Inverse Associations of Helicobacter pylori With Asthma and Allergy

Background  Acquisition of Helicobacter pylori, which predominantly occurs before age 10 years, may reduce risks of asthma and allergy.

Methods  We evaluated the associations of H pylori status with history of asthma and allergy and with skin sensitization using data from 7663 adults in the Third National Health and Nutrition Examination Survey. Adjusted odds ratios (ORs) for currently and ever having asthma, allergic rhinitis, allergy symptoms in the previous year, and allergen-specific skin sensitization were computed comparing participants seropositive for cagA⁻ or cagA⁺ strains of H pylori with those without H pylori.

Results  The presence of cagA⁺ H pylori strains was inversely related to ever having asthma (OR, 0.79; 95% confidence interval [CI], 0.63-0.99), and the inverse association of cagA positivity with childhood-onset (age ≤15 years) asthma was stronger (OR, 0.63; 95% CI, 0.43-0.93) than that with adult-onset asthma (OR, 0.97; 95% CI, 0.72-1.32). Colonization with H pylori, especially with a cagA⁺ strain, was inversely associated with currently (OR, 0.77; 95% CI, 0.62-0.96) or ever (OR, 0.77; 95% CI, 0.62-0.94) having a diagnosis of allergic rhinitis, especially for childhood onset (OR, 0.55; 95% CI, 0.37-0.82). Consistent inverse associations were found between H pylori colonization and the presence of allergy symptoms in the previous year and sensitization to pollens and molds.

Conclusion  These observations support the hypothesis that childhood acquisition of H pylori is associated with reduced risks of asthma and allergy.

Asthma and related allergic disorders are becoming more common in the Western world.¹ The rapid increase in their incidence must reflect changing environmental exposures, but these remain largely undefined, despite active investigation.¹ Some asthma cases are related to gastroesophageal reflux disease (GERD)^2-4; GERD and its sequelae, Barrett esophagus and adenocarcinoma of the esophagus, also have been rising in incidence, a phenomenon occurring later in life but with secular trends similar to those for asthma.⁴

The gastric bacterium Helicobacter pylori is present in all human populations and is so ancient that its genetic variations can be used to trace migrations during the past 100 000 years.^5,6 In developing countries, virtually all adults harbor H pylori, but the prevalence is much lower in industrialized nations.⁷ This difference is due to a birth cohort phenomenon in which H pylori acquisition in industrialized countries has been diminishing with each succeeding generation, at least for the past 60 years.⁷ Since, when present, H pylori is the dominant species colonizing the stomach⁸ and is intimately linked to gastric physiology,⁹ especially the cagA⁺ strains that inject H pylori products into epithelial cells,¹⁰ this disappearance across the population represents a fundamental change in human microecology.⁹

Colonization with H pylori is associated with adenocarcinoma of the distal stomach and peptic ulcer disease.^11-13 In contrast, a substantial body of work^14-20 now shows an inverse relationship between the presence of H pylori, especially cagA⁺ strains, and GERD and its sequelae. These observations suggest that H pylori presence in the stomach protects against GERD and, therefore, could protect against GERD-related asthma. Using data from the Third National Health and Nutrition Examination Survey (NHANES III),²¹ herein we address the hypothesis that H pylori acquisition, which predominantly occurs before age 10 years,^22-24 is associated with reduced subsequent risks of asthma and allergy.

The NHANES III, the seventh health examination survey performed in the United States since 1960,²¹ was conducted between October 18, 1988, and October 15, 1994, in 2 phases, each of which comprised a national probability sample. In the NHANES III, 39 695 persons were studied; of those, 10 120 were adults (≥17 years old) sampled during the first phase (October 18, 1988, to October 24, 1991). All interviewed persons were invited to undergo a medical examination. The survey protocol was approved by the institutional review board of the Centers for Disease Control and Prevention. All the participants gave written informed consent.

