Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 05 and by the nature of the ava ilable relevant data which often had been col lected for other purposes . The risk assessment was con f ined to infections caused by Salmonella species , only a port ion of the problem , because bas ic surve i l l ance data on infections due to other gastrointestinal pathogens of animal origin--such as Campyl9bacter i ejuni , Yersinia enterocolitica , and enterohemorrhagic ï¿½ QQ!i--were not ava i l ab le . The committee hopes that FDA will f ind the risk assessment performed by this committee to be useful in its dec i s i on-making . The committee acknowledges the qual itative and quantitative deficiencies of the primary data and the broad range of estimates used in the assessments of risk . Narrowing the range of these estimates would necess itate ref ining and enlarging the data base used in the risk analys i s . In the committee ' s view , such an e ffort appears reasonable . AHTIMICBOBIAL-BESISTANT SALMONELLA STBAINS AHD THEIR SOQRCES Improved surve i l l ance of salmone l l a isolates in the United States is essential for better understanding and control both of bacterial res istance to antimicrobial agents and o f disease due to sa lmonel lae in humans and animals . Improved surve i l l ance would be easy to put into place , because it would build on , and actua l ly require only a sma l l increment to , a large existing system . Tbe Existing System Each year , hundreds of thousands o f phys icians , analyz ing i l lnesses of mi l l ions of patients , send hundreds o f thousands o f stool specimens t o more than 5 , o o o microbiology laboratories . This is a time-consuming process by which ski l l ed technologists isol ate 4 0 , 0 0 0 or more salmone l l a stra ins and conduct tests for res istance t o antibacterials . Most of the salmonel l a isolates are then forwarded to state reference laboratories , where technologists with spec ial tra ining , ski l l s , and reagents laboriously type them into more than 1 , 0 0 0 poss ible serotypes . Serotype reports are returned to the referring laboratories , where they only rarely contribute to the management of pat ients . Epidemiologi sts in the separate states use the reports of isolated salmonella serotypes to del ineate recogn i z ed outbreaks and detect others . The accumulated reports of a l l the state re ference laboratories are col l ected , tabul ated , and publ i shed by the Centers for Disease Control ( CDC ) . In a s imi lar system , state veterinary laboratories , the National Veterinary Re ference Laboratory , the Food Safety and
2 06 Inspect ion Service ( FS I S ) of the u . s . Department o f Agriculture , and various univers ity diagnost ic laboratories isolate and serotype salmonel lae from specimens taken from animals ; however , these data are unpubl ished and have not been included in the data base used by this committee . Additions to the System The ant imicrobial susceptib i l ities o f each salmonella i sol ate need to be recorded with its serotype . Reference laboratories need to enter their results each day into a networked computer system that analyzes a l l data automatica l ly and comprehensively . Res istant isol ates class i fied by a computer as epidemiological ly important should be forwarded to a laboratory that wi l l catal og the ir pl asmids . Addition of Susceptibility Test Besults One good reason for adding results of susceptibi l ity tests is to obtain accurate and compl ete measurements of prevalence o f res istance , regional variations , trends over time , etc . Al l those would have been valuable to thi s committee , but could only b e pieced together crudely from fragmentary reports of l imited comparabi l ity ( see Chapter V ) . A second , and probably better , reason i s to improve epidemiologic information , which is the only j usti f ication for the present elaborate system . Essential ly , outbreaks are confirmed or recogn ized now by virtue of an excess of isolates of one serotype over the expected inc idence . The outbreak cl one boosts the serotype isolation rate above the threshold of random appearances . However , for the more commonly isolated serotypes (which account for most of the isolates ) , the threshold for detecting excess is high . Genera l ly , only large outbreaks are invest igated , except for outbreaks caused by rare serotypes ; in fact , most outbreaks are overl ooked . Coupl ing antibiotype to serotype permits detect ion of sma l l outbreaks that are due to res istant subclones that belong to common serotypes . Two isolates of a subclone can be recognized to constitute an outbreak : without the ant ibiotype , do zens might be needed . Recognition o f more sma l l outbreaks or of large outbreaks sooner improves the understand ing and control of sa lmonella d isease and of the flow of res istance genes in animals and humans . Recognition of more outbreaks also provides more opportunities to trace cha ins of transmission . A th ird reason for recording ant ib iotypes is that it is the key to the use of plasmid cataloging , as described below , which adds another level of subcl one d iscrimination .
