Abstract
Purpose
The authors recently reported high Helicobacter pylori sero-prevalence among African-Americans of high African ancestry. We sought to determine whether neighborhood-level socio-economic characteristics are associated with H. pylori prevalence and whether this helps explain the link between African ancestry and H. pylori.
Methods
Antibodies to H. pylori proteins were assessed in the serum of 336 African-American and 329 white Southern Community Cohort Study participants. Prevalence odds ratios (ORs) and 95 % confidence intervals (CIs) for CagA+ and CagA− H. pylori were calculated using polytomous logistic regression in relation to 10 Census block group-level measures of socio-economic status.
Results
After adjusting for individual-level characteristics, three neighborhood-level factors were significantly inversely related to CagA+ H. pylori: percent completed high school; median house values; and percent employed (comparing highest to lowest tertile, OR, 0.47, 95 % CI, 0.26–0.85; OR, 0.56, 95 % CI, 0.32–0.99; and OR, 0.59, 95 % CI, 0.34–1.03, respectively). However, accounting for these measures did not attenuate the association between African ancestry and CagA+ H. pylori, with African-Americans of low, medium, and high African ancestry maintaining two-, seven-, and ninefold increased odds, respectively, compared to whites.
Conclusions
Neighborhood-level measures of education, employment, and house values are associated with CagA+ H. pylori sero-prevalence, but do not explain the persistent strong relationship between African ancestry level and CagA+ H. pylori. The findings suggest that neighborhood socio-economic status can help to highlight high-risk areas for prevention and screening efforts and that the link between African ancestry and H. pylori may have a biological basis.
Keywords: Helicobacter pylori, Education, Income, Neighborhood factors, African ancestry
Introduction
Helicobacter pylori, a gram-negative spiral bacterium that resides in the lining of the stomach of approximately half of the world’s population, is one of the leading risk factors for chronic gastritis, peptic ulcer disease, and gastric cancer. While the prevalence of infection with H. pylori in the United States (~30 %) has generally been found to be lower than that in countries in Asia, Africa, and South America (60–90 %) [1], individuals of ethnic minorities in the United States have consistently been found to have higher H. pylori sero-prevalence [2, 3]. We recently found very high H. pylori sero-prevalence among both low-income African-Americans (89 %) and whites (69 %) recruited from community health centers in the southeastern United States [4]. Upon further delineation by African ancestry, even stronger racial differences were observed, with 92 % of individuals in the highest tertile of African ancestry having antibodies to H. pylori. Of note, 74 % of high-African-ancestry individuals (compared to 25 % of whites) had antibodies to the H. pylori protein cytotoxin-associated antigen (CagA) [4], a gastric cancer virulence factor and the most well-established high-risk marker.
In addition to non-white race, low socio-economic status (SES) has consistently been associated with H. pylori infection [5, 6], potentially as a marker of hygiene practices or other lifestyle characteristics (such as household crowding) that would increase transmission rates, and/or related to psychosocial stressors that might impair the immune system. Studies attempting to untangle the effects of race and indicators of SES generally do find a racial disparity in the prevalence of H. pylori infection persisting even after adjusting for individual levels of education and income [4, 7], suggesting that the cofactors resulting in a higher risk for racial minorities are still unknown. It has been suggested that these cofactors may be other, unmeasured characteristics related to low SES, possibly those captured by the neighborhood environment. In the United States, a variety of health conditions, including cardiovascular disease [8], atherosclerosis [9], and obesity [10] have been found to be associated with neighborhood-level characteristics. In China, it has also been reported that village education level is more predictive of an individual’s H. pylori status than the individual’s own education level [11]. An important question therefore is whether infection with H. pylori in the United States is associated with neighborhood-level factors and whether such associations help to explain the high H. pylori sero-prevalence among African-Americans, particularly those of high African ancestry.
To further investigate the relationship between socioeconomic characteristics, race, and H. pylori infection in the United States, our aim was to explore, above and beyond individual SES, the association between neighborhood-level socio-economic characteristics and CagA-specific H. pylori sero-prevalence within the Southern Community Cohort Study. We also sought to investigate whether the combination of neighborhood-level and individual-level socio-economic characteristics could explain the strong association between higher African ancestry and CagA+ H. pylori infection.
Materials and methods
Study population
A detailed description of the Southern Community Cohort Study (SCCS) has previously been published [12]. Briefly, from 2002 to 2009, approximately 86,000 men and women 40–79 years of age from 12 southeastern states were recruited to be participants in the SCCS. The majority (~86 %) were recruited from community health centers, where they completed a comprehensive computer-assisted in-person interview that collected information on demographics, including the individual-level socio-economic status measures of annual household income, educational attainment, marital status, and current occupational status, as well as other lifestyle characteristics, medical history, anthropometrics, and regular diet. The remaining 14 % were recruited by mail, completing the same baseline survey on paper. All participants reported their race using a preprinted card that instructed them to indicate all race/ethnic categories to which they belong. Individuals whose baseline interviews took place at a community health center were asked to donate a venous blood sample (20 mL) that was then refrigerated, shipped overnight to Vanderbilt University, centrifuged the next day, and stored at −80 °C.
For initial studies of biomarkers in the SCCS, 792 individuals were selected from the 12,162 participants who enrolled in the SCCS from March 2002 to October 2004 and donated a blood sample at baseline, based on a 2 × 2 × 3 × 3 factorial design, leading to 22 individuals in each of the 36 strata defined by self-reported race (African-American/white), sex, smoking status (current/never/former), and body mass index (18–24.9/25–29.9/30–45 kg/m2). This design was selected so that there would be a balanced distribution across these factors in consideration of the studies of several blood biomarkers measured, including H. pylori.