Information on demographics and medical history of asthma, allergic rhinitis, and allergy symptoms was collected using in-person interviews.²¹ Participants were asked whether they had ever been diagnosed as having asthma or hay fever by a physician, the age they were first diagnosed, and whether they continued to have asthma or hay fever. They also were asked about allergy symptoms in the previous year (including wheezing; whistle in chest; wheezy/whistling chest sounds; stuffy, itchy, or runny nose; and watery, itchy eyes) that were unrelated to the common cold and about potential exposures that elicited allergic symptoms.

Examinees 20 years and older from phase 1 were tested for H pylori IgG antibodies in 1996 using the H pylori IgG enzyme-linked immunosorbent assay (Wampole Laboratories, Cranbury, NJ) and the CagA IgG enzyme-linked immunosorbent assay developed and standardized at Vanderbilt University, as described elsewhere.²⁵ On the basis of H pylori and cagA results, patients were classified into 3 groups: H pylori positive and cagA positive, H pylori positive and cagA negative, and H pylori negative and cagA negative, as described elsewhere.²⁶ The H pylori⁺cagA⁺ group included all persons with a positive cagA assay regardless of the results of the H pylori assay, based on the utility of the CagA antigen to detect true-positive responses in culture-positive persons in the face of negative or equivocal values in the H pylori serologic assay.²⁷ By definition, all persons in the H pylori⁻ group had negative CagA assays.

Allergy skin tests included evaluation of immediate hypersensitivity reactions to 10 licensed, commercially available, Food and Drug Administration–approved, standardized allergens.²¹ Prick-puncture allergy skin tests to the 10 allergens and positive (histamine) and negative (glycerinated diluent) controls were administered to a random half-sample of the 20- to 59-year-old adults, following a standard allergy testing protocol.²¹ Hypersensitivity reactions were evaluated 15 minutes after administering the allergens on an examinee's forearms; the length and width of each wheal and flare induced were measured. A skin test panel was considered valid if there was at least 1 mm of difference between the wheal diameters of the positive and negative controls.²⁸ An allergen-specific skin test response was considered positive if the skin test was valid, and the differences in the wheal's length and width for the allergen-specific test with the negative control were 3 mm or greater.

We included 7663 participants with valid answers on asthma history and valid serologic testing for H pylori. Analysis of skin reactivity was conducted in a subsample (n = 2385) of these participants who underwent allergy skin testing. We first described distributions of demographics, smoking status, body mass index, and outcomes of interest in the 3 groups with respect to H pylori status (H pylori⁺cagA⁺, H pylori⁺cagA⁻, and H pylori⁻). Unconditional logistic regression models were conducted to estimate adjusted odds ratios (ORs) for asthma and allergy symptoms and ORs for allergy skin reactivity. Multivariate analysis excluded participants with unknown information on any of the covariates. We did not use sampling weights in the analysis for the following reasons: (1) for the present analysis, the internal validity, that is, the relationships of H pylori status with asthma and allergy, is considered to be more important than generalization to the total US population; (2) representative population estimates were not needed; and (3) parameter estimators can vary considerably when the weights are highly variable, especially when a few individuals have very large sample weights.^29,30

All ORs were adjusted for sex, race/ethnicity, age, smoking status, body mass index, and educational attainment. A separate analysis was conducted to additionally adjust for country of birth and region in the United States; the ORs did not change appreciably, and, therefore, the results are not shown. Stratified analysis was conducted based on the median age (43 years) in the overall study population. Age also was adjusted in the stratified analysis by age to control for potential confounding due to differences in age strata. The statistical significance of interaction was determined based on P values of the cross-product terms of H pylori status with age. Participants who ever had asthma or allergic rhinitis were divided into those with childhood onset (age at onset ≤15 years) and those with older onset (>15 years) for comparison with participants who never had asthma or allergic rhinitis. Sensitivity analysis conducted using age 12 or 18 years as the cutoff point yielded results similar to using age 15 years(data not shown). All analyses were conducted using a software program (SAS 9.1.3; SAS Institute Inc, Cary, NC).