207 Determining the res istance of a l l salmonel l a i solates in the United States would be easy and inexpens ive . Four- f i fths of the human isolates exhibit no antimicrobial res istance . The ones that do nearly always have res istance at l east to e ither tetracycl ine , streptomycin , sul fonamide , or amp i c i l l in : dropping four disks on a sma l l plate or part of a plate would screen out the susceptible ( nonresistant ) fourÂ f i fths , leaving 8 , 0 0 0 res istant isolates--an average of less than one per day per state reference l aboratory . Each would need a routine disk-susceptibil ity test plate , which entai l s several minutes o f work and a dol lar ' s worth of suppl ies . Some 4 , 0 0 0 - 5 , 0 0 0 animal i solates of salmonellae would need susceptibi l ity testing each year . An Integrated Computer System S almone l l a outbreaks often appear as a smal l number of isolates scattered across several states . Accordingly , cont inu ing analys is of a l l u . s . data in an integrated system is needed to detect these outbreaks early or at a l l . such a system would become more important as antibiotyping and plasmid cataloging began to discriminate more salmonel l a subclones . It would a l so b e helpful i n devel oping or adapting shared software , such as automatic not i f i cation whenever any clone or subclone in any state or combination o f states exceeded i t s outbreak threshold . Integration should be cons idered soon , before individual reference laboratories acquire various incompatible systems . Plasm id cataloging Plasmid catal oging would be a useful diagnostic tool and should begin with a survey of the restriction endonuclease digestion patterns of res istance plasmids from anima l and human sa lmone l l a isolates representing preval ent antibiotypes for a number of serotypes ( see Table V-2 ) . Experience has shown each set of plasmids from isolates of one antibiotypeÂ serotype combination would be expected to have one or more restriction patterns . Whol ly di fferent restriction patterns in any set would discriminate subclones presumed to be unrel ated . Patterns with sma l l d i fferences undoubtedly represent evolutionary variants of one clone and may be distingui shed with microepidemiologic studies . For exampl e , plasmids from six human isolates from different parts of one state over a two month period had identical restriction patterns , whereas those from humans or anima l s in other states a l l d i f fered s l ightly from those s ix and from one another .
2 08 As the catalog of u . s . salmone l l a plasmids might build , in parallel with a computeri zed isolate data base , the epidemiology of salmonellae in animals and humans could begin to emerge at a new level of detai l and quantitat ion . The distribut ion of speci fic clones and subcl ones among animal populat ions could be del ineated , and rates o f appearance in humans could be established . That would provide better understanding of the epidemiology of salmonel lae in animal and human outbreaks , give a basi s for each of them at the outset to known plasmid famil ies with known prior d istributions , and hence provide early clues to pos s ible chains o f spread . One laboratory could catalog the salmone l l a pl asmids and explore technology for improved cataloging of plasmids ; restriction endonuclease profil ing , although workable now , i s l ikely t o b e supplemented ( i f not replaced ) b y newer methods that would be faster and provide more critical molecular deta i l . In particular , as more i s learned about the stages of spread of res istance through bacterial populations and the molecular changes that accompany those stages , it wi l l be of great value to f ind correlates in the data on salmonel lae . Impl ications o f Change in Use of Animal Feed A4ditiyes It can be quest ioned whether a different program o f salmonella surve i l lance would b e needed i f use o f animal feed additives changed . The program recommended above would probably be a sens itive monitor of such change , as wel l as providing improved observation of the present s ituation . HUMAN MQRBIDITY AND MORTALITY DUE TO ENTERIC PATHOGENS OF FARM ANIMAL ORIGIN " Salmonella deaths" might be construed as deaths due primarily to salmonellosis (with bacteremia and shock , endarteritis , metastatic abscesses , and severe gastroenteritis with dehydration , usua l ly in infants or elderly ) ; those in patients with underlying diseases in whom sa lmonel l a gastroenteritis contributed to the death , but was not the primary cause of it ; and those in patients from whom sa lmonel l ae were isolated , but in whose deaths had other causes . In cases with other causes , hospital i z ation might have been initiated by salmonella infection that had subsided and the ir deaths have had unrelated causes . S imilarly , salmonella gastroenteriti s and asymptomatic carriage can occur as almost incidenta l matters in patients with other maj or medical problems to which they succumb . Often , the above distinctions have not been made in pub l i shed series of cases and reports of outbreaks , and
