduction in the use of antibiotic drugs in food-animal production. Others contend that microbial contamination of animal-food products would increase without the use of these drugs. The following summaries of data and studies suggest that antibiotic use in farm animals is largely beneficial:
• Antibiotic treatment of humans who have enteritis caused by Salmonella is generally contraindicated. General intestinal enteritis usually is self-limiting and resolves relatively quickly; a greater risk is associated with the development of resistant Salmonella in individuals who have used oral antibiotics within a month of Salmonella exposure (Riley et al. 1984). Systemic, invasive Salmonella requires antibiotic intervention, and the newly emerging multidrug-resistant strain of Salmonella, DT-104, could pose an even more significant threat to human health because of the increasing number of treatment failures encountered as
isolates emerge for which treatment options are limited (Wall et al. 1994; Wall et al. 1995).
Treatment of Salmonella infection is widely used in veterinary medicine, particularly for swine. As several investigators (DeGeeter et al. 1976; Gutzmann et al. 1976; Wilcock and Olander 1978; Jacks et al. 1981; Schwartz 1991) reported that vigorous antibacterial therapy (in combination with supportive therapy) early in the course of septicemic salmonellosis significantly reduces the magnitude and the duration of shedding of organisms. These investigators pointed out that if such septicemic cases were not treated, shedding of the organisms would increase, and Salmonella isolations from carcasses (from apparently healthy animals) would increase. The significance of this would be apparent in the greater risk for Salmonella to enter the food chain at slaughter and even more directly contaminate the hog environment, fostering the persistence of the problem.
• Drug therapy is effective in controlling and reducing the spread of a number of zoonotic infections, including leptospirosis in cattle. In one clinical case, proper treatment of that disease eliminated shedding of the organism. Without drug therapy, however, Leptospira can contaminate the environment, including milk and water, to create a health risk for humans (Jackson 1993). Similar reduction in the shedding of pathogens with drug treatment has been shown for Campylobacter fetus (Kotula and Stern 1984; Wokatsch and Bockemuhl 1988; Jackson 1993). Other major food-borne bacterial pathogens that cause significant human health problems associated with contamination of meat products are Streptococcus suis, E. coli, especially O157:H7, Salmonella spp., Enterococcus spp., and Yersinia (Clifton-Hadley 1983; Walton 1985; Tauxe et al. 1987; IOM 1992; CDC 1994). Proper treatment of infections from those pathogens at clinical presentation can reduce or eliminate the spread of infectious agents. “In the absence of evidence to the contrary,” Mackinnon (1993) inferred that use of antibiotic drugs in pigs could reduce the transmission of some of these zoonotic diseases.
• From an economic standpoint, the therapeutic use of antibiotics to combat active infection in individual animals and herds is unquestioned. The economic benefit of subtherapeutic antibiotic use is more often debated—especially by those not aligned with the animal production industries. However, the overall economic benefit is made possible because of a 1 to 15 percent increase in feed efficiency and performance (growth rate, egg production) over similar animals that do not receive antibiotics (see earlier discussion in Chapter 2). The magnitude of the production response to low concentrations of antibiotics is influenced by animal age, diet, stress, duration of drug usage, and general cleanliness of pens, and stocking rates (Fagerberg and Quarles 1979). One could argue that this occurs only because of the impetus to intensify production practices, but this is the way that food-animal production is accomplished, and the economic benefit is apparent for these systems (CAST 1981).
Inspection at slaughter results in rejection of a proportion of carcasses—most commonly for abscesses, arthritis, pneumonia and pleurisy, peritonitis, and fever (including septicemia). Survey results on 1.3 million pigs slaughtered at abattoirs in the United Kingdom (Hill and Jones 1984a,b) indicated that 262,149 kg of meat and 273,080 kg of liver, heart, and lungs were rejected, contributing to millions of dollars lost in the production of the animals and an inability to recoup the investment input. The greater problem was that the pigs that went to market were not visually different from any other pigs that were slaughtered and that had passed inspection. The investigators concluded that many of the rejections were associated with localized lesions and further suggested that this valuable data resource (slaughter rejection data) was substantively underused in the identification of cost-effective practices to enhance animal health.
