Tobacco Documents Online

1 . Z c 0 emiology id Journal of Epidemi Formerly AMERICAN JOURNAL OF HYGIENE s 1978 by The Johns Hopkins University School of Hygiene and Public Health VOL. 108 OCTOBER, 1978 Re~-ieWs and Commentary CAUSATION AND DISEASE: A CHRONOLOGICAL JOURNEY THE THOMAS PARRAN LECTURE'~ Z ALFRED S. EVANS' NO. 4 The invitation to give the Thomas Par- Public Health at the University of Pitts- ran Lecture is both a pleasure and an honor burgh. On retirement from Pittsburgh he for me. I wish first to indicate something of the life and contributions of Thomas Par- ran. Dr. Parran was born in 1892, and re- ceived the Master of Arts degree at the University of Maryland and his medical degree from the Georgetown University, both in 1915. He joined the Public Health Service shortly after graduation, became Chief of the Division of Venereal Disease, and then was appointed Health Commis- sioner of the State of New York by Franklin Delano Roosevelt when he became Gover- nor of that state. When Roosevelt was elected to the Presidency he appointed Dr. Parran to Surgeon General of the Public Health Service. In addition to these impor- tant posts in the practice of public health Dr. Parran was a founder of the World Health Organization and both the founder and first Dean of the Graduate School of ' Delivered at the Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA, De- cember 15, 1977. 2 Adapted, by permission, from "Causation and dis- ease: The Henle-Koch postulates revisited," by A. S. Evans. Yale Journal of Biology and Medicine, Vol- ume 49, pages 175-195, 1976. 3 Professor of Epidemiology and Director of WHO Serum Reference Bank, Department of Epidemiology and Public Health, Yale University School of Medi- cine, New Haven, CT 06510. became President of the Avalon Founda- tion. Some of Dr. Parran's contributions to public health included his courageous at- tack on venereal diseases, his promotion of clinical research at the National Institutes of Health, his emphasis on the right of individuals to good health and medical care, his promotion of improvements in diet, par- ticularly in good milk supplies, and his par- ticipation in international health move- ments, such as the WHO and the Pan American Health Organization. Dr. Par- ran's book on syphilis, entitled Shadow on the Land: Syphilis (published by Reynal and Hitchcock, New York, 1937), in which he presented his program for the control of venereal disease in the United States, was a pioneering publication that dared to bring this problem to the open attention of the public. It should still be of interest to those concerned with the control of infectious diseases in this country. HENLE-KOCH POSTULATES In my presentation on the chronology of causation and disease, I wish to emphasize three points: first, that the Koch postulates should really be termed the "Henle-Koch postulates"; secondly, that those postulates

i 250 ALFRED S. EVANS were not regarded as rigid criteria by Koch himself at the time they were presented and should not be regarded as such, today; and thirdly, that all of our concepts of cau- sation are limited by the technology avail- able to prove them and our understanding of the pathogenesis and epidemiology of the disease at the time of the investigation. A fuller exposition of this chronology has previously been published (1). Let us first turn to the Henle-Koch postulates. Jakob Henle was a German-born physician who went to Zurich at age 31 as Professor of Anatomy where in addition to his brilliant contributions to our understanding of the histology of the retina and the kidney and many other organs he wrote in 1840 an essay, "On Miasmata and Contagie" (2), in which the concept was advanced that "Be- fore microscopic forms can be regarded as the cause of contagion in man they must be constantly found in contagious material. They must be isolated from it and their strength tested." This essay was written some 42 years before the first bacillus, that of tuberculosis, was discovered by Henle's pupil, Robert Koch, in 1882. Henle left Zii- rich, went to Germany and finally became Professor at_the University of Gottingen where Koch studied under him and adopted the concepts of his teacher. Koch's contri- bution to causation was made originally in 1884 and more formally presented in a lec- ture given in 1890 at the International Con- gress in Berlin (3). These postulates have become our classical point of ieference in relating causative agents to disease. The major features were that the parasite must be present in every case of the disease under appropriate circumstances, that it should occur in no other disease as a fortuitous and non-pathogenic parasite, and finally that it must be isolated from the body in pure culture, repeatedly passed, and induce the disease anew. These criteria of causa- tion have been slavishly held up as the only true basis for establishing causation. As recently as the October 29, 1977 issue of Lancet (4) some putative bacterial causes for Crohn's disease were discarded because they did not fulfill these postulates. Yet it is clear that failure to fulfill them fully was not regarded by Koch himself at the time of presentation as being grounds to exclude the possibility of a causal relationship and they certainly should not be so regarded today. Koch recognized that the cause of anthrax, tuberculosis, erysipelas, tetanus and many animal diseases fully fulfilled his criteria.. At the same time