Genetic analysis and ancestry estimation
Laboratory personnel blinded to the status of the samples extracted genomic DNA from buffy coat using QIAamp DNA kits (Qiagen, Valencia, CA) according to manufacturer’s instructions and carried out genotyping using the Illumina GoldenGate genotyping platform (Illumina Inc., San Diego, CA). The inclusion of blinded quality control samples (n = 29) and pairs of duplicate samples (n = 171) revealed a consistency rate of 99.9 %. Using a Bayesian clustering approach within STRUCTURE software (version 2.2.3) [13], a set of 276 single-nucleotide polymorphisms were selected to estimate African and European ancestry levels. For each individual in the present study, STRUCTURE generated an admixture estimate (from 0.00 to 1.00) for both African ancestry and European ancestry.
H. pylori multiplex serology
In preparation for H. pylori assaying, serum samples for each study subject were aliquoted into 50 μL portions. H. pylori multiplex serology was performed using an antibody detection technology based on fluorescent polystyrene beads (Luminex) and recombinant glutathione S-transferase (GST) fusion protein capture [14–16]. All sera were analyzed once within a single assay day. Antigen-specific cut-point values previously determined in a validation study were applied to 15 H. pylori proteins used as antigens (UreA, Catalase, GroEL, NapA, CagA, CagM, Cagδ, HP0231, VacA, HpaA, Cad, HyuA, Omp, HcpC, HP0305), using a bridging panel of 78 previously characterized sera containing 38 H. pylori-negative sera and 40 H. pylori-positive sera. H. pylori sero-positivity was defined as sero-positivity to four or more H. pylori proteins, as this cutoff has shown good agreement with commercial serological assay classification [16].
Neighborhood measures of socio-economic status
To determine neighborhood measures of SES, participant home addresses, provided at study entry, were geocoded and linked to the US Census 2000 block group. Briefly, most (80.2 %) addresses were geocoded with ArcMap 9.3.1 (ESRI, Redlands, CA), using either the ESRI StreetMap USA [17] or TIGER/Line 2008 shapefiles as the reference database [18]. For those addresses that did not geocode by either method, an online geocoding vendor was used to geocode the street address, if possible, or the better of the delivery-weighted centroid of the ZIP+4, ZIP+2, or 5-digit ZIP code. All remaining addresses (9.4 %) were manually processed using the above resources as well as Google Earth. Nearly all (97.6 %) of the non-post office box addresses of SCCS participants were geocoded to the street level. Geocoded participant addresses were then spatially joined to Census 2000 block groups, which are generally comprised of 600–3,000 individuals, and are the lowest level of the census geographic hierarchy for which demographic data are routinely available. The ten block group-level characteristics chosen from the Census to represent the SES of the neighborhood include the measures of the following: income or wealth (median household income, percentage of households in poverty, median value of owner-occupied housing units, and percentage of housing units that are owner-occupied); education (percentage of adults who completed high school and percentage of adults who completed college); occupation or employment (percentage employed and percentage employed in executive, managerial, or professional occupations); and crowding (population density and mean number of occupants per room).
Data analysis
Of the 792 individuals selected for the initial SCCS bio-marker studies, 665 (84.0 %) were included in the present analyses. Reasons for exclusion include the following: depletion of the available serum from previous assays performed (n = 77); unusable samples due to serum handling issues (n = 3); testing sero-positive to CagA but sero-negative to H. pylori (n = 8); missing information on antibiotic use (n = 3); ancestry estimates highly discordant with self-reported race (n = 23); and missing block group-level census information (n = 13). All individuals were classified into the ancestry categories of white (no or minimal African ancestry: 0–17 % African); low African ancestry (50–<85 % African); medium African ancestry (85–<95 % African); and high African ancestry (≥95 % African), based on previously utilized cut-points [19].
To assess differences in demographic and lifestyle characteristics between individuals of differing race or ancestry categories, crude linear regression was used for all ten continuous neighborhood-level measures as well as the individual-level continuous variables of age and years residing in current home. The Mantel–Haenszel chi-square test was used for the remaining individual-level categorical variables.
Prevalence odds ratios (ORs) and 95 % confidence intervals (CIs) for CagA-specific infection with H. pylori were calculated using polytomous logistic regression for the ten neighborhood-level measures of SES, in tertiles based on the distribution in the study population as a whole. The outcomes were categorized as follows: sero-negativity to H. pylori and CagA (H. pylori−, CagA−), sero-positivity to H. pylori but not to CagA (H. pylori+, CagA−), and sero-positivity to both H. pylori and CagA (H. pylori+, CagA+). Adjusted models included variables representing age (continuous); African ancestry (white and low/medium/high African ancestry); sex; marital status (married/single/other); individual level of education (less than high school/high school or GED/more than high school); antibiotic use in the past year (yes/no); and duration, in years, of residency in the current home as reported during the baseline interview (continuous). Individual level of income was not included as it was not significantly associated with H. pylori status, and its inclusion in the models did not change the main results by 10 % or more.
In separate models including each significant neighborhood-level characteristic and in a final model including all significant neighborhood-level characteristics, we examined whether adding these factors attenuated the association between African ancestry and H. pylori sero-positivity.
As it has been suggested that population-level characteristics may modify the relationship between individual-level characteristics and the likelihood of disease [20], we investigated whether neighborhood-level educational status acted as an effect modifier on the strong association between individual-level educational status and CagA sero-positivity. To do so, new variables were created to group each participant into one of the six categories, based on individual-level educational achievement (less than high school/high school or more) and neighborhood level of percent adults who had completed high school (in tertiles: <62.7 %/62.7–<75.9 %/≥75.9 %). Prevalence ORs of CagA-specific H. pylori positivity for each of the six individual and neighborhood categories (with low individual-level education and low neighborhood-level education as the reference) were calculated using polytomous regression, adjusting for African ancestry, sex, age, and duration of residence in the current home (marital status and antibiotic use were not included in these models to preserve power, as their inclusion did not change the main results by >10 %). The associations between the six combined individual- and neighborhood-education-level categories and H. pylori status were also examined in models stratified by race.