The effects of antibiotic drug use in many species are associated with a generalized decrease in health problems in the animals in which they are used (CAST 1981). For example, in the summary prepared for the Council for Agricultural Science and Technology report on antibiotics in animal feeds (CAST 1981), the use of chlortetracycline, oxytetracycline, erythromycin, tylosin, and bacitracin in cattle was associated with a significant reduction in the incidence of liver abscesses. Additional data demonstrate that the decrease in weight gain in abscessed cattle was lower than it was in nonabscessed cattle. All of these subclinical issues add to the expense of raising food-producing animals, and the use of the drugs is associated with improvements in animal health and in economic productivity (CAST 1981). In addition, Mackinnon (1993) summarized data from 12 swine-finishing farms where, throughout the year, a veterinary preventive medicine scheme was implemented to curb the effects of infection on production characteristics and carcass rejections. The introduction of veterinary advice coupled with selective use of medication to eradicate pneumonia and swine dysentary led to a progressive decline throughout the year in offal losses and carcass rejections and decreased carcass rejection variation (Table 3–1).
Among other pathogenic microorganisms cited as food-borne hazards, Erysipelothrix rhusiopathiae (in swine and turkeys) and Listeria monocytogenes (in sheep and cattle) also cause clinical disease in animals that might be treated successfully with antibiotics.
Human health concerns associated with antibiotic use often focus on the more nebulous connections between subtherapeutic use in animals and their consequences, but therapeutic uses also present a set of risk concerns. An assessment of some aspects of the economic consequences of partial or total restriction in subtherapeutic drug use appears in Chapter 7.
POSSIBLE HAZARDS OF ANTIBIOTIC USE
Scientific literature can be cited to support the opinion that antibiotics used in food-animal industries are fundamentally benign to human health (Frappaolo
TABLE 3–1 The Effect of Implementation of a Veterinary Preventive-Medicine Scheme on Offal and Carcass Rejections from 12 Finishing Farms
Carcass Rejectionsc ± SD
9.4 ± 5.0
306 ± 267
8.1 ± 6.9
303 ± 287
6.4 ± 5.0
219 ± 139
5.0 ± 3.1
216 ± 77
aValue of rejected lung, heart, liver, and intestines, pence.
bSD = standard deviation.
cWeight (g) of meat and bone rejected per pig slaughtered.
Source: Mackinnon 1993.
1986; Van den Bogaard 1993). However, the Institute of Medicine (IOM 1989) and the Office of Technology Assessment (OTA 1995) reported on circumstantial evidence linking subtherapeutic use of antibiotic drugs in farm animals to potential human health hazards. The committee members who prepared those reports suggested that caution be used in extrapolating conclusions too generally given the paucity of data on the reviewed issue.
Antibiotic Resistance as a Human Health Risk
Many bacterial species multiply rapidly enough to double their numbers every 20 minutes. With even the simplest bacterial genome, the replication processes are imperfect and, statistically, chromosomal mutations and genetic DNA alterations develop that result in the expression of altered biochemical makeup of some feature of the affected bacterium. The ability for bacterial populations to adapt to changes in their environment and survive otherwise inhospitable conditions often results from the development of favorable mutations that allow for the coding of specific proteins or processes that are not affected by the impinging condition. For example, a hypothetical case can be constructed to suggest how easily an invading bacteria could proliferate to cause disease (Cooper 1991). Suppose a favorable alteration in a bacterial phenotype (the physical expression of the genetic coded information) occurs with the unlikely frequency of 1 in 1 billion. Assume that the average time for bacterial replication is 20 minutes. If an infection were initiated with 1,000 organisms, a first mutational event might occur in one organism after only 7 to 14 hours. Once that occurred, the relative proliferative capacity of the bacteria would allow it to attain significant numbers within 24 to 48 hours, given the longer replicating time in vivo in contrast to in vitro, or in a healthy animal in contrast to one whose immune system is overwhelmed. These events are fundamentally random, and the prolif-
erating numbers are a function of statistics and probability. Therefore, the task of assessing the actual biological consequences is extremely difficult.
Bacterial populations respond to imposed environmental conditions and pressures by adapting and proliferating to become versions of the original populations that are better able to survive in new conditions. The new offspring are strains, and the term applied to the developed ability of the strain to fend off the survival threat is resistance. The factors that allow the resistant organisms to proliferate in the prevailing conditions are selection pressures.