he recognized that there were many causes of illness which did not fulfill the criteria, such as typhoid fever, diphtheria, leprosy, relapsing fever and Asiatic cholera. Cholera, in par- ticular, should be noted because Koch iso- lated the organism of this disease in Egypt and was well aware of its capacity to pro- duce epidemic disease in man. So, even at the time of presentation, Koch emphasized that not all the criteria were necessary for proof and that just the first two were suffi- cient. The second criteria of the postulates became a problem in fulfillment when the carrier state was recognized for diphtheria by Park and Beebe (5) and by the work of Chapin (6) and by Koch himself for typhoid fever (7) in the period 1895-1902. CAUSATION IN VIRUS DISEASES The virus era opened in the 1900s, and by 1930 many viruses had been identified in mice and tissue culture. In 1937, Thomas Rivers, in his Presidential Address to the American Society of Immunology (8), em- phasized the following limitations of the Koch postulates for virus diseases: 1) more than one agent might be needed to produce a given disease; 2) viruses could not be grown on lifeless media but required living cells in contrast to bacteria; and 3) asymp- tomatic carriers existed. Rivers proposed a set of criteria quite similar in essence to those of the Henle-Koch postulates but in- cluded two additional concepts: 1) the re- production of the disease experimentally should exclude the possibility that the lab- oratory animal was infected with some lab- oratory virus and include appropriate con- trols, and 2) that a: should appear durin ness. HUEBNER: GUIDELIP A VIRUS AS CA These concepts of of virus diseases be many viruses were s simultaneously pres, even in healthy peoF Huebner who is now tory of Tumor Virolo of Health, discusse lemma (9). He indict of many viruses-ev, made the identificat a virus of low order tion and one could c tive associations if y sions on this fact alor some infections wer= ruses, and that sorrr produce chronic dise or that viral reactivr ner's contributions in a bill of rights for pr emphasized that the virus should not be r basis for etiology and miologic principles cross-sectional studiE proof in causation. I-I, the disease should bE cific vaccine to the s concept is of great im ing the etiologic relati infectious agent to ti also emphasized that order to accomplish proof and included tl-, for etiologic studies. BROADENING OF '1 CAUSATION FOR These concepts of c. eases were then broad years by the recognitic ical picture or syndror

_iecause . Yet it i11y was ie time exclude hip and =garded :ause of tetanus illed his ,ognized illness such as 2lapsing , in par- och iso- n Egypt to pro- even at ahasized ;sary for :re suffi- )stulates ~hen the phtheria work of typhoid 'ES -00s, and dentified Thomas ss to the (8), em- .s of the : 1) more , produce :i not be °ed living t ) asymp- °oposed a ssence to es but in- i) the re- imentally .t the lab- some lab- -riate con- CAUSATION AND DISEASE 251 trols, and 2) that antibody to the disease by a variety of different agents, that differ- should appear during the course of the ill- ent agents predominated under different ness. HUEBNER: GUIDELINES FOR ESTABLISHING A VIRUS AS CAUSE OF DISEASE These concepts of Rivers on the etiology of virus disease.s became untenable when manv viruses were sometimes found to be simultaneously present in ill persons and even in healthy people. In 1957 Dr. Robert Huebner who is now Chief of the Labora- tory of Tumor Virology, National Institutes of Health, discussed the virologist's di- lemma (9). He indicated that the presence of many viruses-even in normal people- made the identification of the presence of a virus of low order in establishing causa- tion and one could derive spurious causa- tive associations if you based the conclu- sions on this fact alone. He recognized that some infections were due to multiple vi- ruses, and that sometimes viruses could produce chronic diseases, or carrier states, or that viral reactivation occurred. Hueb- ner's contributions included what he called a bill of rights for prevalent viruses. They emphasized that the mere presence of the virus should not be regarded solely as the basis for etiology and he introduced epide- miologic principles by longitudinal and cross-sectional studies as an element of proof in causation. He also emphasized that the disease should be prevented by a spe- cific vaccine to the suspected agent. This concept is of great importance in establish- ing the etiologic relationship of a suspected infectious agent to that disease. Huebner also emphasized that one needed money in order to accomplish the establishment of proof and included this as a ninth criteria for etiologic studies. BROADENING OF THE CONCEPTS OF CAUSATION FOR VIRUS DISEASES These concepts of causation for virus dis- eases were then broadened over the next 10 years by the recognition that the same clin- ical picture or syndrome could be produced epidemiologic circumstances, and that the host response to a given virus would vary from one setting to another. These ideas were expressed by Evans in 1960 (10) and 1967 (11) as they related to acute infectious diseases. It also became clear that the agent alone was a necessary but not sufficient _ factor needed to cause most diseases. Co- factors and the susceptibility of the host were of key importance in the occurrence of clinical illness. IMMUNOLOGIC PROOF OF CAUSATION