2 09 information on death cert i ficates is not of sufficient qual ity to a l l ow such distinctions to be drawn . Such data can probably be developed only with expanded studies of selected count ies o f the type performed by coc . 7 However , it wil l l ikely require more detai l ed analys is o f hospital records and information ( based on chart-di rected recal l ) from attending physic ians to categori ze " salmonella deaths " more def initively . In addit ion , definit ive class i fication o f those deaths as t o contribution of salmone l l a infection ( primary cause , contributing cause , of unknown relevance , or unrelated ) should be strati f ied according to ant imicrobial . susceptib i l ity pattern o f the strain involved ( susceptible , res istant to penici l l in or tetracycl ine , res istant to multiple ant imicrobial s , etc . ) . In the characteri z ation o f stra ins , primary attention should b e o n res istances to antimicrobials known to be R-plasmid-mediated . The selected-counties study of the Salmonella Surve i l l ance System is the only current system of survei l lance o f infection with Salmonella of which the committee i s aware . Additional useful information could be obta ined through the system i f the fol lowing mod i fications were introduced : o Expans ion o f the number o f communities to yield a larger data base . o Continuation of the selected-counties study in the form o f annua l surve i l l ance , rather than monitoring every 4 -5 years . o Determination o f morb idity ( days of diarrhea , hospita l i zation rates , etc . ) associated with infection caused by antimicrob ial-susceptible and antimicrob ial-res istant stra ins . o I nclus ion of infections with Campylobacter i ej uni based on the same kinds o f epidemiologic quest ionnaires (given the increas ingly evident impact o f campyl obacter infection ) . OQANTITIES OF AHTIMICRQBIAL QRQGS USED IN SUBTHEBAPEQTIC CONCENTRATION IN ANIMAL FEEDS Amounts o f antimicrobials in the aggregate ( and o f individual antibiotics , such as penici l l in and the tetracycl ines ) used in animal feed are not known . The best ava ilable informat ion comes from the Animal Health I nstitute and from industrial sources as estimates that are admittedly rough . Valuable data could be provided by monitoring and surve i l l ance to determine actual subtherapeutic use o f
2 10 antimicrobials in animal and poultry feeds , perhaps through sampl ing of farms and feedlots . Sampl ing should reflect characteristics of a cross sect ion of users of antimicrobial feed additives . 0 Cbaracteristics of facilities to be monitored . Sampl ing should include large and sma l l farms , appropriate geographic areas , the maj or animal sources of meat products ( beef cattle , veal cattle , pigs , and poultry ) , and ( in the case of catt l e ) widely d i fferent methods of animal rearing ( confined , h igh-dens ity herds , and open-range gra z ing ) . It should include the spectrum from sma l l local fac i l it ies to huge industrial operations and existing data bases , such as APHIS -USDA . o Bationales for Subtherapeutic use of antimicrobials giyen by user at each Administration . Was it for growth promot ion , disease prevention , or some other purpose? Did the farmer know , in fact , whether feeds contained an antimicrobial additive and , if so , what it was? For what fract ion of the total growth cycle is the regular use o f penici l l in and tetracycl ines ( and other antimicrobial s ) in subtherapeutic concentrations in feed a regular practice? o Concentrations o f penici l l in and tetracyclines Cand other antimicrobials > achieved in feed on farms and in feed !QtÂ§ . Are the prescribed subtherapeutic concentrations being ach ieved a fter mixing on farm and feedlot? Are they being exceeded unwittingly? o Would routine record-keeping of antimicrobial use on farms and feedl ot s be feasible? such records would be of value , after antimicrobial -resistant strains involved in outbreaks of human salmonellosis were traced to the farm source ( by