Epidemiology in Military and Veteran Populations: Proceedings of the Second Biennial Conference, March 7, 1990 (1991)

Leonard B. Seeff ^*

A well-recognized sequel to acute hepatitis B virus (HBV) infection is the development in some infected individuals of the viral carrier state, associated with progression to chronic hepatitis (1, 2). This sequence is best described in the Far East and Sub-Saharan Africa where HBV infection is endemic. In these geographic areas, most instances of transmission of this virus take place at birth through an infected carrier mother. The majority of newborns to such mothers are themselves infected, virtually all become and remain carriers for life, and many of them develop progressive liver disease, advancing from chronic hepatitis to cirrhosis and, ultimately, to primary hepatocellular carcinoma (HCC) (2-4). indeed, HCC is the most common malignant tumor in these areas of the world.

In contrast, in areas of the world nonendemic for HBV, such as the United States, HBV infection occurs predominantly in late adolescence or early adulthood, the carrier state is low, and HCC is an uncommon form of cancer (1). Limited data from such areas have suggested that advancement from acute to chronic hepatitis occurs in 5-10% of acutely infected persons (5, 6). Although it is presumed that such carriers also are at risk of developing HCC, it has been difficult to prove this relationship in low endemicity areas because of the low infection rate and hence of a

sufficiently large population among whom to conduct natural history studies.

An opportunity to undertake just such a study, however, seemed provided by the knowlege of an extensive outbreak of hepatitis that had affected United States Army servicemen during World War II (7-9). This outbreak was first noted in March 1942, prompting an intensive investigation soon after its recognition of the surrounding circumstances. A careful questionnaire survey led to the conclusion that the epidemic appeared attributable to the receipt of presumed hepatitis-contaminated yellow fever vaccine. Accordingly, the Surgeon General ordered on April 14 that the implicated vaccine lots be withdrawn and replaced with vaccine that was serum free. The epidemic reached a peak in June 1942, after which the number of reported cases of “jaundice” slowly returned to the baseline level. It was estimated that there were approximately 50,000 cases of hepatitis involved in this outbreak. Thus, if this epidemic could be shown to be of HBV origin, and if the infected patients could be traced now more than 40 years later, it was believed that the link known to exist in the Far East between chronic HBV infection and HCC might be demonstrable in the United States also.

An essential requirement for such a study would be to have access not only to those who had been overtly infected and developed jaundice in 1942, but also to those who had received the vaccine at that time and had developed inapparent disease, since chronic hepatitis has been believed to evolve more frequently following anicteric than icteric hepatitis (10). Moreover, it would be necessary to evaluate individuals who, during that time, had received non-contaminated vaccine, in order to determine the background prevalence of cirrhosis and HCC in this “control” group. Because it indeed appeared possible to assemble the necessary groups, the present investigation was undertaken with two goals in mind: first, to determine whether HBV was responsible for the epidemic and, second, to establish whether the epidemic, assuming it to be of HBV origin, had increased the rate of development of cirrhosis and/or HCC.

A three-pronged approach was planned (11). The first consideration was to define the responsible virus by serologic means; the second was to establish a link between receipt of the contaminated vaccine and the occurrence of cirrhosis and HCC by means of a cohort mortality study, comparing the rates of these endpoints among the selected cohorts; and the third was to conduct a case-control study, comparing the frequency of receipt of the implicated vaccines among World War II veterans who had died from HCC with the frequency of their receipt among those who had

died from other pre-defined forms of cancer. Presented here are only the serologic and limited case-control mortality data.