The interaction of the animal’s biological host defenses, coupled with the action of antibiotics, even when those antibiotics are used at subtherapeutic concentrations, is often overlooked. It often is either forgotten or dismissed because of the difficulty of assessing in vivo responses compared with the simplicity, cost, and turn-around time of in vitro antimicrobial experiments. The sensitivity of the organism to selection pressure is complex. There are clearer boundaries in vitro to define the effectiveness of antibiotics to achieve killing and conversely to suggest the degree to which a bacterium is sensitive to a given drug. Very low drug concentrations might be ineffective in vitro in incapacitating the growth of a given bacterial population, and high concentrations might be required to be effective. However, as a caveat, the concentration of antibiotic that kills an organism in vitro might not affect the organism’s survival in vivo. Certainly, the ability of the animal’s immune system to interact with a chemotherapeutic agent to clear and eliminate invading organisms must be considered. There are clear data from biomedical research to suggest that the natural host defenses against invading bacteria are increased with the use of antibiotics. Furthermore, several studies illustrate the fact that the use of subtherapeutic concentrations of antibiotics increases specific immunological responses of the host to the invading bacteria (Easmon and Desmond 1982; Veringa and Verhoef 1985; Hand et al. 1989). Although many of these effects are reported for phagocytosis and opsonization of bacteria, the story is far from clear. Other data suggest that some antibiotics, such as the cephalosporins (Gillissen 1982), increase immunoglobulin production but decrease lymphocyte blastogenic capability (Chaperon 1982); still others, such as the rifamycins (Bassi and Bolzoni 1982) affect immunosuppression.
The drug concentrations that can kill a given microbial species also might be toxic to humans or animals. For example, chloramphenicol is highly effective against many pathogenic microorganisms. Although well tolerated in domestic animals, this antibiotic in humans results in the non-dose-related development of aplastic anemia. As a result, chloramphenicol has been banned from use under any circumstance in food-producing animals because of possible residue carryover (Merck Veterinary Manual 1986).
The emergence of resistance in a bacterial population does not automatically signal the emergence of a pathological disease corollary. Similarly, in animal production, the emergence of resistance does not necessarily confer inefficacy on subtherapeutic antibiotic use. However, several cases of human illness from
antibiotic-resistant pathogens that originated in antibiotic-treated livestock have occurred (IOM 1989). Likewise, there is a report in the literature of a Salmonella infection of a mother and nursery infants that was associated with the mother handling sick calves that had recently arrived on the farm from several locations. The resistance patterns of the bacteria (chloramphenicol, sulfa-methoxazole, and tetracycline) were unique, but the calves presumably were infected before coming to the farm and without direct administration of those antibiotics (Lyons et al. 1980).
Recent studies on plasmid transfer between bacteria have suggested that resistance factors can be linked with genes that code for enhanced virulence (the capability to cause disease). Consequently, the potential for animal-to-human transfer in this fashion exists. The risk is greater than zero, but basically incalculable, and the threat is perceived to be significant (WHO 1997; IOM 1998). The use of perceived here is stressed. The threat might be real, and case studies have shown that the passage of resistant organisms from animals to humans can occur and be perpetuated and amplified through food (Spika et al. 1987).
The question remains, How likely is that to happen? The answer is not available and can be addressed only with the development of the proper database and effective risk analysis. The database should be generated jointly by regulatory agencies; animal, pharmaceutical, and health-care industries; and academic basic and clinical science departments. It must be open to all concerned parties.
Antibiotic Resistance Trends
A 1994 Science editorial, “The Biological Warfare of the Future,” described the issue of antibiotic resistance as “a menace of major proportions to the health of the world” (Koshland 1994). Most of the issue in which the editorial appeared was devoted to a discussion of the problems in antibiotic resistance. With current funding restricting the development of new agents (Culotta 1994) and with a paucity of promising new antibiotic drugs for veterinary and human use occurring at a time of emerging multidrug-resistance problems, the health and well-being of the U.S. and European human populations are seriously threatened (Kingman 1994). Microbial resistance to antibiotics is a global issue that amounts to what some health professionals consider a crisis (Kunin 1983 and 1993; Levy 1992; Burke and Levy 1985; Neu 1992; Cohen 1993). This is reflected in the stand taken by the World Health Organization (WHO) in its world health report statement (WHO 1998). Kunin (1993) outlined the response of many multinational groups and their efforts to control the problem, particularly in human use and applications. Many of those efforts involve increased education and broadened awareness of the proper and improper use of these powerful drugs, largely based on documentation of disease in hospitals and health care facilities. Concerns about the agricultural use of antibiotics were raised because of the large amount of the drugs used and the potential for disease to occur in humans—
despite the low rate of documented cases. Witte (1998) reemphasized the human clinical stand on the use of antibiotics in agriculture as a health risk to humans, citing specific examples of avoparcin-related, vancomycin-resistant enterococci disease transfer from animals to humans and the speculation about the relationship between satA-gene-mediated streptogramines-resistance development and the use of virginiamycin in food animals. The concern is that the unwarranted use of antibiotics “can lead to unexpected consequences that limit medical choices.”