plasmid fingerprinting techniques ) , in determining relationship to prior use of subtherapeutic concentrations of antimicrobials in feed given to the farm animals involved . ROLE OF PRIOR EXPQSQRE TO AHTIMICRQBIALS IN INFECTION BY AHTIMICRQBIAL-BESISTAHT StRAINS OF SALKONEIJAE C "ETIOLQGIC fRACTION" > Studies of the role of the " etiologic fraction" in humans should be expanded to provide a larger statistical bas is for the estimates of risk to human health . Further data might be gl eaned both from outbreaks , as has already been done , and from the Salmonella Surve i l lance System for selected count ies . It would also be of interest to devel op data on the morbidity and mortal ity rate in patients who const itute the etiologic fraction . Do these pat ients exhib it
211 more o r less severe i l lness than other patients overa l l ? The select ion of matched control s must be carried out with qreat care , because the patients in the etioloqic fraction are l ikely to d i f fer from the qeneral population at risk of salmonel los i s by virtue of the ir recent intake of antimicrobial druqs , their qreater aqe , their underlyinq i l l nesses , and by their rece ipt of a smal ler inoculum . As far as we are aware , the term " etioloqic fraction" has been appl ied only to infections in humans , but there miqht also be an etioloqic fraction in anima l s - - i . e . , decreased colonization and infection of anima l s by susceptible strains , an effect that miqht be termed a "neqative etioloqic fraction . " Indeed , evidence supports the latter hypothesis . 3 A neqative etioloqic fraction for infections caused by susceptible stra ins wi l l increase the proport ion of infections caused by res i stant stra ins . A more crucial question concerns the net effect o f the subtherapeutic administration of antibiotics on the overal l prevalence of carriaqe o f salmonel lae ( res istant or susceptib l e ) in animals . Studies to evaluate that question could be carried out by pass ive observation o f naturally occurrinq infection durinq the administration of antibiotics or by del iberate feedinq of salmonella stra ins under controlled conditions . The results would probably be h iqhly influenced by the choice of ant ib iotics for administration , as wel l as by the dosaqe and frequency of administration , the deqree of res istance ( or susceptibil ity ) of the infectinq stra ins , and the deqree of res istance o f the normal flora ( possibly includinq both aerobes and anaerobes ) . Sma l l d i fferences in any those variables would influence the subtle interaction between the suppress ive effect of a druq on a bacterial pathoqen and on the normal bacterial flora and the interact ion between these bacteria . A neqative etioloqic fraction miqht also be hypothes i z ed in humans . I ndeed , some have suqqested that the inappropriate use o f ant ib iotics for such i l lnesses as viral pharynqitis , a lthouqh attended by potential adverse effects , miqht have a beneficial effect in preventinq secondary infection or compl ications of infection caused by susceptible pathoqens . Th is committee ' s charqe does not deal with the administration of antibiotics to humans , so that subj ect has not been pursued in thi s report . MORBIDITY AND MORTALITY AS SOCIATED WITH UNREPORTED CA5ES OF SALMONELLQS IS There is rel iabl e evidence that only about 1 - 1 0 % o f cases o f salmonellosis are identi fied in outbreaks : a comparably smal l fraction is thouqht to be reported in sporadic cases . Perhaps more important , it is often commonly
2 12 assumed .that the unrecogni zed or unreported cases are general ly less severe than the recogni zed or reported ones . The committee , in estimating risk , had no data on salmonel los i s morbidity and of course could use data on salmonellosis deaths only i f such deaths were reported . It might be possible for CDC to conduct telephone surveys during outbreaks of salmone l l os i s to detect not only the unreported proportion of cases , but also to obtain data on morb idity in those cases : the results for both mortal ity and morbidity could be compared with data on reported cases . I f morbidity were s imilar in the reported and unreported cases-- i . e . , chance , rather than severity of i l lness , would be determining whether an infection is reported--th i s would suggest that the impact o f salmonellosis in our risk assessment , which is based on 4 0 , 0 0 0 - 6 5 , 0 0 0 cases per year , should be scaled sharply upward to take account of unreported cases . STQDIES TO DETERMINE EXISTENCE OF DIRECT EYIDENCE O F HUMAN