In the 7663 participants, H pylori status varied in relation to demographic and lifestyle factors, reflecting well-recognized trends.³¹ Participants in the H pylori–positive groups (H pylori⁺cagA⁻ and H pylori⁺cagA⁺) were more likely to be men, older, and born in Mexico compared with participants in the H pylori⁻ group (Table 1). Participants in the H pylori⁺cagA⁺ group were more likely to be non-Hispanic blacks compared with those in the H pylori⁺cagA⁻ and H pylori⁻ groups.

There was no overall association between the presence of either a cagA⁻ or a cagA⁺ strain of H pylori and current asthma status (Table 2). However, the association between colonization with H pylori⁺cagA⁺ strains and current asthma differed by age (interaction P = .02); H pylori⁺cagA⁺ strains were inversely related to current asthma in younger participants (OR, 0.68; 95% confidence interval [CI], 0.43-1.07), although the estimate was not significant (P = .09). Participants colonized with H pylori⁺cagA⁺ strains were less likely to have ever been diagnosed as having asthma compared with those without H pylori (OR, 0.79; 95% CI, 0.63-0.99). Colonization with a cagA⁺ H pylori strain was inversely associated with ever having had asthma only in younger participants (age <43 years; interaction P = .04); this inverse association was consistent with the inverse association of childhood onset of asthma (age ≤15 years) and cagA⁺ status (OR, 0.63; 95% CI, 0.43-0.93).

Current allergic rhinitis status and ever having had allergic rhinitis diagnosed were inversely associated with the presence of H pylori, especially cagA⁺ strains (Table 3). Inverse associations between H pylori presence and current allergic rhinitis diagnosis were more significant in younger than in older participants (interaction P = .06 and .07 for cagA⁻ and cagA⁺ strains, respectively). Similarly, inverse associations between H pylori presence and ever having had allergic rhinitis diagnosed also were more significant in younger participants. The presence of H pylori, especially cagA⁺ strains, was inversely associated with childhood-onset allergic rhinitis (OR, 0.68; 95% CI, 0.44-1.05 and OR, 0.55; 95% CI, 0.37-0.82 for cagA⁻ and cagA⁺ strains, respectively).

Colonization with either a cagA⁺ or a cagA⁻ strain of H pylori was inversely associated with having had 1 of a group of specified allergy symptoms in the previous 12 months (Table 4). The associations of H pylori⁺cagA⁻ and cagA⁺ status with allergy were not significantly different (P = .83) and were apparent only in younger adults. The interaction P values indicate that the association of H pylori (either cagA⁻ [P = .04] or cagA⁺ [P = .01]) with allergy differs significantly based on age. Analysis of individual allergy symptoms consistently showed significantly stronger inverse associations in younger participants (interaction P<.01). Among participants younger than 43 years, those with cagA⁺ strains were less likely to have any wheezing, whistling in the chest (OR, 0.86; 95% CI, 0.66-1.13), stuffy, itchy, or runny nose (OR, 0.77; 95% CI, 0.64-0.91), or watery, itchy eyes (OR, 0.86; 95% CI, 0.71-1.03) in the previous year compared with those without H pylori (Appendix 1 [available at http://homepages.nyu.edu/~cheny16/Appendixes.pdf]). We further evaluated the associations of H pylori status with allergy symptoms according to self-reported allergy sources. Compared with participants without H pylori, those with cagA⁺ strains were significantly or marginally significantly less likely to have allergy symptoms due to exposure to pollens (OR, 0.71; 95% CI, 0.71-0.95), animals (OR, 0.63; 95% CI, 0.43-0.88), or house dust (OR, 0.87; 95% CI, 0.73-1.04), especially younger adults (median age <43 years). There were no significant associations with exposures to the work environment, exercise, or cold air (Appendix 2 [available at http://homepages.nyu.edu/~cheny16/Appendixes.pdf]).

We evaluated associations between H pylori status and allergen-specific skin sensitization in the subgroup of 2386 adults who had undergone allergy skin testing. Compared with participants with H pylori–negative status, participants with H pylori, especially with cagA⁺ strains, were less likely to have skin sensitization due to several pollens and molds, especially younger participants (Table 5). There were no significant associations of H pylori status with skin sensitization due to the tested indoor allergens and foods (Appendix 3 [available at http://homepages.nyu.edu/~cheny16/Appendixes.pdf]).