The next problem arose with the isola- tion of a virus now known as the Epstein- Barr virus (12) and with the identification= of an antigen known as the Australia anti- gen gen by Blumberg and his associates (13). These two agents were established as__ causally related to infectious mononucleo- sis and to viral hepatitis B, respectively, mainly by serologic studies. The agents were a particular challenge in terms of the Henle-Koch postulates because they could not be grown in pure culture nor did their injection reproduce the disease in experi- mental animals. This resulted in the estab- lishment of an immunologic proof of cau- sation. The criteria required that antibody to the agent is regularly absent prior to the disease and that it regularly appears during illness. The absence of the antibody should indicate susceptibility to infection and its _ presence indicate immunity. Antibody to no other agent should be similarly associ- ated with the disease unless it is a co-factor. _ These criteria of immunologic proof were the basis for the ass ociation of Epstein-Barr virus and infectious mononucleosis (14). _ The discovery of this relationship was made in 1968 by Dr. Werner Henle, his wife, Dr. Gertrude Henle, and a young associate, Dr. Volker Diehl, in Philadelphia (15). It is a particular irony that Dr. Henle, who is the grandson of Jakob Henle, established the causative relationship of EBV to infectious mononucleosis without fulfilling a single

252 ALFRED S. EVANS one of the postulates set up by his grand- produce difficulties in interpretation in the father and by Robert Koch. This serves to future. There are also some "conventional emphasize that we must change our criteria viruses" associated with chronic neurologic with our technology. diseases, namely measles virus in relation to subacute sclerosing panencephalitis SLOW VIRAL INFECTIONS AND CHRONIC (SSPE) (18) and papova virus in relation to NEUROLOGIC DISEASE progressive multifocal leukoencephalopa- The next challenge to the Henle-Koch thy (PML) (19). In both of these conditions postulates arose with the recognition of reactivation of a virus in which the primary agents called slow viruses. These caused exposure was years before is responsible for kuru, a disease occurring in the Fore tribe the chronic disease observed. In SSPE the in New Guinea, and Jakob-Creutzfeldt dis- main epidemiologic event appears to be ease, a pre-senile dementia. The work on infection with measles virus very early in these agents carried out by Dr. Gajdusek life at a time of inadequate host response and his associates at the National Institutes (20). With PML, alteration in our immune of Health resulted in the award of the Nobel system by immunosuppression either nat - prize to Dr. Gajdusek in 1977 (16). The urally induced, as in Hodgkin's disease, or particular difficulties posed by these agents artifically induced by immunosuppressive in relation to causation were their long in- drugs is responsible for the reactivation of cubation period, their association with this agent (21). Thus the criteria for cau- chronic neurologic diseases and, most im- sation in these diseases depends on the portantly from the standpoint of the Henle- identification of the agent in brain tissue, Koch postulates, the fact that the agents the presence of high antibody titers to the could not be isolated in tissue culture in the virus and, at least for SSPE, the experimen- laboratory. Furthermore, these unique tal reproduction of the disease in a labora- agents did not produce an immune response tory animal (22). and were highly resistant to a great variety EVIDENCE RELATING VIRUSES TO CANCER of physical and chemical agents. A new set __ of criteria for causation was therefore The possibility that viruses might cause needed for this group. cancer required new criteria of causation. Dr. Richard Johnson of The Johns Hop- The problems in establishing this relation- kins University and Dr. Clarence Gibbs, an ship include the long incubation period be- associate of Dr. Gajdusek's at the NIH, tween exposure to the putative agent and proposed guidelines (17) for establishing the cancer and the probability that disease causal relationship which were based on 1) results from reactivation of the virus rather the consistency in animal transmission of than from a primary infection. The agents the agent, 2) the fact that the agent should most commonly recognized as the best on- be serially transmissible in experimental cogenic candidates are the herpes viruses. animals with filtered material and 3) that But proof of causation is difficult because similar results should not be obtained from they are common and ubiquitous viruses, normal tissues. In light of our knowledge of probably require co-factors, and there are the presence of many viruses in the intes- difficulties in reproduction of the cancer in tines and throat of healthy persons, it would animals. In addition, human volunteer not be surprising to me if normal brain studies are not possible. There is also the tissue may not sometimes contain latent probability that the cancer may have dif- agents. It is well recognized that herpes ferent causes in different geographic areas zoster and simplex and papova viruses can or under different epidemiologic setting! remain latent in many tissues, including (23). The most likely oncogenic candidat( nerve tissue. Thus, their third criteria may is Epstein-Barr virus in relationship to Af