Seroepidemiologic. Study of Epstein-Barr Virus Infections in a Rural Community

Transcription

1 THE JOURNAL OF INFECTIOUS DISEASES VOL. 131, NO.4 APRIL by the University of Chicago. All rights reserved. Seroepidemiologic. Study of Epstein-Barr Virus Infections in a Rural Community Ciro Valent Sumaya, Werner Henle,* Gertrude Henle, Margaret H. D. Smith, and Dorothy LeBlanc From the Division ofinfectious Diseases, Department of Pediatrics, UCLA School of Medicine, Los Angeles, California; The Children's Hospital of Philadelphia and the School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania; and the Departments of Epidemiology and Pediatrics, Tulane University School of Medicine, New Orleans, Louisiana The prevalence and titers of antibodies to capsid antigens of Epstein-Barr virus and to the diffuse and restricted components of the Epstein-Barr virus-induced early antigen complex were determined in 109 families of a semirural community in Louisiana. Titers of antibody to the capsid antigens ~ 10 were found in 84% of children aged two to five years, and the prevalence increased with age to nearly ]00%. There was a positive but variable correlation of the prevalence of anti-capsid antigen reactivity with low socioeconomic status and crowding. An overrepresentation of high titers of antibody to capsid antigens was present in individuals with a past history of pneumonia and urinary tract infections. The geometric mean titers of antibody to capsid antigens were highest in early childhood, lowest in adolescence and young adulthood, and high in the elderly. Females in all age groups and tonsillectomized children showed a higher geometric mean titer than their male and nontonsillectomized counterparts, respectively. Antibodies to the early antigen complex were found rarely (8.2%) and only in sera with relatively high titers of antibody to capsid antigens. There is little doubt that the Epstein-Barr virus (EB V), first detected in cultures of African Burkitt's lymphoma cells [1], is the cause of Received for publication July 22, 1974, and in revised form December 11, This work was supported by training grants no. AI and 5T01-CA , grant no. CA-04568, and contracts no. PH and NOI-CP within the Virus Cancer Program, National Cancer Institute. We thank Drs. Guy R. Newell, Peter Mansell, and James D. Cherry for advice and assistance; Ms. Louise Zimmerman and students of the Louisiana State University School of Nursing for assistance in the field work; Dr. Potter Chang for the statistical analyses; the St. James Public Health Unit, local physicians, clergy, news media, and especially the people of the Fifth Ward in St. James Parish, Louisiana for their generous cooperation. Please address requests for reprints to Dr. Ciro V. Sumaya, Department of Pediatrics, UCLA School of Medicine, Los Angeles, California * Recipient of career award no. 5-K6-AI-22 from the National Institutes of Health. heterophile-positive (and sometimes heterophile-negative) infectious mononucleosis [2, 3]. The relationship of EB V to human neoplasms and other disorders is less well defined [4, 5]. Seroepidemiologic surveys have shown that antibodies to EB V are often acquired in early childhood, especially under low socioeconomic conditions, and that the prevalence of antibodies increases with age [6-16]. The majority of EB V infections at an early age are thought either to be asymptomatic or to cause mild illnesses that are not readily distinguishable from many other common viral infections of child; hood. Noticeably lacking are large-scale surveys of randomized American families. In the present study, the prevalence of antibodies to EBV was determined in a systematic household sample of a semirural community. The specific aims were to ascertain an association between antibody titers against EB V and various epidemiologic or demographic factors. 403