Multilevel modeling techniques, such as the use of a random intercept, were not employed in this analysis because there was very little clustering of participants within block groups. The vast majority (88 %, n = 496) of the 556 block groups represented by the 665 individuals in this study had only a single participant, with 8 % (n = 48) of block groups having two participants, 3 % (n = 16) of block groups having three participants, 1 % (n = 5) of block groups having four participants, and only one block group having five participants.
Results
Level of African ancestry was significantly associated with individual marital status and education (Table 1). In particular, among African-Americans, the percentage of African ancestry was inversely associated with individual educational attainment. In univariate linear regression models, categories of African ancestry were also significantly associated with lower levels of eight of the ten neighborhood socio-economic characteristics in the realms of income/wealth, education, occupation/employment, and crowding (Table 1).
Table 1.
Demographic and lifestyle characteristics, by race/ancestry status, of a sample of 665 participants in the Southern Community Cohort Study recruited from 12 southeastern states between 2002 and 2004
| White n = 329 |
African-American
|
|||
|---|---|---|---|---|
| Level of African ancestry:
| ||||
| Low (50–84.9 %) n = 54 |
Med. (85–94.9 %) n = 94 |
High (≥95 %) n = 188 |
||
| Individual-level measures | ||||
| Age (years, mean)* | 53.4 | 48.9 | 52.2 | 52.1 |
| Women (%) | 52.9 | 42.6 | 48.9 | 56.9 |
| Marital status (%)** | ||||
| Married | 47.4 | 24.1 | 16.0 | 30.3 |
| Separated/divorced | 33.1 | 48.2 | 40.4 | 35.6 |
| Widowed | 9.4 | 1.9 | 12.8 | 11.7 |
| Single/never married | 10.0 | 25.9 | 30.9 | 22.3 |
| Education (%)** | ||||
| Less than high school | 26.1 | 11.1 | 37.2 | 39.9 |
| High school or GED | 41.0 | 40.7 | 38.3 | 42.6 |
| More than high school | 32.8 | 48.2 | 24.5 | 17.6 |
| Currently working (%) | 33.0 | 55.6 | 35.9 | 41.6 |
| Household income ($, %) | ||||
| <15,000 | 61.2 | 45.3 | 68.8 | 61.5 |
| ≥15,000–<25,000 | 18.9 | 32.1 | 19.4 | 23.1 |
| ≥25,000 | 20.2 | 22.6 | 11.8 | 13.4 |
| Household size (%) | ||||
| 1–2 | 65.7 | 61.1 | 56.4 | 59.6 |
| 3–4 | 25.2 | 25.9 | 29.8 | 29.3 |
| 5+ | 9.1 | 13.0 | 13.8 | 11.2 |
| Antibiotic prescription in the past year (%)* | 54.7 | 38.9 | 47.9 | 41.5 |
| Years in current home (mean)* | 7.9 | 8.0 | 11.8 | 10.1 |
| Neighborhood-level measures (mean) | ||||
| Income/wealth | ||||
| Median household income ($)** | 33,817 | 31,479 | 25,634 | 25,308 |
| Percent poverty** | 18.3 | 24.4 | 30.8 | 31.9 |
| Median value of owner-occupied housing units($)* | 80,217 | 72,087 | 78,930 | 63,572 |
| Percent of housing units that are owner-occupied** | 67.3 | 56.0 | 51.9 | 54.3 |
| Education (individuals ≥25 years old) | ||||
| Percent completed high school** | 72.4 | 70.7 | 66.5 | 64.9 |
| Percent completed college | 15.0 | 16.2 | 15.2 | 12.9 |
| Occupation/Employment (individuals ≥16 years old) | ||||
| Percent employed* | 58.4 | 59.8 | 56.1 | 54.5 |
| Percent employed in executive, managerial or professional occupations | 24.7 | 22.6 | 22.8 | 22.1 |
| Crowding | ||||
| Population density (all ages within block group)** | 870 | 1,552 | 1,318 | 1,285 |
| Mean number of occupants per room** | 0.42 | 0.46 | 0.46 | 0.47 |
p <0.05 comparing individuals across the four race/African ancestry categories
p <0.0001 comparing individuals across the four race/African ancestry categories
In crude analyses, three of the ten neighborhood-level socio-economic characteristics (median value of owner-occupied housing units, percent of adults who completed high school, and percent of adults who completed college) were significantly and inversely associated with sero-prevalence of CagA− strains of H. pylori, whereas all of the ten neighborhood-level measures except population density were associated with the more virulent CagA+ strains (Table 2). In the final multivariate models, which included adjustment for African ancestry and individual-level socio-economic characteristics, three neighborhood-level socio-economic characteristics were still significantly associated with H. pylori+ CagA+ status: median value of owner-occupied housing units (comparing highest to lowest tertile, prevalence OR: 0.56, 95 % CI: 0.32–0.99); percent of adults who completed high school (comparing highest to lowest tertile, prevalence OR: 0.47, 95 % CI: 0.26, 0.85); and percent employed (comparing middle to lowest tertile, prevalence OR: 0.55, 95 % CI; 0.31, 0.96) (Table 2). In race-stratified analyses, similar results were observed among African-Americans and whites separately.
Table 2.