rican Burkitt lym ryngeal cancer. Tl- consist immunolog higher antibody t healthy, matched : stration of high le~ the cancer; and virc of the viral genom ability of the virus thirdly, (at least t experimental evidE induce a malignant Similar virologic ar for EBV and nasopl this cancer has not perimental animals phasize that co-fact appearance of bot, Burkitt lymphoma appears to be mal< EBV and by the n first year of life ma~ malignancy to deve curring early in life relation of SSPE a ence of high antibc development of Bur cently been demonst spective study of 41" Africa (24). It is impc not all tumors fulfil teria of Burkitt lym with EBV. America for example, is not u high EBV titers, the sent from the tumo cases EBV antibody sent from the blood, infection had never c pharyngeal cancer ge be an important risk tigen 2 and a new ant 2 have both been asso a fivefold higher incic tibility to nasophary Chinese lacking 'thest Another virus of the ated with possible r type 2 in cervical canc

in the itional °ologic 4ation halitis tion to alopa- Jitions rimary ble for ?E the to be arly in sponse nmune er nat- ase, or ressive tion of )r cau- on the tissue, , to the :rimen- labora- ANCER t cause isation. aation- °iod be- mt and disease ; rather agents lest on- Aruses. )ecause viruses, ere are .ncer in lunteer 1so the .ve dif- .c areas ;ettings ididate ) toAf- .,-_.~~ CAUBATION AND DISEASE 253 rican Burkitt lymphoma and to nasopha- ryngeal cancer. The elements of causation consist immunologically of the presence of higher antibody titers in cases than in healthy, matched controls and the demon- stration of high levels of antibody prior to the cancer; and virologically of the presence of the viral genome in tumor cells and the ability of the virus to transform cells; and thirdly, (at least for Burkitt lymphoma), experimental evidence that the virus can induce a malignant lymphoma in animals. Similar virologic and serologic proof exists for EB V and nasopharyngeal cancer, except this cancer has not been reproduced in ex- perimental animals. It is important to em- phasize that co-factors are involved in the appearance of both of these cancers. In Burkitt lymphoma the essential co-factor appears to be malaria. Infection both by EBV and by the malarial parasite in the first year of life may be requirements for a malignancy to develop. This infection oc- curring early in life is reminiscent of the relation of SSPE and measles. The pres- ence of high antibody titers prior to the development of Burkitt lymphoma has re- cently been demonstrated in a massive pro- spective study of 45,000 children in West Africa (24). It is important to recognize that not all tumors fulfilling the histologic cri- teria of Burkitt lymphoma are associated with EBV. American Burkitt lymphoma, for example, is not usually associated with high EBV titers, the genome has been ab- sent from the tumor tissue, and in 'some cases EBV antibody has been entirely ab- sent from the blood, indicating that EBV infection had never occurred (25). In naso- pharyngeal cancer genetic factors seem to be an important risk factor since HLA an- tigen 2 and a new antigen called Singapore 2 have both been associated in Chinese with a fivefold higher incidence of their suscep- tibility to nasopharyngeal cancer than in Chinese lacking 'these characteristics (26). Another virus of the herpes group associ- ated with possible malignancy is herpes type 2 in cervical cancer. Here the virologic and serologic evidence is much less con- vincing than in African Burkitt lymphoma (27). It is clear that at least 40 per cent of cancer of the cervix may occur in the ab- sence of antibody to HSV 2. My own view of the relationship of viruses to cancer is that the host plays a critical role in the development of the malignancy. Such fac- tors as the age at the time of infection, the immunologic state of the host, the presence of co-factors such as concurrent infection, and the genetic attributes as expressed by human leukocyte antigens, all play a role in the pathogenesis of the malignant process. CAUSATION IN CHRONIC DISEASE Let us now turn to the issue of causation in chronic disease. Interest in this began around 1957 and was based on the Henle- Koch postulates. Important in this early work was the contribution of the late Dr. Yerushalmy and the late Dr. Palmer (28) who proposed a series of relationships that include the concepts that 1) the suspected characteristic must be found more fre- quently in persons with the disease than in persons without, 2) that persons possessing the characteristic should develop the dis- ease more frequently than those not pos-' sessing it, and 3) that this association must be tested for its validity by investigating the relationship of the characteristic to other diseases in order to establish the spec- ificity of the characteristic and the disease. These guideposts for implication of an eti- ologic factor in a chronic disease were am- plified by Dr. Abraham Lilienfeld, who pro- posed five additions: 1) the incidence of the disease should increase in relation to the duration and intensity of the suspected fac- tor, 2) its distribution should parallel that of the disease in all relevant aspects, 3) a spectrum of illness should be related to the exposure to this suspected factor, 4) reduc- tion or removal of the factor should reduce or eliminate the disease and 5) human pop- ulations exposed to the factor in controlled or even in natural experiments should de-