The following three cohorts were chosen for serologic and mortality study: Group I were patients who had received the plasma-containing yellow fever vaccine and had been hospitalized with jaundice in 1942; this “symptomatic” group were derived from the microfilmed file of 31,500 military personnel who had been questioned during 1942-1943 in regard to their hospitalization for jaundice. Group II individuals had been vaccinated with one of the seven most “infectious” lots, but had not developed jaundice or been hospitalized, thus being acceptably referred to as the “ asymptomatic” group; their names came from the payroll rosters of enlisted men in units who had received the contaminated vaccine matched against the file of patients from those units hospitalized for jaundice. The use of the lot number as a basis for inclusion was precluded because a fire in 1973 in St. Louis, where personnel folders had been stored, had destroyed the vast majority of them. Group III were men who had entered the Army after July 1942 and hence had received only serum-free vaccine; their names were derived from a 2% file of men who had taken out National Service Life Insurance after August 1942. The selection of persons for serologic evaluation was confined in all three groups to white males alive on January 1, 1946, while Groups II and III were further restricted to men whose names did not appear in the 1942 file of persons hospitalized for jaundice. The total numbers of persons in each group from which the selections were made were 21,415, 19,114, and 25,262, respectively.

For the serologic analysis segment of the study, a sample of men from each of the three groups was selected, the numbers being based on sound statistical principles; approximately 200 persons in each group were believed to be necessary to distinguish a statistically significant difference in the HBV carrier rate among the three groups. A further guiding principle was that the subjects had to live within a 50-mile radius of the six Veterans Administration Medical Centers that had been asked to participate in the study (Boston, Massachussetts; Hines, Illinois; Dallas, Texas; Los Angeles, California; Miami, Florida; and Washington, D.C.) (11). The addresses of the study subjects were acquired with the help of the Internal Revenue Service, the Social Security Administration, and the Health Care Financing Administration. Once located, they were asked to come to one of the six study centers where they were subjected to an historical interview and a brief physical examination, and had a sample of blood drawn for the following hepatitis serologic assays: hepatitis B surface antigen (HBsAg), antibody to hepatitis B surface antigen (anti-HBs), antibody to the hepatitis B core antigen (anti-HBc), antibody to the hepatitis A virus (anti-HAV), and antibody to the hepatitis delta virus

(anti-HDV); testing for the hepatitis B e antigen (HBeAg) and for the presence of hepatitis B virus DNA (HBV DNA) was to be conducted in all HBsAg-positive samples and among a limited number of randomly selected HBsAg-negative samples; in all instances, commercial radio-immunoassays from Abbott Laboratories, N. Chicago, were used. In addition, samples were tested by enzyme-linked immunoassay for the presence and concentration of alpha-fetoprotein (AFP) levels as a marker of incipient HCC.

For the cohort mortality segment of the study, mortality status was established by examining the Veterans Administration Beneficiary Identification and Records Locator System, which has recorded over 90% of deaths of World War II veterans. Cause of death was derived from a review of the death certificates, seeking the classification code number 155 (malignant neoplasm of liver and biliary passages) in the 7th revision of the International Classification of Diseases (ICD) of the World Health Organization. Where possible, confirmation of diagnosis was sought from a review of clinical and pathologic information relating to cause of death.

Blood samples were obtained from 221 Group I, 171 Group II, and 205 Group III subjects. Among the Group I individuals, HBV markers were detected in a total of 216 (97.7%). This consisted of a single person with HBsAg and anti-HBc (0.5%), 199 (90%) who were positive for both anti-HBs and anti-HBc, and 16 (7%) who were positive for anti-HBc alone. The Group II individuals had detectable HBV markers in 131 (77%); this consisted of both anti-HBs and anti-HBc in 120 (70%), anti-HBs alone in 10 (6%), and anti-HBc alone in 1 (1%). In marked contrast was the finding of HBV markers in only 27 (13%) of the Group III subjects--both anti-HBs and anti-HBc in 13 (6%) and anti-HBs alone in 14 (7%). With respect to anti-HBc, this marker was detected in all of those in Group I with positive HBV markers, in 92% of those in Group II with markers, and in 48% of the Group III subjects with markers. A striking statistically significant difference in detected anti-HBs and anti-HBc together was noted between the individuals of both Groups I and II and those of Group III (p<0.0001) and also between the subjects of Group I and those of Group II (p<0.0001).