Prevalence odds ratios (OR) for antibodies to H. pylori and CagA in relation to neighborhood-level characteristics among a sample of 665 participants in the Southern Community Cohort Study recruited from 12 southeastern states between 2002 and 2004
| Neighborhood-level measures |
H. pylori +, CagA−
|
H. pylori +, CagA+
|
||||||
|---|---|---|---|---|---|---|---|---|
| Crude OR | 95 % CI | Adjusted ORa | 95 % CI | Crude OR | 95 % CI | Adjusted ORa | 95 % CI | |
| Income/wealth | ||||||||
| Median household income | ||||||||
| <$23,567 | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| $23,567–<$33,500 | 0.73 | 0.42, 1.28 | 0.82 | 0.45, 1.48 | 0.45 | 0.27, 0.73 | 0.79 | 0.45, 1.39 |
| ≥$33,500 | 1.04 | 0.60, 1.79 | 1.41 | 0.78, 2.53 | 0.35 | 0.21, 0.58 | 0.79 | 0.44, 1.40 |
| Percent poverty | ||||||||
| <14.0 % | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| 14.0–<29.1 % | 1.19 | 0.73, 1.93 | 1.08 | 0.65, 1.79 | 1.62 | 1.01, 2.60 | 1.07 | 0.63, 1.82 |
| ≥29.1 % | 1.47 | 0.85, 2.55 | 1.17 | 0.64, 2.13 | 3.87 | 2.33, 6.44 | 1.59 | 0.89, 2.86 |
| Median value of owner-occupied housing units | ||||||||
| <$52,000 | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| $52,000–<$76,000 | 0.62 | 0.36, 1.08 | 0.70 | 0.39, 1.25 | 0.53 | 0.32, 0.88 | 0.73 | 0.41, 1.29 |
| ≥$76,000 | 0.55 | 0.32, 0.95 | 0.73 | 0.41, 1.28 | 0.35 | 0.21, 0.58 | 0.56 | 0.32, 0.99 |
| Percent of housing units that are owner-occupied | ||||||||
| <50.3 % | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| 50.3–<77.9 % | 1.27 | 0.74, 2.19 | 1.36 | 0.76, 2.42 | 0.72 | 0.44, 1.19 | 1.01 | 0.57, 1.77 |
| ≥77.9 % | 0.93 | 0.55, 1.57 | 1.05 | 0.59, 1.87 | 0.46 | 0.29, 0.75 | 0.91 | 0.51, 1.60 |
| Education | ||||||||
| Percent adults ≥25 years completed high school | ||||||||
| <62.7 % | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| 62.7–<75.9 % | 0.49 | 0.27, 0.87 | 0.52 | 0.28, 0.95 | 0.41 | 0.24, 0.70 | 0.72 | 0.40, 1.30 |
| ≥75.9 % | 0.40 | 0.23, 0.69 | 0.51 | 0.28, 0.91 | 0.25 | 0.15, 0.42 | 0.47 | 0.26, 0.85 |
| Percent adults ≥25 years completed college | ||||||||
| <7.2 % | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| 7.2–<15.5 % | 0.42 | 0.24, 0.72 | 0.45 | 0.26, 0.80 | 0.54 | 0.33, 0.90 | 0.69 | 0.39, 1.21 |
| ≥15.5 % | 0.45 | 0.26, 0.77 | 0.58 | 0.33, 1.01 | 0.51 | 0.31, 0.84 | 0.63 | 0.36, 1.11 |
| Occupation/employment | ||||||||
| Percent employed | ||||||||
| <53.0 % | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| 53.0–<62.6 % | 0.77 | 0.45, 1.34 | 0.73 | 0.42, 1.29 | 0.51 | 0.31, 0.85 | 0.55 | 0.31, 0.96 |
| ≥ 62.6 % | 0.64 | 0.37, 1.09 | 0.75 | 0.43, 1.32 | 0.35 | 0.21, 0.57 | 0.59 | 0.34, 1.03 |
| Percent employed in executive, managerial, or professional occupations | ||||||||
| <17.2 % | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| 17.2–<26.7 % | 0.60 | 0.36, 1.02 | 0.71 | 0.41, 1.23 | 0.58 | 0.35, 0.94 | 0.89 | 0.51, 1.55 |
| ≥26.7 % | 0.61 | 0.36, 1.03 | 0.80 | 0.45, 1.40 | 0.64 | 0.39, 1.04 | 0.95 | 0.54, 1.66 |
| Crowding | ||||||||
| Population density | ||||||||
| <306 | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| 306–<1,370 | 0.90 | 0.54, 1.49 | 0.80 | 0.47, 1.38 | 1.29 | 0.80, 2.10 | 0.88 | 0.51, 1.54 |
| ≥1,370 | 0.67 | 0.40, 1.13 | 0.58 | 0.33, 1.03 | 1.41 | 0.87, 2.26 | 0.81 | 0.46, 1.42 |
| Mean number of occupants per room | ||||||||
| <0.408 | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref | 1.00 | Ref |
| 0.408–<0.466 | 0.84 | 0.51, 1.39 | 0.78 | 0.47, 1.31 | 1.23 | 0.77, 1.95 | 1.13 | 0.67, 1.91 |
| ≥0.466 | 1.48 | 0.87, 2.54 | 1.24 | 0.70, 2.18 | 2.43 | 1.47, 4.03 | 1.47 | 0.83, 2.60 |
Adjusted for individual-level measures of African ancestry, sex, marital status, and education, as well as participant age, whether antibiotics were used in the past year, and duration of residency in the current home as reported during the baseline interview. All ORs relative to persons negative for H. pylori
The univariate association between African ancestry and sero-positivity for the CagA+ strain of H. pylori showed a strong (p <0.0001) and monotonic increase in sero-positivity across categories, with those in the highest category (≥95 % African ancestry) having a tenfold increase in risk compared to whites [OR: 10.05, 95 % CI: 5.70, 17.73] (Table 3). Adjustment for neither the individual-level covariates nor the three significant neighborhood-level covariates resulted in much attenuation of these associations, with African-Americans of low, medium, and high African ancestry maintaining two-, seven-, and ninefold increased odds of CagA+ sero-positivity, respectively, compared to whites. Separate models including each of the other seven non-significant neighborhood-level variables similarly found that they do not act as confounders of the relationship between African ancestry and CagA+ sero-positivity (data not shown). In a fully adjusted model including all significant neighborhood-level characteristics (i.e., model 6 from Table 3) restricted to African-Americans, the prevalence OR for CagA+ sero-positivity for those with high vs. low African ancestry was 3.92 (95 % CI; 1.59, 9.70).