HEALTH HAZARD ASSOCIATE D WITH SUBTUEBAPEUTIC USE OF PENICILLIN AND TETRACYCLINES IN ANIMAL FEEDS Only CDC is in a pos ition to perform such studies . The use of molecular techniques to characteri z e salmone l l a isolates clonally and t o identi fy antimicrobial -resistant stra ins through the food chain from farm animal to human consumer has proved success ful . Do the one or two examples reported constitute clear evidence of a common phenomenon , or are they exceptions? The weak point in the argument is the lack of a direct demonstration of the role of subtherapeutic use of approved antibiot ics , such as penici l l in or tetracycl ines , as the farm origin of the infect ing salmonella clone . More salmonella outbreaks should be studied with molecular fingerprinting techniques and with conventional epidemiologic methods , including documentation o f the qual itat ive and quantitative aspects of antimicrobial use on the implicated farms or feedl ots . OTHER STUDIES AHTIMICRQBIAL BESISTAHCE AMQNG . AND FBEOUENCY OF DISEASE DUE TO . OTHER FQODBOBNE ANIMAL PATHOGENS THAT AFFECT HQMANS Recognition over the last decade that other bacterial enteric pathogens , such as Campylobacter j eiuni and ï¿½ Â£Qli 0 157 : H7 , can spread from animals to humans and that Campylobacter might infect humans more often than any other intestinal pathogen , ra ises the question of whether
2 13 monitoring the frequency with which they cause human and animal infections might i l luminate the issues faced by thi s committee . I nformat ion on those pathogens seems sparse , compared with information on salmonel lae . Serotyping does not yet subdivide them as elaborately as it does salmonel lae and i s i n any case not yet routinely pract iced by a national network of l aboratories . Res istance does not seem to impede therapy often or to be as varied as in salmonel lae . Further work on these pathogens needs to be supported , but survei l l ance data on them are not l ikely to influence antimicrobial use greatly in the near future . SELECTIVE EFFECTS OF THEBAPEQTI C AND SU8THEBAPEQTIC DQSAGES O F ANTIMICROBIAL AGENT Given the paucity of in vitro data of the kinds required to predict the effects of different dosages of drugs on sel ect ion o f ant ib iotic-resistant strains , much more work is recommended . In particular , in a mixed bacterial populat ion , what are the effects of di fferent dosages of antimicrobials on the growth characteristics of res istant and suscept ible bacteria , and on the genetic spread of R plasmids , particularly between species? Such studies require care ful determination of growth rates and conj ugative trans fer rates under wel l -control led cond itions . The resulting information should make it poss ible to develop appropriate computer model s to a l l ow prediction of dosage effects on the emergence of drug-resistant organisms . EFFECT OF ANtiBIOTIC BES I STAHCE ON VIRULENCE OF FOODBQBNE PATHOGENS OR SEYERITY OF DISEASE PRODUCED Recently , some plasmids in salmonel lae have been ident i f ied as be ing involved in the express ion o f virulence . 1 , 2 , 8 Plasmids in ï¿½ typbimu rium ( 9 0 kilobases ) , in ï¿½ dubl in ( 7 5 kb ) , and in ï¿½ enteritidis ( 54 kb ) are necessary for these three serotypes to express virulence in mice . Plasmid- free " cured" strains lose virulence in mice , and re introduction of the pl asmids into the " cured" strain , or into a naturally occurring plasmid- free isol ate , restores virulence . Studies defining the preval ence o f virulence plasmids in the common Salmone l l a serotypes ( antimicrobial -suscept ible stra ins ) isolated from farm animals ( and humans ) not exposed to antib iotics wi l l be helpful in providing basel ine informat ion concerning possible e ffects of subtherapeutic administration of antimicrobials in feed . Virulence plasmids might not be present in most stra ins o f ant imicrobial-
2 14 susceptibl e salmonel l ae in farm animals ; in that case the possible effect of antimicrobial selection or pers istence and expans ion of the population of virulence plasmids in farm animals receiving subtherapeutic concentrat ions of antibiotics warrants investigation . S imilarly , direct comparison of human ( and animal ) isolates of ant ibiot icÂ susceptible and -resistant salmonella isolates o f speci f ic serotypes in mouse-virulence assays bears directly on the question of whether antibiotic res istance in such a foodborne pathogen can a f fect its virulence . BACTERIOLQGIC CONTAMINATION OF ANIMAL FOODSTQFFS WITH ENTERIC PATHOGENS FS I S has the respons ibi l ity for