Anti-HAV was found in 165 (75%) of the individuals in Group I, in 124 (73%) of those in Group II, and in 129 (63%) of Group III. Although a similar prevalence of anti-HAV was detected in all three groups, the figure was significantly lower in Group III.

HBsAg was found in only 1 subject (from Group I), representing an overall positivity rate of 1 in 392 (0.26%) exposed persons (Groups I and II combined). This person was positive also for hepatitis B e antigen (HBeAg) and hepatitis B virus DNA (HBV DNA).

The single HBsAg-positive individual was subtyped as adw2. Sixty-four anti-HBs-positive specimens from Groups I and II were also subtyped, 62 with anti-a, 24 with anti-d, 10 with anti-w, and 7 with anti-w2.

None of the subjects had detectable anti-HDV.

AFP levels ranged from 0.05 to 31.4 ng/ml (median, 2.2). The slightly elevated levels (>10.0 ng/ml) did not differ among the three groups or correlate with serologic findings.

The “concentrations” of anti-HBs and anti-HBc, as judged by the median positive:negative ratios of the radioimmunoassay tests, showed the following: anti-HBs was significantly higher in Group II as compared to the other two groups, while anti-HBc was significantly higher in Group I in comparison to the other two.

Standardized mortality ratios showed that death from all causes did not differ significantly among the three groups. The numbers of deaths attributed to liver and gall bladder cancer were not unexpectedly high (30, 33, and 20 in the three groups, respectively). Standardized mortality ratios for cancer of the liver and gall bladder also did not differ significantly as a set, although there was a significantly higher ratio for Group II than for Group III. Deaths from liver diseases of other types, in particular viral hepatitis and nonalcoholic chronic liver disease, showed no significant differences.

These data indicate that the yellow fever vaccine-related hepatitis outbreak was clearly a result of HBV infection, that the epidemic was associated with an unexpectedly low HBsAg carrier rate, and that the epidemic appears not to have increased the rate of development of the two feared consequences of chronic HBV infection, namely cirrhosis and HCC. The evidence in support of the implication of HBV is the strikingly high prevalence of HBV antibody markers among both those who had received the vaccine and developed jaundice (97%) and those who had received the implicated vaccine but did not develop overt disease (77%). In contrast, the prevalence of these markers in the “control” group (Group III) was comparable to the prevalence in the general adult U.S. population (12). Although the prevalence of antibody to hepatitis A also was high, it was present in similar frequency in all three groups and also similar to that of the general adult U.S. population of an equivalent age (13).

It can be estimated, based on the knowledge that 427,000 doses were administered from the known icterogenic vaccine lots, on the knowledge that 50,000 overt vaccine-related hepatitis cases occurred, and on the anti-HBc results among Group II of the present study, that 78% of all recipients of the known icterogenic lots were probably infected with HBV. Extending the calculations further, it appears that about 330,000 men who received these vaccines developed HBV infection, the ratio of icteric:anicteric hepatitis being 1:7. This epidemic thus is the largest point-source hepatitis B outbreak ever recorded.

Several of the serologic findings are worthy of special mention. First is the evidence that this dramatic hepatitis B epidemic did not appear to be associated with the expected HBV carrier rate. It had been anticipated that approximately 10% of acutely infected persons would become carriers, especially if they had had subclinical infection (5, 6). Instead, the rate as based on the single carrier, whose subtype characteristics conformed to the findings among those with anti-HBs, was less than 0.5%. A second item of importance is the evidence that HBV antibodies that derive from an overt infection appear to persist for life, as indicated by their high “concentration” detected more than 40 years after the original infection. This contrasts with the shorter life span of antibodies induced by hepatitis B vaccination (14).