Table 3.
Prevalence odds ratios (OR) for antibodies to CagA+ H. pylori in relation to race/ancestry, among a sample of 665 participants in the Southern Community Cohort Study recruited from 12 southeastern states between 2002 and 2004
| White | African-American
|
|||||||
|---|---|---|---|---|---|---|---|---|
| Level of African ancestry
| ||||||||
| Low (50–84.9 %) | Medium (85–94.9 %) | High (≥95 %) | ||||||
|
|
|
|
|
|||||
| OR | 95 % CI | OR | 95 % CI | OR | 95 % CI | OR | 95 % CI | |
| Model 1: crude relationship | 1.00 | Reference | 1.95 | 0.96, 3.97 | 8.03 | 3.88, 16.58 | 10.05 | 5.70, 17.73 |
| Model 2: adjusted for individual-level factorsa only | 1.00 | Reference | 2.38 | 1.13, 5.00 | 7.95 | 3.69, 17.11 | 10.01 | 5.47, 18.34 |
| Model 3: adjusted for individual-level factorsa plus neighborhood-level value of owner-occupied housing units | 1.00 | Reference | 2.27 | 1.07, 4.78 | 7.58 | 3.51, 16.37 | 9.52 | 5.18, 17.48 |
| Model 4: adjusted for individual-level factorsa plus neighborhood-level education | 1.00 | Reference | 2.32 | 1.10, 4.90 | 7.43 | 3.43, 16.11 | 9.02 | 4.87, 16.70 |
| Model 5: adjusted for individual-level factorsa plus neighborhood-level employment | 1.00 | Reference | 2.35 | 1.11, 4.95 | 7.61 | 3.53, 16.40 | 9.56 | 5.18, 17.65 |
| Model 6: adjusted for individual-level factorsa plus the neighborhood-level factors included in models 3–5 above | 1.00 | Reference | 2.28 | 1.08, 4.83 | 7.38 | 3.40, 16.03 | 9.29 | 4.99, 17.29 |
Individual-level factors adjusted for included sex, marital status, and education, as well as participant age, whether antibiotics were used in the past year, and duration of residency in the current home as reported during the baseline interview
When examining the joint effect of individual- and neighborhood-level education, the individual-level index was more strongly associated with sero-positivity for CagA+ H. pylori than the neighborhood-level index (Table 4). Furthermore, the neighborhood-level effect seemed evident only among those with at least a high school education, although a significant interaction between individual-level education status and neighborhood-level education status was not found (Pinteraction = 0.27). When stratified by race, generally similar results to Table 4 were found for both whites and African-Americans (data not shown).
Table 4.
Prevalence odds ratios (OR) for antibodies to CagA+ H. pylori in relation to joint individual- and neighborhood-level education, among a sample of 665 participants in the Southern Community Cohort Study recruited from 12 southeastern states between 2002 and 2004
| Neighborhood education level | Individual education level
|
|||
|---|---|---|---|---|
| Less than high school
|
High school/GED or more
|
|||
| ORa | 95 % CI | ORa | 95 % CI | |
| Percent adults ≥25 years completed high school | ||||
| <62.7 % | 1.00 | Reference | 0.37 | 0.14, 0.97 |
| 62.7–<75.9 % | 0.83 | 0.27, 2.61 | 0.25 | 0.10, 0.63 |
| ≥75.9 % | 1.08 | 0.24, 4.84 | 0.16 | 0.07, 0.40 |
Adjusted for African ancestry and sex, as well as participant age and duration of residency in the current home as reported during the baseline interview
Discussion
We found that individuals living in neighborhoods with higher house values and where proportionately more adults had a high school education and were employed had lower odds of being sero-positive for CagA+ H. pylori than those living in neighborhoods with lower levels of these SES measures. These associations with neighborhood-level socio-economic characteristics were observed above and beyond that of the individual’s own risk factors for infection with CagA+ H. pylori and were present for both whites and African-Americans. Of import is that accounting for these individual- and neighborhood-level measures of SES did not diminish nor explain the strong association between African ancestry and sero-prevalence of CagA+ H. pylori, suggesting that it is not mainly through SES that African-Americans are at greater odds of CagA+ H. pylori sero-positivity.
Previous studies worldwide have consistently found that individual measures of low SES are associated with greater odds of H. pylori infection [5, 21–27], but we are aware of only one study [11] that examined the effect of neighborhood-level socio-economic factors on H. pylori sero-prevalence. In this study among Chinese, in which 66 % were sero-positive for H. pylori, an increasing risk was observed with decreasing village education level, but no association was found with individual education or income level. The increased odds for those living in a village of medium or low level of education vs. a high education level was similar to ours (OR: 1.7, 95 % CI: 1.4–2.1, and OR: 2.4, 95 % CI: 2.0–2.9, respectively). The authors suggested this association could be due to a shared water source, although no analyses accounting for the source of drinking water were presented. To our knowledge, no previous studies have evaluated individual- or neighborhood-level measures of SES separately by H. pylori type, which is potentially critical given our strong findings regarding the virulent CagA+ strain.