Using the NHANES III database, we found that colonization with cagA⁺ strains of H pylori was inversely related to ever having had asthma, especially in younger adults and for asthma cases with onset during childhood. Colonization with H pylori also was inversely related to having had allergic rhinitis, allergy symptoms, and skin sensitization due to pollens and molds, especially in younger adults.

Consistent with the increasing prevalence of esophageal diseases as H pylori prevalence has been declining³² has come evidence of inverse associations between H pylori presence, especially cagA⁺ strains, and the risk of GERD and its sequelae.^14-18 Because GERD can be asymptomatic,^33,34 especially in childhood,³⁵ the extent of its increasing incidence may be underestimated. These trends and associations suggest the hypothesis that H pylori, especially cagA⁺ strains, could be protective against asthma, mediated in part by their protection against GERD, possibly due to heightened cagA⁺-induced gastric atrophy.³⁶ However, because the presence of GERD was not evaluated in NHANES III, and few participants reported taking prescription medicines for treating GERD, we did not directly evaluate whether the inverse association between asthma and H pylori is mediated via GERD. Future studies are needed.

The present observations also are consistent with the “hygiene hypothesis” that microbial infections during early childhood may prevent or diminish atopic sensitization and asthma.³⁷ In particular, inadequate microbial stimulation of gut-associated lymphoid tissue, a critical site for maturation of mucosal immunity,³⁸ may be relevant to this mechanism. Consistent with the present findings, asthma and allergic rhinitis were less frequent in NHANES III participants seropositive for hepatitis A virus (HAV), Toxoplasma gondii, and herpes simplex virus 1 than in seronegative persons.³⁹ In Italian cadets, atopy was inversely related to infections (T gondii and HAV) transmitted through gastrointestinal routes but not to those with differing major transmission.⁴⁰

A few studies have reported no associations^41,42 or weak inverse associations^40,43-45 of H pylori colonization with asthma and allergy. However, some of these studies were not adequately powered to detect a moderate association, and they did not address the relationships regarding colonization of H pylori cagA⁺ strains. The findings from the present study confirm and extend a previous analysis (J. Reibman, MD, M. Marmor, PhD, M.-E. Fernandez-Beros, PhD, L. Rogers, MD, G. I. Perez-Perez, DSc, and M.J.B., unpublished data, 2006) of 318 asthmatic patients and 206 controls in New York in which a similar inverse association between colonization of H pylori cagA⁺ strains and risk of asthma was seen. One form of the hygiene hypothesis proposes that particular allergic conditions are increasing because of T_H1- and T_H2-type immune response imbalances due to modern lifestyles.^37,46 Since T_H2 mediators suppress T_H1 responses and T_H1 mediators reciprocally inhibit T_H2 responses, these systems should be balanced in “health.”⁴⁶ Evidence suggests that immune activation after the establishment of H pylori colonization is more pronounced with cagA⁺ strains,^9,36,47 especially in children.^48,49 One explanation for the present findings is that enhanced host cellular responses to cagA⁺H pylori strains^48-50 affect early life equilibria of T_H1- and T_H2-type immune responses, driving the induction of immunoregulatory lymphocytes that prevent immune hyperreactivity states, such as asthma and allergy.⁵¹ Future studies should examine this hypothesis.

The inverse relationships of H pylori, especially cagA⁺ strains, with asthma, allergic rhinitis, and allergies were more pronounced in younger (median age <43 years) than in older individuals. Similarly, H pylori cagA positivity was inversely associated with the onset of asthma and allergic rhinitis at age 15 years or earlier but not after (Tables 2 and 3). The specificity for younger but not older persons is consistent with a birth cohort phenomenon¹ and suggests that secular increases in asthma and atopy in children and young adults may reflect reduced exposures to microbes such as H pylori. Because the acquisition of many microbes has receded more than that of H pylori, the biological effects of H pylori may be easier to distinguish in younger persons. In 2 Finnish cross-sectional studies,⁵² the prevalence of allergen specific IgE increased between 1973 and 1994, mainly in H pylori–negative persons.