A full discussion of the problem of worldwide multidrug resistance is beyond the scope of this report, but in an era of crisis, defining the contributing factors is of paramount importance in designing solutions. There is a great deal of disagreement over who or what is responsible for the spread of antibiotic resistance. Clearly, much evidence suggests that most of clinically important resistant pathogens in humans result from inappropriate uses of antibiotics in human medicine (IOM 1989 and 1998; Amabile-Cuevas 1993; Hickey and Nelson 1997). There are some data that support the idea that antibiotic resistance in agriculture can result from the use of antibiotics in subtherapeutic and therapeutic regimens in the food-animal industry (for example, Berghash et al. 1983; Kobland et al. 1987). The challenge is to determine the extent to which resistant microbes of animal origin affect human health. The challenge addresses the interconnectedness of the respective ecosystems and might not be resolved with current clinical data. If resistance to a drug develops but the microorganism is not a pathogen, is there a propensity for human disease? Similarly, although possible in laboratory settings, the passage of resistance plasmids from clinically benign to pathogenic bacteria might be clinically irrelevant. However, the answer to this concern is incomplete because of very limited data on passage frequency outside the laboratory.
The issue of antibiotic resistance in bacteria from animals is relevant to human health (Dupont and Steele 1987). A component of the concern could arise from the relationship of humans and the farm animal environment (Haapapuro et al. 1997). Levy (1992) voiced concerns regarding antibiotic use in farm animals and the consequences of resistance in humans from environmental exposure to animal manure:
For example, the amount of feces excreted by a cow per day is 100 times more than that of a human each day. If an animal is given an antibiotic, the fecal bacteria that survive the antibiotic treatment are resistant to it. Hence, via their excrement, animals are contributing a large amount of resistant bacteria to the natural environment, much [more] than are people. (P. 140)
Clearly, the use of antibiotics in food animals has been associated with the development of human antibiotic resistance. The development of resistant microbes with antibiotic use is regarded as a fundamental underlying assumption of antimicrobial chemotherapy. The increase in resistance with the assumptions of
antimicrobial chemotherapy and use in agriculture was cited in the report from a Rockefeller University workshop on antibiotic resistance as a threat to human health because of the increased propensity for this practice to set up conditions favorable to the selection of resistant bacteria (Tomasz 1994). In that report, however, the conclusion regarding agricultural use of antibiotics as a threat to human health was derived from a single previous review of the issue (Dupont and Steele 1987). The report failed to critically assess data that would take the conclusion to the next logical step—a substantive review of the actual development of disease (incidence, severity) directly related to antibiotic resistance in bacteria of food animals, and not to the mere potential for this to occur.
Threlfall (1992) reviewed the issue of drug resistance and antibiotic use with regard to selection of food-borne pathogens. He concluded that the prophylactic and therapeutic use of such antibiotics contributed substantially to the emergence of multidrug-resistant strains. He cited many examples of the emergence of such organisms from poultry, dairy calves, and pigs that he believed resulted in human disease. Conversely, Shah et al. (1993) reviewed the major pathogens involved in antibiotic-resistant human infections and their resistance patterns, compared them with the organisms and resistance patterns isolated from animals, and concluded that the veterinary pool has not contributed substantially to the overall profile of clinically significant antibiotic-resistant infection in humans. Wiedmann (1993) summarized the monitoring and origin of resistant organisms in humans and suggested that development of resistance could not be generalized but had to be discussed on the basis of specific drugs, bacterial species, or locations. Although he stated that the use of antibiotics in food-animal production had minimal consequences for the treatment of human infections in hospitals, those conclusions must be viewed from the perspective that the effects were minimal because there were alternative antibiotics that could be used to treat the infections.