2 404 Sumaya et at. Materials and Methods Study population. The area sampled was Ward 5, St. James Parish, Louisiana, a sernirural, predominantly farming (sugar cane) and oil-refining community, abutting the Mississippi River. The population of the ward was 2,788 (1970 census), of which 23.6% was white and 76.4% was nonwhite. The first household was picked randomly, and then every fourth house.was systematically asked to participate in the study. The visit consisted of an interview of all.household members (parents answered for the children) and. single blood samples from each individual. The interview questionnaire was divided into three sections: (I) general vital statistics, (2) medical history, and (3) testing of exposure to chemicals, other products, and animals. The visits were made between August 1972 and February The Hollingshead social position method (sum of the numerical scale of occupation [x7] and education [x4] figured on the head of the household) was used to determine socioeconomic status. For our purposes, the divisions were high, middle, low, and very low. Serology. Blood samples were kept refrigerated until serum was separated. The sera were stored at -70 C until tested. Antibodies to EB viral capsid antigens (VCA) were determined by the indirect immunofluorescence technique [17, 18}. In brief, the EB3 line of Burkitt's lymphoma cells was propagated on medium RPMI 1640 supplemented with 10% fetal calf serum. Three to four days before preparation of smears, the cells were changed to arginine-free Eagle's basal medium supplemented with 25% fetal calf serum. Cell smears were fixed with acetone and stored at -20 C. For staining, the cell smears were overlayed with the diluted test serum and incubated at 37 C for 45 min. The smears were then washed with phosphatebuffered saline solution, were overlayed with fluorescein isothiocyanate-conjugated goat antibodies to human IgG, and were again incubated at 37 C for 45 min. They were then washed and fixed onto slides with polyvinyl alcohol-glycerol mounting medium and examined under ultraviolet illumination. Antibodies to the diffuse and restricted com- ponents of the early antigen complex were titrated in a similar fashion; Raji cells (a Burkitt's lymphoma line normally free of VCA- and early antigen-producing cells) were used 48 hr after exposure to EB V [19, 20]. The titers of antibody to VCA and early antigen are reported as the reciprocal of the serum dilution. Results Survey data. A total of 109 households (462 individuals) were recruited into the study; 85% of the households had every member participating. The age span among participants was 30 days to 90 years. Serological results. (l) Personal and socioeconomic factors. Table 1 shows that the percentage of individuals with titers of antibody to VCA of ~ 10 increased from 45.5% at one year of age to 84.3% in the two- to five-year-old group. (A titer < 10 indicates susceptibility to a primary EBV infection and EBV-associated disease [21].) The percentage thereafter increased gradually with age to nearly 100%. The peak geometric mean titer of the positive sera was noted in the group aged seven to 12 months, but the number in this group was small. The geometric mean titer then declined to the lowest level in the year-old group and subsequently increased progressively with advancing age. An asymptomatic 74-year-old black female had the highest recorded titer, 1,280. Thirty-eight sera (8.2%) had antibodies to the early antigen complex. These were found mainly in sera with high anti-vca titers (~160) but occasionally also in sera with anti- VCA titers of 80 (six) and 40 (one). Most titers of antibody to the early antigen were low; i.e., 10 in 24 cases and ~4O in three. Although the incidence of sera with antibody titers ~ 10 was the same for both sexes, females consistently had significantly higher geometric mean titers of anti- VCA (13-22 points) than males throughout the various age groups. (Comparison of the different age groups by t-test gave the following values for P: zero to five years, P < 0.005; years, P < 0.005; above 31 years, P < 0.1; total, P < ) Although the geometric mean titers of anti-