ensuring the safety o f meat and poultry products . FS IS has a n overal l goal of ensuring that meat , poul try , and their products are whol esome , unadulterated , and properly labeled and do not constitute a health ha zard to the consumer . In 1 9 8 3 , FS IS asked NRC to evaluate the scienti fic bas is o f the current system for inspecting meat poultry , and meat and poultry products . The 1 9 8 5 report 4 that resulted from that request of fered numerous recommendations , not all of which are directly rel evant to the present study . However , one of the maj or conclus ions of the report states that Salmonella and Campyl obacter spec ies are maj or causes of di seases transmissible to humans through the consumption of meat and poultry products and that current postmortem inspection methods are not adequate to detect these organisms . The report recommended that efforts to control and el iminate contamination with microorganisms include evaluation of rapid diagnostic procedures for detect ing Salmonella and Campyl obacter especially . Postmortem inspect ion methods have been relatively e ffective for the detection of unwholesome meats ; before these methods are abandoned , FS IS should determine the effectiveness of the methods that would replace them . At the request of FS IS , another study was done S to evaluate the current FS IS poultry inspection programs in the framework of a ri sk-assessment model incorporat ing stat istical procedures . The report of that study drew several conclus ions and offered recommendations that are rel evant to the current study . "There is conclus ive evidence that microorganisms pathogenic to humans ( such as Sa lmonel la and Campylobacter ) are present on poultry at the time of slaughter and at reta i l . " The report stated that " the critical control po ints at which known pathogenic microorganisms such as Salmone l l a and Campyl obacter may be introduced into the poultry system should be identi fi ed and monitored , pre ferably as a part of an HACCP ( Ha zard Analys is
2 15 Crit ical Control Point ) proqram" and that FS I S begin to lay the groundwork for a more comprehens ive program with statist ical ly based sampl ing that modi fies the traditiona l bird-by-b ird inspection . In 1 9 6 9 , the NRC Committee on Sa lmonel l a 6 recognized that there was no way to be absolutely certain that a particular l ot of nonsterile food is free of salmonellae . It recommended the development o f a sampl ing plan to provide adequate assurance that the number of salmonel l ae present , i f any , i s bel ow a statistica l ly defined l imit that reflects minimal haz ard to the consumer . A subcommittee of that committee supported the HACCP concept as an e ffective and rational approach to the assurance of safety . FS IS is now evaluating the effectiveness o f the HACCP procedure for analyz ing the sl aughter of poultry and the procedures for handl ing finished products . FS IS plans to determine where the critical control points are and what procedures would be best for controll ing bacterial contamination from the time when the animal reaches the process ing plant to the time when the finished product goes to market . Today there is no routine microbiologic sampl ing of meat or poultry . To help in evaluating the critical control points in process ing of poultry , an agency p i l ot plant in Puerto Rico is gathering basel ine data at each critical control point in process ing . Investigators wil l look at microbiologic contamination be fore anima l s reach the process ing plant and during process ing . The present committee recommends that , with respect to microbiologic sampl ing of meats , cons ideration be given not only to the identi f ication of pathogenic organi sms , but also to the ir testing for antimicrobial susceptib i l ities . The resulting microbiologic data should be avai l able to researchers for use in studying the relationship of antibiotic use to drug res istance in pathogens isolated from foodborne outbreaks of human disease . BEFEBENCES 1. Beninger , P . R . , G . Chikami , K . Tanabe , c . Roudier , J . Fierer , and D . G . Guiney . Physical and genet ic mapping of the Salmonella dublin virulence plasmid p . S DL2 . Relationship to plasmids from other Salmone l l a stra ins . J . Cl in . Invest . 8 1 : 1 3 4 1- 1 3 4 7 , 1 9 8 8 . 2. Heffernan E . J . , J . Fierer , G . Ch ikami , and D . G . Guiney . Natural history of oral Salmonel l a dubl in infection in BALB/c Mice : E ffect of an 8 0 -kilobase-pa ir plasmid on virulence . J . Infect . Dis . 1 5 5 : 1 2 5 4 - 1 2 5 9 , 1987 .