That H pylori cagA⁺ status was inversely related to the presence of allergic rhinitis, self-reported allergy symptoms, and skin sensitization suggests that the protection related to H pylori (especially cagA⁺ strains) may not be specific for asthma but extends to other allergic conditions. A recent study⁵³ of Russian adults also found that H pylori presence was inversely associated with atopy, consistent with the present findings. Skin sensitization examined in NHANES III also was inversely associated with HAV seropositivity.³⁹ For this reason, we performed additional analyses to evaluate the relationships of skin sensitization and risk of asthma with joint status of HAV and H pylori seropositivity. The inverse associations with H pylori status remained similar in persons negative for serum antibody to HAV (data not shown), suggesting the independent effect of H pylori colonization. The H pylori cagA⁺ strains were inversely associated with skin sensitization due to pollens and molds but not with other tested antigens. Skin sensitization may vary in an allergen-specific manner; increased domestic exposure to dust mite and cockroach may be more etiologically relevant⁵⁴ to sensitization to these allergens compared with H pylori status.

The present study and others collectively provide evidence that colonization with H pylori, especially cagA⁺ strains, may reduce risks of asthma and allergy. Although we conducted a cross-sectional analysis, we consider that the observed associations are not due to “reverse causation.” Helicobacter pylori is acquired almost exclusively in childhood and usually persists for life unless antimicrobial therapy is given.^55,56 Although H pylori acquisition age was not assessed in this study, in most cases it precedes the reported onsets of asthma, allergic rhinitis, and allergy.^22,55 One issue is whether asthma or allergy would promote H pylori loss, for example, due to heightened antibiotic drug use. The stronger inverse associations between H pylori colonization and asthma and allergic rhinitis in younger adults provide evidence against the latter hypothesis since older persons should have had increased cumulative duration of asthma or atopy and exposure to medications (eg, antibiotics) compared with younger persons.

Treatment of H pylori was not evaluated in the NHANES III, and, therefore, some participants might have had previous treatment. However, no evidence has suggested that specific treatment to eradicate H pylori differs by affective status of asthma and allergy. Although laboratory-confirmed cases of asthma would be ideal for evaluating the putative association, errors in reporting history of asthma should not differ by H pylori status. These nondifferential misclassifications of outcome and exposure would, in general, bias toward the null, indicating that the true association may be greater.

In summary, these findings provide evidence that the continuing disappearance of H pylori in developed countries⁵⁶ is related to the increase in asthma and atopic disorders.⁵⁷ How the lack of H pylori might contribute to the pathogenesis of these disorders is not known but could relate to immunologic imbalance. Additional studies are needed to confirm these observations and to identify the mechanisms.

Correspondence: Martin J. Blaser, MD, Department of Medicine, New York University School of Medicine, 550 First Ave, OBV-A606, New York, NY 10016 (martin.blaser@med.nyu.edu).

Accepted for Publication: December 21, 2006.

Author Contributions:Study concept and design: Chen and Blaser. Acquisition of data: Chen and Blaser. Analysis and interpretation of data: Chen and Blaser. Drafting of the manuscript: Chen and Blaser. Critical revision of the manuscript for important intellectual content: Chen and Blaser. Statistical analysis: Chen. Obtained funding: Chen and Blaser. Administrative, technical, and material support: Blaser. Study supervision: Blaser.

Financial Disclosure: Dr Blaser, as a co-discoverer of cagA at Vanderbilt University, can receive royalties from the commercial exploitation of cagA. No diagnostic tests for cagA are currently licensed.

Funding/Support: This research was supported by grant ES000260 from the National Institute of Environmental Health Sciences, grant CA016087 from the National Cancer Institute, grant RO1GM63270 from the National Institutes of Health, the Diane Belfer Program in Human Microbial Ecology, and the Senior Scholar Award of the Ellison Medical Foundation.