All of these studies reached valid conclusions based on the interpretation of their data; however, none fully accounted for the issues of interconnectivity between species, genera of bacteria, or human and animal ecosystems. There are studies that critically examine the extent or mechanisms by which microbes pass from animal to human populations. Some microorganism transfers between animals and humans are clinically significant and result in invasive infections. There is no doubt that the passage of antibiotic-resistant bacteria from animals to humans occurs and that it can result from direct contact with animals or their manure (as might occur with workers on the farm [Holmberg et al. 1984b; Bates et al. 1994; Haapapuro et al. 1997]), through indirect exposure to food contaminated with animal-derived bacteria (Witte and Klare 1995), or from person-to-person contact after a primary exposure of nonfarm persons (Lyons et al. 1980). The passage of microorganisms from animals to humans probably also occurs without clinically overt disease in humans or animals, or more frequently, with self-limiting disease that is untreated. Clinically relevant diseases also can be
misdiagnosed with respect to the source or nature of the infection. Chalker and Blaser (1988) suggested that, for each case of salmonellosis that is confirmed by cultural methods, there are as many as 100 undocumented cases (see also, ERS 1996b). Perhaps more insidious to unraveling the causes and effects of the relationship between animal drug use, resistance emergence, and the potential for human disease are the inherent problems of the tests of antibiotic sensitivity and the interpretation of results (Murray 1994).
The resistance of microorganisms arising from subtherapeutic use of penicillin, tetracyclines, and sulfa drugs in agriculture is suggested by WHO (WHO 1997) to be a high- priority issue. WHO would phase out the use of antibiotics—particularly penicillin, tetracyclines, and others used to treat human diseases—as subtherapeutic-concentration growth promoters in food animals. Arguments persist that even if low-level resistance to antibiotics exists in bacteria from treated food animals, illness resulting from infection by organisms resistant to these drugs could easily be controlled by newer medications available for humans or animals strictly by prescription (AHI 1998). Levy (1998) suggested that even low-level drug resistance is a factor that predisposes bacteria to develop resistance more easily to other antibiotics. For some people, alternative antibiotic therapy might not be viable because of physiological or even economic limitations, and for these individuals some level of assurance and accommodation might need to be in place. Until more accurate data on animal antibiotic use, patterns and rates of resistance transfer to humans, occurrence of actual disease emergence, and mechanisms of resistance are available, actions aimed at regulating antibiotics cannot be implemented through a science-driven, well-validated, justified process.
The consequences of inappropriate use and accountability of antibiotics in human and veterinary medicine and in agriculture are (1) a shortened lifespan of an antibiotic’s usefulness, (2) additional complications in surveillance, (3) the ability to predict resistance patterns, and (4) the consequences for human health. Certainly, over-the-counter availability of antibiotics for domestic animals and the absence of professional oversight in many uses contribute to the frustration encountered by regulatory officials for the lack of accountability (Scott 1987) and limit the ability to make a true estimate of the magnitude of resistance problems that threaten human and animal health. Records of sales do not necessarily imply proper use, and there is no centralized repository of records of antibiotic use by animal species. Newer generation antibiotics are available only by prescription and this facilitates control over these drugs. In contrast, ethical issues of illegal and black market drug use in agriculture as well as in human medicine could pose an undocumentable risk.
HUMAN HEALTH RISKS FROM DRUG RESIDUES IN FOODS
The toxicity of drugs is an inherent part of all uses of medication, and there
are differences from one animal or human to another, especially in allergic reactions. Residues of drugs or their metabolites in food products from treated food animals are major considerations in the safety of drugs approved for use in food animals. FDA approval of drug dosages, routes of administration, durations of treatment, withdrawal times, and residue tolerances is designed to ensure the safety of foods derived from treated animals.