3 Seroepidemiology of EBV 405 OO-NO-NM Oao.--lrlo\'>Oool""ivi"':"':oOt--= - trlor) ~ tri...,. or) r- vca were not appreciably different according to race, significantly more blacks under 15 years of age than whites of the same age group had antibodies (P < 0.01) (figure 1). The anti- VCA titers among households of the four socioeconomic levels were not significantly different from one another. On the other hand, children from high- or middle-status families had a lower prevalence of anti- VCA titers than those from the low- or very low-status households (P < 0.01). There was no consistent correlation between anti-vca titers and household size. Another measure of crowding, i.e., schooling, revealed that all 17 children in the Head Start and kindergarten programs had anti- VCA titers ~ 10. Among the one- to three-year-old children (not yet in any school), 11 of 34 still had no antibodies detectable at a titer of ~ 10. The variation of anti- VCA titers among the members of any one household was smaller than the variation among the household means (ratio was greater than the 95 percentile of the F distribution, with degrees of freedom of 60 and 120). This may indicate some aggregation of titers within members of the household. (2) Diseases. None of the individuals in the study had a history of infectious mononucleosis-like illness or of noticeable lymphadenopathy. No difference was noted between anti-vca status among persons from households in which tumors had occurred and that in members of households without such cases. A history of frequent or chronic respiratory tract, gastrointestinal, skin, or other infection was not related to anti-yea titers, except in those individuals who had had pneumonia or a urinary tract infection. A significantly higher percentage of those in the latter two groups had anti-vca titers ~160 than was found among those in the nonpneumonia or non-urinary tract infection group (P < 0.05), although the percentage of those whose sera had titers ~ 10 was similar in the two groups. A previous tonsillectomy was associated with a significantly larger number of individuals with anti- VCA titers ~ 160 in the six- to 17-year-old groups (P < 0,01) but not in those under 17 years of age. None of the children under five had had a tonsillectomy.

4 406 Sumaya et al. (3) Environmental factors. Of the various occupations, previous bagasse plant workers had the highest percentage of anti- VCA titers ;3160 (10 of 21 individuals); next in order were commercial business workers, teachers, and others. Households close to chemical plants or sugar cane fields where insecticide and herbicide spraying is common had results similar to those of households farther away. History of cigarette smoking and exposure to domestic animals had no correlation with anti- VCA titers. Discussion ence of antibody to the diffuse component of early antigen is generally taken as a sign of a current or recent primary EB V infection [20]. The presence of antibody to the restricted component of early antigen is considered to be a result of the extent of the viral carrier state, reflected also in relatively high anti- VCA titers [23]. There was no difference between the prevalence of anti- VCA in males and that in females, nor have significant differences between the sexes been observed in the incidence of infectious mononucleosis [24]. However, females in all age groups (not just young girls [15]) consistently had higher geometric mean titers of anti- VCA than males. Similarly, females have shown a greater antibody responsiveness to enteric bacterial antigens, to rubella vaccine [25], and to cytomegalovirus [26], and females have had somewhat higher levels of IgM than males [27]. The difference between the prevalence of anti- VCA in white persons under 15 years of age and that in black individuals of the same age group (figure 1) is probably not related to race but rather is a reflection of socioeconomic conditions. The white children in this study were mostly of a higher social class than the black children. The relatively lower percentage of anti- VCA-positive children from families in a high social position confirms previous reports [7, 12, 14, 28]. However, the facts that primary EB V infections occur mainly under low socioeconomic conditions, predominantly early % WITH 60 ANTI BODIES NUMBER TESTED 57 II AGE GROUP (yrs.) Figure 1. Racial differences in the percentage of persons with titers of antibody to the viral capsid antigen of Epstein-Barr virus. 