Several previous studies have evaluated the association of neighborhood-level SES with markers of immune response to inflammation, potential indicators of infections, and findings from these studies provide some support to the results obtained in the present study. A study of adults in southwest Pennsylvania found that, independent of individual-level socio-economic risk factors, individuals living in communities of lower SES had higher circulating levels of interleukin-6 and C-reactive protein than individuals in more affluent communities [28]. The authors suggested that the association may be related to differential access to health resources, including exercise and medical facilities, as well as to shared environmental factors that either encourage or discourage health-related behaviors, all of which may relate to inflammation susceptibility. Another US study found significant associations between both low individual- and neighborhood-level measures of SES, and increased levels of fibrinogen and white blood cell count and interestingly found weaker, less consistent results among African-Americans as compared to whites [29]. It had previously been noted however that in that cohort, African-Americans lived in neighborhoods of much lower SES than whites, so that the most advantaged of the African-American neighborhoods were comparable to the least advantaged white neighborhoods [9].
One of the primary strengths of the SCCS is that African-Americans and whites were drawn primarily from the same low-income populations, increasing the power and validity to make comparisons within the same range of socio-economic characteristics. In fact, while the neighborhoods represented by the SCCS are in a relatively narrow, low end of the SES scale, it is a strength of the present study that this hard-to-reach population was able to be studied and attests to the power of the association between neighborhood-level SES and H. pylori sero-positivity that significant results were found. However, a weakness of the present study is our limited ability to assess all relevant correlates of neighborhood-level SES. We were only able to use those variables collected and presented by the US Census to describe neighborhood-level SES. Thus, other potential factors related to neighborhood SES that might also explain differing H. pylori prevalence—such as sanitation practices, water source, refrigeration, and cooking and eating habits—were not accounted for, although these factors are most likely also associated with individual- and neighborhood-level education and income categories. It has also been suggested that poor oral health is associated with gastric cancer [30] and precancerous lesions [31], although the association with H. pylori infection specifically has not been found to be significant above and beyond individual SES parameters [32, 33]. Unfortunately, we did not have sufficient data on oral health to address this question in our population. Another limiting factor is that chronic H. pylori infection generally begins in childhood, and the participants in this study were on average around 50 years old, having lived in their current address for approximately 10 years. Thus, we are measuring adult, and not childhood, SES. While studies have generally found positive correlations between childhood and adult SES, it does appear that this association over the lifecourse differs by race, whereby among those living in poor neighborhoods, African-Americans are more likely than whites to have lived there for a longer time span, and among those who leave high-poverty neighborhoods, African-Americans are more likely to reenter them [34]. In the present study, the association with measures of lower levels of neighborhood SES was found not just for African-American individuals compared to whites, but also for individuals of increasing level of African ancestry. Whether the differences in associations of neighborhood SES over the life-course also differ by percent African ancestry among African-Americans is unknown.
Our finding that an increasing percentage of African ancestry is associated with increased odds of infection with CagA+ H. pylori, independent of significant individual- and neighborhood-level measures of SES, suggests the possibility that the biology of the bacteria and/or the host is responsible for the association. In fact, conservation of H. pylori genotype by race in the United States has been found before, and findings indicate that transmission occurs primarily through the family; thus, genotypes of the bacteria could be conserved through multiple human generations [35, 36]. In addition, the host genetics in families will be more similar. It is possible, then, that individuals of greater percentage of African ancestry have also had greater conservation of their H. pylori genotype. Others have suggested that African-American race is associated with higher H. pylori sero-prevalence above and beyond socioeconomic risk factors due to a shorter “generational distance” than whites from being very economically disadvantaged [37]. This association may also apply to increasing percentage of African ancestry. Finally, in terms of host factors, African-Americans, particularly those of greater African ancestry, may be more susceptible to initial and/or chronic H. pylori infection. Additionally, increasing African ancestry in our study was strongly associated not only with the prevalence (yes/no) of sero-positivity but also with antibody titer level among those sero-positive for CagA+ H. pylori (data not shown), potentially indicating a greater inflammatory response.
Conclusions
Beyond individual-level socio-economic factors, neighborhood levels of education, employment, and house values are associated with infection with CagA+ H. pylori, highlighting high-risk areas for targeted prevention and screening efforts. Furthermore, the strong relationship between African-American race, particularly individuals of a high level of African ancestry, and CagA+ H. pylori sero-prevalence is not diminished nor explained by individual- or neighborhood-level socio-economic characteristics, suggesting the possibility of a biological basis—of the bacteria and/or the host—for the association between African ancestry and H. pylori infection.
Acknowledgments
We would like to thank Mr. Michael Mumma for his oversight of the geocoding process and derivation of neighborhood-level variables for the SCCS. The Southern Community Cohort Study is funded by a grant from the National Cancer Institute (R01 CA092447) and by supplemental American Recovery and Reinvestment Act funding (3R01 CA092447-08S1). Analysis of the samples in this study was funded by an American Cancer Society—Institutional Research Grant to Vanderbilt University (ACS–IRG-58-009-50). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or National Institutes of Health.
Contributor Information
Meira Epplein, Email: meira.epplein@vanderbilt.edu, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center and Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, 2525 West End Avenue, 6th floor, Nashville, TN 37203-1738, USA.
Sarah S. Cohen, International Epidemiology Institute, Rockville, MD, USA
Jennifer S. Sonderman, International Epidemiology Institute, Rockville, MD, USA
Wei Zheng, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center and Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, 2525 West End Avenue, 6th floor, Nashville, TN 37203-1738, USA.