t 254 ALFRED S. EVANS velop the disease more commonly than those not exposed (29). The contributions of the committee ap- pointed by the Surgeon General to deline- ate the relationship between smoking and health should also be recognized (30). They included five features of association be- tween a putative cause and the disease; namely, consistency, strength, specificity, temporal relationship, and the coherence of the association. CRITERIA FOR CAUSATION: A UNIFIED CONCEPT A review of these various concepts of causation in infectious diseases and in chronic diseases has led me to try and for- mulate a unified concept (1). It combines the elements of many of the guidelines pre- viously mentioned. The main features are 1) the prevalence of the disease should be higher in those exposed than in those not exposed, 2) that exposure to the putative cause should be present more commonly in those with the disease than in those without the disease, 3) that the incidence should be higher in persons who are so exposed than in those not exposed as shown in prospec- tive studies, 4) that exposure to the sus- pected factor should precede the disease, 5) that there should be a measurable biologic spectrum of host responses, 6) that experi- mental reproduction of the disease should be demonstrated, 7) that elimination of the putative cause should decrease the inci- dence of the disease, and 8) that prevention or modification of the host response should decrease or eliminate the expression of the disease. These concepts are not original with me but put together a concept of cau- sation applicable to both infectious and non-infectious diseases. Just as the Henle- Koch postulates cannot be regarded with any finality, so too, these concepts should be taken only as guidelines, subject to our changing knowledge of technology and cau- sation. In summary, the original Henle-Koch postulates had many limitations, some rec- ognized then, and others later (table 1). Fulfillment of the postulates is certainly reasonable grounds for accepting a causal role of the putative agent but lack of fulfill- ment of the postulate should not exclude such a relationship. This chronologic view has focused primarily on the relation of a single cause to a disease. This is a simplistic view because most infectious agents are a necessary but not sufficient cause of dis- ease; indeed many viral infections are in- apparent. Causation in both infectious and non-infectious disease involves a complex interplay of agents, environmental, and host factors. The latter include the host's immunologic status, genetic background, socioeconomic level, hygienic practices, be- havioral patterns, age at the time of expo- sure and the presence of co-existing disease. Different qualitative and quantitative mixes of the agent, environment, and host may result in the same clinical and patho- logical diseases under different circum- stances. These more sophisticated ap- proaches to causality are discussed in re- cent books (31-33) and articles (34-39). Another view of the causal relationship of an agent to disease might be framed in legal terms. Table 2 presents some of these comparisons (35). In criminal law, the pres- ence of the criminal at the scene of the TABLE I Limitations of Koch's' postulates of causation 1) Not applicable to all pathogenic bacteria. 2) May not be applicable to viruses, fungi, parasites. 3) They do not include the following concepts: a) The asymptomatic carrier state. b) The biologic spectrum of disease. c) Epidemiologic elements of causation. d) Immunologic elements of causation. e) Prevention of disease by elimination of putative cause as element of causation. f) Multiple causation. g) One syndrome has different causes at different settings. h) Reactivation of latent agents as cause of disease. i) Immunologic processes as cause of disease. ` More properly termed the Henle-Koch postulates: Mayhem or murder and c 1) Criminal present at the scer 2) Premeditation 3) Accessories involved in the 4) Severity or death related to 5) Motivation-the crime mu terms of gain to the crimina 6) No other suspect could hal crime in the circumstances , 7) The proof of the guilt mu beyond a reasonable doubt. I crime would be equivaler + of the agent in a lesio: Premeditation would be quirement that the causa precede the onset of the ~ ence of accessories at the might be compared to tl- ~ factors and/or multiple diseases. The severity ot consequence of death equivalent to susceptibi responses which determi the illness. The motivat crime should make sense to the criminal, just _R_q r` agent should make biui=,: sence of other suspects , tion in a criminal trial w that of the exclusion c causes in human illness. the proof of guilt must yond a reasonable doubt both criminal justice an; sation. What will the future hc infectious diseases in wh: essary to look for cau First, we have those ne might be produced by ch or its point of entry. Rt will continue to emerge v to antibiotics and other ~ or will appear as a result, ~ of genetic factors of resist t13 w ,