In the United States today, residues of carcinogenic chemicals or their genotoxic metabolites are rare in meat and meat products. FDA regulations have effectively prevented allergenic, toxic, and carcinogenic animal drug residues from entering the food supply. A review of the medical literature from 1966 to 1994 (National Library of Medicine 1994) yielded no evidence in short- or long-term studies of human cancers traceable to carcinogenic animal drug residues in foods. Chronic toxicity related to drug residues might be manifested by mutagenic, teratogenic, or carcinogenic potential. FDA operates under the 1958 congressional mandate that “no proven carcinogen should be considered suitable for use as a food additive in any amount.” Many other countries and international organizations apply the same stipulation to prevent carcinogenic residues in foods (FAO/WHO 1961, 1988). Although FDA approves new animal drugs and permits the continuance of approvals of animal drugs that have potential carcinogenic properties in food animals, it does so under strict guidelines: (1) The compound must be used only at authorized concentrations. (2) The compound must have no demonstrated carcinogenicity in the target animal species. (3) No carcinogenic residues can be detected in the edible animal tissues or products after a suitable drug withdrawal time (FDA 1992). Some drugs, such as diethylstilbestrol, nitroimidazole, internal-use nitrofurans, and quinoxaline di-N-oxides, have not been approved or have been removed from use in food animals because they have demonstrated a carcinogenic and mutagenic potential (nitrofurazone as a topical ointment is permitted).
Maximum residue concentrations for these drugs vary from 0 to 10 ppm. In 1993, FDA proposed a maximum safe concentration of 1 ppm in the total daily diet for noncarcinogens; 2 to 3 ppm would therefore be permitted in meat, assuming meat would constitute only one-third of the daily diet. FDA states this concentration has no adverse effects on intestinal ecology (Kidd 1994).
Some 30 antibiotic drugs are approved by FDA for oral administration in food animals. Several are antiprotozoal coccidiostats and anthelmintics for control of intestinal parasites. The rest are systemic or nonsystemic antibiotics. Systemic antibiotics are absorbed from the intestines in substantial amounts and include tetracycline, penicillin, erythromycin, and lincomycin. Nonsystemic antibiotics are not absorbed or are absorbed in trace amounts. This group includes bacitracins, neomycin, streptomycin, tylosin, oleandomycin, novobiocin, virginiamycin, and the bambermycins. When drugs are supplied to animals in feed or water, only those that are absorbed from the alimentary tract can induce residues in edible animal products.
In 1994, residue-monitoring tests for 9 antibiotics in food animals sampled at slaughter plants were positive at violative concentrations in 0.5 percent of 3,595 cattle; 0.3 percent of 960 sheep and goats; 0.2 percent of 1,298 swine; and 0.3 percent of 2,112 poultry. The most frequently detected antibiotics were tetracycline (27 percent of total), penicillin (27 percent), gentamicin (16 percent), and neomycin (20 percent) (FSIS 1994a).
Residues of drugs used in food animals can enter the human diet directly (as compounds of edible animal tissues and products) or indirectly (from the environment). The possible clinical implications of consuming residues of antibiotics are: toxicity, allergenicity, and infection by drug-resistant disease-causing microorganisms. Drug residues are considered unintentional food additives and thus come under regulatory scrutiny, as do other chemicals added to or entering the food supply. The Food Safety Inspection Service (FSIS) of the U.S. Department of Agriculture (USDA) conducts and coordinates an intensive program of residue screening, detection, and research, and publishes annual summaries of those data (Domestic Residue Data Book, USDA, Washington, D.C.).
Most antibiotic drugs administered in therapeutic and subtherapeutic form to domestic animals also are approved for human use. The drugs have been shown to be relatively safe as based on the therapeutic index of the drug and largely through the historic database that can be used to link adverse responses to residue concentrations. Patterns, distribution, and residue concentrations in food animal tissues vary according to how the drug is administered. Treatment through water or feed avoids the potential complications of high localized concentrations that might accumulate at the site of injection, where intramuscular or subcutaneous routes of administration could be needed or used. Injection sites can pose special concern in regard to residues. Care should be exercised to ensure that the smallest possible amount is left at injection sites. Strict adherence to withdrawal times and suggested withdrawal intervals is critical, and sometimes removal and discarding of the tissue at and surrounding the injection or treatment site is required.
Acute and chronic toxicities have been evaluated and are well documented. In most cases, the amount ingested by an individual who consumes the drugs as tissue residue will be considerably less than that consumed as a primary drug (Wilson 1994). The likelihood of direct toxicity from antibiotics or their metabolites in animal tissues is extremely low, as indicated by the lack of cases documented in the literature (Corry et al. 1983; Black 1984). There is exception in chloramphenicol, a drug that produces toxic aplastic anemia that is not related to dosage. Chloramphenicol has been implicated as the causative agent in several cases of fatal aplastic anemia (in one case, a 73-year-old woman died after receiving chloramphenicol) after its use as an ophthalmic drug at an estimated total dose of only 82 mg (Fraunfelder et al. 1982). In another study, chloramphenicol
residues were found in 13 calves of 3,020 tested (Settepani 1984), confirming that the residues can be consumed in human food. That finding led to a ban on the use of chloramphenicol in food animals in the United States.