The reason for the unexpectedly low carrier rate is not apparent from this study. Conceivably, it could be accounted for by the loss to follow-up of carriers who had died from cirrhosis or cancer. The fact that the mortality study failed to detect excess mortality from cirrhosis and/or HCC, however, tends to disprove this view. On the other hand, it could be low because of certain host-related or virus-related factors. Host-related

factors that are known to be of importance in regard to the rate of chronicity are age at which the initial infection occurs, the immune status of the infected person, gender, and race. The first two items appear to be the most compelling correlative factors (15, 16). Data from highly endemic areas where perinatal transmission is the predominant mode of virus propagation indicate that there is a strong inverse relationship between the age of the primary infection and the likelihood of becoming a carrier, the rate declining dramatically with advancing age (17-19). The outbreak under discussion involved young, generally healthy white males, permitting the conclusion that this population group uncommonly become carriers following acute HBV infection. How then to reconcile this information with the prevailing view that 5-10% of acutely HBV-infected persons in the United State become carriers?

A critical analysis of the extant literature indicates that the figure of 5-10% derives from studies of markedly heterogeneous populations (5, 6). Such studies have involved groups that have included multiply exposed homosexual men, intravenous drug abusers, and persons of Asian descent among whom a subclinical HBsAg carrier state is relatively common. Many of these individuals may present for the first time with a bout of apparent acute hepatitis B, HBsAg persisting after the acute illness has subsided, leading to the mistaken conclusion that chronic hepatitis has just evolved. In actual fact, they may actually be unrecognized pre-existing carriers who are given a diagnosis of acute hepatitis when spontaneous or induced viral reactivation or superimposed infection with another undetected virus (A, non-A/non-B, or D) occurs (1, 20, 21). Because acute hepatitis B in young, healthy, immunocompetent, singly exposed persons is relatively uncommon, it has been difficult to develop accurate data on the frequency of development of the carrier state in this population group.

Virus-related factors conceivably also could account for the lower-than-anticipated carrier rate, factors such as the size of the inoculum, viral subtype or strain differences, or other factors that might interfere with viral replication. There is little to support the former two items with respect to altering the frequency of chronicity, but there is evidence that viral interference can occur and change the natural history. For example, it is apparent that a reduction in viral replication results when the HBsAg carrier is superinfected with other known hepatitis viruses, such as those of hepatitis A, non-A/non-B hepatitis, and hepatitis D (1). It is possible, therefore, that the virus of yellow fever in the vaccine could have altered the outcome. Similarly, anti-HBs that almost certainly was present in the original pooled plasma could have affected both the incubation period and the carrier state. The opportunity to study these potential interfering factors is, unfortunately, thwarted by the inability to trace the original contaminated yellow fever vaccine lots.

It is perhaps not surprising that this dramatic and extensive epidemic of hepatitis B was not associated with a correspondingly high rate of cirrhosis and HCC. The data seem indisputable that HBsAg carriers are at unusually high risk of developing these serious sequelae (22). However, since this epidemic did not culminate in a recognizably high carrier rate, these sequelae did not emerge. Further information in this regard is anticipated from the case-control study.

1. Seeff LB Koff RS : Evolving concepts of the clinical and serologic consequences of hepatitis B virus infection Sem Liv Dis 1986 ; 6 : 11-22

2. Popper H Shafritz DA Hoofnagle JH : Relation of the hepatitis B virus carrier state to hepatocellular carcinoma Hepatology 1987 ; 7 : 764-772

3. Szmuness W : Hepatocellular carcinoma and the hepatitis B virus: evidence for a causal association Prog Med Virol 1978 ; 24 : 40-69

4. Kew MC : The possible etiologic role of hepatitis-B in hepatocellular carcinoma: evidence from Southern Africa In: Chisari FV, ed. Advances in Hepatic Research New York : Masson Publishing USA, Inc., 1984 : 203-215

5. Redeker AG : Viral hepatitis, clinical aspects Am J Med Sci 1975 ; 270 : 9-16

6. Hoofnagle JH Seeff LB Bales ZB Gerety RJ Tabor E Serologic responses in HB In: Vyas GN, Cohen SN, Schmid R eds. Viral Hepatitis: A Contemporary Assessment of Etiology, Epidemiology, Pathogenesis and Prevention Franklin Institute Press 1978 : 219-242