ne rec- ble 1). rtainly causal ' fulfill- :xclude ic view )n of a -lplistic s are a of dis- are in- ;us and )mplex il, and : host's round, :es, be- f expo- lisease. titative id host patho- ;ircum- =d ap- 11 in re- 39). ionship med in )f these ie pres- of the >ation a-rasites. ts: putative different f disease. ' ase. '~stulates. CAUSATION AND DISEASE TABLE 2 Rules of evidence: criminality and causality Nfavhem or murder and criminal law _ 1) Criminal present at the scene of the crime 2) Premeditation 3) Accessories involved in the crime. 4) Severity or death related to state of victim. 5) Motivation-the crime must make sense in terms of gain to the criminal. 6) No other suspect could have committed the crime in the circumstances given. 7) The proof of the guilt must be established beyond a reasonable doubt. _ crime would be equivalent to the presence of the agent in a lesion of the disease. Premeditation would be similar to the re- quirement that the causal exposure should precede the onset of the disease. The pres- ence of accessories at the scene of a crime might be compared to the presence of co- factors and/or multiple causes for human diseases. The severity of the crime or the consequence of death might be loosely equivalent to susceptibility and the host responses which determine the severity of the illness. The motivation involved in a crime should make sense in terms of reward to the criminal, just as the role of a causal agent should make biologic sense. The ab- sence of other suspects and their elimina- tion in a criminal trial would be similar to that of the exclusion of other putative causes in human illness. Finally, need that the proof of guilt must be established be= yond a reasonable doubt would be true for both criminal justice and for disease cati= sation. What will the future hold in terms of new infectious diseases in which it may be nec- essary to look for causal relationships? First, we have those new diseases which might be produced by changes in the agent or its point of entry. Resistant organisms will continue to emerge which are resistant to antibiotics and other therapeutic drugs, or will appear as a result of plasmid transfer of genetic factors of resistance. Recently we Morbidity, mortality and causality Agent present in lesion of the disease. Causal events precede onset of disease. Co-factors and/or multiple causality involved. Susceptibility and host response determine sever- ity._ The role of the agent in the disease must make biologic and common sense. No other agent could have caused the disease under the circumstances given. The proof of causation must be established beyond reasonable doubt or role of chance. _ have recognized the plasmid transfer of re- sistance of typhoid organisms to chloram- phenical, of the gonococcus to penicillin and most recently of penicillin resistant forms of the pneumococcus. All these indi- cate that there is an epidemic potential of these highly resistant organisms. It is also frightening to think that this resistance might be transferred in vivo to other orga- nisms. For example, is it possible that a resistant gonococcal infection of the mouth might lead to the transfer of that plasmid to a streptococcus or to a meningococcus present in the same environment? Then we have new agents in old portals. The recog- nition of new agents of infectious diseases -probably transmitted by the respiratory routes, such as Lassa fever and Marburg fever, are examples of this. We also have old agents in new portals, such as the trans- fer of herpes type 2 virus from its normal habitat beneath the diaphragm to a new habitat in the oropharyngeal cavity as a result of changing sexual practices. New agents will be identified for old diseases like the common cold of which we now only recognize about half of the causative agents or for encephalitis in which only 25-50 per cent have a recognized cause or for diar- rheal diseases. There are important new advances in the discovery of rotavirus in infant diarrhea, of toxogenic Escherichia coli in the traveler, and of Giardia lamblia in the camper, but the majority of common

a a=a 256 ALFRED S. EVANS tivate and cancer increase as our normal immunologic controls diminish in efficacy. The control of umbrella diseases like ma- laria and schistosomiasis in tropical settings may permit recognition of host responses which were previously hidden by the co- _ existence of these infections. Changing host susceptibility may occur as a consequence of new environmental stresses, of immuno- logic manipulations, or possibly even of ge- netic engineering. Changing social and cul- _ tural habits may result in new patterns of illness. So we may be faced in the future with a variety of new and unrecognized infectious diseases whose etiology must be established. Proof of this relationship must be based on common sense, good guidelines of causation appropriate to existing tech- gastroenteritides remain unexplained. Fi- nally, there are latent infections that are being reactivated which were formerly un- recognized as human pathogens. For ex- ample, reactivation of latent papova virus causes progressive multifoccal leukoence- phalopathy but we have not recognized any disease syndrome associated with the com- mon primary infection with this agent. There are also environmental factors which may result in new diseases. The ex- panded tourist travel, research exploration, and military excursions to the remote cor- ners of the globe may bring us in contact with new pathogens not previously identi- fied. Our exploration of the ocean bottoms and of outer space may expose us to new microbial forms. Our own creation of spe- cial environments, such as intensive units or kidney dialysis units may new and unusual host responses. care nology and a keen sense of the biologic produce basis of disease. Changing