The effect of age and occupation on the seroprevalence of Helicobacter pylori infection

Serological studies in developed and developing countries using enzyme-linked immunosorbent assays have validated this technique as a rapid, noninvasive method for the diagnosis of Helicobacter pylori infections. The prevalence of serum antibodies to H pylori was studied in 473 Canadian blood donors from Manitoba, 212 healthy Japanese. and 226 healthy Americans. As expected, the seroprevalence rose progressively with age in the three populations and reached its peak (greater than 55%) in subjects 60 years of age and older. The seroprevalence did not decrease in elderly persons (60 to 99 years), indicating a persistent immune response. More detailed analysis was perfom1ed on the Canadian population. Age-adjusted prevalence rates in men and women were similar. Among young adults (15 to 29 years). farmers had a significantly higher seroprevalence rate than white-collar or blue-collar workers. but in older persons occupational rates were similar. A multiple linear regression analysis of the data confirmed that age and occupation in young adults were both significantly associated with seroprevalence of H pylori infections.

T HE PRESENCE OF' H ELICOBACTER (FO RM ERLY CAMP\'LO- bacter [11) pylori in the stomach of persons with active chronic gastritis and peptic ulce r djsease h as led to s ubstantial inte rest in thls bacterium (2 -4).An increasing body of evidence indicates th at H pylori is involved in the etiology of active chronic gastritis and possibly in duodenal ulceration (3) .Virtually a ll infected persons m ount a specific systemic humoral immune response to the organism (4-7) .a phenomenon that has allowed the development of diagnostic tests U1at do not require endoscopic examination of the s ubject.In a ddition to their u se in diagnosing infection , serological tests can be used to gain knowledge of the epidemiology of the infection.Although much has been learned recently about H pylori epidemiology (5.8 -14), the use of a variety of different assays , not a ll of which have been rigorou sly validated .has limited in terpretation of the data.
This study used a single.previously validated enzyme-linked immunosorbent assay (ELISA) of high sensitivi ty and specificity {4,5, 10) to compare H pylori antibody prevalence in populations of a dults from three developed countries: Canada, th e United States of America and Japan.In the Canadian population.the effects of age, gender and occupation on tl1e prevalence of H pylori antibodies were also examined .

MATERIALS AND METHODS
Populatio n : Serum specimens were collected from 901 healthy persons.There were 473 blood donors from Manitoba.Winnipeg (aged 19 to 65 years.including 26 persons who were older tl1an 60 years).The blood don ors were from farms (34.2%), towns (48.9%) or cities (15.9%).as previously descri bed (15).From this population , information on gender and occupation was also tab ulated.There were 212 persons from Hyogo, Japan (aged 15 to 90 years, including 20 older than 60 years).There were 226 persons from Colorado (aged 15 to 59 years including 101 nursing home residents older than 60 years) .The populations from Hyogo and Colorado were from the Kobe and Denver metropolitan areas , respectively, and both represented urban popu lations.No endoscopic information was availa ble a bou t any of the subjects.All sera had been stored at -20°C for at least six months.ELISA for H pylori antibodies: Sera were examined for H pylori-s pecific IgG antibodies by ELISA, as previously described (5) .Th e screening serum dilution was 1 :800.To establish a threshold for positivity in the IgG ELISA.criteria previously described we re used (10).In brief.the mean optical density for sera from 35 healtl1y children under 10 years of age was determined.A p ositive result for an unknown sample was then defined as having an optical density value more than the m ean plus tl1ree intervals of standard deviation of the values obtained for the children's sera.Using linear regression, a threshold was calculated for each day's run based on Age, occupation and H pylori infection U1e mean optical density valu e of fo ur con trol sera that had been simultaneo us ly ru n with the children 's sera.as previou s ly desciibed (10).If the ratio of the optical density value of the unknown sample to the calculated thres h old for that day•s run wa s greate r tl1an 1.0, tl1e sam ple was considered positive .All assays were don e in duplicate on at least two sep arate days .Th e in tra-a n d interassay varia tions in optical dens ity of positive and n egative control sera were less than 5% .
Statistical m e t hods: The x2 test was used for univariate analys is of statistical s ignifica n ce of different rates of positivity among s ubgroups of the population studied.Multiple analysis of covariance ( 16) was used to determine correlates of H pylori a ntibod ies while cont.rollingfor other potentially confounding factors.