7. Sawyer WA Meyer KF Eaton MD Bauer JH Putnam P Schwentker FF : Jaundice in Army personnel in the western region of the United States and its relation to vaccination against yellow fever Am J Hyg 1944 ; 39 : 337-430

8. Idem Jaundice in Army personnel in the western region of the United States and its relation to vaccination against yellow fever Am J Hyg 1944 ; 40 : 35-107

9. Walker DW : Some epidemiological aspects of infectious hepatitis in the U.S. Army Am J Trop Med 1945 ; 25 : 75-82

10. Alward WLM McMahon BJ Hall DB e al : The long-term serologic course of asymptomatic hepatitis B virus carriers and the development of primary hepatocellular carcinoma J Infect Dis 1985 ; 151 : 604-609

11. Seeff LB Beebe GW Hoofnagle JH et al : A serologic follow-up of the 1942 epidemic of post-vaccination hepatitis in the United States Army New Engl J Med 1987 ; 316 : 965-970

12. McQuillan GM Townsend TR Fields HA Carroll M Leahy M Polk BF : Seroepidemiology of hepatitis B virus infection in the United States: 1976 to 1980 Amer J Med 1989 ; 87(suppl 3A) : 5S-10S

13. Szmuness W Dienstag JL Purcell RH Harley EJ Stevens CE Wong DC : Distribution of antibody to hepatitis A antigen in urban adult populations N Engl J Med 1976 ; 295 : 755-759

14. Hadler SC : Are booster doses of hepatitis B vaccine necessary? Ann Intern Med 1988 ; 109 : 657-458

15. McMahon BJ Alward WLM Hall DB et al : Acute hepatitis B virus infection: relation of age to the clinical expression of disease and subsequent development of the carrier state J Infect Dis 1985 ; 151 : 599-603

16. Szmuness W Prince AM Grady GF et al : Hepatitis B infection: a point-prevalence study in 15 US hemodialysis centers JAMA 1974 ; 227 : 901-906

17. Beasley RP Hwang L-Y Lin C-C et al : Hepatitis B immune globulin (HBIG) efficacy in the interruption of perinatal transmission of hepatitis B virus carrier state Lancet 1981 ; 2 : 388-393

18. Beasley RP Hwang L-Y Lin C-C et al : Incidence of hepatitis B virus infections in preschool children in Taiwan J Infect Dis 1982 ; 146 : 198-204

19. Beasley RP Hwang L-Y Lin C-C Ko Y-C Twu S-J : Incidence of hepatitis among students at a University in Taiwan Am J Epidemiol 1983 ; 117 : 213-222

20. Davis GL Hoofnagle JH : Reactivation of chronic type B hepatitis presenting as acute viral hepatitis Ann Intern Med 1985 ; 102 : 762-765

21. Rizzetto M Verme G : Delta hepatitis - present status J Hepatol 1985 ; 1 : 187-193

22. Beasley RP Hwang L-Y Lin C-C Chien C-S : Hepatocellular carcinoma and hepatitis B virus: a prospective study od 22,707 men in Taiwan Lancet 1981 ; 2 : 1129-1133

I am indebted to Drs. Gilbert W. Beebe (National Cancer Institute), James E. Norman (National Heart, Lung, and Blood Institute), and Jay H. Hoofnagle (Division of Digestive Diseases and Nutrition, NIH), without whom this study would not have been initiated or conducted. I would also like to thank the following study participants: Zelma Buskell-Bales, B.S., Jeanne G. Waggoner, B.A., Neil Kaplowitz, M.D., Raymond S. Koff, M.D., John L. Petrini, Jr., M.D., Eugene R. Schiff, M.D., James Shorey, M.D., and Malcolm M. Stanley, M.D.