Scott M. Williams, Division of Human Genomics, Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, TN, USA
William J. Blot, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center and Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, 2525 West End Avenue, 6th floor, Nashville, TN 37203-1738, USA. International Epidemiology Institute, Rockville, MD, USA
Lisa B. Signorello, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center and Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, 2525 West End Avenue, 6th floor, Nashville, TN 37203-1738, USA. International Epidemiology Institute, Rockville, MD, USA
References
- 1.Marshall BJ. The 1995 Albert Lasker Medical Research Award. Helicobacter pylori. The etiologic agent for peptic ulcer. JAMA. 1995;274(13):1064–1066. doi: 10.1001/jama.274.13.1064. [DOI] [PubMed] [Google Scholar]
- 2.Brown LM. Helicobacter pylori: epidemiology and routes of transmission. Epidemiol Rev. 2000;22(2):283–297. doi: 10.1093/oxfordjournals.epirev.a018040. [DOI] [PubMed] [Google Scholar]
- 3.McQuillan GM, Kruszon-Moran D, Kottiri BJ, Curtin LR, Lucas JW, Kington RS. Racial and ethnic differences in the sero-prevalence of 6 infectious diseases in the United States: data from NHANES III, 1988–1994. Am J Public Health. 2004;94(11):1952–1958. doi: 10.2105/ajph.94.11.1952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Epplein M, Signorello LB, Zheng W, Peek RM, Jr, Michel A, Williams SM, Pawlita M, Correa P, Cai Q, Blot WJ. Race, African ancestry, and Helicobacter pylori infection in a low-income United States population. Cancer Epi Biomarkers Prev. 2011;20(5):826–834. doi: 10.1158/1055-9965.EPI-10-1258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Moayyedi P, Axon AT, Feltbower R, Duffett S, Crocombe W, Braunholtz D, Richards ID, Dowell AC, Forman D. Relation of adult lifestyle and socioeconomic factors to the prevalence of Helicobacter pylori infection. Int J Epidemiol. 2002;31(3):624–631. doi: 10.1093/ije/31.3.624. [DOI] [PubMed] [Google Scholar]
- 6.Aiello AE, Diez-Roux A, Noone AM, Ranjit N, Cushman M, Tsai MY, Szklo M. Socioeconomic and psychosocial gradients in cardiovascular pathogen burden and immune response: the multi-ethnic study of atherosclerosis. Brain Behav Immun. 2009;23(5):663–671. doi: 10.1016/j.bbi.2008.12.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zajacova A, Dowd JB, Aiello AE. Socioeconomic and race/ethnic patterns in persistent infection burden among US adults. J Gerontol A Biol Sci Med Sci. 2009;64(2):272–279. doi: 10.1093/gerona/gln012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Diez Roux AV, Merkin SS, Arnett D, Chambless L, Massing M, Nieto FJ, Sorlie P, Szklo M, Tyroler HA, Watson RL. Neighborhood of residence and incidence of coronary heart disease. N Engl J Med. 2001;345(2):99–106. doi: 10.1056/NEJM200107123450205. [DOI] [PubMed] [Google Scholar]
- 9.Borrell LN, Diez Roux AV, Rose K, Catellier D, Clark BL. Neighbourhood characteristics and mortality in the Atherosclerosis Risk in Communities Study. Int J Epidemiol. 2004;33(2):398–407. doi: 10.1093/ije/dyh063. [DOI] [PubMed] [Google Scholar]
- 10.Leal C, Bean K, Thomas F, Chaix B. Are associations between neighborhood socioeconomic characteristics and body mass index or waist circumference based on model extrapolations? Epidemiology. 2011;22(5):694–703. doi: 10.1097/EDE.0b013e3182257784. [DOI] [PubMed] [Google Scholar]
- 11.Brown LM, Thomas TL, Ma JL, Chang YS, You WC, Liu WD, Zhang L, Pee D, Gail MH. Helicobacter pylori infection in rural China: demographic, lifestyle and environmental factors. Int J Epidemiol. 2002;31(3):638–645. doi: 10.1093/ije/31.3.638. [DOI] [PubMed] [Google Scholar]
- 12.Signorello LB, Hargreaves MK, Steinwandel MD, Zheng W, Cai Q, Schlundt DG, Buchowski MS, Arnold CW, McLaughlin JK, Blot WJ. Southern community cohort study: establishing a cohort to investigate health disparities. J Natl Med Assoc. 2005;97(7):972–979. [PMC free article] [PubMed] [Google Scholar]
- 13.Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics. 2000;155(2):945–959. doi: 10.1093/genetics/155.2.945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Waterboer T, Sehr P, Michael KM, Franceschi S, Nieland JD, Joos TO, Templin MF, Pawlita M. Multiplex human papillomavirus serology based on in situ-purified glutathione S-transferase fusion proteins. Clin Chem. 2005;51(10):1845–1853. doi: 10.1373/clinchem.2005.052381. [DOI] [PubMed] [Google Scholar]
- 15.Waterboer T, Sehr P, Pawlita M. Suppression of non-specific binding in serological Luminex assays. J Immunol Methods. 2006;309(1–2):200–204. doi: 10.1016/j.jim.2005.11.008. [DOI] [PubMed] [Google Scholar]
- 16.Michel A, Waterboer T, Kist M, Pawlita M. Helicobacter pylori multiplex serology. Helicobacter. 2009;14(6):525–535. doi: 10.1111/j.1523-5378.2009.00723.x. [DOI] [PubMed] [Google Scholar]
- 17.Esri. StreetMap USA in Esri Data and Maps [machine-readable datafiles] 2005. [Google Scholar]
- 18.US Census Bureau. 2008 TIGER/Line Shapefiles [machine-readable data files] 2008. [Google Scholar]