The responsibility for monitoring food for violations of animal-drug-residue limits is shared by USDA (meat, poultry, and eggs) and FDA (milk and seafood). All standards are set and enforced by FDA. Details of the residue-monitoring program are discussed in Chapter 5.
The nitrofurans, quinoxalinedinoxides, and nitroimidazoles require restrictions as carcinogens, mutagens, or inducers of DNA synthesis, but the inherent hazards of their genotoxicity could be overcome by appropriate use and adherence to conservative withdrawal protocols (Somogyi 1984). For example, a conservative withdrawal period might be increased two- or three-fold from the last drug administration to ensure that any potential residues would have been eliminated. Such use and withdrawal regimens would preserve the value of these drugs in animal infection control. The toxicity of the sulfonamides in thyroid gland stimulation (Swarm et al. 1973) and phenotypically variable detoxification rates in the liver (Peters et al. 1990) require restrictions in food-animal use and continuation of residue monitoring.
Sulfonamides have been used widely at subtherapeutic and therapeutic concentrations in food-animal production, but increasing concern over their carcinogenic and mutagenic potential and their thyroid toxicity has led to decreased use, longer withdrawal times, and tighter residue monitoring. The sulfonamides approved for use in food animals are sulfamethazine, sulfadimethoxine, sulfaquinoxaline, sulfachlorpyridazine, sulfathiazole, sulfacetamide, and sulfanilamide (Compendium of Veterinary Products 1993).
A literature search of published records and clinical epidemiological testing indicates that allergic reactions in humans from ingesting antibiotic-contaminated foods of animal origin are rare. Most reactions resulted from ß-lactam antibiotic residues in milk or meat. The allergic reactions occurred in people exposed to the antibiotic drug residues in the foods. Many of the people went through prior medical treatment and were hypersensitized to a degree that subsequent oral exposure evoked a response (Dayan 1993). Dayan (1993) and Dewdney and Edwards (1984) presented several biochemical and biological reasons that antibiotic residues present in animal-derived foods are considered a relatively small health risk to humans: (1) The molecular weight of the free antibiotics is too low to make them immunogenic by themselves; (2) when complexed to larger molecular weight proteins that would make them immunogenic, the number of immunogenic epitopes per protein molecule is extremely low (less than 0.01 epitopes per protein molecule), which minimizes the ability of such residues to initiate a hypersensitivity reaction; (3) heating as would occur in food preparation
further degrades residue epitopes and reduces the potential for allergic response; and (4) sensitizing reactions are more directly related to intramuscular drug administration than to oral administration and the epitope distribution of protein-bound drug is so low as to be relatively insignificant as a potential cause for initiating and sensitizing responses when they are eaten. A summary of those rarely reported allergic reactions follows, with a commentary on conditions resulting in the adverse responses.
Four reports (two from the United States and two from England) of allergic reactions in persons previously sensitized to penicillin were identified between 1958 and 1969, when milk residues of penicillin were more prevalent. Vickers et al. (1958), Zimmerman (1958), Borrie and Barrett (1961), and Wicher et al. (1969) reported patients with dermatitis, urticaria, and subacute eczematous eruptions after drinking milk that contained residues of penicillin. Dewdney et al. (1991) cast doubt on (haptenized) penicillin residues as the causative factor in development of penicillin hypersensitivity. They argued that the immunogenicity, epitope density, and overall concentration were too low to contribute to allergy development. However, they did not point out that oral consumption of penicillin was less sensitizing than was parenteral administration. Questions still exist regarding the ability of parenteral administration to be the sensitizing stimulus and regarding the consumption of penicilloyl residues as a trigger for hypersensitivity reaction.
Other cases of allergic reactions reported between 1972 and 1980 were traced to consumption of penicillin-residue-containing meat. One reaction was to residues in pork, which originated from swine treated with penicillin 3 days before being butchered. Another reaction was to the beef in a frozen dinner, which subsequently was found to contain penicillin residues (Tscheuschner 1972; Schwartz and Sher 1984). Two patients experienced pruritus on the face and fingers, and one suffered an anaphylactic reaction. No deaths occurred.