100 Studies in Uganda [15] and Mexico [16] revealed that by the age of one year >80% of infants had antibodies to EB V. In the present survey and in one done in Japan [9], this percentage of antibody-positive individuals was not reached before the age of two to five years. At the other extreme, 31% of Swedish girls under 10 years of age still had no detectable antibodies [10], and about 80% of the children from wellto-do American families were negative at that age [14]. It was not surprising that the peak geometric mean titer of anti- VCA was observed in early childhood, since primary EBV infections preceded the collection of serum by relatively short intervals. A decline in the geometric mean titer in later childhood and adolescence as observed here was noted also by Porter et al. [12]. It is known that anti- VCA titers, after reaching peaks in the early acute phase of infectious mononucleosis, decline to lower levels which then persist for years, if not for life [2, 13, 14]. These antibody levels are probably maintained by the persistent viral carrier state that regularly becomes established after primary EB V infections [2]. It is speculated that persistent infection is occasionally reactivated [21], perhaps more frequently in older individuals after a decline in immunologic responsiveness [22]. This could explain the increased geometric mean titer of anti- VCA in the older age group observed in our study or the almost constant titers noted by Tischendorf et al. [13]. The same explanation may hold for the detection of antibody to early antigen in donors of advanced age. The pres-

5 Seroepidemiology of EBV 407 in life, that they remain silent or are not specifically diagnosed, and that they induce a persistent immunity provide an explanation for the absence of histories of infectious mononucleosis in the study population and the infrequency of this disease among blacks [29]. It has been reported [11, 16] but not confirmed [30] that EB V spreads most readily under conditions of crowding. In our study, depending on the criteria chosen in evaluating crowding (Le., school attendance or household size, or even the method of analyzing socioeconomic position), different results are obtained. Thus it appears that, although the spread of EB V infections is facilitated to a variable extent by low socioeconomic status and crowding, such effects are not as evident as with other viruses of the herpes group [12]. The overrepresentation of elevated anti- VeA titers in individuals with a history of previous pneumonia or urinary tract infection or in tonsillectomized children is difficult to interpret. In tonsillectomized children, it can be speculated that the absence of tonsils may allow greater viral replication or more persistent infection at other lymphoidal sites. It is interesting that exposure to other respiratory irritants, such as insecticides, herbicides, liquid chemicals, or cigarette smoke, was not related to the anti- VCA titers. Bagasse workers (none of whom were clinically ill) had an increased prevalence of high titers. This could be related to that material's effect on lung tissue. A similar lack of correlation with exposure to domestic animals was not unexpected, since only certain nonhuman primates acquire infections with EB V or a closely related virus under natural conditions [31]. The height or difference in antibody titers associated with various epidemiologic factors in this study, even if statistically significant, may be considered marginal. To minimize methodologic differences [12], the results from a single laboratory in Philadelphia are presented. However, similar results were obtained at Tulane on the same serum tested blindly. Prospective studies are needed in selected groups of individuals to corroborate and elucidate the epidemiologic associations with EBV presented in this report. References 1. Epstein, M. A., Achong, B. G., Barr, Y. M. Virus particles in cultured lymphoblasts from Burkitt's lymphoma. Lancet 1: , Henle, W., Henle, G. Epstein-Barr virus: the cause of infectious mononucleosis. In I. M. Biggs, G. de-the, and L. N. Payne [ed.]