2 16 3. Kiser , J . s . Transmission o f food-borne diseases . Implications o f subtherapeutic use of antimicrobial s , pp . 2 03 at seq . (Appendix G ) . In National Research Council , Committee to Study the Human Health E f fects o f Subtherapeutic Antibiotic Use in Animal Feeds . Washington , D . C . : National Academy Press , 1 9 8 0 . 4. National Research Council . Committee on the Scient i fic Basis of the Nation ' s Meat and Poultry Inspection Program . Meat and Poultry Inspection : The Scienti fic Bas i s of the Nat ion ' s Program . Washington , D . C . : National Academy Press , 19 8 5 . 5. National Research Council . Committee on Publ ic Health Risk Assessment of Poultry Inspection Programs . Poultry Inspect ion : The Basis for a Risk-Assessment Approach . Washington , D . C . : National Academy Press , 1 9 8 7 . 6. National Research Council . Committee on Salmonella . An Evaluat ion of the Salmonella Problem . Washington , D . C . : Nat ional Academy of Sciences , 1 9 6 9 . 7. Riley , L . w . , M . L . Cohen , J . E . Seal s , M . J . Blaser , K . A . Birkness , N . T . Hargrett , s . M . Martin , and R . A . Feldman . Importance o f host factors in human salmonellosis caused by multiresistant stra ins o f Salmonella . J . Infect . Dis . 149 : 8 7 8 - 8 8 3 , 19 8 4 . 8. Wi l l iamson , c . M . , G . D . Ba ird , and E . J . Manning . A common virulence region on pl asmids from eleven serotypes of Salmonella . J . Gen . Microbiol . 1 3 4 : 9 7 5- 982 , 1988 .
216 3. Kiser, J. S. Transmission of food-borne diseases. Implications of subtherapeutic use of antimicrobials, pp. 203 et seg. (Appendix G). In National Research Council, Committee to Study the Human Health Effects of Subtherapeutic Antibiotic Use in Animal Feeds. Washington, D.C.: National Academy Press, 1980. 4. National Research Council. Committee on the Scientific Basis of the Nation's Meat and Poultry Inspection Program. Meat and Poultry Inspection: The Scientific Basis of the Nation's Program. Washington, D.C.: National Academy Press, 1985. 5. National Research Council. Committee on Public Health Risk Assessment of Poultry Inspection Programs. Poultry Inspection: The Basis for a Risk-Assessment Approach. Washington, D.C.: National Academy Press, 1987. 6. National Research Council. Committee on Salmonella. An Evaluation of the Salmonella Problem. Washington, D.C.: National Academy of Sciences, 1969. 7. Riley, L. W., M. L. Cohen, J. E. Seals, M. J. Blaser, K. A. Birkness, N. T. Hargrett, S. M. Martin, and R. A. Feldman. Importance of host factors in human salmonellosis caused by multiresistant strains of Salmonella. J. Infect. Dis. 149:878-883,1984. 8. Williamson, C. M., G. D. Baird, and E. J. Manning. A common virulence region on plasmids from eleven serotypes of Salmonella. J. Gen. Microbiol. 134:975- 982, 1988.