economic factors are more likely to result in the re-emergence of old diseases than the creation of new ones. Funds which are now provided for the sur- veillance and immunization of important diseases, like smallpox, cholera, polio and measles may not be appropriated as these diseases are brought under control and lose their emotional and political clout. In- creased economic development and better hygiene may delay exposure to common agents like polio, infectious mononucleosis, and hepatitis from early childhood to young adult life. This may result in the occurrence of more clinical illnesses due to these agents because the host response is more severe in older life. There are also host factors which may play a role in the emergence of new dis- eases. These include those which are asso- ciated with our increasing use of new ther- apies, such as immunosuppressive drugs, new antibiotics, organ transplants and tis- sues from man and perhaps other primates, ande th_e use of synthetic organs and valves. With a large proportion of our population now living into the sixth, seventh and eighth decade, latent infections may reac- CAUSATION OF HEALTH This presentation has reviewed our changing concepts of the causation of dis- ease. Perhaps it is time we also look at the causation of health (table 3). Let us direct our attention not only to those factors that produce disease but also to those that pro- duce health. Let us change our "don'ts" for "do's" in medical practice and public TABLE 3 Postulates for the causation of health 1) The preventive factor must be consistently present in persons of good health or free of a particular --- disease. 2) The factor must be isolatable in a pure form, (i.e., can be identified as causal). 3) The extent to which the factor is effectively applied must parallel an increase in good health and/or freedom from that disease. 4) Experimental application of the factor to one seg- ment of a population should significantly increase their good health as compared with matched con- _ trols. 5) Withdrawal of the preventive factor should be as- sociated with an increase of disease associated with that factor. 6) The effect of the factor shall be measured in terms t+~ of lower morbidity and mortality, longer life, and p lower medical costs. O C.3 - -R r.t ~ L3 health. Let us approac sation of health with a tistical, and biologic or Let us do studies of vive and not only of wh us do case/control stu sons with ill persons ae prospective studies of I those who remain in than only those who recognize that there wi. for health just as there for disease and that he biologic spectrum. Let establishing the proof health will be more diffi than the causation of begin. REFEREi 1. Evans AS: Causation a Koch postulates revisit 49:175-195, 1976 2. Henle J: On Miasmata aw from German by G Rosei Baltimore, 1938 3. Koch R: Ueber bakteri( Verh. X. International 2A 1890. p 35, 1892 4. Ward M: The pathogen Lancet 2:903-990, 1977 5. Park WH, Beebe AL: diphtheria. Med Rec 46:2 6. Chapin CV: The Sources New York, Wiley, 1910 7. Koch R: Die Bekampfui gehalten in der Sitzung Senats bei der Kaiser Wi November, 1902 8. Rivers TM: Viruses and I teriol 33:1-12, 1937 9. Huebner RJ: The virolo~ Acad Sci 67:430-445, 195, 10. Evans AS: Common clin tious disease. I. Introduc ratory diseases. II. Comrr III. Common infections system. IV. Common diar J 59:508-512, 553-556, 65` 11. Evans AS: Clinical syndrc respiratory infection. Mec 1967 12. Epstein MA, Achong BG, in cultured.lymphoblasts Lancet 1:702-703, 1964 13. Blumberg BS, Sutnick . Australia antigen and he 282:349-354,1970

< normal !fficacy. ike ma- settings sponses the co- ing host ~quence nmuno- ! n of ge- I and cul- terns of future -ognized riust be = ,ip must .idelines ig tech- biologic ed our z of dis- k at the is direct ;ors that hat pro- n'ts" for public Ith .ly present particular form, (i.e., rly applied th and/or 0 one seg- y increase ched con- uld be as- iated with i in terms r life, and CAUSATION AND DISEASE 257 health. Let us approach the proof of cau- 14. Evans AS: New discoveries in infectious mononu- sation of health with a keen scientific, sta- cleosis. Mod Med 42:18-24, 1974 15. Henle G, Henle W, Diehl V: Relation of Burkitt's_ tistical, and biologic orientation. _ _ tumor-associated herpes-type virus to infectious Let us do studies of what makes us sur- mononucleosis. Proc Natl Acad Sci USA vive and not only of what makes us die. Let 59:94-101, 1968 -_ 16. Gajdusek DC: Kuru and Creutzfeldt-Jakob dis- us do case/control studies of healthy per- ease. Ann Clin Res 5:254-261, 1973 sons -with ill persons as the control; let our prospective studies of populations focus on those who remain in good health rather than only those who become ill. Let us recognize that there will be multiple causes for health just as there are multiple causes for disease and that health like disease is a biologic spectrum. Let us recognize that establishing the proof of the causation of health will be more difficult and challenging than the causation of disease. Butt let us begin. REFERENCES Johnson RT, Gibbs CJ Jr: Editorial. Koch's pos- tulates and slow infections of the nervous system. Arch Neurol 30:36-38, 1974 Sever JL, Zeman W (Editors): Conference on mea- sles virus and subacute sclerosing panencephalitis. Bethesda, Maryland, September 19, 1967. Neurol- ogy 19:30, 1968 Zu Rhein GM: Association of papova-virions with a human demyelinating disease (progressive mul- tifocal leukoencephalopathy). Prog Med Virol 11:185-247, 1969 Detels R, Brody JA, McNem J, et al: Further epidemiologic studies of subacute sclerosing pan- encephalitis. Lancet 2:11-14, 1973 Narayan 0, Penney JB, Johnson RT, et al: Etiol- ogy of progressive multifocal leukoencephalopa- 17. 