. Oncogenesis and herpes viruses. International Agency for Research on Cancer, Lyon, 1972, p Henle, W., Henle, G. Epstein-Barr virus and infectious mononucleosis. N. Engl. J. Med. 288: , Klein, G. Herpes viruses and oncogenesis. Proc. Natl. Acad. Sci. U.S.A. 69: , Henle, W., Henle, G. Evidence for an oncogenic potential of the Epstein-Barr virus. Cancer Res. 33:14I9-1423, Levy, J. A., Henle, G. Indirect immunofluorescence tests with sera from African children and cultured Burkitt lymphoma cells. J. Bacteriol. 92: , Henle; G., Henle, W. Immunofluorescence, interference, and complement fixation technics in the detection of the herpes-type virus in Burkitt tumor cell lines. Cancer Res. 27: , Goldman, M., Reisher, J. I., Bushar, H. F. Serumantibodies to Burkitt cell virus. Lancet 1:1156, Hinuma. Y., Ohta-Hatano, R., Suto, T., Numazaki, Y. High incidence of Japanese infants with antibody to a herpes-type virus associated with cultured Burkitt lymphoma cells. Jap. J. Microbiol. 13: , Dernissie, A., Svedmyr, A. Age distribution of antibodies of EB virus in Swedish females as studied by indirect immunofluorescence on Burkitt cells. Acta Pathol. Microbiol. Scand. 75: , Pereira, M. S., Blake, J. M., Macrae. A. D. EB virus antibody at different ages. Br. Med. J. 4: , Porter, D. D., Wimberly. I., Benyesh-Melnick, M. Prevalence of antibodies to EB virus and other herpes viruses. J.A.M.A. 208: , B. Tischendorf, P., Shramek, G. J., Balagtas, R. C., Deinhardt, F., Knospe, W. H., Noble. G. R., Maynard. J. E. Development and persistence of immunity to Epstein-Barr virus in man. J. Infect. Dis. 122: , Henle, G., Henle, W. Observations on childhood infections with Epstein-Barr virus. J. Infect. Dis. 121 : , Kafuko, G. W., Henderson, B. E., Kirya, B. G., Munube, G. M. R., Tukei, P. M., Day, N. E., Henle, G., Henle, W., Morrow, R. H., Pike, M. c., Smith, P. G.. Williams, E. H. Epstein-Barr virus antibody levels in children from West Nile District of Uganda. Report of a field study. Lancet 1: , Golubjatnikov, R., Allen, V. D., Steadman, M.. Olmos B1ancarte, M. P., Inhorn, S. L. Prevalence of antibodies to Epstein-Barr virus, cytomegalovirus,

6 408 Sumaya et al. and toxoplasma in a Mexican highland community. Am. J. Epidemiol. 97: , Henle, G., Henle, W. Immunofluorescence in cells derived from Burkitt's lymphoma. J. Bacteriol. 91: , Henle, G., Henle, W., Clifford, P., Diehl, V., Kafuko, G. W., Kirya, B. G., Klein, G., Morrow, R. H., Munube, G. M. R., Pike, P., Tukei, P. M., Ziegler, J. L. Antibodies to Epstein-Barr virus in Burkett's lymphoma and control groups. J. Natl. Cancer Inst. 43: , Henle, W., Henle, G., Zajac, B. A., Pearson, G., Waubke, R., Scriba, M. Differential reactivity of human serums with early antigens induced by Epstein-Barr virus. Science 169: , Henle, G., Henle, W., Klein, G. Demonstration oftwo distinct components in the early antigen complex of Epstein-Barr virus infected cells. Int. J. Cancer 8: , Evans, A. S. The spectrum of infections with Epstein-Barr virus: a hypothesis. J. Infect. Dis. 124: , Burnet, F. M. Immunological surveillance. Pergamon, New York, Henle, W., Henle, G. Epstein-Barr virus-related serology in Hodgkin's disease. Natl. Cancer Inst. Monogr. 36:79-84, Carter, R. L., Penman, H. G. (ed.]. Infectious mononucleosis: Blackwell, Oxford, 1969, p Michaels, R. H., Rogers, K. D. A sex difference in immunologic responsiveness. Pediatrics 47: , Luby, J. P., Shasby, D. M. A sex difference in the prevalence of antibodies to cytomegalovirus. J.A.M.A. 222: , Stoop, J. W., Zegers, B. J. M., Sander, P. c.. Ballieux, R. E. Serum immunoglobulin levels in healthy children and adults. Clin. Exp. Immunol. 4: , Niederman, J. C., McCollum, R. W., Henle, G., Henle, W. Infectious mononucleosis: clinical manifestations in relation to EB virus antibodies. J.A.M.A. 203: , Niederman, J. C. Infectious mononucleosis. In M. M. Wintrobe, G. W. Thorn, R. D.Adams, E. Braunwald, K. J. Isselbacher, and R. G. Petersdorf (ed.]. Harrison's principles of internal medicine. McGraw-Hill, New York, Joncas, J. H. Clinical significance of the EB herpesvirus infection in man. Prog. Med. Virol. 14: , Landon, J. C,; Malan, L. B. Seroepidemiologic studies of Epstein-Barr virus antibody in monkeys. J. Natl. Cancer Inst. 46: , 1971.