Relative Risks: Residues versus Microbial Contamination
Microbial contamination of food is a major health problem worldwide. Great difficulty exists in ensuring that foods are free of microbial contamination, and there are many points in the chain of processing, storage, sale, and preparation that provide opportunities for microorganisms to proliferate in food. Initializing contamination events might be innocuous, but under conditions that permit these organisms to proliferate, the build-up of pathogenic bacteria and toxins will contribute significantly to food-borne illness (Altekruse et al. 1997). Surveillance and monitoring of contamination and disease outbreaks associated with microorganism-based food-borne illness is spread across several federal agencies, including FSIS, FDA, and the Centers for Disease Control and Prevention (CDC). There are now 10 organisms identified and tracked by the federal agencies under a collaborative interagency Pathogen Reduction Task Force that pro-
duces updated Sentinel Site Study reports. Of these 10 pathogens, FSIS has identified Campylobacter, Salmonella and Shigella as the 3 most frequently encountered pathogens causing reportable diarrheal disease in humans (FSIS 1997). Surveys of disease incidence data between 1980 and 1994 (Bryan 1980; Bean and Griffin 1990; CDC 1994) demonstrate that, of almost 5,000 food-borne illness outbreaks, fewer than 10 percent were traced and confirmed to have arisen from meat or meat products.
Protection of the public from animal products contaminated with animal-drug residues that could cause human toxic reactions could be considered much more effective than protection from products contaminated with microorganisms. This is because there is little chance of residues entering the food after the point of slaughter and because so much of the opportunity for bacteria to multiply in an animal-derived food occurs long past the time when federal inspectors can monitor contamination and take action. Inspection at food-processing facilities can detect and monitor residues with accuracy, and inspectors can respond to violations quickly. But after a product is beyond the live animal, the risk of microbial contamination and microbial load increase with time. The number of handling steps and the care retailers and consumers use in preserving the integrity of the product affect the potential for bacteria to increase. Human infections and intoxications by food-borne microorganisms originating from infected food animals are commonly from commensal organisms of carrier animals. Prevention and elimination of carrier states in food animals requires an armamentarium of drugs and vaccines, professional decisions on their administration, and measures to ensure the safety of their products for human consumption. Safety of foods from animals that have been given medical treatment requires that the therapy eliminates primary or secondary infectious agents that might remain in carrier and shedder states. Antibiotics are needed for specific application in eliminating carrier states in food animals subclinically infected with agents that are infectious to human consumers.
SUMMARY OF FINDINGS
There appears to be a hierarchy of concerns regarding animal-drug use and human health. Principles of animal microbiology, antibiotic use, and food processing and preparation all relate to human health. Antibiotic resistance is a global problem found in human and animal environments, and is fostered by overuse, inadequate oversight, and inappropriate use in all areas of human and animal medicine. Only a multilateral effort can contain resistance. Inappropriate use of antibiotics must be controlled in all environments. Although resistance will develop in any animal, including humans, in which antibiotics are administered, the resistance itself cannot automatically be linked to a disease state. Current evidence indicates that microbial contamination of food causes many more
cases of human illness than are caused by antibiotic-resistant organisms transmitted from animals to humans.
There is no doubt that the passage of antibiotic-resistant bacteria from food animals to humans occurs. It can result from direct contact with animals or manure, from indirect exposure to food contaminated with animal-derived bacteria, from person-to-person contact, and from the use of antibiotics in food animals. A demonstrable link can be found between the use of antibiotics in food animals, development of resistant microorganisms in those animals, and zoonotic spread of pathogens to humans. Although occurrence is historically rare, the data are woefully inadequate to show whether changes in disease rate are occurring. It is difficult to establish whether an increase in resistance detection is the result of increased antibiotic use in food animals or the result of the perpetuation of resistant species in food animals, the environment, or other reservoirs. Thus, a significant limitation is that the real number of incidents of zoonotic antibiotic-resistant passage to humans that resolve in clinical disease might not be well documented or even trackable.
Although therapeutic and subtherapeutic antibiotic treatment might be effective in decreasing a small percentage of the microbial load of food animals at harvest, the greatest proliferation of organisms occurs during inappropriate handling and processing after slaughter. A concern is that available data for critical review are scarce and that the information that is available is used opportunistically to support or refute claims by interested groups. In contrast to microbial contamination of food, drug residues appear to constitute a relatively lower risk as assessed by the available monitoring data.