RESULTS
Pre v a lenc e of H pylo ri-specific serum IgG antib odies: In each of the populations tested.there was a relatively low seroprevalence of H pylori infection d uring U1e first three decades of life.However.seroprevalen ce rose progressively with age in each popu la tion , reachincr the highest point by the s ixth decade of life (Figure 1).
Using a x2 test of trend tl1e a utl1ors found that a s ignificant trend in the prevalence of infection was present in each of the three countries (P<0.02).If the last age categoiy in Figure 1 is excluded from analys is, no s ign ificant trend was found in any cou ntry, but if the data a r e combined from the three countries, the trend is borderline s ignificant (P=0.056).Per sis t ence of H pylori-specific antibodies in elderly per sons: The prevalence of specific IgG did not decrease in persons aged 60 to 99 years (Table 1).Eld erly persons (older tl1an 60 years) living in nursing homes or in fan1ily settin gs had similar seroprevalences (70 versus 65%).Moreover.the age-specific seroprevalen ce ra tes were similar in tl1e tl1ree popu la tions (Japan .Canada and USA) (data not shown) .----+ ------+ -----1 -----  sociated with H pylori infection, and that the farmers were at risk for infection at a younger age than the other groups (Figure 2).DISCUSSION Age has been the single most important determinant of the prevalence of H pylori antibodies in studies of healthy populations (9 , 12, 17, 18).However, comparisons of different populations has been difficult because a variety of serological assays were employed.In this study using a single, well validated assay (4,5 ,9, 10) to determine seropositivity in populations from three developed countries, the authors observed a progressive age-related increase in the prevalence of H pylori antibodies that was similar in the three populations.Although subjects under 15 years of age were not included, the trend of seroprevalence by age in the populations studied was similar to the rates observed in asymptomatic volunteers from several developed countries who underwent endoscopy (4 , 19-2 1), corroborating the lack of significant artefact in the volunteer studies .In developed countries, H pylori infection is uncommon early in life; the progressive increases suggest that most transmission probably occurs during tl1e third through sixth decades of life.In developing countries, transmission is apparently more intense and occurs earlier (8 ,9).Similar to the previous studies (5,6) , the present authors found no gender-related differences in seroprevalence of H pylori infection when a multiple variant analysis was performed.
A decline in the prevalence of detectable antibodies, as well as a decline in natural in1munity to some infectious agents , has been reported to occur with ageing (22 ,23) .However, the present study found that the prevalen ce of antibodies to H pylori remained high and stable in persons 60 to 99 years old.This finding s uggests that H pylori infection persists for many years , perhaps for life, in most infected persons, and that infection is associated with chronic antigenic stimulation.The alternative hypothesis is that antibodies persist despite clearance of infection.Although this may be possible for brief periods following clearance, studies of antibiotic treatment of H pylori infection indicate that when the infection is eradicated antibodies diminish (2 4), as conventional B cell theory suggests (25).
Several a u thors have suggested that certain occupations may be associated with an increased risk of H pylori infection (26)(27)(28)(29)(30). Vaira et al (31) demonstrated that abattoir workers had significantly high er infection rates than other workers, but no information was provided on ethn icity or other socioeconomic factors.As previously reported in blood donor populations (26)(27)(28)(29)(30)(31)(32).age was strongly associated witl1 tl1e presence of H pylori infection.In the present study, occupation was associated with the presence of infection after controlling for gender and age, suggesting that certain activities may be associated with a higher risk of H py lori infection than others.Using multiple analysis of covariance, farmers were more likely to be infected than were the other occupational groups studied, an effect due mostly to high rates among young farmers.There is n o information on the specific natu,e of the farm exposures to determine the potential source for tl1is early transmission.Ethnic status (8,9, 13, 14,33,34) may be related to the seroprevalence of H pylori infection, or perhaps residence (ie, urban, suburban, rural) and not occupation is the important factor in acquisition of H pylori infection.Unfortunately, the present study lacked the demographic data to evaluate this hypothesis: however, the population of Manitoba is predominantly Caucasian.Nevertheless, a possible confounding role of etl1nicity and residence on occupational association with H pylori infection cannot be ruled out.

TABLE 1
Prevalence of Helicobacter pylori-specific serum lgG in 147 elderly persons in Japan , Canada and the United StatesRelation of H pylori infection with gender: Of the 4 73 Canadian blood donors, the seroprevalence of H pylori infection was studied in 469 subjects in whom gender was known (256 men and 213 women).A significantly higher proportion of women were positive for H pylorispecific IgG than were men (41 versu s 28%; P=0 .007);however,when data from the groups were age-adjusted no significant difference remained.Influence of occupation on H pylori infection rates: Among the Canadian subjects, information regarding employment and occupation were available for 466 persons.Not surprisingly, based on their age, retired persons had the highest seroprevalence of H pylori infection (Table2) .The homemaker group had a high seroprevalence similar to that of the farmer group , but when results were age-adjusted there was no significant association.Among working age persons 15 to 39 years old, the highest seroprevalence of H pylori infection was in farmers (52%).
Figure 2) Risk of injection with Helicobacter pylori among 466 healthy blood donors from Manitoba according to age and occupation: w hite-collar b=0.04±0.01(oJ;blue-collar b=0.05±0.01(t,.J; farmer b=0.008±0.01(OJ;homemaker b=0.06±0.01(OJ.T he b values represent the incremental gain in injection rate with each year•s increment in age 136TABLE 2 Comparison of He/icobacter pylori-specific lgG antibody prevalence among Manitoban blood donors in relation to occupation REFERENCESl.Goodwin CS.Armstrong JA.Chilvers T, et al .Transfer of Campylobacter pylori and Campylobacter mustelae to Helicobacter gen.nov.as Helicobacter pylori comb.nov.and Helicobacter mustelae comb.nov.respectively .Int J Syst Bacteriol 1989:39:397-405.2. Marshall BJ, Warren JR.Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration.Mitchell.HM. Lee A. Berkowicz J, Borody T. The use of serology to diagnose active Campylobacter pylori infection .Med J Aust 1988: 149:604-8.12. Jones OM, Eldridge J , Fox AJ.Sethi P, Whorwell PJ .Antibody to tl1e gastric can1pylobacte r-like organ ism (Campylobacter pyloridis) -Clinical correlations and distribution in the nonnal population.J Med Microbial 1986:22 :57 -62.13.Dwyer B. Kalclor J. Tee W. Marakowski E, Raios K. Antibody response to Campy lobacter pylori in diverse CAN J INFECT D rs VOL 3 No 3 MAY/JUNE 1992 Age, occupation and H pylori infection