- 19.Signorello LB, Williams SM, Zheng W, Smith JR, Long J, Cai Q, Hargreaves MK, Hollis BW, Blot WJ. Blood vitamin D levels in relation to genetic estimation of African ancestry. Cancer Epidemiol Biomarkers Prev. 2010;19(9):2325–2331. doi: 10.1158/1055-9965.EPI-10-0482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Diez Roux AV, Aiello AE. Multilevel analysis of infectious diseases. J Infect Dis. 2005;191(Suppl 1):S25–S33. doi: 10.1086/425288. [DOI] [PubMed] [Google Scholar]
- 21.Rosenstock SJ, Andersen LP, Rosenstock CV, Bonnevie O, Jorgensen T. Socioeconomic factors in Helicobacter pylori infection among Danish adults. Am J Public Health. 1996;86(11):1539–1544. doi: 10.2105/ajph.86.11.1539. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Russo A, Eboli M, Pizzetti P, Di Felice G, Ravagnani F, Spinelli P, Hotz AM, Notti P, Maconi G, Franceschi S, Ferrari D, Bertario L. Determinants of Helicobacter pylori seroprevalence among Italian blood donors. Eur J Gastroenterol Hepatol. 1999;11(8):867–873. doi: 10.1097/00042737-199908000-00010. [DOI] [PubMed] [Google Scholar]
- 23.Santos IS, Boccio J, Santos AS, Valle NC, Halal CS, Bachilli MC, Lopes RD. Prevalence of Helicobacter pylori infection and associated factors among adults in Southern Brazil: a population-based cross-sectional study. BMC Public Health. 2005;5:118. doi: 10.1186/1471-2458-5-118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Naja F, Kreiger N, Sullivan T. Helicobacter pylori infection in Ontario: prevalence and risk factors. Can J Gastroenterol. 2007;21(8):501–506. doi: 10.1155/2007/462804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Naous A, Al-Tannir M, Naja Z, Ziade F, El-Rajab M. Fecoprevalence and determinants of Helicobacter pylori infection among asymptomatic children in Lebanon. J Med Liban. 2007;55(3):138–144. [PubMed] [Google Scholar]
- 26.Zaterka S, Eisig JN, Chinzon D, Rothstein W. Factors related to Helicobacter pylori prevalence in an adult population in Brazil. Helicobacter. 2007;12(1):82–88. doi: 10.1111/j.1523-5378.2007.00474.x. [DOI] [PubMed] [Google Scholar]
- 27.Shi R, Xu S, Zhang H, Ding Y, Sun G, Huang X, Chen X, Li X, Yan Z, Zhang G. Prevalence and risk factors for Helicobacter pylori infection in Chinese populations. Helicobacter. 2008;13(2):157–165. doi: 10.1111/j.1523-5378.2008.00586.x. [DOI] [PubMed] [Google Scholar]
- 28.Petersen KL, Marsland AL, Flory J, Votruba-Drzal E, Muldoon MF, Manuck SB. Community socioeconomic status is associated with circulating interleukin-6 and C-reactive protein. Psychosom Med. 2008;70(6):646–652. doi: 10.1097/PSY.0b013e31817b8ee4. [DOI] [PubMed] [Google Scholar]
- 29.Pollitt RA, Kaufman JS, Rose KM, Diez-Roux AV, Zeng D, Heiss G. Cumulative life course and adult socioeconomic status and markers of inflammation in adulthood. J Epidemiol Community Health. 2008;62(6):484–491. doi: 10.1136/jech.2006.054106. [DOI] [PubMed] [Google Scholar]
- 30.Abnet CC, Kamangar F, Dawsey SM, Stolzenberg-Solomon RZ, Albanes D, Pietinen P, Virtamo J, Taylor PR. Tooth loss is associated with increased risk of gastric non-cardia adenocarcinoma in a cohort of Finnish smokers. Scand J Gastroenterol. 2005;40(6):681–687. doi: 10.1080/00365520510015430. [DOI] [PubMed] [Google Scholar]
- 31.Salazar CR, Francois F, Li Y, Corby P, Hays R, Leung C, Bedi S, Segers S, Queiroz E, Sun J, Wang B, Ho H, Craig R, Cruz GD, Blaser MJ, Perez–Perez G, Hayes RB, Dasanayake A, Pei Z, Chen Y. Association between oral health and gastric pre-cancerous lesions. Carcinogenesis. 2012;33(2):399–403. doi: 10.1093/carcin/bgr284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Pearce MS, Steele JG, Campbell DI, Thomas JE. Tooth loss and Helicobacter pylori seropositivity: the Newcastle Thousand Families Cohort Study at age 49–51 years. Helicobacter. 2005;10(1):90–94. doi: 10.1111/j.1523-5378.2005.00296.x. [DOI] [PubMed] [Google Scholar]
- 33.Stolzenberg-Solomon RZ, Dodd KW, Blaser MJ, Virtamo J, Taylor PR, Albanes D. Tooth loss, pancreatic cancer, and Helicobacter pylori. Am J Clin Nutr. 2003;78(1):176–181. doi: 10.1093/ajcn/78.1.176. [DOI] [PubMed] [Google Scholar]
- 34.Quillian L. How long are exposures to poor neighborhoods? The long-term dynamics of entry and exit from poor neighborhoods. Popul Res Policy Rev. 2003;22(3):221–249. [Google Scholar]
- 35.McNulty SL, Mole BM, Dailidiene D, Segal I, Ally R, Mistry R, Secka O, Adegbola RA, Thomas JE, Lenarcic EM, Peek RM, Jr, Berg DE, Forsyth MH. Novel 180- and 480-base-pair insertions in African and African-American strains of Helicobacter pylori. J Clin Microbiol. 2004;42(12):5658–5663. doi: 10.1128/JCM.42.12.5658-5663.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Yamaoka Y, Reddy R, Graham DY. Helicobacter pylori virulence factor genotypes in children in the United States: clues about genotype and outcome relationships. J Clin Microbiol. 2010;48(7):2550–2551. doi: 10.1128/JCM.00114-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Malaty HM, Evans DG, Evans DJ, Jr, Graham DY. Helicobacter pylori in Hispanics: comparison with blacks and whites of similar age and socioeconomic class. Gastroenterology. 1992;103(3):813–816. doi: 10.1016/0016-5085(92)90011-m. [DOI] [PubMed] [Google Scholar]