18. 19. 20. 21. thy. N Engl J Med 289:1278-1282, 1970 22. Byington DP, Johnson KP: Experimental sub- acute sclerosing panencephalitis in the hamster. 1. Evans AS: Causation and disease: The Henle- Correlation of age with chronic inclusion-cell en- Koch postulates revisited. Yale J Biol Med cephalitis. J Infect Dis 126:18-26, 1972 49:175-195, 1976 23. Evans AS (Editorial): The spectrum of infections 2. Henle J: On Miasmata and Contagie. (Translated with Epstein-Barr virus: A hypothesis. J Infect from German by G Rosen). Johns Hopkins Press, Dis 124:330-337, 1971 Baltimore, 1938 24. de The G, Geser A: A prospective seroepidemio- 3. Koch R: Ueber bakteriologische Forschung. In logical study to investigate the role of EBV in Burkitt's lymphoma. In Unifying Concepts of Leu- kemia. Bull Haematol No 39 (edited by RM Durcher, L Chieco-Bianchi), Karger, Basel, 1973, p448 25. Pagano JS, Huang CH, Levine P: Absence of Epstein-Barr viral DNA in American Burkitt's 6. Chapin CV: The Sources and Modes of Infection. lymphoma. N Engl J Med 289:1395-1399, 1973 New York, Wiley, 1910 -26. Simons MJ, Wee GB, Day NE, et al: Immunoge- Verh. X. International Medical Congress;: Berlin_ 1890. p 35, 1892 4. Ward M: The pathogenesis of Crohn's disease. Lancet 2:903-990, 1977 5. Park WH, Beebe AL: Diphtheria and pseudo- diphtheria. Med Rec 46:385-401, 1894 7. Koch R: Die Bekampfung des Typhus. Vortrag gehalten in der Sitzung des wissenschaftlichen Senats bei der Kaiser Wilhelms Akademie am 28 November, 1902 8. Rivers TM: Viruses and Koch's postulates. J Bac- teriol 33:1-12, 1937 9. Huebner RJ: The virologist's dilemma. Ann NY Acad Sci 67:430-445, 1957 10. Evans AS: Common clinical syndromes of infec- tious disease. I. Introduction and common respi- ratory diseases. II. Common infectious exanthem. III. Common infections of the central nervous system. IV. Common diarrheal diseases. Wis Med J 59:508-512, 553-556, 656-660, 700-702, 1960 11. Evans AS: Clinical syndromes in adults caused by respiratory infection. Med Clin N Am 51:803-815, 1967 12. Epstein MA, Achong BG, Barr YM: Virus particles in cultured lymphoblasts from Burkitt lymphoma. Lancet 1:702-703, 1964 13. Blumberg BS, Sutnick Al, London WT, et al: Australia antigen and hepatitis. N Engl J Med 282:349-354, 1970 netic aspects of nasopharyngeal carcinoma. IV. Probable identification of HLA second locus anti- gen associated with high risk of nasopharyngeal carcinoma. Lancet 1:142-143, 1975 27. Melnick JL, Adam E, Rawls WE: The causative role of herpesvirus type 2 in cervical cancer. Can- cer 34:1375-1385, 1974 28. Yerushalmy J, Palmer CE: On the methodology of investigations of etiologic factors in chronic dis- eases. J Chronic Dis 10:27-40, 1959 29. Lilienfeld AM: On the methodology of investiga- tions of etiologic factors in chronic disease-Some comments. J Chronic Dis 10:41-46, 1959 30. Surgeon General, Advisory Committee of the USPHS: Smoking and Health. PHS Publication - No 1103. Washington DC, US GPO, 1964 31. Lilienfeld A: Fundamentals of Epidemiology. New York, Oxford University Press, 1976 32. MacMahon B, Pugh TF, Ipsen J: Epidemiologic Methods. Boston/Toronto, Little, Brown, 1960 33. Susser MW: Causal Thinking in the Health Sci- ences; Concepts and Strategies of Epidemiology. New York, Oxford University Press, 1973, p 181

I 258 ALFRED S. EVANS 34. Bollet AJ: On seeking the cause of disease. Clin Res 12:305, 1964 35. Rothman KJ: Causes. Am J Epidemiol 38. 104:587-592, 1976 36. Koopman JS: Causal models and sources of inter- action. Am J Epidemiol 106:439-444, 1977 39. 37. Cassel J: The contribution of the social environ- ment to host resistance. Am J Epidemiol 104:107-123, 1976 Miettinen OS: Proportion of disease caused or prevented by a given exposure, trait, or interven- tion. Am J Epidemiol 99:325-332, 1974 Stewart GT: Limitations of the germ theory. Lan- cet 1:1077-1088, 1968 AMERICAN JOURNAL OF EPIDEMIOLOGY Copyright © 1978 by The Johns Hopkins All rights reserved C VENTILATORY LU PAUL BROWN,' I Brown, P., D. Sadov and Communicative Di tory lung function stu 108:259-265, 1978. In the first study of v observations in 182 nc ages of 15 and 80 yea FEV,, and FEV, %. Mee those reported for Ca Negroes and Chinese. three Oceanic races st of FEV, % with increa& significance of this var tain, the definition o, pulmonary airway dise age factors; forced lung volume measurerr In the course of long-te growth, development, and d in several Western Caroline tions, we have been imp amount of chronic respii among the Micronesian p_ observation prompted a rec the islands for the specific : umenting the extent and v Received for publication Decem final form April 4, 1978. Abbreviations: BTPS, water-sa perature and pressure; FEV,, forc.- ratory volume; FEV,%, ratio of F forced vital capacity. ' Central Nervous System Studd tional Institute of Neurological a. Disorders and Stroke, National Ii Bethesda, MD 20014. 2 Office of Biometry and Epid Institute of Neurological and Cor ders and Stroke, National Institt: thesda, MD 20014. The authors thank Jesus Malr rui, and Jesus Tamel for assistan the government of the Trust Ten Islands for permission to conduc cronesian people.